A constitutively active NFATc1 mutant induces a transformed phenotype in 3T3-L1 fibroblasts.

The calcineurin/nuclear factor of activated T cells (NFAT) signaling pathway is best known for its role in T lymphocyte activation. However, it has become increasingly apparent that this signaling pathway is also involved in the regulation of cell growth and development in a wide variety of different tissues and cell types. Here we have investigated the effects of sustained NFATc1 signaling on the growth and differentiation of the murine 3T3-L1 preadipocyte cell line. Remarkably, we find that expression of a constitutively active NFATc1 mutant (caNFATc1) in these immortalized cells inhibits their differentiation into mature adipocytes and causes them to adopt a transformed cell phenotype, including loss of contact-mediated growth inhibition, reduced serum growth requirements, protection from growth factor withdrawal-induced apoptosis, and formation of colonies in semisolid media. Furthermore, we find that caNFATc1-expressing cells acquire growth factor autonomy and are able to proliferate even in the complete absence of serum. We provide evidence that this growth factor independence is caused by the NFATc1-dependent production of a soluble heat-labile autocrine factor that is capable of promoting the growth and survival of wild type 3T3-L1 cells as well as potently inhibiting their differentiation into mature adipocytes. Finally, we demonstrate that cells expressing caNFATc1 form tumors in nude mice. Taken together, these results indicate that deregulated NFATc1 activity is able to induce the immortalized 3T3-L1 preadipocyte cell line to acquire the well established hallmarks of cellular transformation and thereby provide direct evidence for the oncogenic potential of the NFATc1 transcription factor.

The calcineurin/nuclear factor of activated T cells (NFAT) signaling pathway is best known for its role in T lymphocyte activation. However, it has become increasingly apparent that this signaling pathway is also involved in the regulation of cell growth and development in a wide variety of different tissues and cell types. Here we have investigated the effects of sustained NFATc1 signaling on the growth and differentiation of the murine 3T3-L1 preadipocyte cell line. Remarkably, we find that expression of a constitutively active NFATc1 mutant (caNFATc1) in these immortalized cells inhibits their differentiation into mature adipocytes and causes them to adopt a transformed cell phenotype, including loss of contact-mediated growth inhibition, reduced serum growth requirements, protection from growth factor withdrawal-induced apoptosis, and formation of colonies in semisolid media. Furthermore, we find that caNFATc1-expressing cells acquire growth factor autonomy and are able to proliferate even in the complete absence of serum. We provide evidence that this growth factor independence is caused by the NFATc1-dependent production of a soluble heat-labile autocrine factor that is capable of promoting the growth and survival of wild type 3T3-L1 cells as well as potently inhibiting their differentiation into mature adipocytes. Finally, we demonstrate that cells expressing caNFATc1 form tumors in nude mice. Taken together, these results indicate that deregulated NFATc1 activity is able to induce the immortalized 3T3-L1 preadipocyte cell line to acquire the well established hallmarks of cellular transformation and thereby provide direct evidence for the oncogenic potential of the NFATc1 transcription factor.
The nuclear factor of activated T cells (NFAT) 1 family of transcription factors is composed of four calcium-responsive family members (NFATc1 (NFAT2/NFATc), NFATc2 (NFAT1/ NFATp), NFATc3 (NFAT4/NFATx), and NFATc4 (NFAT3)) and is best known for its role in the regulation of the T cell immune response (1)(2)(3). The principal function of NFAT proteins in T cells is to couple stimulation of the T cell antigen receptor to changes in the expression of a number of cytokine and other immunologically important genes. NFAT proteins are regulated primarily at the level of their subcellular localization through the actions of the calcium/calmodulin-dependent serine/threonine phosphatase, calcineurin (1)(2)(3). In resting cells, NFAT family members are normally located in the cytoplasm in a hyperphosphorylated latent form. However, following an increase in the intracellular calcium concentration, activated calcineurin directly dephosphorylates NFAT proteins, inducing their rapid nuclear import and increasing their intrinsic DNA binding activity. Once located in the nucleus, NFAT proteins are then free to bind to their target promoter elements and activate the transcription of specific NFAT target genes, either alone or in combination with other nuclear partners. This calcineurin-mediated activation pathway is strongly opposed by a number of specific NFAT kinases, which act to directly rephosphorylate NFAT proteins, thereby directly antagonizing NFAT activity by inhibiting their DNA binding activity and promoting their rapid export back into the cytoplasm (4 -10). Consequently, NFAT-dependent transcription is highly dynamic and exquisitely sensitive to changes in the intracellular calcium concentration (11,12). This aspect of NFAT regulation is likely to be significant, since both the extent and duration of NFAT activity have recently been shown to influence the qualitative pattern of NFAT-dependent gene expression induced during T cell activation (13).
