Cyclin D1 Is a Bona Fide Target Gene of NFATc1 and Is Sufficient in the Mediation of Injury-induced Vascular Wall Remodeling^*

Introduction

Nuclear factors of activated T cells (NFATs)² are a family of transcriptional factors that belong to the Rel/NFκB group (1). Among the five members of this family of transcriptional factors identified to date, NFATc1 to -c4 require Ca²⁺/calcineurin for activation and play a pivotal role in transcriptional regulation of cytokine genes (2–4). On the other hand, NFAT5, which is also called the tonicity enhancer-binding protein, although it belongs to the same family of transcriptional factors, appears to be Ca²⁺/calcineurin-independent and regulates genes associated with hypertonicity (5). The calcineurin-NFAT signaling has been shown to play a role in various developmental aspects, including cardiac valve formation (6), skeletal muscle growth (7), neuronal and osteoclast differentiation (8, 9), and vascular development (10). In addition, a large body of evidence indicates a role for calcineurin-NFAT signaling in the regulation of pathological cardiac hypertrophy (11, 12). In contrast, suppression of an endogenous calcineurin inhibitor, DSCR1 (Down syndrome candidate region-1), in endothelial cells leads to hyperactivation of calcineurin, which results in precocious endothelial cell apoptosis, thereby inhibiting tumorigenesis (13). Furthermore, in infantile hemangiomas (the tumors of endothelial cell origin), increased VEGFR2 signaling was found to be linked to suppression of NFAT-dependent expression of VEGFR1 (14). Thus, NFATs appear to be involved in the regulation of both positive and negative cell growth.

Dedifferentiation of vascular smooth muscle cells (VSMCs), followed by their migration from media to intima and proliferation in intima, contribute to the pathophysiology of renarrowing of the artery after angioplasty (15, 16). Previously, we have shown that NFATs play a role in the regulation of VSMC growth and migration (17, 18). Towards understanding the mechanisms of NFATs in the regulation of VSMC migration, we have shown that these transcriptional factors mediate interleukin-6 expression as one of the effector molecules in receptor tyrosine kinase (RTK) and G protein-coupled receptor (GPCR) agonist-induced VSMC motility (18). In addition, we have reported the involvement of NFATs in injury-induced neointima formation (19). Recently, we have also demonstrated that activation of NFATc1 causes up-regulation of the cell cycle dependent gene, cyclin A2, leading to increased CDK2 activity in VSMCs and the progression of these cells through the cell cycle (20). Because cell cycle progression is a complex multistep process orchestrated by different cyclins and their partner cyclin-dependent kinases (21–23), it is quite possible that NFATc1 may also have a role in the regulation of other cell cycle-dependent genes. In this work, we identified cyclin D1 as a target gene for NFATc1 in human aortic smooth muscle cells (HASMCs), and it is sufficient to mediate the progression of these cells through the cell cycle as well as their motility downstream to NFATc1. In addition, adenovirus-mediated transduction of cyclin D1 was found to be sufficient in overcoming the inhibitory effect of NFAT blockade on balloon injury-induced SMC migration from medial to luminal surface and their proliferation in the intimal region forming neointima. These observations together with our previous findings provide convincing evidence for the role of NFATc1 in vascular wall remodeling following injury.

