Chondrocytes as a specific target of ectopic Fos expression in early development

Abstract

The Finkel–Biskis–Jinkins murine sarcoma virus, which carries v-fos, induces osteosarcomas, whereas high-level expression of exogenous c-fos in transgenic and chimeric mice leads to postnatal development of osteogenic and chondrogenic tumors, respectively. To test whether such target cell specificity of an oncogene can be detected even in early development, we induced ectopic expression of fos in chicken limb buds by microinjecting replication-competent retrovirus into the presumptive leg field of stage 10 embryos. This caused cartilage truncation of all the long bones of the injected leg, which was mainly attributable to chondrodysplasia due to severe retardation of differentiation of the proliferating chondrocytes into mature or hypertrophic chondrocytes, as well as a slight delay in precartilagenous condensation. Expression of genes for all the other known members of chicken AP-1, which include such transforming genes as c-jun and fra-2, however, caused no macroscopic abnormalities in limb formation, indicating a specific function of Fos proteins in embryonic endochondral bone differentiation. The extent of truncation was stronger with v-Fos than with c-Fos, and comparative analysis of these proteins, as well as v-Fos mutants, revealed that strong transforming activity of Fos protein is necessary to cause dysplasia, suggesting that common molecular mechanisms are involved in both embryonic chondrodysplasia and bone tumor formation in postnatal mice.

Functional analysis of oncogenes isolated from RNA tumor viruses and their cellular counterparts (protooncogenes) in cell culture has revealed several important molecular mechanisms involved in cellular proliferation, differentiation, and transformation. For example, the c-fos protooncogene (1) was originally identified as the cellular counterpart of the viral oncogene (2) carried by Finkel–Biskis–Jinkins (FBJ) murine sarcoma virus (MSV) (3) and is now known to belong to a multigene family that includes fra-1 (4), fra-2 (5), and fosB (6). The fos gene family encodes nuclear proteins that dimerize with the Jun family proteins, such as c-Jun, JunB, and JunD, to form the transcription factor complex AP-1 (7). Dimerization occurs specifically through a leucine zipper structure; Jun family members can form low-affinity homodimers and high-affinity heterodimers with the Fos family, whereas Fos-related proteins do not form stable homodimers. Although these hetero- and homodimers bind to similar sites on DNA (TGACTCA, AP-1-binding sites) through the basic domains of both proteins, each dimer has a distinct transcriptional regulatory function (8), so that transcription can be either positively or negatively modulated.

High-level expression of most members of the fos or jun gene families has been reported to cause cellular transformation of chicken embryo fibroblasts (CEF) (5, 9–13). JunD, however, has no transforming activity at all, but it can acquire transforming potential as a result of spontaneous mutations in the viral genome (14). These results indicate that uncontrolled expression and qualitative changes of any component of AP-1 could induce cellular transformation.

From the standpoint of tumor formation, it is interesting that FBJ-MSV and Finkel–Biskis–Reilly MSV, which carry v-fos, produce primarily osteosarcomas in mice (3). c-fos is expressed at high levels in murine and human osteosarcomas (15), further suggesting a possible association of fos expression with tumor formation in bone-forming cells. Wagner and coworker have presented several studies on fos transgenic mice (16–18). One c-fos transgenic family, which expresses high levels of c-fos from the H-2K^b class major histocompatibility complex promoter, developed bone tumors with high efficiency (18). The tumors were typical osteosarcomas and developed in all bones after transgene expression, which is undetectable in the embryos but appears 2–3 weeks after birth in a variety of tissues. Because exogenous fos was not expressed in the embryos of any transgenic mice, embryonic stem cells were used as a means of overexpressing c-fos genes at a late stage of embryogenesis (19), and it was demonstrated that embryonic development is not affected in embryonic stem cell chimeras, but chondrosarcomas develop in young mice. No clear effect of c-Fos in early embryonic development has been described because of the lack of an appropriate testing system.

In this study, we have ectopically expressed fos, as well as other members of the fos/jun gene family, in one entire chicken limb bud by microinjecting replication-competent retrovirus vectors carrying various members of the fos/jun family. The limb bud, composed of undifferentiated mesenchyme surrounded by an ectodermal jacket, is considered to be a good model of morphogenesis and cellular differentiation involving such cells as myoblasts, chondroblasts, fibroblasts, and epithelial cells (20). Since the results clearly indicated that only fos among all the members of the fos/jun family of genes caused macroscopic effects on the limb and, further, that the major target cells are the proliferating chondrocytes, we next compared this fos function with its cellular transforming activity.

