Pharmac. Ther. Vol. 48, pp. 157-188, 1990 Printed in Great Britain. All rights reserved
0163-7258/90 $0.00 + 0.50 © 1990 Pergamon Press pie
Associate Editor: D. GRUNBERGER
REGULATION OF GENE EXPRESSION BY TUMOR PROMOTERS HANS J. RAHMSDORF a n d PETER HERRLICH Kernforschungszentrum Karlsruhe, Institut fiir Genetik und Toxikologie, Postfach 3640, D-7500 Karlsruhe 1, F.R.G. Abstraet--Tumor promoters change the program of genes expressed in cells in culture and in the multicellular organism. The growing list of genes that are induced or repressed includes protooncogenes, transcription factors, secreted proteases and viruses. Most of the regulation is at the level of transcription. Several of the cis-acting promoter elements mediating regulation, the transcription factors binding to these elements and their post-translational activation, as well as some of the initial steps of the interaction of cells with tumor promoters have been characterized. The components of the signal transduction chain to the nucleus are, however, still unknown. Mutant and inhibitor studies suggest that the activation or inactivation of certain genes constitute the basis for the development of the tumor promotion phenotype.
CONTENTS 1. 2. 3. 4.
Introduction Definition of Tumor Promotion and Tumor Promoters Cells in Culture Respond to Tumor Promoters Action of Tumor Promoters on Genes 4.1. Responsive genes 4.2. Mechanisms of TPA induced RNA accumulation or decrease 4.2.1. In most cases TPA acts by changing the transcription of the responsive gene 4.2.2. Phorbol esters activate a responsive gene either 'directly' or by inducing the prior synthesis of a regulatory protein 4.2.3. TPA responsive DNA elements (TREs) 4.2.4. Proteins which bind to the TPA responsive elements 5. Phorbol Ester Induced Signal Transduction 5.1. Protein kinase C as a receptor for tumor promoting phorbol esters 5.2. Immediate phorbol ester induced changes: changes in concentrations of intracellular ions, release of arachidonic acid, protein phosphorylation 5.3. Preliminary identification of other components in the signal transfer 5.4. Signal transduction is transient and strictly counter-regulated 5.5. Glucocorticoid hormones interfere with transcription factor activity 6. From Induced Genes to the Tumor Promotion Phenotype Acknowledgements References
1. I N T R O D U C T I O N T u m o r promoters by definition enhance carcinogeninduced tumor formation in animals and men. The phenotype of tumor promotion may be characteristic of multiceilular organisms, and indeed be a multicellular phenomenon. Whether a clonal cell population can undergo promotion is still debated. In cell culture, promotion-like processes have been observed which suggests that clonal populations can indeed be promoted. A tissue that had been treated with a tumor promoter, shows dramatic morphological changes (e.g. a sun-burn-like inflammation of the mouse skin and an extensive proliferative response of the epidermis); similar proliferative and differentiative processes can
157 158 158 159 159 163 163 164 164 167 172 172 172 173 174 175 175 176 176
be induced in clonal cells in culture. The intracellular correlate of these visible changes is a massive change in the program of genes expressed. It is likely although not yet proven that these activations and inactivations of genes by tumor promoters represent the primary and decisive events in tumor promotion in that they include genes causally involved in the generation of tumorigenic steps. These tumorigenic steps must ultimately result in clonally hereditable changes such as point mutations and chromosomal aberrations. In this review article we will concentrate on the mechanisms of the activation and inactivation of genes. This focus is justified since most advances in understanding gene regulation by tumor promoters have been made over the last three years. N o attempt will be made to 157
158
H. J, RAHMSDORFand P. HERRLICH
describe all genes affected. Rather, examples will be selected to illustrate individual mechanistic arguments. The last paragraph will attempt to interpret the data obtained on the regulated genes, for the generation of the promotion phenotype.