Whereas the calcineurin/NFAT signaling pathway is certainly best known for its role in the regulation of the immune response, it has become increasingly apparent that this pathway also plays an important role in the regulation of a wide variety of cellular responses in a number of other tissues (2,14). Thus, important roles for distinct NFAT proteins have been demonstrated in the development and function of the cardiovascular system, including regulation of cardiac valve morphogenesis, cardiac hypertrophy, early patterning of the vasculature, and angiogenesis in mature blood vessels (15)(16)(17)(18)(19). In addition, NFAT proteins have also been shown to be involved in the regulation of chondrocyte growth and differentiation (20) as well as to play important roles in skeletal muscle, where they have been shown to differentially regulate myogenesis, muscle fiber type gene expression, and myocyte hypertrophy (21)(22)(23)(24). These findings indicate that the calcineurin/NFAT signaling pathway is likely to play a much broader role in the regulation of cell growth and development than previously appreciated.
Previous studies have revealed the expression of NFAT proteins in the murine 3T3-L1 preadipocyte cell line, suggesting a potential role for these proteins in the regulation of adipocyte differentiation and function (25). The 3T3-L1 preadipocyte cell line is a well established in vitro model of adipocyte differentiation that has been extensively used to investigate the molecular processes that control adipocyte growth and development (26,27). When confluent, growth-arrested 3T3-L1 cells are exposed to the adipogenic hormones methylisobutylxanthine, dexamethasone, and insulin (collectively known as MDI), they synchronously enter the cell cycle and undergo a defined genetic program of terminal differentiation. This process requires the ordered expression of a number of transcription factors, including members of the CCAAT/enhancer-binding protein (C/EBP) family and the peroxisome proliferatoractivated receptor ␥ (PPAR␥) before permanently exiting the cell cycle and giving rise to mature morphologically distinct adipocytes containing large cytoplasmic triglyceride depots (26,27).
The current study was prompted by our recent demonstration that calcineurin is a critical effector of a calcium-dependent signaling pathway that acts to inhibit adipocyte differentiation (28). Given this role of calcineurin and the known expression of NFAT proteins in preadipocytes (25), we wished to examine the potential role of the NFAT signaling pathway in the regulation of adipocyte differentiation. To address this question, we took advantage of a previously characterized constitutively active NFATc1 mutant (caNFATc1) that is known to constitutively localize to the nucleus, bind DNA with high affinity, and activate endogenous chromatin-embedded NFAT target genes (29). We have used this constitutively active mutant to examine the effects of sustained NFATc1 activation on the adipocyte differentiation of 3T3-L1 cells. Remarkably, we find that whereas ectopic expression of this NFATc1 mutant in 3T3-L1 cells is able to potently inhibit their ability to differentiate into mature adipocytes, it is also sufficient to induce these immortalized cells to acquire the well established hallmarks of cellular transformation (30). Hence, we find that cells expressing caNFATc1 (a) lose contact-mediated growth inhibition, (b) exhibit reduced serum growth requirements, (c) are protected from growth factor deprivation-induced apoptosis, (d) gain growth factor autonomy and continue to proliferate in the absence of serum as a result of an NFATc1-induced autocrine regulatory growth loop, (e) undergo anchorage-independent cell growth in semisolid media, and (f) form tumors in athymic nude mice. Taken together, these results indicate that sustained NFATc1 activity is able to subvert the mechanisms that regulate the normal cell growth and differentiation of 3T3-L1 cells and is sufficient to induce these immortalized cells to adopt a transformed cell phenotype, thereby establishing the oncogenic potential of the NFATc1 transcription factor.

EXPERIMENTAL PROCEDURES
Cell Culture and Adipocyte Differentiation-3T3-L1 preadipocytes (ATCC) were cultured in growth medium: Dulbecco's modified Eagle's medium (DMEM) containing 4.5 g/liter glucose (Invitrogen) supplemented with 10% (v/v) fetal calf serum (FCS; Hyclone), 100 units/ml penicillin G, and 100 g/ml streptomycin (Invitrogen). To induce adipocyte differentiation, cells were grown until 2 days postconfluence (day 0), then treated for 2 days with growth medium plus MDI (0.5 mM methylisobutylxanthine, 1 M dexamethasone, and 10 g/ml insulin; all from Sigma). The cells were refed with growth medium containing 10 g/ml insulin at day 2 and every 2 days thereafter with growth medium alone. After 10 days, cells were fixed with formalin and stained with the lipophilic dye Oil Red O (Sigma) as previously described (28). Stained cells were either photographed directly or counterstained with Giemsa and visualized by bright field microscopy.