MATERIALS AND METHODS

Viruses.

For the production of retrovirus vectors (subgroup A), a SalI digest of pREP (containing the 5′ half of the provirus) was ligated to a SalI digest of pDS3 (which contains the 3′ half of the provirus without an oncogene) (9), pDS3-FC1 (carrying chicken c-fos), pDS3-F2C1 (carrying chicken fra-2) (5), pDS3H-JC1 (carrying chicken c-jun), pDS3H-JDC1 (carrying chicken junD), pDS3-FM4 (carrying mouse c-fos) (9, 21), pDS3-FJ2 (carrying v-fos of FBJ-MSV) (9, 21), pDS3-CD3 (carrying a v-fos mutant, CD3; Fig. 1), or pDS3-IR2-CD3 (an internal ribosomal entry site from encephalomyocarditis virus was inserted between the env gene and the CD3 coding region of pDS3-CD3) to form the replication-competent provirus structure. For the production of subgroup B virus, pREP was replaced with pREP-B (13, 22). All the recombinant viruses were obtained by the transfection of the ligated DNA (2 μg) into CEF and subsequent recovery from the culture fluids as described previously (9, 23). To obtain high-titer virus stocks of replication-defective avian retrovirus, NK24 (24), nonproducer CEF transformed by this virus were isolated and infected with RAV-1, a helper virus of subgroup A.

Figure 1.

Protein structures of Fos proteins used in this study. ∗, A single amino acid exchange from c-Fos; ▤, the leucine zipper motif; ▨, a frame-shift mutation generated by a deletion present in v-Fos of FBJ-MSV. The mutant of v-Fos (FBJ), CD3, terminates at amino acid residue 202, whereas the wild type is composed of 381 amino acid residues.

Microinjection.

Microinjection into chicken embryos was performed essentially as described previously (25, 26). Fertilized eggs (specific pathogen free, C/O) were incubated at 38.5°C until stage 10. The eggshells were opened, and a few drops of Tyrode’s solution were added to prevent drying. Virus stock solutions were concentrated 20-fold by volume with an Ultrafree C3KH (Millipore) just before injection. The concentrated virus solutions were mixed with Methyl green (final concentration of 0.05%) and injected into the presumptive right leg or wing field (50–100 nl per injection) with a microinjector IM-3 (Narishige, Tokyo). For the injection of subgroup B virus, polybrene was added to the virus solutions at the final concentration of 0.4 μg/ml (25). After the injection, the eggs were sealed, incubated at 38.0°C and isolated on the day indicated. For the measurement of long bone length, all the live embryos were fixed and doubly stained with Alcian blue and Alizarin red, and the length ratio of each of the long bones in the injected and uninfected limbs of the same embryo was determined after photographs were taken.

In Situ Hybridization and Immunohistochemical Analysis.

Whole mount in situ hybridization to the transcripts of proteoglycan H or v-fos was performed as described previously (26) using digoxygenin-labeled RNA probes transcribed in vitro (Boehringer Mannhein). The template DNAs for proteoglycan H and v-fos genes were pPG525 (27) and pRPA-Fos (13), respectively. For the immunohistochemical analysis, paraffin sections prepared by standard procedures were treated with 1:200 diluted anti-chicken type X collagen antiserum (28) (the rabbit was supplied by M. Pacifici, University of Pennsylvania Dental School, Philadelphia) overnight at 4°C, and an ABC detection kit (Vector Laboratories) was employed.

Northern Blot Analysis of Primary Chondrocytes.

The caudal sternum was dissected from 18-day embryos (29), cultured in high-glucose DMEM containing 10% defined fetal calf serum and infected with recombinant retrovirus vectors (subgroup B) in the presence of polybrene (2 μg/ml). Five days after infection, cells were disrupted, and total RNA was isolated by the acid guanidinium thiocyanate-phenol-chloroform method (30) using ISOGEN (Wako Biochemicals, Osaka), electrophoresed under denaturing conditions and transferred to Photogene membranes (GIBCO) by capillary blotting. The cDNA probes for type II collagen and type X collagen were the 0.8-kb BamHI–PstI fragment of pCs2 (encoding the C-terminal propeptide) (31) and the 0.6-kb EcoRI–XhoI fragment encoding the C-terminal noncollagenous region (32), respectively. DNA fragments were labeled with ³²P with the RTS Random Primers DNA Labeling System (GIBCO) and used as probes. Hybridized membranes were washed, and radioisotopes were visualized by autoradiography.