2. D E F I N I T I O N O F T U M O R P R O M O T I O N A N D T U M O R PROMOTERS The terms 'tumor initiation' and 'tumor promotion' have originally been defined in experiments with mice. Skin tumors were induced by a single treatment of skin with an initiating agent (often applied as a subcarcinogenic dose of 7,12-dimethylbenz[a]anthracene) and subsequent repeated treatments with a tumor promoter. While neither the initiator nor the promoter alone (at the doses given) led to neoplastic transformation, combined treatment caused the appearance of many skin papillomas in about 6 weeks, some of which progressed to skin carcinoma (for review see Diamond et al., 1980; Yuspa, 1984; Diamond, 1987; Hecker, 1987a,b). Tumor formation, thus, requires two operationally distinct steps, initiation and promotion. The first promoting condition discovered was, in fact, quite physiologic: repeated healing reactions after mechanical wounding of mouse skin that had been pretreated with carcinogenic tars, yielded a greatly enhanced frequency of tumor formation (Deelman, 1927). Chemicals that replaced the physical wounding permitted the subdivision of the promotion process into stage I and stage II. Repeated treatment with a ('complete') tumor promoter (e.g. TPA: 12-O-tetradecanoylphorbol-13-acetate) could be replaced by a single treatment followed by repeated treatments with an 'incomplete' (stage II) promoter which on its own could not promote (e.g. turpentine, Boutwell, 1964; the diterpene ester, mezerein, Slaga et al., 1980a; or 12-retinoylphorbol-13acetate, Ffirstenberger et al., 1981). According to the refined protocol three types of agents cooperate: (i) carcinogens generate the initiated cell; (ii) a stage l tumor promoter converts it into a 'latent tumor cell'; (iii) stage II tumor promoters lead to the 'promotion' of these cells. By varying the order and time course of treatment it could be shown, that the process of initiation is practically irreversible, that the 'latent tumor cell phenotype' brought about by a single dose of a complete tumor promoter, decays with a half life of several weeks, and that the originally described time schedule of treatment can be reversed; the tumor promoter can be given with similar results prior to 'initiation' of the cells (Ffirstenberger et al., 1983, 1985). Potent skin promoters have been isolated from plants and marine organisms. The tumor promoter TPA was originally purified from a methanolic extract of crotonoil (made from Croton tiglium [Euphorbiaceae]; van Duuren and Orris, 1965; Hecker, 1968), mezerein from Daphne mezereum (Sharkey et al., 1989), the indole alkaloid teleocidin from the mycelia of Streptomyces mediocidicus, the polyacetate aplysiatoxin from the blue green alga Lyngbya majuscula (Fujiki and Sugimura, 1987), palytoxin from the marine colenterate Palythoa (Fujiki et al., 1986) and
okadaic acid from a black sponge, Halichondria okadai (Suganuma et al., 1988). In the broad definition of promotion being a process that enhances the yield of tumors but does not suffice for tumorigenesis, tumor formation in many species in addition to mice and in many tissues probably profits from promoting treatments. Experimental evidence exists only for a few systems. These systems reveal what might have been expected: tumor promoters may be tissue-specific. For instance phorbol derivatives that are inactive on skin, as well as TPA given intra peritoneum act as promoters of leukemia and of tumors of the lung, liver and mammary gland in initiated rodents. Phenobarbital is a potent promoter of liver tumorigenesis in rats initiated with 2-acetylaminofluorene. Two dietary sweeteners, saccharine and cyclamate, are promoters of bladder tumor formation in rats that had been initiated with N-methyl-N-nitrosourea (for review see Diamond et al., 1980; Diamond, 1987). Occupational exposure to asbestos, which does not seem to be mutagenic on its own, leads to an increased incidence of bronchogenic carcinoma in man (for references see Marsh and Mossman, 1988; Mossman et al., 1990). Of course the possibility exists that many physiologic agents can act as tumor promoters. This possibility is plausible for steroid hormones which can act as promoters on certain tissues (Henderson et al., 1988) and for inflammatory mediators which may imitate the wounding condition mentioned above. It is advisable to maintain some reservation with respect to the simplicity of the initiation-promotion hypothesis. Carcinogens may turn out to act as initiators and promoters at the same time and promotion may include the generation of mutagenic events. Thus, the distinction between initiation and promotion may depend on a narrow window of agent's action. We will return to an interpretation offered by the overlap of phorbol ester-induced and D N A damage-induced responses, at the end of this review.