Production of Recombinant Retroviruses and Infection of 3T3-L1 Cells-The retroviral expression vectors pMSCV-GFP and pMSCV-caNFATc1 have been previously described (29). Recombinant retroviruses were produced by co-transfecting either the pMSCV-GFP or pM-SCV-caNFATc1 proviral vectors together with pVSV-G (Clontech), encoding the vesicular stomatitis virus-glycoprotein, into the GP293 pantrophic packaging cell line (Clontech) using LipofectAMINE Plus (Invitrogen). Medium was replaced after 24 h, and viral supernatants were harvested 2 days post-transfection and stored at Ϫ80°C. For infections, 5 ϫ 10 4 3T3-L1 cells were plated per well of a six-well plate. The next day, medium was replaced with 2 ml of viral supernatant containing 8 g/ml polybrene (Sigma), and plates were centrifuged at 2000 rpm for 1.5 h at room temperature. After removal of the viral supernatant, cells were expanded in growth medium for subsequent analysis and typically used within 5-7 days of infection. To ensure reproducibility, each experiment was repeated using cells derived from independent viral infections and independently derived retroviral stocks. Flow cytometric analysis of green fluorescent protein routinely revealed that greater than 95% of cells were virally infected.
Immunoblot and Northern Blot Analysis-Protein extracts prepared from cells harvested at the indicated times following induction of differentiation or withdrawal of serum were resolved by SDS-PAGE and subjected to immunoblot analysis with the relevant antibody (Ab). The Abs PPAR␥ (H-100), C/EBP␣ (14AA), C/EBP␤ (H-7), C/EBP␦ (C-22), cyclin D (H-295), c-Myc (N-262), and actin (C-2) were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA); Rb (G3-245) was purchased from BD Pharmingen. For Northern blot analysis, total RNA was isolated from cells using Trizol (Invitrogen) on the indicated day after the induction of differentiation. RNA samples (10 g) were separated using 1.2% agarose, 2.2 M formaldehyde gel electrophoresis and transferred to Hybond-N membrane (Amersham Biosciences). Immobilized RNA was hybridized with a 32 P-radiolabeled murine aP2 probe cDNA probe (ATCC) and visualized by exposure to Eastman Kodak Co. X-AR film. Membranes were stripped and reprobed with a glyceraldehyde-3-phosphate dehydrogenase probe as a loading control.
Focus Formation and Methylene Blue Staining-For focus-forming studies, 3T3-L1 cells infected with either the control MSCV-GFP or the MSCV-caNFATc1 viruses were diluted 1:30 with uninfected wild type 3T3-L1 cells. These were plated at a final density of 1 ϫ 10 4 cells/well of a six-well plate, and the growth medium was changed every 2 days. After 2 weeks, the cells were visualized by both phase-contrast microscopy and fluorescence microscopy for detection of GFP expression. Subsequently, the plates were rinsed in phosphate-buffered saline and fixed with methanol and then stained with 0.4% methylene blue in 0.5 M sodium acetate to visualize foci.
Cell Proliferation Studies-3T3-L1 cells infected with either the control MSCV-GFP or the MSCV-caNFATc1 virus were plated in triplicate. After 24 h, the cells were rinsed three times with DMEM and cultured in DMEM supplemented with either 10, 0.5, or 0% (v/v) FCS. At the indicated times after the medium change, cells were trypsinized and enumerated with a Coulter particle counter set to record events between 7 and 25 m.
Cell Cycle Analysis-On the indicated day following medium exchange, cells were trypsinized, washed once with growth medium and twice with sample buffer (Dulbecco's phosphate-buffered saline containing 1 g/liter glucose), and fixed in ice-cold 70% ethanol. At least 24 h later, 5 ϫ 10 5 fixed cells were stained for 30 min in 0.5 ml of sample buffer containing 50 g/ml propidium iodide and 100 units/ml RNase A. Analysis of cellular DNA content was determined by collecting 10,000 events using a FACScaliber flow cytometer and CELLQuest software (BD Biosciences).
Conditioned Medium Experiments-3T3-L1 cells infected with either the control MSCV-GFP or the MSCV-caNFATc1 virus were grown until they had achieved 80% confluence and then washed three times with DMEM and cultured in DMEM lacking serum for a further 2 days. The medium from each cell population was then collected and centrifuged at 2000 ϫ g for 10 min to yield a debris-free supernatant of serum-free conditioned medium. Where indicated, the conditioned medium was heated to 94°C for 20 min prior to use. To determine the effects of serum-free conditioned medium on cell growth and survival, wild type 3T3-L1 cells at 20% confluence were washed three times in DMEM and then cultured in a 1:1 mix of the appropriate conditioned medium and an equal volume of fresh DMEM. For adipocyte differentiation assays, serum-free conditioned medium was combined with an equal volume of DMEM containing 20% FCS, which was then used to induce adipocyte differentiation as described above.
Methylcellulose Growth and Nude Mouse Injections-To assay cells for growth in methylcellulose medium, 60-mm Petri dishes were coated with 1% agarose to resist cell adhesion. Cells were trypsinized, and 1 ϫ 10 5 cells were resuspended in 8 ml of growth medium containing 1.8% methylcellulose and allowed to grow in the Petri dishes for 4 weeks, after which representative colonies were photographed, and colonies were counted. For analysis of tumorigenic potential in vivo, cells that had been infected with either the control MSCV-GFP or the MSCV-caNFATc1 virus were trypsinized, washed, and resuspended in phosphate-buffered saline at a concentration of 2.5 ϫ 10 6 cells/ml. Athymic nu/nu mice (Harlan) were injected subcutaneously in each flank with 1 ϫ 10 6 cells, and tumor volume was determined after 16 days using a previously described formula (31).

