¹Research Institute of Molecular Pathology (I.M.P.), Dr. Bohr-Gasse 7, A-1030 Vienna, Austria

²Institute of Clinical Pathology, University Hospital, Schmelzbergstrasse 12, CH-8091 Zurich, Switzerland

Correspondence to: EF Wagner, Research Institute of Molecular Pathology (I.M.P.), Dr. Bohr-Gasse 7, A-1030 Vienna, Austria

Genetically modified mice have provided important insights into the biological functions of the dimeric transcription factor complex AP-1. Extensive analyses of mice and cells with genetically modified Fos or Jun proteins provide novel insights into the physiological functions of AP-1 proteins. Using knock-out strategies it was found that some components, such as c-Fos, FosB and JunD are dispensable, whereas others, like c-Jun, JunB and Fra-1 are essential in embryonic development and/or in the adult organism. Besides the specific roles of AP-1 proteins in developmental processes, we are beginning to obtain a better molecular understanding of the cell-context dependent function of AP-1 in cell proliferation and apoptosis, in bone biology as well as in multistep tumorigenesis. Oncogene (2001) 20, 2401-2412.

AP-1; Fos proteins; Jun proteins; development; bone biology; tumorigenesis

Introduction

The transcription factor AP-1 consists of a variety of dimers composed of members of the Fos, Jun and ATF families of proteins (Angel and Karin, 1991). While the Fos proteins (c-Fos, FosB, Fra-1, Fra-2) can only heterodimerize with members of the Jun family, the Jun proteins (c-Jun, JunB, JunD) can both homo- and heterodimerize with Fos members to form transcriptionally active complexes. Some members of the ATF and CREB families of proteins are also part of AP-1 complexes (see Shaulian and Karin; Angel et al., this issue). AP-1 converts extracellular signals into changes in the expression of specific target genes which harbour AP-1 binding site(s) in their promoter or enhancer regions. AP-1 has been implicated in a large variety of biological processes including cell differentiation, proliferation, apoptosis and oncogenic transformation. AP-1 activity is modulated by interactions with other transcriptional regulators and is further controlled by upstream kinases that link AP-1 to various signal transduction pathways (see Herrlich; Rincon et al.; Shaulian and Karin, this issue). Here we will summarize the current understanding of AP-1 functions during development and tumorigenesis obtained from the analysis of genetically modified mice and cells in which distinct fos and jun genes have been ectopically expressed, inactivated, mutated or replaced by each other.

AP-1 in development and organogenesis

Most important insights regarding the specific functions of AP-1 proteins in mouse development were obtained from loss-of-function experiments using ES cell technology. A summary of the phenotypes of mice harbouring genetic modifications of the different fos and jun genes is given in Tables 1 and 2. These analyses have revealed that each AP-1 component has specific functions during embryogenesis and organogenesis. With regard to embryonic development, the Fos and Jun proteins can be grouped into two categories: some e.g. c-Fos, FosB and JunD are dispensable, whereas others e.g. c-Jun, JunB and Fra-1 are essential for embryonic development. In the following we will discuss the specific phenotypes of these knock-out mice, as well as studies in transgenic mice expressing individual AP-1 components.

c-Fos

Mice lacking c-Fos are viable and fertile but lack osteoclasts resulting in an osteopetrotic phenotype (Johnson et al., 1992; Wang et al., 1992). This phenotype is strain-dependent, since on a 129/sv background fewer pups are born and reach weaning age, when compared to C57B1/6 mutant mice (Matsuo K, unpublished observations). Mutant mice also show abnormalities of the haematopoietic system including extramedullary haematopoiesis and lymphopenia, which are both secondary to the bone phenotype (Okada et al., 1994). Although c-Fos is rapidly induced in T-cells upon activation and regulates the transcription of various cytokines, including interleukin 2, the analysis of c-Fos deficient mice has shown that c-Fos is not required for the differentiation and activity of peripheral T-cells (Jain et al., 1994). To further study c-Fos functions in cell differentiation and development, c-Fos was ectopically expressed in various cell types and transgenic mice. Embryonic stem cells overexpressing c-Fos efficiently contribute to the development of chimeric mice indicating that high c-Fos levels do not affect the differentiation potential of embryonic stem cells (Wang et al., 1991). However, c-Fos overexpression in an in vitro model of chondrogenesis inhibits the differentiation of chondrocytes (Thomas et al., 2000) and chimeric mice obtained from c-Fos overexpressing embryonic stem cells develop chondrogenic tumours (Wang et al., 1991). Ectopic c-Fos expression from a ubiquitous promoter in transgenic mice has no noticeable effects on cell differentiation, but results in the transformation of osteoblasts leading to osteosarcomas (Grigoriadis et al., 1993).

FosB

Mice lacking FosB develop normally (Brown et al., 1996; Gruda et al., 1996). However, in one study adult mutant females display a profound nurturing defect which correlates with the absence of FosB expression in a hypothalamic region critical for nurturing behaviour (Brown et al., 1996). Broad overexpression of FosB in transgenic mice does not cause an overt phenotype (Grigoriadis et al., 1993), whereas overexpression of the second FosB protein, FosB, interferes with normal cell differentiation. FosB is an alternatively spliced form of FosB which lacks transactivation activity but binds Jun proteins and DNA with similar efficiency as FosB (Mumberg et al., 1991; Nakabeppu and Nathans, 1991; Yen et al., 1991). Etopic expression of FosB promotes differentiation in osteoblasts and overexpression of FosB in thymocytes disrupts normal T cell differentiation (Carrozza et al., 1997; Sabatakos et al., 2000).

Fra-1

Inactivation of Fra-1 results in embryonic lethality around day 10 of development due to defects in the placenta and the yolk sac (Schreiber et al., 2000). The labyrinth layer of mutant placentas is reduced in size and largely avascular suggesting that the invasion of allantoic vessels into the chorionic plate is impaired in the absence of Fra-1. The development of mutant foetuses can be rescued up to birth by providing wild-type extra-embryonic tissues using tetraploid blastocyst injection. These rescued Fra-1 deficient pups display no overt morphological abnormalities suggesting that Fra-1 is dispensable for the differentiation along most, if not all, cell lineages in the foetus (Schreiber et al., 2000). Ectopic Fra-1 expression in transgenic mice accelerates osteoblast differentiation in vitro and in vivo and is able to fully rescue the lethality of Fra-1 deficient mice (Schreiber et al., 2000; Jochum et al., 2000).

Fra-2

The phenotype of mice lacking Fra-2 has not yet been reported. However, based on the broad expression pattern during late embryonic development one might expect defects in numerous tissues (Carrasco and Bravo, 1995; Foletta et al., 1994). Transgenic mice broadly overexpressing Fra-2 display occular malformations due to disrupted development of anterior eye structures (McHenry et al., 1998).

c-Jun

Mice lacking c-Jun die between day 12.5 and 13.5 of embryonic development (Hilberg et al., 1993; Johnson et al., 1993). At E12.5, all c-Jun deficient foetuses show defects of the interventricular septum in the heart and incomplete separation of the aorta and the pulmonary artery, which results in persistent truncus arteriosus indicating that c-Jun is essential for the development of a normal cardiac outflow tract (Eferl et al., 1999). Mutant foetuses also show abnormalities in the liver, which include areas of haemorrhaging and generalized oedema as well as increased numbers of apoptotic hepatoblasts and haematopoietic cells (Eferl et al., 1999; Hilberg et al., 1993). However, the abnormalities are not intrinsic to the haematopoetic compartment since lethally irradiated mice can be reconstituted with c-Jun deficient foetal liver cells (Eferl et al., 1999). Furthermore, c-Jun mutant embryonic stem cells efficiently contribute to the development of lymphatic tissues in recombination-activating gene-2 (RAG-2)-deficient mice (Chen et al., 1994). Although these chimeric mice exhibited altered thymic structures and changes in the proportions of different thymocyte populations, these data indicate that c-Jun is dispensable for the differentiation of both mature T and B lymphocytes. In contrast, c-Jun deficient foetal hepatoblasts show increased apoptosis and impaired proliferation in vitro (Eferl et al., 1999). A function for c-Jun in hepatic development is also supported by the finding that c-Jun mutant embryonic stem cells efficiently contribute to the development of all tissues in chimeric mice except for the liver (Hilberg et al., 1993). In contrast to inactivation, broad overexpression of c-Jun in transgenic mice does not result in an overt phenotype (Grigoriadis et al., 1993).