3. CELLS IN C U L T U R E RESPOND TO T U M O R PROMOTERS Phorbol esters rapidly cause inflammation and hyperplasia in the mouse skin irrespective of prior treatment with an initiating agent: 24hr after the application of a tumor promoter, the skin is heavily infiltrated by leukocytes (Frei and Stephens, 1968). In addition, skin contains several other types of cells which has made an analysis of the promotion step difficult. This has generated the motivation for developing systems with cloned cells in culture that mimic part of the promotion phenotype. With the information obtained with cultured cells, some studies have then been done on mouse skin and similar molecular processes have been found (e.g. Krieg et al., 1988; Hashiba et al., 1989). Several cell systems have been established, in which tumor promoters enhance dramatically the neoplastic transformation of both primary and established cells in culture treated with low doses of carcinogens (for review see Diamond, 1987). Upon subcarcinogenic doses of carcinogenic hydrocarbons, u.v.-irradiation
Regulation of gene expression by tumor promoters or X-irradiation, all followed by TPA, 'transformed' fibroblastoid or epithelial cell clones are generated (Mondal and Heidelberger, 1976; Mondal et al., 1976; Kennedy et al., 1978; Steele et al., 1980). TPA also causes primary rat embryo fibroblasts transfected with a 'transforming' oncogene (ras) (not so when transfected with an 'immortalizing' oncogene, e.g. myc), to grow out and to form foci (Dotto et al., 1985; Lopez et al., 1989). Only when unphysiologically high TPA concentrations are used, rnyc-transfected cells also form foci (Connan et al., 1985). Since only a combination of a transforming and an immortalizing oncogene is capable of transforming primary cells (Land et al., 1983), ras expression may substitute for the initiating carcinogen, while TPA treatment may imitate the action of one of the immortalizing oncogenes. Also in the mouse skin, the treatment with an initiating agent can be replaced by infection of the epithelial cells with Harvey- or Balb-murine sarcoma viruses (Brown et al., 1986). In line with the idea that myc substitutes for the promoting treatment, overexpression of an exogenous m y c gene leads to an increase in radiation-induced transformation of cells in culture (Sorrentino et al., 1987). To equate TPA action with a myc-like function is probably too simplistic--as are current functional descriptions of oncogene action. Just as ras can be coupled to proliferative and non-proliferative signal pathways depending on cell type, tumor promoters may exert positive and negative effects on the proliferation of cells in culture (Yuspa et al., 1976; Driedger and Blumberg, 1977; Dicker and Rozengurt, 1978; Kaufmann and Schwartz, 1981; Kinzel et al., 1981), and they may promote or inhibit differentiation of cells in culture (Rovera et al., 1977; Cohen et al., 1977; Lowe et al., 1978; Miao et al., 1978; for review see Diamond, 1987). One of the most intensely studied phorbol ester-induced differentiation processes causes promyelocytic HL60 cells to mature along the mononuclear phagocytic lineage (Rovera et al., 1979). Using selection systems for phorbol ester-induced end points, resistant cell variants have been isolated including variants of epithelial cells (Colburn et al., 1983; Bernstein and Colburn, 1989), fibroblasts (Herschman, 1985), promyelocytic cells (Kiss et al., 1987; Leftwich et al., 1987) and T-lymphocytes (Mills et al., 1988). Tabuse et al. (1989) have even isolated phorbol ester-resistant mutants of the nematode Caenorabditis elegans. These variants may be blocked in various parts of the signal transduction pathways and will be valuable tools in unravelling the signal chains initiated by phorbol esters.