caNFATc1 Inhibits Adipocyte Differentiation in 3T3-L1
Cells-In a previous study, we identified calcineurin as a critical intrinsic component of a Ca 2ϩ -dependent signaling pathway involved in the inhibition of adipocyte differentiation (28). Since NFAT proteins are known to be directly activated by calcineurin and are also known to be expressed in 3T3-L1 cells (25), we initially examined whether increased NFAT activity could affect adipogenesis. For this experiment, we took advantage of our previously characterized caNFATc1 mutant that is known to be constitutively localized to the nucleus, able to bind DNA with high affinity, and capable of activating endogenous gene expression (29). To facilitate efficient gene transfer, the caNFATc1 cDNA was introduced into the MSCV-GFP retroviral vector under the transcriptional control of the MSCV-long terminal repeat and upstream of an internal ribosome entry site-GFP expression cassette, thereby allowing the expression of both caNFATc1 and GFP from a single bicistronic mRNA. We have previously shown that we are able to generate high titer virus capable of routinely infecting Ͼ95% of 3T3-L1 cells by pseudotyping our recombinant retroviruses with the vesicular stomatitis virus glycoprotein-G (28). This high level of infection obviates the process of isolating and expanding clonal cell lines and allows analysis of bulk populations of cells immediately after infection.
To determine the effects of sustained NFATc1 signaling on adipocyte differentiation, 3T3-L1 preadipocyte cells were infected with either a retrovirus encoding caNFATc1 (MSCV-caNFATc1) or the control MSCV-GFP retrovirus and then induced to differentiate using a well established adipocyte differentiation protocol that involves exposure to the MDI adipogenic mixture. As shown in Fig. 1A, cells infected with the control MSCV-GFP virus and treated with MDI readily differentiated into morphologically distinct fat-laden adipocytes, as assessed by staining with the neutral lipid-specific dye, Oil Red O. In contrast, 3T3-L1 cells expressing caNFATc1 failed to differentiate into mature adipocytes, based upon both morphological criteria and a lack of specific staining with Oil Red O. This lack of differentiation was confirmed by analyzing cells for their expression of both the adipocyte-specific marker gene aP2 and the critical proadipogenic transcription factors PPAR␥ and C/EBP␣. Whereas control cells expressed aP2, PPAR␥, and C/EBP␣ following treatment with MDI, the expression of these gene products was undetectable in MDI-treated, caNFATc1expressing cells (Fig. 1, B-D). However, ectopic expression of caNFATc1 did not completely disrupt all responses of 3T3-L1 cells to the MDI adipogenic stimuli. Treatment of caNFATc1expressing cells with MDI induced expression of C/EBP␤ and C/EBP␦, two early transcription factors in the adipogenic cascade, with similar kinetics to those observed in control cells (Fig. 1, E and F). Taken together, these data suggest that sustained activation of the NFATc1 signaling pathway inhibits the differentiation of 3T3-L1 cells into mature adipocytes by preventing the expression of the critical proadipogenic transcription factors PPAR␥ and C/EBP␣.

Sustained NFATc1 Activity Induces Morphological Changes and Loss of Contact-dependent Growth Inhibition in 3T3-L1
Cells-When we initially analyzed 3T3-L1 cells expressing caNFATc1, we noticed that they exhibited a very different cellular morphology compared with either control 3T3-L1 cells or cells infected with the control MSCV-GFP virus. Whereas control 3T3-L1 cells and cells infected with control MSCV-GFP virus were relatively uniform in size and shape and represented a typical fibroblast morphology, cells expressing caN-FATc1 adopted a wider range of morphologies, including many tracts of highly refractile spindle-like cells interspersed with occasional large, flat cells ( Fig. 2A). Moreover, unlike control cells that stopped growing once they had reached confluence, we found that cells expressing caNFATc1 overgrew the monolayer and continued to proliferate beyond confluence (Fig. 2B). In fact, visual inspection of caNFATc1-expressing cells under the microscope revealed that they accumulated in dense mutilayers of cells growing on top of one another. These observations suggested that expression of caNFATc1 had caused these 3T3-L1 cells to lose contact-mediated growth inhibition, one of the known hallmarks of cellular transformation (30). To more clearly visualize this phenotype, we performed a focus-forming assay in which a small number of either MSCV-GFP-or MSCV-caNFATc1-infected cells were mixed together with an excess of uninfected wild type 3T3-L1 cells and then grown in culture for 10 -14 days and analyzed for the formation of transformed foci by either methylene blue staining or direct microscopy.