Western Blot Analysis.

The total extracts of CEF infected with viruses were prepared under denaturing conditions, and 15-μg protein samples were resolved by electrophoresis on 10% SDS polyacrylamide gels. Fos proteins were detected by Western blotting using anti-Fos Pep-1 antiserum (5). Bands were visualized by use of the ECL Western Blotting Detection System (Amersham).

RESULTS

Cartilagenous Truncation Observed in Limbs Exogenously Expressing Fos.

As chicken fos/jun family genes, c-fos (33), fra-2 (5) (fos family), c-jun (34), and junD (35) (jun family) have been detected and cloned so far. We exogenously introduced each of these chicken genes as well as v-fos derived from FBJ-MSV and its cellular counterpart mouse c-fos (the viruses used are summarized in Table 1). For this purpose, we constructed a series of replication-competent retroviruses (subgroup A), each carrying one of these genes, and microinjected each virus into the prospective right leg field of stage 10 chicken embryos. In all cases, the virus spread and induced exogenous gene expression in the entire right leg region within 3 days, whereas the left leg remained uninfected (Fig. 2A) until at least day 11.5 (data not shown). When fra-2, c-jun, or junD was introduced, no macroscopic morphological abnormalities were detected at least up to day 9.5 and the length ratio of every long bone (femur, fibula, tibia, and metatarsus) between the infected and uninfected limbs was very close to 1.00 (Table 1), as was also the case for the limbs infected with the empty vector, DS3, or mock-infected.

Table 1.

Length ratio of all the long bones in hindlimbs expressing exogenous proteins to those in uninfected hindlimbs

Protein	Virus	Length ratio of long bones
		Expression by subgroup A vector	Expression by another vector
	DS3	1.00  ±  0.03
Transforming proteins
c-Fos (mouse)	FM4	0.91  ±  0.08	0.82  ±  0.08*
v-Fos (FBJ-MSV)	FJ2	0.50  ±  0.15
v-Fos-CD3	FJ-CD3	0.99  ±  0.07	0.90  ±  0.04†
c-Fos (chicken)	FC1	0.95  ±  0.05
v-Fos (NK24)	NK24		0.85  ±  0.11‡
Fra-2 (chicken)	F2C1	0.98  ±  0.05	0.99  ±  0.08§
c-Jun (chicken)	JC1	0.99  ±  0.02
Nontransforming protein
JunD (mouse)	JDM1	0.99  ±  0.05

Data, such as those shown in Fig. 4, were averaged for all the long bones and are shown with the standard deviation. Stocks of all the replication-competent transforming virus of subgroup A have a titer of more than 3.0 × 10⁶ focus-forming units/ml. 

^*A subgroup B virus encoding c-fos, FM4(B) (7.0 × 10⁶ focus-forming units/ml). 

^†A subgroup A virus with internal ribosomal entry site (IR2-CD3) (3.0 × 10⁶ focus-forming units/ml). 

^‡The natural replication-defective chicken c-fos virus, NK24, with a helper virus (RAV-1) (8.0 × 10⁵ focus-forming units/ml). 

^§A subgroup B virus encoding fra-2, F2C1(B) (5.0 × 10⁶ focus-forming units/ml). 

Figure 2.

Effect of v-fos virus expression in limb buds. A concentrated virus stock of recombinant replication-competent retrovirus (subgroup A), FJ-2, which encodes v-fos (originated from FBJ-MSV), was microinjected into the presumptive right leg or wing field at stage 10. Pairs of uninfected left limb (A–F Left) and infected right limb (A–F Right) from the same embryos are shown. The distribution of exogenous v-fos RNA (A) and proteoglycan H mRNA (B) is shown in the embryos at day 4.5, as detected by whole-mount in situ hybridization. (C–F) Dorsal view of wing cartilage (D Upper) or leg cartilage (C and D Lower, and E and F) in limbs at day 6 (C), day 8.5 (D), day 9.5 (E), and day 11.5 (F) after double staining with Alcian blue and Alizarin red. [Bars = 1 mm (A, for A–C) and 3 mm (F, for D–F).]