MATERIALS AND METHODS

RESULTS

Previously, we have demonstrated that NFATc1 mediates PDGF-BB-induced expression of cyclin A2 in VSMCs (20). Because sequential activation of cyclin D-CDK4, cyclin E-CDK2, cyclin A-CDK2, and cyclin B-CDK1 is required for orderly cell cycle progression (21–23), we asked the question of whether NFATc1 besides cyclin A2 targets other cyclins in the regulation of PDGF-BB-induced VSMC proliferation. To address this question, we have studied the role of NFATs in PDGF-BB-induced expression of cyclin D1, because it plays a determinant role in cell cycle progression from G₀ to G₁ (21). PDGF-BB (20 ng/ml) induced cyclin D1 expression both at mRNA and protein levels, as measured by semiquantitative RT-PCR and Western blot analysis, respectively, in a time-dependent manner with maximum increases at 8 and 16 h in HASMCs (Figs. 1A and 2A). To examine the role of NFATs in PDGF-BB-induced cyclin D1 expression, we tested the effect of CsA, a potent inhibitor of calcineurin (1), and VIVIT, a pentapeptide that competes with NFATs for binding to calcineurin and inhibits NFATs (30). Both CsA (10 μm) and VIVIT (80 MOI) blocked PDGF-BB-induced cyclin D1 expression at mRNA and protein levels (Figs. 1B and 2B and supplemental Fig. 1, A and B). Because cyclin D1 forms a complex with and activates CDK4, we next studied the effect of PDGF-BB on CDK4 activity. Consistent with its effect on cyclin D1 levels, PDGF-BB (20 ng/ml) induced CDK4 activity in a time-dependent manner with maximum 2–3-fold increases at 8 and 16 h (Fig. 2A). In addition, CsA and VIVIT inhibited PDGF-BB-induced CDK4 activity (Fig. 2B and supplemental Fig. 1B). To find whether PDGF-BB-induced cyclin D1-CDK4 activity correlates with phosphorylation of endogenous Rb protein, we also studied the time course effect of PDGF-BB on Rb phosphorylation levels. Consistent with its effect on cyclin D1-CDK4 activity, PDGF-BB-induced Rb (Ser⁷⁹⁵) phosphorylation in a time-dependent manner with maximum effects at 8 and 16 h, and VIVIT significantly blunted this effect (Fig. 2, A and B). Because cyclin D1-CDK4 plays an important role in cell cycle progression, we next examined the effects of CsA and VIVIT on PDGF-BB-induced DNA synthesis and cell cycle progression. PDGF-BB induced DNA synthesis by 4-fold, as measured by [³H]thymidine incorporation, and this effect was blocked by both CsA and VIVIT (Fig. 3A and supplemental Fig. 2A). Fluorescence-activated cell sorting analysis showed that treatment with PDGF-BB induced about 45% of quiescent HASMCs to enter S phase from G₁ phase of the cell cycle as compared with untreated cells (Fig. 3B and supplemental Fig. 2B). Consistent with their effect on DNA synthesis, both CsA and VIVIT abrogated PDGF-BB-induced HASMC progression from G₁ to S phase (Fig. 3B and supplemental Fig. 2B). Our previous studies have demonstrated a role for NFATc1 in the regulation of VSMC growth by both GPCR and RTK agonists (17, 20). Therefore, we asked the question of whether NFATc1 also plays a role in PDGF-BB-induced cyclin D1 expression in HASMCs. We used NFATc1 siRNA molecules to down-regulate NFATc1 levels and determined its effects on PDGF-BB-induced cyclin D1 expression. NFATc1 siRNA, while having no effect on NFATc3 levels, depleted NFATc1 levels by more than 90% (Fig. 4A). NFATc1 siRNA also abolished PDGF-BB-induced cyclin D1 expression at both mRNA and protein levels in HASMCs (Fig. 4, B and C) and their DNA synthesis and migration (Fig. 4, D and E). To find whether other members of the NFAT family of transcription factors have any role in PDGF-BB-induced cyclin D1 expression, we also studied the role of NFATc2. Depletion of the steady-state levels of NFATc2 by its siRNA had no noticeable effect on PDGF-BB-induced cyclin D1 expression at mRNA or protein levels (Fig. 5, A–C).