4. ACTION OF T U M O R PROMOTERS ON GENES We will now turn to the major subject of this review: the regulation of genes by tumor promoters. While the tumorigenic outcome of tumor promotion concerns only a minority of cells (10 -6 or less), the massive changes of gene expression occur in all cells. The last paragraph of this review will summarize why the changes in gene expression are, nevertheJPT 48/2--D
159
less, likely to be responsible for the promotion phenotype. To approach the mechanism of gene regulation we will order the experimental data into the following frames: (i) define responsive genes; (ii) examine the type of regulation (transcriptional versus other modes; direct action of tumor promoters versus involvement of newly synthesized proteins); (iii) analyze responding genes with respect to cis-acting sequences and protein factors binding to these sequences; and (iv) analyze the primary event, that is the target of TPA interaction, and the components immediately depending on the primary event and feeding into the signal transduction from the plasma membrane to the gene. 4.1. RESPONSIVEGENES An ever growing number of genes has been found to respond to phorbol esters (Table 1). We can group them according to type of regulation. For instance, many genes which are activated within minutes after treatment with tumor promoters, are likely to be under direct phorbol ester control. Interestingly these genes are also induced upon treatment of cells with serum or isolated growth factors, their expression is usually transient and their induction is not affected by inhibitors of protein synthesis. Rather they are overexpressed if protein synthesis is blocked. The increased abundance of mRNAs in cells after such treatment, has been exploited to isolate cDNA clones (Lau and Nathans, 1987; Almendral et al., 1988; Sukhatme et al., 1988). Some of these have attracted particular attention, especially those coding for cellular homologues of viral nuclear oncogenes, c-fos, c-jun and c-rnyc are among the immediate-response genes in fibroblasts (for references see Table 1). Their transcription after phorbol ester treatment is, however, not simultaneous, but rather follows a defined order: c-fos mRNA is detected first, within a few minutes, followed by an increased accumulation of c-jun RNA and c-myc RNA (Greenberg and Ziff, 1984; Miiller et al., 1984; and references of Table 1). Another group of immediate-response genes codes for cytoplasmic and secreted oncoproteins. For instance, in HL60 cells, in glioblastoma cells and in microvascular endothelial cells mRNA coding for the B-chain of platelet-derived growth factor (c-sis) is produced, preceding differentaition (Colamonici et al., 1986; Pantazis et al., 1986; Starksen et al., 1987). c-fins mRNA that codes for the receptor of the mononuclear phagocyte colony stimulating factor, CSF1, is elevated in HL60 cells (Rettenmier et al., 1986). Also cellular genes, which as yet have not been related to viral oncoproteins, are transiently activated upon phorbol ester treatment of fibroblasts: among those are the actin genes, Jun B, EGR1 (synonymous with Krox 24 and Zif 268), TIS 8 that codes for a 'zink-finger' protein and nut/77 ( = T I S 1), a member of the hormone receptor gene family (for references see Table 1). Other related genes have to our knowledge not been shown to be inducible by phorboi esters although their responsiveness is likely (Krox 20: Chavrier et al., 1988; and the two c-fos-related proteins Fra-I and fos B: Cohen and Curran, 1988; Zerial et al., 1989).
160
H.J. RAHMSDORF and P. HERRLICH
'~,
~
o e~
0", i ~
OX
O0
-u .-~o~ _
:~
~
~
~
~
~
~
~
~
~
~
"~
~
0~.~ ~
~. ~~ o- ~
.
~
'~'=
~
~ , ~1:~ 0
~
O0
.~
"~
~
~ ~,.
~.-,~
~
0
"~
~ 0 0 ~
,o
,~ ~
,~ ~
~
~ . .,~ , , - ~ . " ~ - ~ ~ ~ ~ ~
~
~..~ , , . ~ I . ~
"~ ~ . ~
.~ ~ o ~ ~ "
~ ~.~'--
~
~
~
e~
~
~.~_ 0
~"~
~
~°~
~.-
I.
~ ~ ~.~
~ ~ o ~,
^ ~
"-
~~
•~ -.~
'~ I..-, m
~
~--~-~
~ ~ ~"'o ~
E
0
0
0
0
0
0
•.~ ~ .~
~ ~,.~'" ...~ ~ '~..~
h"
0
0
0
0
0
~" ~
0
162
H . J . RAHMSDORF and P. HERRLICH
e,m
.@
.o
~d
~,~
~,~,~,~,
.
~-~,
N
~_~-
-g
~=
_
~
o
o
~
o
=
~
•"~
,
~
~
~
0
o
= .o
~-~-~ ml
g~
,o
~ .=
8
0
o 0
e.,
0
0
× < Z
e~ o
"o
._~ ~
.
-~
.