In the embryos infected with virus carrying v-fos from FBJ-MSV (Fig. 2A), clear truncation of the entire right leg was observed at day 8.5–9.5 when compared with the left limb of the same embryo (Figs. 2D Lower and 3). In the right leg, the length of the cartilage of every long bone was reduced (Fig. 4A), i.e., the average length ratio of the four long bones in the infected limbs to those in the uninfected limbs was 0.50 ± 0.15, although no change in the skeletal pattern was observed. When this virus was microinjected into the presumptive right wing field, all the long bones (humerus, radius, ulna, and metacarpus) were truncated (Fig. 2D Upper). When the chicken c-fos and mouse c-fos genes (Fig. 1) were microinjected, truncation of the cartilage of all the long bones was also detected (Fig. 4), but the average extent of truncation was more drastic in the case of v-fos (FBJ-MSV) (Table 1).

Figure 3.

(A) Suppression of calcification of several long bones is seen in the v-fos virus (FJ-2)-infected limb (right) as compared with the uninfected limb (left) of the same 8.5-day embryo. Cartilage was stained with Alcian blue (blue) and mineralized cartilage with Alizarin red (red). (B) Immunohistochemical analysis of type X collagen in sections near the central part of the tibia in the v-fos virus (FJ-2)-infected (Right) and uninfected (Left) limbs of the same 8.5-day embryo. Sections were counterstained with hematoxylin. The lower part is closer to the center. (Bar = 100 μm.)

Figure 4.

Length ratio of each long bone in the infected right limb to that in the uninfected left limb of the same embryo (day 8.5–9.5). (A) Limbs infected with virus carrying c-fos (mouse), FM4 (•), or virus carrying v-fos derived from FBJ-MSV, FJ-2 (○), as well as mock-infected limbs (□). (B) Limbs infected with virus carrying c-fos (chicken), FC-1 (•), or NK24 (RAV-1), which carries v-fos (○).

NK24, which was originally isolated as a replication-defective virus in a chicken nephroblastoma, encodes an N-terminally truncated c-Fos fused to a part of Gag protein on its N terminus (Fig. 1) (24). When NK24 was introduced into limb buds (Fig. 4B) with a helper virus (RAV-1), it also caused truncation of cartilage, like v-fos of FBJ-MSV, to a greater extent than its cellular counterpart (Table 1). A possible explanation of the apparently reduced and unstable biological effects of NK24 compared with virus carrying v-fos from FBJ-MSV (Fig. 4B) is that the cell population in the injected limb was not fully infected with this replication-defective v-fos-carrying virus.

v-Fos Induces Accumulation of Proliferating Chondrocytes.

We next examined in detail the time course of cartilagenous development in hindlimb expressing v-Fos derived from FBJ-MSV because it exhibited the most distinct biological effect among the Fos proteins. The injected hindlimbs showed a slight but reproducible delay in the onset of precartilagenous condensation, as shown by in situ hybridization with a probe of the proteoglycan H gene (Fig. 2B), which is specifically expressed in chondrocytes (26). The truncation was detectable as early as day 6.5 by cartilage staining (Fig. 2C) and subsequently became much more distinct (Fig. 2 D–F). Long bone truncation was detected in 17 out of 17 embryos injected with virus carrying v-fos from FBJ-MSV, indicating the stability and reproducibility of this biological function of v-fos. Calcium staining with Alizarin red indicated that mineralization of the cartilage in all the long bones was completely suppressed in the injected limbs at day 9.5 (Figs. 2E and 3A). This suppression of mineralization by exogenous v-fos expression was not complete, however, because delayed and retarded mineralization was detectable in tibia and metatarsus in the infected leg on day 11.5 (Fig. 2F).

The femurs of these infected limbs at day 9.5 were rich in very small chondrocytes with the morphology of proliferating chondrocytes and did not contain either mature or hypertrophic chondrocytes, which were readily detectable in the uninfected femur. The tibia or metatarsus of these infected limbs contained mature chondrocytes, but hypertrophic chondrocytes were rare (Fig. 3B). The reduction of cellular volumes by v-fos expression would partly explain the severe truncation of these long bones. In these long bones at day 9.5, the expression of type X collagen gene (32), a maturation-associated marker of hypertrophic chondrocytes that precedes mineralization, was detectable by immunohistochemical analysis in uninfected limbs but was marginal in v-Fos-expressing limbs (Fig. 3B).