Whereas cells infected with the control MSCV-GFP virus did not give rise to any foci as detected by methylene blue staining, we found that caNFATc1-expressing cells formed numerous foci (Fig. 2C). Because caNFATc1 and GFP are expressed from the same bicistronic mRNA in the MSCV-caNFATc1 retroviral vector, both caNFATc1-expressing cells and cells infected with the control retrovirus can be readily discriminated from wild type uninfected 3T3-L1 cells on the basis of their GFP expression. This allowed us to identify cells infected with either the MSCV-caNFATc1 or control MSCV-GFP retrovirus, which in turn allowed us to evaluate their cellular and colony morphologies. As a result of this analysis, we found that all of the caNFATc1-expressing cells in the mixed cultures formed readily detectable foci, whereas cells infected with the control MSCV-GFP virus were morphologically indistinguishable from their uninfected neighboring cells (Fig. 2D). Together, these results suggest that expression of caNFATc1 in 3T3-L1 cells promotes the loss of contact-mediated growth inhibition and induces these cells to form foci in culture.
Sustained NFATc1 Activity in 3T3-L1 Cells Promotes Cell Cycle Progression in the Presence of Reduced Serum Concentrations-Having demonstrated that ectopic expression of caN-FATc1 promotes the proliferation of 3T3-L1 cells in the presence of 10% FCS, we next investigated whether caNFATc1 could affect cell proliferation under reduced serum conditions. As shown in Fig. 3, control 3T3-L1 cells cultured in medium containing only 0.5% FCS fail to proliferate (Fig. 3A) and accumulate in the G 1 phase of the cell cycle after a 24-h period (Fig. 3B). In contrast, we found that caNFATc1-expressing cells steadily continue to proliferate under these reduced serum conditions (Fig. 3A) and maintain a cell cycle profile typical of actively dividing cells (Fig. 3B).
Since cells expressing caNFATc1 appeared to overcome the G 1 cell cycle arrest induced by low serum conditions, we next examined the effects of caNFATc1 on the expression of the cell cycle-related genes cyclin D, retinoblastoma protein (Rb), and c-myc, which are all known to play a role in cell cycle progression. Within 24 h of shifting to reduced serum conditions, the expression of the D-type cyclin proteins, which is known to be dependent upon the presence of mitogenic growth factors (32), was significantly diminished in control cells. In contrast, the levels of cyclin D1 and D2 were maintained at a high level in caNFATc1-expressing cells, even after 48 h in the presence of reduced serum (Fig. 3C). Since cyclin D-dependent protein kinases are known to contribute toward cell cycle progression by directly phosphorylating and inhibiting the growth-inhibitory activity of the Rb cell cycle regulatory protein (33), we next examined the effects of caNFATc1 on the relative distribution of phosphorylated Rb polypeptide species. In resting cells, Rb is found in a hypophosphorylated active form (Rb). Following mitogenic stimulation and the activation of cyclin-dependent protein kinases, Rb becomes hyperphosphorylated and consequently inactive (pRb), as detected by a more slowly migrating form on SDS-PAGE. As shown in Fig. 3C, shifting control cells to low serum conditions results in the disappearance of the hyperphosphorylated Rb forms. Conversely, these inactive Rb isoforms are still readily detectable in caNFATc1-expressing cells maintained in the presence of 0.5% FCS for more than 48 h. Finally, we examined expression of the c-Myc oncoprotein expression in both control and caNFATc1-expressing cells. Interestingly, caNFATc1-expressing cells consistently exhibited high levels of c-Myc protein even under these reduced serum conditions (Fig. 3C). Taken together, these results indicate that ectopic expression of caNFATc1 in 3T3-L1 cells is able to reduce serum growth requirements by promoting cell cycle progression even under conditions of limiting serum growth factors. Sustained NFATc1 Activity Protects Cells from Growth Factor Withdrawal-induced Apoptosis-We next tested the effects of caNFATc1 on the cellular response to complete serum withdrawal, which is known to induce normal fibroblasts to undergo apoptosis. As determined by phase microscopy and direct cell counts, transfer of control MSCV-GFP retrovirally infected cells from growth medium containing 10% FCS into serum-free medium resulted in a rapid and pronounced decrease in the number of viable cells (Fig. 4, A and B). In contrast, we observed that cells infected with the MSCV-caNFATc1 retrovirus not only maintained viability but also continued to proliferate after a brief lag phase despite the complete absence of serum.
As indicated by flow cytometric analysis of the cellular DNA profile of control MSCV-GFP-infected cells, complete serum withdrawal resulted in the accumulation of a significant proportion of cells exhibiting a sub-G 1 DNA content that is known to be indicative of cells undergoing apoptosis. Conversely, we found that cells infected with the MSCV-caNFATc1 virus maintained a typical cell cycle DNA profile with little evidence of apoptosis (Fig. 4C). These data suggest that expression of caN-FATc1 is able to protect 3T3-L1 cells from undergoing apoptosis in response to growth factor withdrawal.