To understand the mechanisms of NFATc1 involvement in the regulation of cyclin D1 expression, we analyzed about 3.0 kb of human cyclin D1 promoter for transcriptional factor binding sites by TFsearch. Five NFAT regulatory elements were found in the 3.0-kb cyclin D1 promoter, one each at the −1254, −1333, −2293, −2801, and −2902 nt positions relative to the transcription start site (31). In order to understand which of these NFAT-binding elements are involved in PDGF-BB-induced cyclin D1 expression, we cloned a 1.5-kb promoter region that contains the first two NFAT-binding sites from the transcription start site with serial 5′-deletions into pGL3 basic vector and determined their effects on luciferase reporter gene activity (Fig. 6, A and B). PDGF-BB induced luciferase reporter gene activity only with the 1.5-kb cyclin D1 promoter that contains the first two NFAT-binding sites. Deletion of these two NFAT-binding sites in the promoter failed to respond to PDGF-BB in the induction of luciferase activity (Fig. 6B). Furthermore, mutation in the NFAT-binding site at nt −1333 but not at nt −1254 abolished PDGF-BB-induced luciferase activity, suggesting that the second NFAT-binding site from the transcription start site is essential for cyclin D1 expression by PDGF-BB (Fig. 6B). Serial deletions of AP-1, NFκB, and E2F transcriptional factor binding sites had no effect on PDGF-BB-induced promoter activity. To obtain further evidence for the role of the NFAT-binding site at nt −1333 in PDGF-BB-induced cyclin D1 expression, we next studied the time course effect of PDGF-BB on NFAT DNA binding activity using the −1333 NFAT element as a ³²P-labeled oligonucleotide probe. PDGF-BB induced protein-DNA binding activity in a time-dependent manner with an about 2-fold increase at 2 h using this element as a ³²P-labeled probe (Fig. 7A). The specificity of NFAT binding activity to this site was further confirmed by lack of binding of a ³²P-labeled mutant −1333 NFAT oligonucleotide probe (Fig. 7B). To check if NFATs bind preferentially to the −1333 NFAT-binding site compared with those at the −1254, −2293, −2801, and −2902 nt positions in response to PDGF-BB, we used a 10-fold molar excess of the cold oligonucleotide probes representing these NFAT sites in a competition binding assay with the ³²P-labeled −1333 NFAT oligonucleotide probe. Only the cold −1333 NFAT-binding oligonucleotides and not the other NFAT-binding site oligonucleotides competed with the ³²P-labeled −1333 NFAT-binding oligonucleotide probe for binding to nuclear extracts of PDGF-BB-induced HASMCs (Fig. 7C). In addition, CsA and VIVIT completely inhibited PDGF-BB-induced −1333 NFAT DNA binding activity (Fig. 7, D and E). To obtain additional evidence for the role of NFATs in the regulation of cyclin D1, we also measured NFATc1 binding to the cyclin D1 promoter using a ChIP assay. ChIP analysis showed that NFATc1 binds to the cyclin D1 promoter region containing NFAT sites at −1254/−1333 nt positions in a time-dependent manner in response to PDGF-BB (Fig. 7F). In contrast, although NFATc2 binds to NFAT elements at −1254/−1333 nt at the basal level, its binding was not influenced by PDGF-BB, suggesting that NFATc1 binding to these sites is specific (Fig. 7G). To test whether NFATc1 binds to other NFAT-binding sites, we also performed a ChIP assay using primers specific for these sites. As shown in Fig. 7H, NFATc1 preferentially binds to NFAT sites at −1254 and −1333 nt but not to NFAT sites at −2293, −2801, and −2902 nt positions. Furthermore, PDGF-BB-induced NFATc1 binding to −1254/−1333 NFAT sites was inhibited by VIVIT (Fig. 7I). As a negative control, we designed primers for the cyclin D1 gene 2.5 kb downstream of the −1333 NFAT binding site, and these primers did not give any amplification product with the anti-NFATc1 immunoprecipitated and purified DNA as a template. These results