When primary chondrocytes were prepared from chicken caudal sternum, which is reported to be rich in mature and hypertrophic chondrocytes (29), they retained the expression of the type and type X collagen genes in monolayer culture as detected by Northern blotting, whereas expression of these two genes was undetectable in CEF (Fig. 5A). In v-fos (FBJ-MSV) virus-infected chondrocytes, the type X collagen gene expression was marginal, whereas expression of the type collagen gene, a marker of chondrocytic lineage, was retained. Expression of the c-fos gene (either chicken or mouse) in chondrocytes caused a significant reduction in type X gene expression, indicating that v-Fos expression causes a more potent transcriptional inhibition of specific genes in these chondrocytes. These results are consistent with the observation that phenotypic changes, such as accumulation of proliferating chondrocytes and delay in type X collagen expression or calcification, were less clear in c-fos virus-infected limbs. It was also shown that the growth rate of the primary chondrocytic culture was increased by the expression of either of the Fos proteins. These results also indicate that the suppression of type X collagen gene expression, as well as accumulation of proliferating chondrocytes in the limbs expressing exogenous fos (Fig. 3B), is a direct effect of the exogenous fos gene expression in chondrocytes, not an indirect effect involving other cell types.

Figure 5.

(A) Suppression of the type X collagen gene expression in primary chondrocytes by the expression of exogenous fos genes. Chicken chondrocytes from the upper sternum (lanes 1–4) as well as CEF (lane 5) were infected just after preparation with a subgroup B virus carrying no exogenous gene [DS3(B), lane 1], or carrying chicken c-fos [FC1(B), lane 2], mouse c-fos [FM4(B), lane 3], or v-fos from FBJ-MSV [FJ-2(B), lane 4]. Total RNAs (8 μg) were prepared, separated on agarose gels, and hybridized with DNA probes encoding chicken type collagen (II) and type X collagen (X). (B) Western blot analysis of Fos proteins in CEF infected with subgroup A viruses such as DS3 (empty vector, lane 1), FM4 (c-fos virus, lane 2), FJ2 (v-fos virus, lane 4), CD3 (lane 5), and IR2-CD3 (lane 6), or a subgroup B virus, FM4(B) (c-fos virus, lane 3). Five days after infection, cells were disrupted under denaturing conditions, and each protein sample (20 μg) was separated in 10% polyacrylamide gel and analyzed. Molecular masses of c-Fos, v-Fos, and CD3 were 60, 55, and 35 kDa, respectively.

Comparative Analysis of Cartilagenous Truncation and Cellular Transforming Activity Among Fos Proteins.

To examine whether the effect of Fos on cartilagenous differentiation is correlated with its cellular transforming activity, we analyzed the molecular basis of the quantitative difference in biological activity between v-Fos and c-Fos. When the absolute amounts of Fos proteins in CEF fully infected with Fos viruses were analyzed by Western blotting, the expression level of v-Fos (FBJ-MSV) was more than twice that of c-Fos (Fig. 5B), suggesting that the higher protein stability of v-Fos can at least partly explain the enhanced activity for cartilage truncation by this protein, as well as the clearer focal morphology in CEF (9). This view was further supported by the finding that a vector of subgroup B [FM4(B)] generates higher expression levels of c-Fos than FM4 (c-fos mouse virus of subgroup A) in CEF (Fig. 5B). FM4(B) caused more severe cartilagenous truncation than FM4 (Table 1), although its biological activity was not as strong as that of FJ2. On the other hand, Fra-2 did not cause any truncation even when expressed by the subgroup B vector (Table 1).

We next introduced a v-Fos mutant (CD3) which is truncated at a C-terminal region (Fig. 5B) that includes the major sequence differences between v-Fos and c-Fos. CD3 formed weaker foci compared with wild-type v-Fos (FBJ) or c-Fos (mouse), although the titer of transforming virus was comparable to that of the wild type (see Table 1 footnotes). Injection of this mutant had no effect in the limb bud (Table 1), indicating that fos transforming activity is not sufficient to cause chondrodysplasia. When CD3 expression levels were enhanced more than 8-fold by inserting an internal ribosomal entry site just in front of the CD3 coding sequence of the vector (IR2-CD3 virus), focal morphology was enhanced, and moderate truncation of the cartilage was seen (Fig. 1 and Table 1). These results indicate that the threshold level of Fos expression for chondrodysplasia is higher than that for cellular transformation assayed by CEF.