0
e~ i
~
----'- 0
" ~ . ~ r~
0163-7258/90 $0.00 + 0.50 © 1990 Pergamon Press pie
Associate Editor: D. GRUNBERGER
REGULATION OF GENE EXPRESSION BY TUMOR PROMOTERS HANS J. RAHMSDORF a n d PETER HERRLICH Kernforschungszentrum Karlsruhe, Institut fiir Genetik und Toxikologie, Postfach 3640, D-7500 Karlsruhe 1, F.R.G. Abstraet--Tumor promoters change the program of genes expressed in cells in culture and in the multicellular organism. The growing list of genes that are induced or repressed includes protooncogenes, transcription factors, secreted proteases and viruses. Most of the regulation is at the level of transcription. Several of the cis-acting promoter elements mediating regulation, the transcription factors binding to these elements and their post-translational activation, as well as some of the initial steps of the interaction of cells with tumor promoters have been characterized. The components of the signal transduction chain to the nucleus are, however, still unknown. Mutant and inhibitor studies suggest that the activation or inactivation of certain genes constitute the basis for the development of the tumor promotion phenotype.
CONTENTS 1. 2. 3. 4.
Introduction Definition of Tumor Promotion and Tumor Promoters Cells in Culture Respond to Tumor Promoters Action of Tumor Promoters on Genes 4.1. Responsive genes 4.2. Mechanisms of TPA induced RNA accumulation or decrease 4.2.1. In most cases TPA acts by changing the transcription of the responsive gene 4.2.2. Phorbol esters activate a responsive gene either 'directly' or by inducing the prior synthesis of a regulatory protein 4.2.3. TPA responsive DNA elements (TREs) 4.2.4. Proteins which bind to the TPA responsive elements 5. Phorbol Ester Induced Signal Transduction 5.1. Protein kinase C as a receptor for tumor promoting phorbol esters 5.2. Immediate phorbol ester induced changes: changes in concentrations of intracellular ions, release of arachidonic acid, protein phosphorylation 5.3. Preliminary identification of other components in the signal transfer 5.4. Signal transduction is transient and strictly counter-regulated 5.5. Glucocorticoid hormones interfere with transcription factor activity 6. From Induced Genes to the Tumor Promotion Phenotype Acknowledgements References
1. I N T R O D U C T I O N T u m o r promoters by definition enhance carcinogeninduced tumor formation in animals and men. The phenotype of tumor promotion may be characteristic of multiceilular organisms, and indeed be a multicellular phenomenon. Whether a clonal cell population can undergo promotion is still debated. In cell culture, promotion-like processes have been observed which suggests that clonal populations can indeed be promoted. A tissue that had been treated with a tumor promoter, shows dramatic morphological changes (e.g. a sun-burn-like inflammation of the mouse skin and an extensive proliferative response of the epidermis); similar proliferative and differentiative processes can
157 158 158 159 159 163 163 164 164 167 172 172 172 173 174 175 175 176 176
be induced in clonal cells in culture. The intracellular correlate of these visible changes is a massive change in the program of genes expressed. It is likely although not yet proven that these activations and inactivations of genes by tumor promoters represent the primary and decisive events in tumor promotion in that they include genes causally involved in the generation of tumorigenic steps. These tumorigenic steps must ultimately result in clonally hereditable changes such as point mutations and chromosomal aberrations. In this review article we will concentrate on the mechanisms of the activation and inactivation of genes. This focus is justified since most advances in understanding gene regulation by tumor promoters have been made over the last three years. N o attempt will be made to 157
158
H. J, RAHMSDORFand P. HERRLICH
describe all genes affected. Rather, examples will be selected to illustrate individual mechanistic arguments. The last paragraph will attempt to interpret the data obtained on the regulated genes, for the generation of the promotion phenotype.