clearly indicate that NFATc1 binds preferentially to the NFAT sites at −1254/−1333 nt in response to PDGF-BB. We should also point out that although it is technically difficult to differentiate between these two sites to which one NFATc1 binds, the mutagenesis and EMSA data clearly point out that NFATc1 binds to the −1333 NFAT element. Besides its involvement in cell cycle regulation, cyclin D1 has been reported to play a role in cell migration (32). In this aspect, we recently reported that STAT-5B-dependent expression of cyclin D1 plays a role in PDGF-BB-induced VSMC migration (24, 33). Because NFATc1 plays a role in PDGF-BB-induced HASMC DNA synthesis and migration, and it mediates PDGF-BB-induced cyclin D1 expression, we were intrigued to find whether cyclin D1 has any role in HASMC migration downstream of NFATc1. Depletion of cyclin D1 levels by the use of its siRNA inhibited PDGF-BB-induced HASMC DNA synthesis and migration (Fig. 8, A–C). In contrast, depletion of cyclin A2 while inhibiting DNA synthesis had no effect on PDGF-BB-induced HASMC migration (Fig. 8, B and C). To find whether depletion of cyclin D1 levels triggers cellular apoptosis and thereby accounts for decreased DNA synthesis and migration, we tested cyclin D1 siRNA effect on apoptosis. Inhibition of cyclin D1 levels had no effect on HASMC apoptosis, suggesting that the decreased HASMC DNA synthesis and migration by cyclin D1 depletion were not due to increased apoptosis (Fig. 8D). To determine whether forced expression of cyclin D1 alone is sufficient to rescue PDGF-BB-induced DNA synthesis and migration from inhibition by NFATc1 depletion, HASMCs that were transfected with NFATc1 siRNA and transduced with various titers of adenovirus expressing cyclin D1 and quiesced were tested for PDGF-BB-induced DNA synthesis and migration. Adenovirus-mediated expression of cyclin D1, while alone inducing HASMC DNA synthesis and migration, ablated the inhibitory effect of NFATc1 siRNA on PDGF-BB-induced DNA synthesis and migration in a dose-dependent manner with maximum effect at 40 MOI (Fig. 9, A and B). However, at higher doses of adenovirus (80 MOI), cyclin D1 expression alone is causing maximal DNA synthesis, and under these circumstances the cells appear to be less responsive to a subsequent challenge by PDGF-BB. To validate these findings in vivo, we used a rat aortic balloon injury model. BI induced cyclin D1 expression and CDK4 activity 3 days postinjury, and adenovirus-mediated transduction of VIVIT inhibited these effects (Fig. 10A). In addition, double immunofluorescence staining of the cross-sections of the artery at 3 days post BI for cyclin D1 and SMCα-actin revealed that the induction of cyclin D1 expression occurs in SMC and that it is mediated by NFAT activation (Fig. 10B). BI caused migration of SMCs from the medial to luminal region 3 days postinjury as determined by staining of longitudinally opened BI arteries with anti-histone H1 antibodies and counting the positively stained cells (Fig. 11A). Similarly, BI induced SMC proliferation in the intimal region at 1 week postinjury as measured by immunostaining of the artery cross-sections for PCNA expression (Fig. 11B). In addition, H/E staining and morphometry analysis of the cross-sections of the arteries at 2 weeks post-BI showed substantial neointimal growth (Fig. 12). Adenovirus-mediated expression of VIVIT significantly blocked BI-induced SMC migration from the medial to luminal surface and SMC proliferation in the intimal region, resulting in reduced neointima (Figs. 11 (A and B) and 12). It is interesting to find that adenovirus-mediated transduction of cyclin D1 soon after BI, while alone enhancing the effects of BI on SMC migration from media to intima and their proliferation in intima, leading to neointima formation, rescued these responses from inhibition by VIVIT (Figs. 11 (A and B) and 12).