Using the Cre-loxP recombination system mice carrying conditional alleles of c-jun have recently been generated and used to study c-Jun functions in various tissues and developmental processes. Mice having c-Jun inactivated in hepatocytes during postnatal life are viable and show no overt biochemical or histological liver abnormalities, but display impaired liver regeneration in response to partial hepatectomy (Behrens et al., submitted). Inactivation of c-Jun in developing chondrocytes using a Coll2a-Cre transgene results in a severe scoliosis phenotype which is due to impaired formation of intervertebral discs and vertebral arches suggesting that c-Jun also regulates sklerotome differentiation (Behrens et al., in preparation).

The transactivation activity of c-Jun is augmented by amino (N)-terminal phosphorylation at serines 63 and 73 by the Jun N-terminal kinases (Binetruy et al., 1991; Smeal et al., 1991). Knock-in mice carrying c-Jun alleles with serines 63 and 73 mutated to alanines (JunAA mice) are viable and develop normally indicating that N-terminal phosphorylation of c-Jun is not required for embryonic development and organogenesis although neuronal apoptosis is impaired in these mice (Behrens et al., 1999). Surprisingly, T-cell proliferation and differentiation was found to be independent of c-Jun N-terminal phosphorylation by JNK, however, efficient T-cell receptor-induced thymocyte apoptosis was affected. (Behrens et al., 2001). We could further show that JNK signalling differentially uses c-Jun and NF-AT as molecular effectors in thymocyte apoptosis and T-cell differentiation (Behrens et al., 2001 and Mácian et al., this issue).

JunD

Mice lacking JunD are viable and only mutant males show impaired growth, hormone imbalance and age-dependent defects in reproduction due to impaired spermatogenesis (Thepot et al., 2000).

JunB

JunB deficient embryos die between day 8.5 and 10.0 of embryonic development due to vascular defects in the extra-embryonic tissues (Schorpp-Kistner et al., 1999). Mutant placentas lack a vascularized labyrinth layer, probably resulting from a failure of vessels of the chorio-allantoic plate to grow into the labyrinth trophoblast. In addition, yolk sac vascularization is impaired and the expression of proliferin, matrix metalloproteinase-9 and urokinase plaminogen activator are deregulated in the trophoblast. Defects in placentation are the cause of the embryonic lethality, since the life span of JunB mutant embryos can be extended up to E12.5 by providing wild-type extra-embryonic tissues using tetraploid blastocyst injection (Schorpp-Kistner et al., 1999).

Mutant foetuses can also be rescued by crossing with transgenic mice broadly expressing JunB in multiple embryonic and adult tissues (Schorpp et al., 1996; Schorpp-Kistner et al., 1999). Like c-Jun, JunB also appears to be dispensable for haematopoietic differentiation, since lethally irradiated mice can be fully reconstituted with JunB deficient foetal liver cells (Passegué et al., 2001). Furthermore, mice specifically lacking JunB expression in the myeloid lineage develop a myeloid leukaemia with increased numbers of segmented neutrophils (Passegué et al., 2001). Whereas broad overexpression of JunB did not result in an overt phenotype (Schorpp et al., 1996), targeted overexpression of JunB in T lymphocytes of transgenic mice interferes with T helper cell differentiation (Li et al., 1999). Transgenic mice ectopically expressing JunB in Th1 cells show increased expression of several Th2 cytokines, including interleukin-4, which positively regulates Th2 differentiation. Together with the observation that JunB is the only Jun protein that is selectively upregulated in developing Th2 helper cells (Li et al., 1999), these results suggest that JunB induction may contribute to the differentiation of naïve T-helper cells into functional subsets of T-lymphocytes.

Lessons learned from knock-out studies

The different phenotypes resulting from the inactivation of single AP-1 genes are quite remarkable. Only the absence of JunB and Fra-1 gives rise to a similar phenotype, thereby suggesting that these two proteins may have a common function during the placentation process. The observation that the lethality of Fra-1 deficient foetuses can be rescued by a JunB transgene, although with a low efficiency, also supports this notion (Schreiber et al., 2000). In contrast, inactivation of the other AP-1 members leads to distinct phenotypes, thereby allowing the identification of specific and novel functions. A major interest of these in vivo analyses is to prove or to disprove concepts derived from in vitro studies. Whereas numerous in vitro data have implicated AP-1 as a critical transcription factor in the immune system, none of the c-fos or c-jun knock-out mice show altered B or T cell production and/or function and only few transgenic mice, e.g. FosB and JunB, display minor defects in T-cell development.

An additional surprising observation is the lack of an overt phenotype in mice overexpressing c-Jun or JunB, whereas these two genes, together with Fra-1, are the only AP-1 members essential for embryonic development. Overexpression of Fra-1 and of the other AP-1 components, which do not play a critical function during development, e.g. c-Fos, FosB, Fra-2, results in a strong phenotype in transgenic mice. Therefore it is possible that during vertebrate evolution some of the AP-1 components may have acquired or lost gene dosage regulation.

Although no compensatory upregulation of other family members has been observed in any of the AP-1 knock-out mice, gene replacement strategies have been initiated to investigate the functional equivalence of different AP-1 proteins. Knock-in mice having c-jun replaced by junB are obtained in Mendelian ratio, develop up to birth, but die a few hours after birth due to a malformed cardiac outflow tract (Passegué et al., in preparation). Interestingly, these knock-in mice display normal livers indicating that JunB can complement for c-Jun in hepatic development but not in cardiac development. Similarly, Fra-1 can compensate for the absence of c-Fos in vivo since knock-in mice having fra-1 expressed from the c-fos locus do not develop the bone disorder osteopetrosis and display normal light-induced apoptosis of retinal photoreceptors (Fleischmann et al., 2000). These studies therefore imply that not only c-Jun and JunB, but also c-Fos and Fra-1, have maintained some functional equivalence during vertebrate evolution.

AP-1 in bone cell biology

The growth and the maintenance of the skeleton depends on the coordinated function of osteoblasts and osteoclasts, the two principal cell types of bone tissue (Karsenty, 1999). Osteoblasts, which derive from mesenchymal progenitors, produce the extracellular matrix of the bone which later undergoes mineralization. In contrast, osteoclasts belong to the monocyte/macrophage lineage and reduce bone mass by resorbing the mineralized extracellular matrix. Recent studies have shown that AP-1 components, mainly members of the Fos family, have important functions in both of these cell types.

Osteoblasts

In osteoblasts, AP-1 activity can be induced by transforming growth factor-, parathyroid hormone and 1,25-dihydroxy vitamin D, which are potent regulators of osteoblast differentiation and proliferation (Karsenty, 1999). The various components of the AP-1 complex are differentially expressed during osteoblast maturation in vitro (McCabe et al., 1996). During osteoblast proliferation, the levels of all Fos and Jun proteins are high. Subsequently, during the period of extracellular matrix production and mineralization, their levels decline, and Fra-2 and JunD become the principal components of the AP-1 complex in fully differentiated osteoblasts. The pattern of c-Fos expression during development suggest a critical role in endochondral ossification (Sandberg et al., 1988), however, the analysis of c-Fos deficient mice indicates that c-Fos is dispensable for the differentiation of osteoblasts (Johnson et al., 1992; Wang et al., 1992). On the other hand, osteoblasts are susceptible to oncogenic transformation by c-Fos in vivo (Grigoriadis et al., 1993). Transgenic mice overexpressing Fra-1 in osteoblasts, as well as in other cell types, show increased bone formation and develop osteosclerosis of the entire skeleton (Jochum et al., 2000). This phenotype is due to a cell autonomous increase in the number of mature osteoblasts indicating that Fra-1 enhances osteoblast differentiation. A similar osteosclerosis phenotype is observed in transgenic mice expressing FosB in osteoblasts (Sabatakos et al., 2000). These mice also show reduced adipogenesis, a phenotype that is not observed in Fra-1 transgenic mice. Although Fra-1 has osteogenic activity when ectopically expressed, Fra-1 deficient mice obtained by tetraploid rescue experiments display no skeletal abnormalities up to birth and harbour mature osteoblasts (Schreiber et al., 2000). However, it cannot be excluded that Fra-1 has a critical role in osteoblast homeostasis during postnatal life.