2. D E F I N I T I O N O F T U M O R P R O M O T I O N A N D T U M O R PROMOTERS The terms 'tumor initiation' and 'tumor promotion' have originally been defined in experiments with mice. Skin tumors were induced by a single treatment of skin with an initiating agent (often applied as a subcarcinogenic dose of 7,12-dimethylbenz[a]anthracene) and subsequent repeated treatments with a tumor promoter. While neither the initiator nor the promoter alone (at the doses given) led to neoplastic transformation, combined treatment caused the appearance of many skin papillomas in about 6 weeks, some of which progressed to skin carcinoma (for review see Diamond et al., 1980; Yuspa, 1984; Diamond, 1987; Hecker, 1987a,b). Tumor formation, thus, requires two operationally distinct steps, initiation and promotion. The first promoting condition discovered was, in fact, quite physiologic: repeated healing reactions after mechanical wounding of mouse skin that had been pretreated with carcinogenic tars, yielded a greatly enhanced frequency of tumor formation (Deelman, 1927). Chemicals that replaced the physical wounding permitted the subdivision of the promotion process into stage I and stage II. Repeated treatment with a ('complete') tumor promoter (e.g. TPA: 12-O-tetradecanoylphorbol-13-acetate) could be replaced by a single treatment followed by repeated treatments with an 'incomplete' (stage II) promoter which on its own could not promote (e.g. turpentine, Boutwell, 1964; the diterpene ester, mezerein, Slaga et al., 1980a; or 12-retinoylphorbol-13acetate, Ffirstenberger et al., 1981). According to the refined protocol three types of agents cooperate: (i) carcinogens generate the initiated cell; (ii) a stage l tumor promoter converts it into a 'latent tumor cell'; (iii) stage II tumor promoters lead to the 'promotion' of these cells. By varying the order and time course of treatment it could be shown, that the process of initiation is practically irreversible, that the 'latent tumor cell phenotype' brought about by a single dose of a complete tumor promoter, decays with a half life of several weeks, and that the originally described time schedule of treatment can be reversed; the tumor promoter can be given with similar results prior to 'initiation' of the cells (Ffirstenberger et al., 1983, 1985). Potent skin promoters have been isolated from plants and marine organisms. The tumor promoter TPA was originally purified from a methanolic extract of crotonoil (made from Croton tiglium [Euphorbiaceae]; van Duuren and Orris, 1965; Hecker, 1968), mezerein from Daphne mezereum (Sharkey et al., 1989), the indole alkaloid teleocidin from the mycelia of Streptomyces mediocidicus, the polyacetate aplysiatoxin from the blue green alga Lyngbya majuscula (Fujiki and Sugimura, 1987), palytoxin from the marine colenterate Palythoa (Fujiki et al., 1986) and
okadaic acid from a black sponge, Halichondria okadai (Suganuma et al., 1988). In the broad definition of promotion being a process that enhances the yield of tumors but does not suffice for tumorigenesis, tumor formation in many species in addition to mice and in many tissues probably profits from promoting treatments. Experimental evidence exists only for a few systems. These systems reveal what might have been expected: tumor promoters may be tissue-specific. For instance phorbol derivatives that are inactive on skin, as well as TPA given intra peritoneum act as promoters of leukemia and of tumors of the lung, liver and mammary gland in initiated rodents. Phenobarbital is a potent promoter of liver tumorigenesis in rats initiated with 2-acetylaminofluorene. Two dietary sweeteners, saccharine and cyclamate, are promoters of bladder tumor formation in rats that had been initiated with N-methyl-N-nitrosourea (for review see Diamond et al., 1980; Diamond, 1987). Occupational exposure to asbestos, which does not seem to be mutagenic on its own, leads to an increased incidence of bronchogenic carcinoma in man (for references see Marsh and Mossman, 1988; Mossman et al., 1990). Of course the possibility exists that many physiologic agents can act as tumor promoters. This possibility is plausible for steroid hormones which can act as promoters on certain tissues (Henderson et al., 1988) and for inflammatory mediators which may imitate the wounding condition mentioned above. It is advisable to maintain some reservation with respect to the simplicity of the initiation-promotion hypothesis. Carcinogens may turn out to act as initiators and promoters at the same time and promotion may include the generation of mutagenic events. Thus, the distinction between initiation and promotion may depend on a narrow window of agent's action. We will return to an interpretation offered by the overlap of phorbol ester-induced and D N A damage-induced responses, at the end of this review.