Sustained NFATc1 Activity in 3T3-L1 Cells Induces the Production of One or More Autocrine Heat-labile Prosurvival/Antiadipogenic Factors-In the previous experiments, we observed that caNFATc1-expressing cells were able to proliferate

FIG. 4. Sustained NFATc1 activity protects cells from growth factor withdrawal-induced apoptosis.
A, phase microscopy of control and caNFATc1-expressing cells on the indicated day following complete withdrawal of serum. B, cells infected with either the control or caNFATc1 retroviruses were plated in triplicate in medium lacking serum, and their number was determined at daily intervals with a Coulter particle counter. The data are presented as the percentage change in cell number compared with the number of cells present at day 0, which was arbitrarily set to 100%. The S.D. values are indicated. C, effects of caNFATc1 on apoptosis induced by serum growth factor withdrawal. Cells infected with either the control or caNFATc1 retroviruses were plated in triplicate in medium lacking serum and were analyzed by PI staining followed by flow cytometric analysis at daily intervals. The data are presented as percentage of apoptotic cells, as determined by the percentage of cells in the sub-G 1 peak compared with the number of cells in the cell cycle (left panel). Representative profiles from day 4 are also shown (right panels). even in the complete absence of exogenously derived serum growth factors (see Fig. 4, A and B). The underlying mechanism for this effect could be the result of an intrinsic cell-autonomous effect of caNFATc1 on the expression of genes involved in the regulation of cell cycle progression. Alternatively, the effect could be non-cell-autonomous and result from the caNFATc1dependent production of one or more prosurvival/promitogenic factors that act in an autocrine fashion to promote cell survival and proliferation. In order to gain initial insights into the underlying molecular mechanism, we first assayed serum-free conditioned medium from either control or caNFATc1-expressing cells for the presence of survival/mitogenic factors. Thus, serum-free conditioned medium collected from either control or caNFATc1-expressing cells were tested for their effects on the survival of wild type 3T3-L1 cells cultured in medium lacking serum. As shown in Fig. 5A, serum-free conditioned medium isolated from control MSCV-GFP-infected cells did not support the survival of wild type 3T3-L1 cells, which appeared to rapidly succumb to apoptosis. In contrast, serum-free conditioned medium isolated from caNFATc1-expressing cells promoted both the survival and proliferation of wild type 3T3-L1 cells (Fig. 5A). Further investigation revealed that the activity of the survival/mitogenic factor present in the serum-free conditioned medium isolated from caNFATc1-expressing cells was sensitive to heat treatment, potentially suggesting the involvement of a heat-labile polypeptide.
Since mitogenic growth factors have previously been shown to inhibit adipocyte differentiation (34,35), we next investigated whether the production of a caNFATc1-induced autocrine factor might also explain the observed blockade of adipocyte differentiation in caNFATc1-expressing cells (Fig. 1). Thus, we examined the ability of normal wild type 3T3-L1 cells to undergo MDI-induced adipocyte differentiation in the presence of conditioned medium isolated from either control or caNFATc1-expressing cells. Whereas the conditioned medium isolated from control cells did not affect the ability of 3T3-L1 cells to efficiently differentiate into mature adipocytes, as determined by staining with Oil Red O, we found that the conditioned medium isolated from caNFATc1-expressing cells potently inhibited adipogenesis (Fig. 5B). As with the mitogenic activity described above, the antiadipogenic activity present in conditioned medium from caNFATc1-expressing cells was found to be heat-labile (Fig. 5B). Taken together, these data indicate that caNFATc1 induces the production of one or more heat-labile factors that act in an autocrine fashion to promote both cell survival and proliferation of 3T3-L1 cells as well as potently inhibiting their ability to undergo terminal differentiation into mature adipocytes.

Sustained NFATc1 Activity in 3T3-L1 Cells Promotes Anchorage-independent Cell Growth and the Formation of Tumors in Athymic Nude
Mice-Collectively, our data indicate that sustained NFATc1 activity causes 3T3-L1 cells to adopt many of the well established hallmarks of transformed cells. To further analyze the transforming potential of caNFATc1, we next tested the ability of caNFATc1 to induce anchorage-independent cell growth, another well established property of transformed cells. Thus, either control or caNFATc1-expressing cells were plated in semisolid methylcellulose medium and monitored for their ability to form colonies in the absence of a solid substratum. We found that control cells remained in suspension as single cells and never formed colonies, whereas the caNFATc1-expressing cells readily formed large colonies (Fig.  6, A and B).