The role of the Jun proteins in osteoblast differentiation is not clear. However, neither JunD deficient mice nor mice overexpressing c-Jun or JunB in bone show an overt skeletal phenotype (Grigoriadis et al., 1993; Schorpp et al., 1996; Thepot et al., 2000).

Osteoclasts

Osteoclast differentiation critically depends on c-Fos expression in progenitor cells (Grigoriadis et al., 1994). c-Fos deficient mice have no osteoclasts, but harbour increased numbers of bone marrow macrophages. The absence of osteoclasts results in the metabolic bone disorder osteopetrosis which is characterized by increased bone mass due to reduced bone resorption (Grigoriadis et al., 1994; Johnson et al., 1992; Wang et al., 1992). Recent data using retroviral gene transfer into c-Fos mutant osteoclast precursors in vitro have shown that all Fos proteins, most efficiently Fra-1, can complement for the absence of c-Fos (Matsuo et al., 2000). Further structure-function analysis also demonstrated that the major C-terminal transactivation domains of c-Fos and FosB are dispensable for the rescue of osteoclast formation. Interestingly, Fra-1 which lacks a transactivation domain, has the highest rescue activity. Moreover, the osteoclast differentiation factor RANKL (also known as ODF, OPGL, TRANCE) induces Fra-1 transcription in a c-Fos dependent manner, thereby establishing a link between RANK signalling and the expression of AP-1 proteins during osteoclast differentiation (Matsuo et al., 2000).

In vivo the osteopetrotic phenotype can be partly cured in c-Fos mutant mice by expressing a Fra-1 transgene (Matsuo et al., 2000). Total restoration of osteoclast differentiation was achieved in knock-in mice generated by the insertion of the fra-1 gene into the c-fos locus (Fleischmann et al., 2000). In addition to its effects on osteoblasts, ectopic Fra-1 expression also enhances the differentiation of osteoclasts, both in progenitor cell lines and primary osteoclast progenitors (Matsuo et al., 2000; Owens et al., 1999), although this osteoclastogenic effect is not apparent in Fra-1 transgenic mice (Jochum et al., 2000). However, osteoclast differentiation appears not to require the presence of Fra-1 since the skeleton of rescued Fra-1 deficient foetuses contains presumably functional osteoclasts at birth (Schreiber et al., 2000).

The role of Jun proteins in osteoclast differentiation and function is largely unknown. JunD deficient mice have no skeletal abnormalities (Thepot et al., 2000) and lethally irradiated mice reconstituted with foetal liver cells lacking c-Jun or JunB do not develop osteopetrosis suggesting that Jun proteins are dispensable for osteoclast differentiation (Eferl et al., 1999; Passegué et al., 2001). Whether haematopoietic progenitor cells lacking individual Jun proteins can efficiently form osteoclasts in vitro must be addressed in future experiments.

AP-1 in cell proliferation

A role for AP-1 in the control of cell proliferation has been suggested based on observations that AP-1 activity is induced upon mitogenic stimulation and that the various Fos and Jun proteins have distinct expression patterns during cell cycle progression which can be inhibited by intracellular injection of neutralizing antibodies (Angel and Karin, 1991; Kovary and Bravo, 1991; Lallemand et al., 1997). Recent analysis using cells or mice either lacking or overexpressing single or multiple AP-1 components has demonstrated that AP-1 proteins have opposite functions in the regulation of cell proliferation and has identified some AP-1 target genes which directly link AP-1 expression to the cell cycle machinery (Figure 1, Mechta-Grigoriou et al.; Shaulian and Karin, this issue).

Although rapidly induced in response to growth factors, c-Fos, FosB and Fra-1 appear to be dispensable for cell cycle progression, since fibroblasts and embryonic stem cells lacking these components have no proliferation defect (Brüsselbach et al., 1995; Field et al., 1992; Gruda et al., 1996; Hu et al., 1994; Schreiber et al., 2000). However, c-fos/fosB double mutant fibroblasts proliferate slower, at least partly due to ineffective cyclin D1 induction upon serum stimulation suggesting that c-Fos and FosB may complement for each other in the single mutant cells (Brown et al., 1998).

c-Jun is a positive regulator of cell proliferation. c-Jun deficient fibroblasts have a marked proliferation defect due to defective cell cycle progression and undergo premature senescence in vitro (Johnson et al., 1993; Schreiber et al., 1999; Wisdom et al., 1999). The absence of c-Jun leads to elevated expression of the tumour suppressor p53 and its target gene p21, an inhibitor of several cyclin-dependent kinases, thereby disturbing S phase entry (Schreiber et al., 1999). c-Jun also regulates the expression of cyclin D1, a positive regulator of the cell cycle, and reduced cyclin D1 expression in mutant primary fibroblasts also accounts for impaired proliferation (Bakiri et al., 2000; Wisdom et al., 1999). In addition, immortalized fibroblasts lacking c-Jun undergo a prolonged UV-induced growth arrest suggesting that c-Jun is also necessary for cell cycle re-entry after DNA damage (Shaulian et al., 2000). Interestingly, fibroblasts carrying c-jun alleles with serines 63 and 73 mutated to alanines show a proliferation defect but do not enter premature senescence (Behrens et al., 1999). Foetal hepatoblasts derived from c-Jun deficient foetuses also display reduced proliferation in vitro (Eferl et al., 1999), and liver regeneration after partial hepatectomy is impaired in mice lacking c-Jun specifically in adult hepatocytes (Behrens et al., submitted). Unlike other mutant cell types, c-Jun deficient embryonic stem cells grow normally in culture and contribute to most tissues in chimeric mice (Hilberg et al., 1993; Hilberg and Wagner, 1992) indicating that the functions of c-Jun in mediating cell proliferation are cell type specific.

In contrast to c-Jun, both JunB and JunD negatively regulate cell proliferation. Fibroblasts derived from JunB overexpressing mice display reduced proliferation due to an extended G1 phase and undergo premature senescence (Passegué and Wagner, 2000). These effects result from elevated levels of the cyclin-dependent kinase inhibitor p16^INK4a, which inhibits cyclinD1/CDK4-6 kinase activity, reduces pRb phosphorylation and delays G1 to S phase progression. In addition, JunB appears to antagonize the c-Jun-mediated expression of cyclin D1 leading to reduced cyclin D1 levels in fibroblasts overexpressing JunB (Bakiri et al., 2000). Additional evidence for an inhibitory effect of JunB on proliferation comes from in vivo observations. Absence of JunB in the myeloid lineage of transgenic mice leads to increased proliferation of myeloid progenitor cells and a leukaemia-like disease characterized by the increased production of neutrophilic granulocytes (Passegué et al., 2001). Similarly to JunB, constitutive overexpression of JunD in immortalized fibroblasts inhibits cell proliferation by increasing the number of resting cells (Pfarr et al., 1994). Consistently, JunD deficient immortalized fibroblasts show increased proliferation (Weitzman et al., 2000). In contrast, JunD mutant primary fibroblasts display reduced proliferation and premature senescence, which depends on p53 and increased p19^ARF expression (Weitzman et al., 2000). These results indicate that the functions of JunD in proliferation control are dependent on the cellular context and suggest that JunD protects cells from senescence by acting as a modulator of the signalling pathways that link Ras to p53.

In addition to their direct effects, JunB and c-Jun also indirectly regulate cell proliferation. In skin fibroblasts JunB and c-Jun antagonistically control the expression of KGF and GM-CSF, which both act as keratinocyte growth factors by stimulating proliferation and differentiation of epidermal cells (Szabowski et al., 2000, Angel et al., this issue).