DISCUSSION

The important observations of the present study are as follows. 1) PDGF-BB induced cyclin D1 expression and CDK4 activity in HASMCs in an NFATc1-dependent manner. 2) Depletion of NFATc1 blocked PDGF-BB-induced HASMC DNA synthesis and HASMC cell cycle progression from G₁ to S phase. 3) Down-regulation of NFATc1 also inhibited PDGF-BB-induced HASMC migration. 4) TransFac analysis of the human cyclin D1 promoter revealed the presence of five NFAT binding elements spanning an about 3.0-kb region starting from the transcription start site, and the site at −1333 nt was found to be sufficient to mediate PDGF-BB-induced promoter activity. 5) NFATc1-mediated expression of cyclin D1 was found to be required for PDGF-BB-induced HASMC DNA synthesis and migration. 6) Adenovirus-mediated transduction of cyclin D1 into artery, while enhancing the BI-induced SMC migration from the medial to luminal surface and their proliferation in the intimal region forming neointima, rescued these vascular wall remodeling responses from inhibition due to blockade of NFATs by VIVIT. Together, these observations provide a mechanistic evidence for the role of NFATc1 in PDGF-BB-induced HASMC proliferation and migration in vitro and BI-induced vascular wall remodeling in vivo.

In addition to their well characterized function in the regulation of cytokine gene expression in immune cells, NFATs have been shown to play a role in skeletal muscle growth (7), neuronal and osteoclast differentiation (8, 9), cardiac valve formation (6), vascular development (10), and pathological cardiac hypertrophy (11, 12). It was demonstrated that constitutively active forms of NFATc1 and NFATc2 induce two distinct phenotypes in NIH 3T3 cells; whereas NFATc1 increases their proliferation capacity and transformation (34, 35), NFATc2 causes their cell cycle arrest and apoptosis (35). NFATc1 has also been shown to mediate VEGF-induced endothelial cell proliferation (36). It was reported that conditional expression of NFATc1 enhances pancreatic β-cell proliferation (37). In addition to these observations, work from our laboratory showed that NFATs mediate VSMC growth and migration in response to both RTK and GPCR agonists (17, 18). A role for NFATs in the regulation of VSMC growth has also been supported by the findings that forced expression of constitutively active sarco/endoplasmic reticulum calcium ATPase inhibits VSMC proliferation through inactivation of calcineurin and its target transcription factor NFAT (38). In contrast to these reports, many studies have also provided evidence for a negative role of NFATs in the regulation of cell proliferation. Specifically, it was shown that NFATc1 via repressing CDK4 maintains hair follicle stem cells in quiescence (39). In infantile hemangiomas, suppression of NFAT-dependent VEGFR1 expression was reported to be associated with altered VEGFR2 signaling (14). In addition, depletion of calcineurin inhibitor, DSCR1, has been shown to prevent tumor growth (13). Based on these findings, it appears that NFAT isoforms carry out differential roles in the regulation of cell proliferation and/or differentiation in different cell types. In this aspect, it may be pointed out that NFATc2^−/− lymphocytes grow faster than wild type cells and express higher levels of cyclins (40). NFATc2 has also been shown to repress cyclin A2 and CDK4 expression in T cells (41, 42). In contrast, work from our laboratory showed a positive role for NFATs, particularly NFATc1, in the regulation of cyclins in VSMCs (20). It was well established that cyclin D1-CDK4 plays a role in cell cycle progression from G₁ to S phase (43, 44). Similarly, cyclin A2-CDK2 is involved in the progression of a cell through the S phase of the cell cycle (45). The role of cyclins in the regulation of cell proliferation can be appreciated by the capacity of cyclin D1-CDK4 to trigger a proliferative response even in a terminally differentiated cell type, such as cardiomyocytes (46). In the present study, we identified cyclin D1, a signature marker for cell proliferation, as a downstream effector molecule of NFATc1 in mediating HASMC DNA synthesis in response to PDGF-BB. The differential effects of NFATc1 and NFATc2 on the regulation of cyclin D1 levels as observed in the present study also correlate well with their differential effects on cell proliferation as reported by other studies (34, 35, 40–42). The characterization of the cyclin D1 promoter revealed a role for NFAT-binding site at nt −1333 in PDGF-BB-induced cyclin D1 expression in HASMCs.

Besides its involvement in cell cycle regulation, cyclin D1 has been reported to play a role in cell migration (32). Because NFATs were found to be involved in the regulation of migration and cyclin D1 expression in VSMCs, it was intriguing to ask the question of whether this cell cycle regulator has any role in cell migration downstream to NFATc1. The finding that a reduction in cyclin D1 but not cyclin A2 levels blocked PDGF-BB-induced migration and that adenovirus-mediated expression of cyclin D1 rescued PDGF-BB-induced HASMC migration from inhibition by NFATc1-depletion clearly point out a selective involvement of cyclin D1 in the regulation of cell migration. Previously, we have demonstrated that NFATs mediate both RTK and GPCR agonist-induced VSMC migration via induction of expression of interleukin-6 (18). Based on these observations, it may be suggested that NFATs target several genes, including interleukin-6 and cyclin D1, in the mediation of cell migration, perhaps different aspects associated with cell motility. We have also reported that cyclin D1 mediates both RTK and GPCR agonist-induced VSMC migration downstream to STAT-5B (24, 33). Based on these observations, one can postulate that several mechanisms may be involved in the regulation of cyclin D1 by RTK and GPCR agonists, and such concerted actions of various transcriptional factors in the regulation of expression of a gene such as cyclin D1 may be needed in tilting the balance between the anti- and promigratory state to the promotility state of the cell.

It is also exciting to find that adenovirus-mediated expression of cyclin D1, while enhancing BI-induced SMC migration from the medial to luminal surface and their proliferation in the intimal region leading to neointima formation, rescued these injury-induced vascular wall remodeling responses from inhibition by VIVIT. These findings reinforce the role of cyclin D1 in the regulation of SMC migration and proliferation downstream of NFAT, most likely NFATc1, in vivo as well. In recent years, it has been reported that the knockdown of any single cyclin or cyclin-dependent kinase gene does not affect the development variably (47, 48), but deletion of a combination of these molecules is lethal (49). Because NFATs regulate more than one cyclin in VSMCs that seem to be needed for the mediation of proliferation, migration, or both, these transcriptional factors could be better targets for the development of therapeutic agents for the control of vascular diseases, such as renarrowing of the coronary artery after angioplasty.

Supplementary Material