As a final examination of the transforming potential of caNFATc1, we tested the ability of caNFATc1 expression to promote the formation of tumors in athymic nude mice, the sine qua non of cellular transformation. Thus, athymic nude mice were injected in each flank with 10 6 3T3-L1 cells infected with either control MSCV-GFP virus or the MSCV-caNFATc1 virus, and the animals were monitored for tumor formation. Whereas all of the mice injected with caNFATc1-expressing cells displayed large bilateral flank tumors ( Fig. 6C; mean tumor volume ϭ 700 Ϯ 170 mm 3 ), none of the mice injected with control cells exhibited detectable tumors. These results indicate that ectopic expression of caNFATc1 is able to promote anchorageindependent cell growth and is sufficient to cause cells to form tumors in nude mice. Taken together with our other data, these results clearly highlight the oncogenic potential of the NFATc1 transcription factor. DISCUSSION In the current study, we provide evidence that deregulated NFATc1 activity is able to subvert the mechanisms that regulate the normal cell growth and differentiation of 3T3-L1 preadipocytes and is sufficient to induce these immortalized cells to acquire the well established hallmarks of cellular transformation (30). Thus, we find that cells expressing caNFATc1 lose contact-mediated growth inhibition, exhibit reduced serum growth requirements, and are fully protected from apoptosis following growth factor deprivation. Furthermore, we observe that caNFATc1-expressing cells acquire full growth factor au- FIG. 5. Enforced expression of caNFATc1 in 3T3-L1 cells induces the production of an autocrine heat-labile promitogenic/ antiadipogenic factor(s). A, phase microscopy of wild type 3T3-L1 cells incubated for either 0 or 4 days with serum-free conditioned medium (CM) isolated from either control or caNFATc1-expressing cells. Where indicated, the conditioned medium was first heated to 94°C for 20 min. B, Oil Red O staining of wild type 3T3-L1 preadipocytes induced to undergo adipocyte differentiation by treatment with MDI together with conditioned medium from either control or caNFATc1-expressing cells. Where indicated, the conditioned medium was heat-treated at 94°C for 20 min. tonomy and are able to continue to proliferate in the complete absence of exogenously added serum growth factors. In fact, we provide evidence that this growth factor independence is caused by the NFATc1-dependent production of one or more autocrine factors that are capable of promoting both the growth and survival of wild type 3T3-L1 cells as well as inhibiting their differentiation into mature adipocytes. Finally, we demonstrate that, unlike control cells, cells expressing caNFATc1 establish colonies in semisolid medium and form tumors when injected into athymic nude mice. Collectively, these results indicate that sustained NFATc1 activity in 3T3-L1 preadipocytes induces an autocrine regulatory growth loop and is sufficient to induce the immortalized 3T3-L1 preadipocyte cell line to adopt a transformed cell phenotype, thereby establishing the oncogenic potential of the NFATc1 transcription factor in vitro and raising the possibility that deregulated NFATc1 activity may play a causative role in tumorigenic progression in vivo.
Significant insight into the potential molecular mechanisms underlying the transforming effects of caNFATc1 was gained by our observation that enforced expression of caNFATc1 is sufficient to promote growth factor autonomy, allowing cells to grow and survive in the complete absence of exogenously added serum growth factors. We provide evidence that this independence from exogenous growth factors is mediated by a non-cellautonomous mechanism, in which caNFATc1 induces the expression of a soluble prosurvival/promitogenic factor(s) that is capable of acting in an autocrine fashion to promote cell growth and survival. Thus, we find that whereas serum-free condi-tioned medium isolated from control cells is unable to support the serum-free survival and proliferation of wild type 3T3-L1 cells, conditioned medium isolated from caNFATc1-expressing cells contains a mitogenic factor(s) capable of promoting both the survival and proliferation of normal uninfected 3T3-L1 cells. In addition to its effects on cell survival and proliferation, we also find that conditioned medium isolated from caNFATc1expressing cells is able to potently inhibit adipocyte differentiation. This antiadipogenic activity of the caNFATc1-conditioned medium is probably related to its mitogenic activity, since mitogenic growth factors are known to inhibit adipocyte differentiation by preventing preadipocytes from permanently exiting the cell cycle (34,35). Further analysis revealed that both the mitogenic and antiadipogenic activities present in caNFATc1-conditioned medium are sensitive to heat treatment, suggesting the involvement of a polypeptide factor(s). Based upon these observations, we propose that the growth factor autonomy and antiadipogenic activity induced by caNFATc1 is most likely caused by a direct transcriptional effect of NFATc1 on the expression of an endogenous factor(s) capable of acting in an autocrine fashion to promote both cell proliferation and the inhibition of adipocyte differentiation. Remarkably, the ability of caNFATc1 to induce an autoregulatory growth loop in 3T3-L1 cells has a striking parallel with the known function of NFAT proteins in the regulation of the T cell immune response. There, NFAT proteins are known to induce the expression of the primary T cell growth factor, interleukin-2, and a component of its high affinity receptor, the interleukin-2 receptor ␣-chain gene (1,36,37), thereby establishing an interleukin-2-dependent autocrine growth loop that is known to promote T cell clonal expansion (38). Although we cannot completely rule out a cell autonomous role for NFATc1, our data certainly favor a mechanism in which sustained NFATc1 activity promotes cell cycle progression by inducing the expression of an autocrine growth factor(s). Since the autocrine production of mitogenic growth factors is a common mechanism employed by tumor cells to escape the normal regulatory constraints that restrict the proliferation and survival of normal nontransformed cells (39), we believe that this NFATc1-induced autocrine regulatory growth loop induced in 3T3-L1 cells is likely to play an important role in the transforming effects of caNFATc1.