AP-1 in apoptosis

Programmed cell death is a physiological mechanism to eliminate cells during development, tumorigenesis and the immune response. Moreover, cells undergo apoptosis in response to growth factor and cytokine withdrawal, stress stimuli and DNA damage. A function for AP-1 in apoptosis has initially been proposed based on the observation that c-fos and c-jun mRNAs are rapidly and transiently induced in cytokine-dependent lymphoid cells upon growth factor withdrawal (Colotta et al., 1992). Furthermore, treatment with antisense oligonucleotides directed against c-fos and c-jun transcripts results in increased survival of growth factor-deprived lymphoid cells, which indicates that AP-1 may have a pro-apoptotic function. However, recent studies have revealed that the same AP-1 component may have pro- and anti-apoptotic functions depending on the cell type and the apoptotic stimulus (see also Shaulian and Karin and Herdegen and Waetzig, this issue).

Overexpression of c-Fos induces cell death in an embryonic Syrian hamster cell line, which is resistant to apoptosis under low-serum conditions (Preston et al., 1996). A similar effect has been observed in a human colorectal carcinoma cell line, in which c-Fos induced apoptosis requires the expression of functional p53 protein. Ectopic c-Fos expression also triggers cell death in immature B lymphocytes and in a myeloid leukemia cell line (Hu et al., 1996; Liebermann et al., 1998). Further evidence for c-Fos being a positive regulator of apoptosis comes from in vivo observations. In fos-lacZ reporter mice, constitutive c-Fos expression is observed in cell populations undergoing terminal differentiation and at sites of naturally occurring cell death (Smeyne et al., 1992a, 1993). c-Fos expression is also induced during the involution of the mammary gland after weaning and castration-induced regression of the prostate, two processes requiring programmed cell death (Buttyan et al., 1988; Marti et al., 1994). Moreover, c-Fos is stimulated in the nervous system as part of the early response to stress stimuli leading to neuronal cell death including ischaemia and kainate-induced excitotoxicity (Pennypacker, 1998; Smeyne et al., 1992b). These results suggest that c-Fos mediates pro-apoptotic signals not only during development, but also during tissue remodelling and stress responses. However, apoptosis occurs normally in c-Fos deficient mice during embryonic development (Roffler-Tarlov et al., 1996) and cell death is efficiently induced by etoposide in c-Fos mutant thymocytes and spleen cells (Gajate et al., 1996). In contrast, light-induced apoptosis of retinal photoreceptors and castration-induced regression of prostatic epithelium are dependent on the presence of c-Fos suggesting that c-Fos is essential for programmed cell death in these cell types (Feng et al., 1998; Hafezi et al., 1997). However, c-Fos can also protect cells against apoptotic cell death. c-Fos deficient fibroblasts are more susceptible to the cytotoxic effects of ultraviolet irradiation, as indicated by increased cell death and prolonged recovery time from UV-induced cell cycle arrest (Schreiber et al., 1995). These results are compatible with the observation that CD4/CD8 double positive thymocytes lacking c-Fos are hypersensitive to various apoptotic stimuli including dexamethasone and forskolin (Ivanov and Nikolic-Zugic, 1997)

c-Jun also appears to be both a positive and a negative modulator of apoptosis. Evidence for a pro-apoptotic function stems from the in vitro observation that the inhibition of c-Jun activity by neutralizing antibodies or a dominant-negative c-Jun mutant protects sympathetic neurons against NGF withdrawal-induced apoptosis (Estus et al., 1994; Ham et al., 1995). c-Jun is also essential for apoptosis in response to the alkylating agent methyl methanesulphonate in fibroblasts, in which c-Jun appears to regulate the expression of the pro-apoptotic gene Fas ligand (Kolbus et al., 2000). Overexpression of c-Jun is sufficient to induce apoptosis in sympathetic neurons and fibroblasts (Bossy-Wetzel et al., 1997; Ham et al., 1995). The pro-apoptotic activity of c-Jun has recently been confirmed in vivo in the context of stress-induced apoptosis. Mice carrying a mutant c-Jun with serines 63 and 73 mutated to alanines are resistent to kainate-induced neuronal cell death of hippocampal and cortical neurons identifying c-Jun phosphorylation by JNK3 as essential for neuronal apoptosis (Behrens et al., 1999; Yang et al., 1997). However, under certain conditions c-Jun appears to have also anti-apoptotic functions. The liver of c-Jun mutant fetuses harbours increased numbers of apoptotic cells (Eferl et al., 1999) and primary embryonic fibroblasts lacking c-Jun show enhanced sensitivity to UV-induced apoptosis (Wisdom et al., 1999). These findings differ from the results obtained in immortalized c-Jun mutant fibroblasts where reduced sensitivity to UV-induced apoptosis has been observed (Shaulian et al., 2000).

Recent analysis has revealed that JunD may also participate in an anti-apoptotic pathway. JunD deficient fibroblasts display increased p53-dependent apoptosis in response to UV-irradiation and hepatocytes of JunD mutant mice are sensitized to TNF-a induced cell death (Weitzman et al., 2000). In contrast, JunB seems to be part of a pro-apoptotic pathway since inactivation of JunB in myeloid cells leads to reduced apoptosis associated with increased expression of the two anti-apoptotic genes Bcl₂ and Bcl_xl (Passegué et al., 2001). This finding suggests that JunB can function as a potential tumour suppressor gene during myelopoiesis in mice.

AP-1 in tumour development

Tumorigenesis is a multistep process which involves cell transformation, invasive growth, angiogenesis and tumour spread to distant sites (Hanahan and Weinberg, 2000). Components of the AP-1 transcription factor have been implicated in many of these processes which are depicted in Figure 2.

Oncogenic potential of the AP-1 components

c-fos is the cellular homologue of v-fos, the transforming oncogene of the murine osteosarcoma viruses FBJ-MuSV and FRB-MuSV (Curran et al., 1983). Like v-Fos, c-Fos also has oncogenic activity, which is dependent on its ability to heterodimerize with Jun proteins and to bind to DNA (Jenuwein and Müller, 1987; Miller et al., 1984; Schuermann et al., 1989). The oncogenic activity of c-Fos also requires the integrity of several structural motifs present in the N- and C-terminal transactivation domains (Funk et al., 1997; Jooss et al., 1994; Wisdom and Verma, 1993). Transformation by c-Fos further depends on a continuous period of c-Fos expression and on DNA 5-methylcytosine transferase activity (Miao and Curran, 1994, Bakin and Curran, 1999). When overexpressed in vivo, c-Fos transforms chondroblasts and osteoblasts, identifying these two cell types of the skeleton as cellular targets of c-Fos-induced tumorigenesis (Grigoriadis et al., 1993; Rüther et al., 1989; Wang et al., 1991). In c-Fos-induced osteosarcomas high levels of cyclin D1 are observed suggesting that modulated expression of cell cycle regulators may contribute to c-Fos induced transformation in vivo (Sunters et al., 1998).

FosB transforms fibroblasts in culture (Kovary et al., 1991; Schuermann et al., 1991). Like for c-Fos, structure-function analysis of FosB has identified the presence of essential motifs for oncogenic activity (Schuermann et al., 1991; Wisdom et al., 1992), but full transforming activity also requires phosphorylation within a cluster of serine residues in the C-terminal domain (Skinner et al., 1997). Despite its oncogenic potential in vitro, FosB transgenic mice do not develop tumours (Grigoriadis et al., 1993), and the transforming potential of FosB is still enigmatic (Mumberg et al., 1991; Nakabeppu and Nathans, 1991; Wisdom et al., 1992).