3. CELLS IN C U L T U R E RESPOND TO T U M O R PROMOTERS Phorbol esters rapidly cause inflammation and hyperplasia in the mouse skin irrespective of prior treatment with an initiating agent: 24hr after the application of a tumor promoter, the skin is heavily infiltrated by leukocytes (Frei and Stephens, 1968). In addition, skin contains several other types of cells which has made an analysis of the promotion step difficult. This has generated the motivation for developing systems with cloned cells in culture that mimic part of the promotion phenotype. With the information obtained with cultured cells, some studies have then been done on mouse skin and similar molecular processes have been found (e.g. Krieg et al., 1988; Hashiba et al., 1989). Several cell systems have been established, in which tumor promoters enhance dramatically the neoplastic transformation of both primary and established cells in culture treated with low doses of carcinogens (for review see Diamond, 1987). Upon subcarcinogenic doses of carcinogenic hydrocarbons, u.v.-irradiation
Regulation of gene expression by tumor promoters or X-irradiation, all followed by TPA, 'transformed' fibroblastoid or epithelial cell clones are generated (Mondal and Heidelberger, 1976; Mondal et al., 1976; Kennedy et al., 1978; Steele et al., 1980). TPA also causes primary rat embryo fibroblasts transfected with a 'transforming' oncogene (ras) (not so when transfected with an 'immortalizing' oncogene, e.g. myc), to grow out and to form foci (Dotto et al., 1985; Lopez et al., 1989). Only when unphysiologically high TPA concentrations are used, rnyc-transfected cells also form foci (Connan et al., 1985). Since only a combination of a transforming and an immortalizing oncogene is capable of transforming primary cells (Land et al., 1983), ras expression may substitute for the initiating carcinogen, while TPA treatment may imitate the action of one of the immortalizing oncogenes. Also in the mouse skin, the treatment with an initiating agent can be replaced by infection of the epithelial cells with Harvey- or Balb-murine sarcoma viruses (Brown et al., 1986). In line with the idea that myc substitutes for the promoting treatment, overexpression of an exogenous m y c gene leads to an increase in radiation-induced transformation of cells in culture (Sorrentino et al., 1987). To equate TPA action with a myc-like function is probably too simplistic--as are current functional descriptions of oncogene action. Just as ras can be coupled to proliferative and non-proliferative signal pathways depending on cell type, tumor promoters may exert positive and negative effects on the proliferation of cells in culture (Yuspa et al., 1976; Driedger and Blumberg, 1977; Dicker and Rozengurt, 1978; Kaufmann and Schwartz, 1981; Kinzel et al., 1981), and they may promote or inhibit differentiation of cells in culture (Rovera et al., 1977; Cohen et al., 1977; Lowe et al., 1978; Miao et al., 1978; for review see Diamond, 1987). One of the most intensely studied phorbol ester-induced differentiation processes causes promyelocytic HL60 cells to mature along the mononuclear phagocytic lineage (Rovera et al., 1979). Using selection systems for phorbol ester-induced end points, resistant cell variants have been isolated including variants of epithelial cells (Colburn et al., 1983; Bernstein and Colburn, 1989), fibroblasts (Herschman, 1985), promyelocytic cells (Kiss et al., 1987; Leftwich et al., 1987) and T-lymphocytes (Mills et al., 1988). Tabuse et al. (1989) have even isolated phorbol ester-resistant mutants of the nematode Caenorabditis elegans. These variants may be blocked in various parts of the signal transduction pathways and will be valuable tools in unravelling the signal chains initiated by phorbol esters.