Based upon our observation that sustained NFATc1 activity is sufficient to induce the transformation of the immortalized 3T3-L1 cell line, it is tempting to speculate that increased and prolonged NFATc1 activity may directly contribute toward tumor progression in certain cell lineages in vivo. At present, there is no evidence that the well known mechanisms of proto-oncogene activation, such as gain-offunction mutations, chromosomal translocations, and gene amplification, are involved in deregulating the activity of any NFAT family member. However, one potential mechanism by which NFATc1 activity could be sustained in vivo is by the continuous presence of an external stimuli that is capable of directly stimulating signal transduction pathways leading to the prolonged induction of NFATc1 activity. In this regard, it is interesting to note that signaling through the insulin-like growth factor (IGF)-1 receptor is known to enhance both the expression and activity of NFATc1 in a number of different cell types, including preadipocytes (23,24,40,41). This observation is of particular interest, since both IGF-1 and IGF-2 are known to play an important role in the etiology and progression of a wide variety of different tumors (42). In fact, IGF-1 and IGF-2 are found to be overexpressed in more than 90% of liposarcomas (43), suggesting a potential mechanism by which IGF-1/IGF-2 autocrine signaling may lead to sus- tained NFATc1 activity. However, whether IGF-1/IGF-2-induced NFATc1 activity contributes toward the development of liposarcomas and other tumors of the adipose lineage remains to be seen. In addition to tyrosine kinase growth factor receptors such as the IGF-1 receptor, several other signaling pathways, including those regulated by G-protein-coupled receptors, integrins, and the receptors for the Wnt family of proteins, are known to induce NFAT activity in a wide variety of different cell types (44 -47). Since excessive signaling through each of these different pathways is known to influence various aspects of the tumorigenic phenotype (48 -51), it will be interesting to determine whether any of these effects are mediated through the actions of NFATc1.
Finally, although our results are the first to directly demonstrate the oncogenic potential of an NFAT family member, other recent studies have implicated NFAT proteins in the regulation of other distinct aspects of the cancer phenotype. Based upon their observation of uncontrolled proliferation of abnormal extraarticular cartilage cells in NFATc2-null animals, Glimcher and colleagues (20) have proposed a potential tumor suppressor role for NFATc2 in the chondrocyte lineage. When isolated and placed in culture, these NFATc2-deficient cells were found to proliferate rapidly, lose their ability to undergo contact-mediated growth inhibition, exhibit an increase in aneuploidy, and form tumors in nude mice. Support for an inhibitory role of NFATc2 in the regulation of cellular growth control is provided by the recent observation that this NFAT family member acts to repress the expression of the key cell cycle regulatory kinase, cyclin-dependent kinase-4, via the direct recruitment of transcriptional co-repressors (52). Hence, it appears that NFATc1 and NFATc2 are likely to play opposing roles in the tumorigenic process, with NFATc1 exhibiting the properties of an oncogene and NFATc2 appearing to function as a tumor suppressor. The opposing effects of these two NFAT family members on the transformation process reflect their respective roles in the regulation of the immune response, since deficiency in NFATc1 results in a defect in immune cell function and T cell proliferation (53,54), as would be expected for a positively acting transcription factor, whereas a deficiency in NFATc2 has been shown to lead to T and B cell hyperproliferation (55,56). Importantly, the role of NFAT family members in the regulation of the tumorigenic phenotype may not be restricted to the early stages of cell growth control, since a number of recent studies have suggested roles for NFAT proteins in angiogenesis and metastasis, processes that occur during the later stages of tumorigenesis. NFAT proteins have been shown to play a role in the regulation of vascular endothelial growth factor-mediated angiogenesis via direct regulation of cyclooxygenase-2 gene expression (19), an enzyme that is known to play a pivotal role in neovascularization (57). This activity would be predicted to play an important role in tumor progression, since successful tumor colonization in vivo is known to require the recruitment of an adequate blood supply in order to provide the tumor with oxygen and essential nutrients (30). More recently, a role for NFAT proteins in the regulation of tumor cell invasion has also been proposed (45). In this case, NFAT proteins have been shown to be expressed in an active form in human breast carcinoma cell lines, and the ability of these cells to undergo integrin-mediated invasion of an extracellular matrix barrier has been shown to depend on the activity of the calcineurin/NFAT signaling pathway. Based upon these multiple lines of evidence, it appears that the NFAT family of transcription factors is likely to contribute to many distinct aspects of the tumorigenic phenotype, including the initial dysregulation of cellular growth control, the recruitment of an adequate blood supply, and the regulation of tumor me-tastasis. Further investigation into the emerging role of NFAT family members in the regulation of the tumorigenic process is clearly warranted.