The transforming activity of Fra-1 and Fra-2 proteins is weak in comparison to the other Fos proteins. Fra-1 overexpression in fibroblasts does not lead to obvious morphological transformation or focus formation, although it results in anchorage-independent growth in vitro and tumour formation in nude mice (Bergers et al., 1995; Wisdom and Verma, 1993). Fra-2 has transforming activity in chicken embryonic fibroblasts, but not in rat fibroblasts (Foletta et al., 1994; Nishina et al., 1990). These in vitro data are consistent with the absence of tumour development in transgenic mice overexpressing Fra-2 in a broad range of organs (McHenry et al., 1998), although some lung tumours have been observed in the osteosclerotic Fra-1 transgenic mice (Jochum et al., 2000).

c-Jun is the cellular homologue of v-jun, the transforming oncogene of the avian sarcoma virus 17 (Maki et al., 1987, see also Vogt, this issue). Transgenic mice expressing oncogenic v-Jun develop no spontaneous tumours, but form fibrosarcomas at sites of wounding (Schuh et al., 1990). c-Jun can lead to oncogenic transformation in mammalian cells, but only when coexpressed with an activated oncogene such as Ras or Src (Schütte et al., 1989a). Transformation by c-Jun also requires an intact transactivation domain (Alani et al., 1991) and N-terminal phosphorylation at serines 63 and 73 augments c-Jun-mediated transformation (Behrens et al., 2000; Binetruy et al., 1991; Smeal et al., 1991). Despite its oncogenic potential in vitro, c-Jun overexpression in transgenic mice does not result in the development of tumours (Grigoriadis et al., 1993). However, c-Jun cooperates with c-Fos in the formation of skeletal osteosarcomas (Wang et al., 1995) and efficient c-Fos-induced transformation of osteoblasts depends on c-Jun phosphorylation (Behrens et al., 2000). The genetic program of c-Jun induced transformation has recently been studied in more detail in chick embryo fibroblasts using c-Jun mutants that can efficiently dimerize only with a restricted set of partner proteins (van Dam and Castellazzi, this issue). A mutant selective for Fos family members induced anchorage-independent growth, but no growth factor independence. In contrast, a mutant with preference for ATF2-like proteins caused growth factor-independence, but no growth in soft agar. Coexpression of both c-Jun mutants re-established the normal transformation programme as induced by wild-type c-Jun indicating that c-Jun-dependent transformation is a combination of at least two distinct processes, which most likely correspond to the activation of different sets of target genes.

Unlike c-Jun, both JunB and JunD lack transforming activity (Castellazzi et al., 1991; Schorpp et al., 1996), although oncogenic mutants of JunD have been reported (Kameda et al., 1993).

AP-1 as mediator of oncogenic transformation

A function for AP-1 in mediating transforming signals has initially been suggested based on the observation that AP-1 activity is induced by various oncogenic signals, including chemicals, such as TPA, growth factors and oncogenic viruses (Angel and Karin, 1991). Furthermore, transformation by activated Ras (Mechta et al., 1997), Raf (Cook et al., 1999) or Mek1 (Treinies et al., 1999) induces distinct patterns of AP-1 protein expression which are typically characterized by the induction of Fra-1, Fra-2, c-Jun and JunB, while no changes in the levels of c-Fos and FosB were observed.

Although induced by many different oncogenes, c-Fos appears to be dispensable for cell transformation. Oncogenic transformation by v-src, v-raf and Ha-ras, among others, is comparable between wild-type and c-Fos mutant fibroblasts (Hu et al., 1994). Furthermore, v-abl induced B cell lymphomas and chemically induced skin tumours develop with similar efficiency in wild-type and c-fos knock-out mice (Saez et al., 1995a,b).

In fibroblasts, Fra-1 is the predominant component of the AP-1 complex induced by activated Ras (Mechta et al., 1997). Similarly, transformation of thyroid cells by v-mos stimulates Fra-1 expression, which is essential for the development of the malignant phenotype (Vallone et al., 1997). Consistent with a function in thyroid tumorigenesis, Fra-1 is strongly expressed in human thyroid tumours (Battista et al., 1998).

c-Jun is required for cellular transformation by oncogenic Ras in vitro since transformation is suppressed in fibroblasts lacking c-Jun or expressing a truncated form of c-Jun which acts as a dominant-negative protein (Brown et al., 1993; Johnson et al., 1996; Lloyd et al., 1991). In addition, transformation by activated Ras partially depends on c-Jun phosphorylation at serines 63 and 73, since fibroblasts harbouring c-Jun alleles with serines 63 and 73 changed to alanines can be efficiently transformed by v-Ras, but show reduced tumorigenicity in nude mice (Behrens et al., 2000). Skin tumour development caused by constitutive activation of the Ras pathway by expressing an activated form of SOS-K5-SOS-F-is impaired in junAA/AA mice further supporting the in vivo significance of c-Jun N-terminal phosphorylation in Ras induced tumorigenesis (Behrens et al., 2000). Consistently, reducing AP-1 activity by targeted expression of a dominant-negative c-Jun (TAM67) in the basal keratinocytes of transgenic mice inhibits the development of chemically induced papillomas (Young et al., 1999).

AP-1 in tumour suppression

In vitro experiments using immortalized fibroblasts have revealed that ectopic JunB or JunD expression inhibits oncogenic transformation by activated Ras indicating that these Jun proteins may act as anti-oncogenes (Passegué and Wagner, 2000; Pfarr et al., 1994; Schütte et al., 1989b). For JunB this notion is supported by the observation that mice lacking JunB in the myeloid lineage develop a transplantable disease resembling human chronic myeloid leukaemia (Passegué et al., 2001). The myeloproliferative phenotype is due to a cell autonomous increase in the numbers of granulocyte progenitors, which display enhanced GM-CSF-mediated proliferation and extended survival. At the molecular level, JunB deficient myeloid progenitors display increased levels of the GM-CSF receptor and of the anti-apoptotic proteins Bcl₂ and Bcl_xl, whereas the expression of the cell cycle inhibitor p16 is decreased. These results suggest that JunB inhibits proliferation and promotes apoptosis in myeloid progenitors thereby suppressing leukaemogenesis.

AP-1 in invasive growth and angiogenesis

Ectopic expression of c-Fos in mammary epithelial cells causes loss of epithelial polarity, epithelial-fibroblastoid cell conversion and invasive growth in collagen gels suggesting that c-Fos is able to modulate the potential of tumour cells for invasive growth (Reichmann et al., 1992). In contrast, c-Jun overexpression in the same mammary epithelial cells induces only loss of epithelial polarity, but is not sufficient to cause invasive growth (Fialka et al., 1996), although expression of the dominant negative mutant of c-Jun, TAM67, inhibits the invasiveness of a variety of cell types including murine and human squamous cell carcinomas and rat fibroblasts (Lamb et al., 1997; Malliri et al., 1998). The notion that c-Fos promotes invasive growth is further supported by the observation that the progression of chemically induced papillomas to invasive squamous cell carcinomas is impaired in c-Fos deficient mice (Saez et al., 1995b). Absence of c-Fos is associated with reduced expression of the matrix metalloproteinases (MMP) stromelysin (MMP-3) and collagenase-1 (MMP-1) (Saez et al., 1995b) suggesting that c-Fos participates in the regulation of MMPs which promote the invasive growth of cancer cells by degrading the adjacent extracellular matrix (Werb, 1997). A regulatory role for c-Fos is further supported by the observations that the induction of stromelysin-1 and collagenase-1 by PDGF and EGF is completely abrogated in c-Fos deficient fibroblasts (Hu et al., 1994) and that the expression of collagenase-1 is increased in normal tissues and in bone tumours of c-Fos transgenic mice (Gack et al., 1994). Transformation by v-Fos induces invasive growth without stimulating cell cycle progression (Hennigan et al., 1994). Using v-Fos transformed rat fibroblasts, various groups have identified additional potential Fos target genes involved in invasive growth including CD44 (Lamb et al., 1997), cathepsin L (Hennigan et al., 1994), Mts-1 (Hennigan et al., 1994), Krp1 (Spence et al., 2000), TSC-36/Frp (Johnston et al., 2000) and the microfilament-associated proteins ezrin and tropomyosin-2 and -5B (Jooss and Müller, 1995). Using c-Fos deficient fibroblasts, VEGF-D has also been identified as a c-Fos target gene (Orlandini et al., 1996). In addition to its effects on proliferation and cell morphology, VEGF-D induces angiogenesis thereby possibly establishing a link between AP-1 and angiogenesis (Marconcini et al., 1999).