4. ACTION OF T U M O R PROMOTERS ON GENES We will now turn to the major subject of this review: the regulation of genes by tumor promoters. While the tumorigenic outcome of tumor promotion concerns only a minority of cells (10 -6 or less), the massive changes of gene expression occur in all cells. The last paragraph of this review will summarize why the changes in gene expression are, nevertheJPT 48/2--D
159
less, likely to be responsible for the promotion phenotype. To approach the mechanism of gene regulation we will order the experimental data into the following frames: (i) define responsive genes; (ii) examine the type of regulation (transcriptional versus other modes; direct action of tumor promoters versus involvement of newly synthesized proteins); (iii) analyze responding genes with respect to cis-acting sequences and protein factors binding to these sequences; and (iv) analyze the primary event, that is the target of TPA interaction, and the components immediately depending on the primary event and feeding into the signal transduction from the plasma membrane to the gene. 4.1. RESPONSIVEGENES An ever growing number of genes has been found to respond to phorbol esters (Table 1). We can group them according to type of regulation. For instance, many genes which are activated within minutes after treatment with tumor promoters, are likely to be under direct phorbol ester control. Interestingly these genes are also induced upon treatment of cells with serum or isolated growth factors, their expression is usually transient and their induction is not affected by inhibitors of protein synthesis. Rather they are overexpressed if protein synthesis is blocked. The increased abundance of mRNAs in cells after such treatment, has been exploited to isolate cDNA clones (Lau and Nathans, 1987; Almendral et al., 1988; Sukhatme et al., 1988). Some of these have attracted particular attention, especially those coding for cellular homologues of viral nuclear oncogenes, c-fos, c-jun and c-rnyc are among the immediate-response genes in fibroblasts (for references see Table 1). Their transcription after phorbol ester treatment is, however, not simultaneous, but rather follows a defined order: c-fos mRNA is detected first, within a few minutes, followed by an increased accumulation of c-jun RNA and c-myc RNA (Greenberg and Ziff, 1984; Miiller et al., 1984; and references of Table 1). Another group of immediate-response genes codes for cytoplasmic and secreted oncoproteins. For instance, in HL60 cells, in glioblastoma cells and in microvascular endothelial cells mRNA coding for the B-chain of platelet-derived growth factor (c-sis) is produced, preceding differentaition (Colamonici et al., 1986; Pantazis et al., 1986; Starksen et al., 1987). c-fins mRNA that codes for the receptor of the mononuclear phagocyte colony stimulating factor, CSF1, is elevated in HL60 cells (Rettenmier et al., 1986). Also cellular genes, which as yet have not been related to viral oncoproteins, are transiently activated upon phorbol ester treatment of fibroblasts: among those are the actin genes, Jun B, EGR1 (synonymous with Krox 24 and Zif 268), TIS 8 that codes for a 'zink-finger' protein and nut/77 ( = T I S 1), a member of the hormone receptor gene family (for references see Table 1). Other related genes have to our knowledge not been shown to be inducible by phorboi esters although their responsiveness is likely (Krox 20: Chavrier et al., 1988; and the two c-fos-related proteins Fra-I and fos B: Cohen and Curran, 1988; Zerial et al., 1989).
160
H.J. RAHMSDORF and P. HERRLICH
'~,
~
o e~
0", i ~
OX
O0
-u .-~o~ _
:~
~
~
~
~
~
~
~
~
~
~
"~
~
0~.~ ~
~. ~~ o- ~
.
~
'~'=
~
~ , ~1:~ 0
~
O0
.~
"~
~
~ ~,.
~.-,~
~
0
"~
~ 0 0 ~
,o
,~ ~
,~ ~
~
~ . .,~ , , - ~ . " ~ - ~ ~ ~ ~ ~
~
~..~ , , . ~ I . ~
"~ ~ . ~
.~ ~ o ~ ~ "
~ ~.~'--
~
~
~
e~
~
~.~_ 0
~"~
~
~°~
~.-
I.
~ ~ ~.~
~ ~ o ~,
^ ~
"-
~~
•~ -.~
'~ I..-, m
~
~--~-~
~ ~ ~"'o ~
E
0
0
0
0
0
0
•.~ ~ .~
~ ~,.~'" ...~ ~ '~..~
h"
0
0
0
0
0
~" ~
0
162
H . J . RAHMSDORF and P. HERRLICH
e,m
.@
.o
~d
~,~
~,~,~,~,
.
~-~,
N
~_~-
-g
~=
_
~
o
o
~
o
=
~
•"~
,
~
~
~
0
o
= .o
~-~-~ ml
g~
,o
~ .=
8
0
o 0
e.,
0
0
× < Z
e~ o
"o
._~ ~
.
-~
.
0
e~ i
~
----'- 0
" ~ . ~ r~