DISCUSSION

Compared with the analysis using transgenic or chimeric mice, the microinjection system of avian retrovirus vectors used here has the advantage that we can introduce genes into a limited region of the early embryos and observe phenotypic changes within a few days, even when the ectopic expression is fatal in the later stages of embryonic development. We observed a unique form of chondrodysplasia characterized by a severe delay in chondrocyte differentiation and embryonic endochondral ossification when an entire limb was infected with fos-containing virus (Fig. 4 and Table 1). The dysplasia observed was specific for fos because limbs expressing exogenous fra-2, c-jun, or junD were phenotypically normal. At the same time, it is noteworthy that such transforming genes as fra-2 (5) and c-jun (10, 11) did not affect mesenchymal tissues in ovo at all, although these oncogenes readily transform cells, such as CEF, which are derived from mesenchymal tissues. Based on an analysis of limbs expressing v-src oncogenes, it was previously proposed that the embryonic environment strongly attenuates oncogenesis by oncogenes in mesenchymal and epithelial tissues (36).

It is interesting that chondrocytic cells are the specific target of ectopic expression of fos even at such an early stage of development. The major effect was severe retardation of the differentiation step from proliferating chondrocytes to mature or hypertrophic chondrocytes. Because clear truncation of the entire long bones was observed even when v-Fos was expressed not homologously expressed throughout the NK24-infected limbs because of clonal infection (unpublished results), ectopic expression of Fos might cause secretion of some factor(s) that functions in a paracrine manner to prevent the expression of type X collagen and calcification. Wagner and coworkers (19) used embryonic stem cells to investigate the target cell specificity of c-fos expressed exogenously, and they found no macroscopic abnormalities at all during embryonic development, although exogenous c-fos mRNA was detected at day 13.5 post coitum However, a high frequency of cartilage tumors developed as early as 3–4 weeks of age. The cartilagenous truncation observed here after ectopic c-fos expression in an entire limb was not detected in embryonic stem cell chimeras. This was probably because the induction of c-fos in chicken limb buds occurred at an earlier stage or because the phenotype was too subtle to be detected in chimeras.

Comparative analysis among Fos proteins suggested that strong transforming activity is necessary to cause chondrodysplasia. We speculate that some of the specific Fos functions in early development detected here share common molecular mechanisms with those in chondrogenic and osteogenic tumor formation in postnatal mice. If chondrogenic progenitor cells present in the long bones of adults exhibited aberrant expression of endogenous c-Fos, prevention of normal development, as well as uncontrolled accumulation of proliferating chondrocytes, would occur.

In mouse day-17 embryos, endogenous c-fos expression become detectable in perichondrial growth regions (37) or in the flat inner layer of chondrogenic cells (38). Analysis of fos promoter-lacZ transgenic mice indicated that hypertrophic chondrocytes of the growth plate expressed-galactosidase in the vertebrae of adult mice (39). While we were not able to detect endogenous c-fos expression in the entire limb of day 5 chicken embryos, c-fos was moderately expressed in proliferating or hypertrophic chondrocytes at day 8 (unpublished results). From these observations, we suggest that endogenous c-Fos in proliferating chondrocytes regulates the rate of differentiation into mature or hypertrophic chondrocytes.

The molecular basis for the specific effects of Fos on chondrocytes remains to be solved, but it might involve the distinct transcriptional function of the heterodimer(s) with Jun family proteins via certain AP-1-binding sites. It is interesting that ATF-2-deficient mice (39) or Ets2 transgenic mice (40) have recently been reported to develop similar chondrodysplasia, which causes skeletal abnormalities. So, an alternative possibility is that Fos alone, and not other Fos family proteins, interacts with some transcriptional factor(s) other than AP-1 components, which in turn results in the modulation of ATF-2 and/or Ets2 function. We believe the limb bud system, as well as the primary chondrocyte culture derived from sternum, presented here is a useful model to study the functional partner of constitutively expressed Fos protein and to search for the target genes for these differentiation processes and for chondrogenic tumor formation.

ABBREVIATIONS

FBJ

Finkel–Biskis–Jinkins

MSV

murine sarcoma virus

CEF

chicken embryo fibroblasts