Fra-1 also appears to promote invasive growth of tumour cells. Mammary adenocarcinoma cell lines overexpressing Fra-1 display increased motility and enhanced invasiveness in vitro (Kustikova et al., 1998). Like c-Fos, Fra-1 induces the expression of extracellular matrix proteinases, but preferentially those of the urokinase plasminogen activator system (Kustikova et al., 1998).

Conclusions and outlook

The analysis of AP-1 components in transgenic mice and cells has identified specific functions in development and/or in the adult organism. Initial studies have largely focused on c-Fos and c-Jun, but have recently been extended to other family members. Since the absence of c-Jun, JunB and Fra-1 results in embryonic lethality, the analysis of these proteins during later stages of development and in postnatal life is still missing. A detailed investigation of their functions will depend on the conditional inactivation of AP-1 genes in a cell type-specific and/or inducible manner using the established Cre/loxP and tet-regulated systems. Finally, we are still awaiting the analysis of gene inactivation of Fra-2, which is the only Fos/Jun component not yet studied by loss-of-function approaches. Similarly, the consequences of JunD overexpression have not been analysed in transgenic mice.

It is remarkable that some of the phenotypes have only become apparent in vitro when cells are exposed to culture or stress conditions and are no longer subject to the homeostasis in the organism. For instance, the effects of c-Jun and JunB on cell proliferation are specifically observed in cultured fibroblasts, whereas for JunB no apparent effect on fibroblast proliferation is seen in transgenic mice. A worthwhile study would be to analyse the molecular mechanism accounting for such in vitro vs in vivo discrepancies. It is equally surprising that deregulated AP-1 through overexpression of individual components can fully rescue the complex developmental phenotypes in knock-out embryos, implying that precise regulation of AP-1 is not essential.

Recent results have also demonstrated that a given AP-1 component may act both as a transcriptional activator or repressor depending on the target gene and the cellular context thereby introducing an additional level of complexity. Further biochemical studies will therefore be necessary to understand the specific consequences of AP-1 binding to a given promoter in a given cell type. Considerable progress has been made in the identification of AP-1 target genes, mainly with regard to proliferation control in fibroblasts and to invasive growth. However, for most other processes the precise composition of the AP-1 complexes and the critical target genes remain to be defined. It is well possible that by applying 'digital biology' using microarrays from different cell types and under defined conditions, we will soon be confronted with a wealth of data from which to extract significant and relevant conclusions (see also Vogt, this issue). This will be a formidable task, where the art of bioinformatics is called upon.

Finally, the role of the individual AP-1 proteins in the process of oncogenesis, and in particular their involvement in the pathogenesis of human tumours, is not clear. None of the AP-1 proteins is apparently involved in cancer-specific translocations, nor are they amplified or unusually overexpressed in any particular tumour. However, the fact that AP-1 is positioned as a signal responsive transcription factor complex at the end of a large number of signalling cascades, makes it very likely that AP-1 components provide the missing link between growth factor signalling and the cell cycle machinery. AP-1 proteins are certainly important participants and possibly determinant factors in the diverse mechanisms that contribute to the development of human cancer, although a causal proof for these functions is yet to be established.

Acknowledgements

The authors thank A Fleischmann, A Grigoriadis, K Matsuo, M Sibilia and M Schorpp-Kistner for critical comments on the manuscript. The authors acknowledge the financial support from the EC TMR network.

Alani R, Brown P, Binetruy B, Dosaka H, Rosenberg RK, Angel P, Karin M, Birrer MJ. (1991). Mol. Cell. Biol. 11, 6286-6295. MEDLINE

Angel P, Karin M. (1991). Biochim. Biophys. Acta. 1072, 129-157. MEDLINE

Bakin AV, Curran T. (1999). Science 283, 387-390. Article MEDLINE

Bakiri L, Lallemand D, Bossy-Wetzel E, Yaniv M. (2000). EMBO J. 19, 2056-2068. MEDLINE

Battista S, de Nigris F, Fedele M, Chiappetta G, Scala S, Vallone D, Pierantoni GM, Megar T, Santoro M, Viglietto G, Verde P, Fusco A. (1998). Oncogene 17, 377-385. MEDLINE

Behrens A, Jochum W, Sibilia M, Wagner EF. (2000). Oncogene 19, 2657-2663. MEDLINE

Behrens A, Sibilia M, Wagner EF. (1999). Nat. Genet. 21, 326-329. Article MEDLINE

Behrens A, Sabapathy K, Graef I, Cleary M, Crabtree GR, Wagner EF. (2001). PNAS in press.

Bergers G, Graninger P, Braselmann S, Wrighton C, Busslinger M. (1995). Mol. Cell. Biol. 15, 3748-3758. MEDLINE

Binetruy B, Smeal T, Karin M. (1991). Nature 351, 122-127. MEDLINE

Bossy-Wetzel E, Bakiri L, Yaniv M. (1997). EMBO J. 16, 1695-1709. MEDLINE

Brown JR, Nigh E, Lee RJ, Ye H, Thompson MA, Saudou F, Pestell RG, Greenberg ME. (1998). Mol. Cell. Biol. 18, 5609-5619. MEDLINE

Brown JR, Ye H, Bronson RT, Dikkes P, Greenberg ME. (1996). Cell. 86, 297-309. MEDLINE

Brown PH, Alani R, Preis LH, Szabo E, Birrer MJ. (1993). Oncogene 8, 877-886. MEDLINE

Brüsselbach S, Möhle-Steinlein U, Wang ZQ, Schreiber M, Lucibello FC, Müller R, Wagner EF. (1995). Oncogene 10, 79-86. MEDLINE

Buttyan R, Zakeri Z, Lockshin R, Wolgemuth D. (1988). Mol. Endocrinol. 2, 650-657. MEDLINE

Carrasco D, Bravo R. (1995). Oncogene 10, 1069-1079. MEDLINE

Carrozza ML, Jacobs H, Acton D, Verma I, Berns A. (1997). Oncogene 14, 1083-1091. MEDLINE

Castellazzi M, Spyrou G, La Vista N, Dangy JP, Piu F, Yaniv M, Brun G. (1991). Proc. Natl. Acad. Sci. USA 88, 8890-8894. MEDLINE

Chen J, Stewart V, Spyrou G, Hilberg F, Wagner EF, Alt FW. (1994). Immunity 1, 65-72. MEDLINE

Colotta F, Polentarutti N, Sironi M, Mantovani A. (1992). J. Biol. Chem. 267, 18278-18283. MEDLINE

Cook SJ, Aziz N, McMahon M. (1999). Mol. Cell. Biol. 19, 330-341. MEDLINE

Curran T, MacConnell WP, van Straaten F, Verma IM. (1983). Mol. Cell. Biol. 3, 914-921. MEDLINE

Eferl R, Sibilia M, Hilberg F, Fuchsbichler A, Kufferath I, Guertl B, Zenz R, Wagner EF, Zatloukal K. (1999). J. Cell. Biol. 145, 1049-1061. MEDLINE

Estus S, Zaks WJ, Freeman RS, Gruda M, Bravo R, Johnson Jr EM. (1994). J. Cell. Biol. 127, 1717-1727. MEDLINE

Feng Z, Joos HJ, Vallan C, Muhlbauer R, Altermatt HJ, Jaggi R. (1998). Oncogene 17, 2593-2600. MEDLINE

Fialka I, Schwarz H, Reichmann E, Oft M, Busslinger M, Beug H. (1996). J. Cell. Biol. 132, 1115-1132. MEDLINE

Field SJ, Johnson RS, Mortensen RM, Papaioannou VE, Spiegelman BM, Greenberg ME. (1992). Proc. Natl. Acad. Sci. USA 89, 9306-9310. MEDLINE

Fleischmann A, Hafezi F, Elliott C, Reme CE, Rüther U, Wagner EF. (2000). Genes Dev. 14, 2695-2700. MEDLINE

Foletta VC, Sonobe MH, Suzuki T, Endo T, Iba H, Cohen DR. (1994). Oncogene 9, 3305-3311. MEDLINE

Funk M, Poensgen B, Graulich W, Jerome V, Müller R. (1997). Mol. Cell. Biol. 17, 537-544. MEDLINE

Gack S, Vallon R, Schaper J, Rüther U, Angel P. (1994). J. Biol. Chem. 269, 10363-10369. MEDLINE

Gajate C, Alonso MT, Schimmang T, Mollinedo F. (1996). Biochem. Biophys. Res. Commun. 218, 267-272. Article MEDLINE

Grigoriadis AE, Schellander K, Wang ZQ, Wagner EF. (1993). J. Cell. Biol. 122, 685-701. MEDLINE

Grigoriadis AE, Wang ZQ, Cecchini MG, Hofstetter W, Felix R, Fleisch HA, Wagner EF. (1994). Science 266, 443-448. MEDLINE

Gruda MC, van Amsterdam J, Rizzo CA, Durham SK, Lira S, Bravo R. (1996). Oncogene 12, 2177-2185. MEDLINE

Hafezi F, Steinbach JP, Marti A, Munz K, Wang ZQ, Wagner EF, Aguzzi A, Reme CE. (1997). Nat. Med. 3, 346-349. MEDLINE

Ham J, Babij C, Whitfield J, Pfarr CM, Lallemand D, Yaniv M, Rubin LL. (1995). Neuron 14, 927-939. MEDLINE

Hanahan D, Weinberg RA. (2000). Cell. 100, 57-70. MEDLINE

Hennigan RF, Hawker KL, Ozanne BW. (1994). Oncogene 9, 3591-3600. MEDLINE

Hilberg F, Aguzzi A, Howells N, Wagner EF. (1993). Nature 365, 179-181. MEDLINE

Hilberg F, Wagner EF. (1992). Oncogene 7, 2371-2380. MEDLINE

Hu E, Mueller E, Oliviero S, Papaioannou VE, Johnson R, Spiegelman BM. (1994). EMBO J. 13, 3094-3103. MEDLINE

Hu L, Hatano M, Rüther U, Tokuhisa T. (1996). J. Immunol. 157, 3804-3811. MEDLINE

Ivanov VN, Nikolic-Zugic J. (1997). J. Biol. Chem. 272, 8558-8566. MEDLINE

Jain J, Nalefski EA, McCaffrey PG, Johnson RS, Spiegelman BM, Papaioannou V, Rao A. (1994). Mol. Cell. Biol. 14, 1566-1574. MEDLINE

Jenuwein T, Müller R. (1987). Cell. 48, 647-657. MEDLINE

Jochum W, David JP, Elliott C, Wutz A, Plenk Jr H, Matsuo K, Wagner EF. (2000). Nat. Med. 6, 980-984. Article MEDLINE

Johnson R, Spiegelman B, Hanahan D, Wisdom R. (1996). Mol. Cell. Biol. 16, 4504-4511. MEDLINE

Johnson RS, Spiegelman BM, Papaioannou V. (1992). Cell. 71, 577-586. MEDLINE

Johnson RS, van Lingen B, Papaioannou VE, Spiegelman BM. (1993). Genes Dev. 7, 1309-1317. MEDLINE

Johnston IM, Spence HJ, Winnie JN, McGarry L, Vass JK, Meagher L, Stapleton G, Ozanne BW. (2000). Oncogene 19, 5348-5358. MEDLINE

Jooss KU, Funk M, Müller R. (1994). EMBO J. 13, 1467-1475. MEDLINE

Jooss KU, Müller R. (1995). Oncogene 10, 603-608. MEDLINE

Kameda T, Akahori A, Sonobe MH, Suzuki T, Endo T, Iba H. (1993). Proc. Natl. Acad. Sci. USA 90, 9369-9373. MEDLINE

Karsenty G. (1999). Genes Dev. 13, 3037-3051. Article MEDLINE

Kolbus A, Herr I, Schreiber M, Debatin KM, Wagner EF, Angel P. (2000). Mol. Cell. Biol. 20, 575-582. MEDLINE

Kovary K, Bravo R. (1991). Mol. Cell. Biol. 11, 2451-2459. MEDLINE

Kovary K, Rizzo CA, Ryseck RP, Noguchi T, Raynoschek C, Pelosin JM, Bravo R. (1991). New. Biol. 3, 870-879. MEDLINE

Kustikova O, Kramerov D, Grigorian M, Berezin V, Bock E, Lukanidin E, Tulchinsky E. (1998). Mol. Cell. Biol. 18, 7095-7105. MEDLINE

Lallemand D, Spyrou G, Yaniv M, Pfarr CM. (1997). Oncogene 14, 819-830. MEDLINE

Lamb RF, Hennigan RF, Turnbull K, Katsanakis KD, MacKenzie ED, Birnie GD, Ozanne BW. (1997). Mol. Cell. Biol. 17, 963-976. MEDLINE

Li B, Tournier C, Davis RJ, Flavell RA. (1999). EMBO J. 18, 420-432. Article MEDLINE

Liebermann DA, Gregory B, Hoffman B. (1998). Int. J. Oncol. 12, 685-700. MEDLINE

Lloyd A, Yancheva N, Wasylyk B. (1991). Nature 352, 635-638. MEDLINE

Maki Y, Bos TJ, Davis C, Starbuck M, Vogt PK. (1987). Proc. Natl. Acad. Sci. USA 84, 2848-2852. MEDLINE

Malliri A, Symons M, Hennigan RF, Hurlstone AF, Lamb RF, Wheeler T, Ozanne BW. (1998). J. Cell. Biol. 143, 1087-1099. MEDLINE

Marconcini L, Marchio S, Morbidelli L, Cartocci E, Albini A, Ziche M, Bussolino F, Oliviero S. (1999). Proc. Natl. Acad. Sci. USA 96, 9671-9676. Article MEDLINE

Marti A, Jehn B, Costello E, Keon N, Ke G, Martin F, Jaggi R. (1994). Oncogene 9, 1213-1223. MEDLINE

Matsuo K, Owens JM, Tonko M, Elliott C, Chambers TJ, Wagner EF. (2000). Nat. Genet. 24, 184-187. Article MEDLINE

McCabe LR, Banerjee C, Kundu R, Harrison RJ, Dobner PR, Stein JL, Lian JB, Stein GS. (1996). Endocrinology 137, 4398-4408. MEDLINE

McHenry JZ, Leon A, Matthaei KI, Cohen DR. (1998). Oncogene 17, 1131-1140. MEDLINE

Mechta F, Lallemand D, Pfarr CM, Yaniv M. (1997). Oncogene 14, 837-847. MEDLINE

Miao GG, Curran T. (1994). Mol. Cell. Biol. 14, 4295-4310. MEDLINE

Miller AD, Curran T, Verma IM. (1984). Cell. 36, 51-60. MEDLINE

Mumberg D, Lucibello FC, Schuermann M, Müller R. (1991). Genes Dev. 5, 1212-1223. MEDLINE

Nakabeppu Y, Nathans D. (1991). Cell. 64, 751-759. MEDLINE

Nishina H, Sato H, Suzuki T, Sato M, Iba H. (1990). Proc. Natl. Acad. Sci. USA 87, 3619-3623. MEDLINE

Okada S, Wang ZQ, Grigoriadis AE, Wagner EF, von Rüden T. (1994). Mol. Cell. Biol. 14, 382-390. MEDLINE

Orlandini M, Marconcini L, Ferruzzi R, Oliviero S. (1996). Proc. Natl. Acad. Sci. USA 93, 11675-11680. Article MEDLINE

Owens JM, Matsuo K, Nicholson GC, Wagner EF, Chambers TJ. (1999). J. Cell. Physiol. 179, 170-178. Article MEDLINE

Passegué E, Wagner EF. (2000). EMBO J. 19, 2969-2979. MEDLINE

Passegué E, Jochum W, Schorpp-Kistner M, Möhle-Steinlein U, Wagner EF. (2001). Cell. 104, 21-32. MEDLINE

Pennypacker K. (1998). Int. Rev. Neurobiol. 42, 169-197. MEDLINE

Pfarr CM, Mechta F, Spyrou G, Lallemand D, Carillo S, Yaniv M. (1994). Cell. 76, 747-760. MEDLINE

Preston GA, Lyon TT, Yin Y, Lang JE, Solomon G, Annab L, Srinivasan DG, Alcorta DA, Barrett JC. (1996). Mol. Cell. Biol. 16, 211-218. MEDLINE

Reichmann E, Schwarz H, Deiner EM, Leitner I, Eilers M, Berger J, Busslinger M, Beug H. (1992). Cell. 71, 1103-1116. MEDLINE

Roffler-Tarlov S, Brown JJ, Tarlov E, Stolarov J, Chapman DL, Alexiou M, Papaioannou VE. (1996). Development 122, 1-9. MEDLINE

Rüther U, Komitowski D, Schubert FR, Wagner EF. (1989). Oncogene 4, 861-865. MEDLINE

Sabatakos G, Sims NA, Chen J, Aoki K, Kelz MB, Amling M, Bouali Y, Mukhopadhyay K, Ford K, Nestler EJ, Baron R. (2000). Nat. Med. 6, 985-990. Article MEDLINE

Saez E, Oppenheim H, Smoluk J, Andersen JW, Van Etten RA, Spiegelman BM. (1995a). Cancer Res. 55, 6196-6199.

Saez E, Rutberg SE, Mueller E, Oppenheim H, Smoluk J, Yuspa SH, Spiegelman BM. (1995b). Cell. 82, 721-732. MEDLINE

Sandberg M, Vuorio T, Hirvonen H, Alitalo K, Vuorio E. (1988). Development 102, 461-470. MEDLINE

Schorpp M, Jager R, Schellander K, Schenkel J, Wagner EF, Weiher H, Angel P. (1996). Nucleic Acids. Res. 24, 1787-1788. MEDLINE

Schorpp-Kistner M, Wang ZQ, Angel P, Wagner EF. (1999). EMBO J. 18, 934-948. Article MEDLINE

Schreiber M, Baumann B, Cotten M, Angel P, Wagner EF. (1995). EMBO J. 14, 5338-5349. MEDLINE

Schreiber M, Kolbus A, Piu F, Szabowski A, Möhle-Steinlein U, Tian J, Karin M, Angel P, Wagner EF. (1999). Genes Dev. 13, 607-619. MEDLINE

Schreiber M, Wang Z, Jochum W, Fetka I, Elliott C, Wagner EF. (2000). Development 127, 4937-4948. MEDLINE

Schuermann M, Jooss K, Müller R. (1991). Oncogene 6, 567-576. MEDLINE

Schuermann M, Neuberg M, Hunter JB, Jenuwein T, Ryseck RP, Bravo R, Müller R. (1989). Cell. 56, 507-516. MEDLINE

Schuh AC, Keating SJ, Monteclaro FS, Vogt PK, Breitman ML. (1990). Nature 346, 756-760. MEDLINE

Schütte J, Minna JD, Birrer MJ. (1989a). Proc. Natl. Acad. Sci. USA 86, 2257-2261. MEDLINE

Schütte J, Viallet J, Nau M, Segal S, Fedorko J, Minna J. (1989b). Cell. 59, 987-997. MEDLINE

Shaulian E, Schreiber M, Piu F, Beeche M, Wagner EF, Karin M. (2000). Cell. 103, 897-907. MEDLINE

Skinner M, Qu S, Moore C, Wisdom R. (1997). Mol. Cell. Biol. 17, 2372-2380. MEDLINE

Smeal T, Binetruy B, Mercola DA, Birrer M, Karin M. (1991). Nature 354, 494-496. MEDLINE

Smeyne RJ, Curran T, Morgan JI. (1992a). Brain. Res. Mol. Brain. Res. 16, 158-162.

Smeyne RJ, Schilling K, Robertson L, Luk D, Oberdick J, Curran T, Morgan JI. (1992b). Neuron 8, 13-23. MEDLINE

Smeyne RJ, Vendrell M, Hayward M, Baker SJ, Miao GG, Schilling K, Robertson LM, Curran T, Morgan JI. (1993). Nature 363, 166-169. MEDLINE

Spence HJ, Johnston I, Ewart K, Buchanan SJ, Fitzgerald U, Ozanne BW. (2000). Oncogene 19, 1266-1276. MEDLINE

Sunters A, McCluskey J, Grigoriadis AE. (1998). Dev. Genet. 22, 386-397. MEDLINE

Szabowski A, Maas-Szabowski N, Andrecht S, Kolbus A, Schorpp-Kistner M, Fusenig NE, Angel P. (2000). Cell. 103, 745-755. MEDLINE

Thepot D, Weitzman JB, Barra J, Segretain D, Stinnakre MG, Babinet C, Yaniv M. (2000). Development 127, 143-153. MEDLINE

Thomas DP, Sunters A, Gentry A, Grigoriadis AE. (2000). J. Cell. Sci. 113, 439-450. MEDLINE

Treinies I, Paterson HF, Hooper S, Wilson R, Marshall CJ. (1999). Mol. Cell. Biol. 19, 321-329. MEDLINE

Vallone D, Battista S, Pierantoni GM, Fedele M, Casalino L, Santoro M, Viglietto G, Fusco A, Verde P. (1997). EMBO J. 16, 5310-5321. MEDLINE

Wang ZQ, Grigoriadis AE, Möhle-Steinlein U, Wagner EF. (1991). EMBO J. 10, 2437-2450. MEDLINE

Wang ZQ, Liang J, Schellander K, Wagner EF, Grigoriadis AE. (1995). Cancer Res. 55, 6244-6251. MEDLINE

Wang ZQ, Ovitt C, Grigoriadis AE, Möhle-Steinlein U, Rüther U, Wagner EF. (1992). Nature 360, 741-745. MEDLINE

Weitzman JB, Fiette L, Matsuo K, Yaniv M. (2000). Mol. Cell. 6, 1109-1119. MEDLINE

Werb Z. (1997). Cell. 91, 439-442. MEDLINE

Wisdom R, Johnson RS, Moore C. (1999). EMBO J. 18, 188-197. MEDLINE

Wisdom R, Verma IM. (1993). Mol. Cell. Biol. 13, 7429-7438. MEDLINE

Wisdom R, Yen J, Rashid D, Verma IM. (1992). Genes Dev. 6, 667-675. MEDLINE

Yang DD, Kuan CY, Whitmarsh AJ, Rincon M, Zheng TS, Davis RJ, Rakic P, Flavell RA. (1997). Nature 389, 865-870. Article MEDLINE

Yen J, Wisdom RM, Tratner I, Verma IM. (1991). Proc. Natl. Acad. Sci. USA 88, 5077-5081. MEDLINE

Young MR, Li JJ, Rincon M, Flavell RA, Sathyanarayana BK, Hunziker R, Colburn N. (1999). Proc. Natl. Acad. Sci. USA 96, 9827-9832. MEDLINE

Figure 1 Effects of AP-1 proteins on cell cycle regulation. AP-1 proteins control cell cycle progression by regulating the expression of key components of the cell cycle machinery. c-Jun stimulates G1 to S phase transition by inducing cyclin D1 and repressing p53, which in turn reduces p21 levels. c-Fos and FosB have redundant functions in the stimulation of S phase entry and the induction of cyclin D1 expression. JunB inhibits G1 to S phase progression by inducing p16 and repressing cyclin D1. JunD inhibits S phase entry and increases the numbers of resting cells by modulating the Ras/p53 pathway

Figure 2 AP-1 functions in multistep tumorigenesis. AP-1 proteins participate in tumorigenesis by regulating oncogenic transformation, proliferation, apoptosis, invasive growth and angiogenesis through modulating the expression of some critical target genes

Table 1 Analysis of fos and jun knock-out mice

Table 2 Analysis of fos and jun transgenic mice