<<

Oncogene (2008) 27, 366–377 & 2008 Nature Publishing Group All rights reserved 0950-9232/08 $30.00 www.nature.com/onc ORIGINAL ARTICLE The AP-1 regulates breast cancer cell growth via and factors

Q Shen, IP Uray, Y Li, TI Krisko, TE Strecker, H-T Kim and PH Brown

Breast Center, Departments of Medicine and Molecular and Cellular Biology, Baylor College of Medicine, Houston, TX, USA

The activating -1 (AP-1) transcription factor Introduction transduces growth signals through signal transduction pathways to the nucleus, leading to the expression of The activating protein-1 (AP-1) family of transcription involved in growth and malignant transformation in many factors consists of multiple Jun (cJun, JunB and JunD) cell types. We have previously shown that overexpression and Fos (cFos, FosB, Fra-1 and Fra-2) members (Angel of a dominant negative form of the cJun proto-oncogene, a and Karin, 1991). The Jun members either homodimer- cJun dominant negative mutant (Tam67), blocks AP-1 ize with Jun itself or heterodimerize with different Fos transcriptional activity, induces a G1 block and members, whereas Fos members heterodimerize only inhibits breast cancer cell growth in vitro and in vivo.We with Jun members. Thus, Jun is the predominant found that AP-1 blockade byTam67 in MCF-7 breast partner to form AP-1 complexes such as Jun:Jun or cancer cells downregulates D1 transcriptional Jun:Fos dimers. The AP-1 complex converges multiple activitybyat least two mechanisms: bysuppressing trans- growth signals at the transcriptional level, making it a cription at the known AP-1 binding site (À934/À928) critical connecting node for many signal transduction and bysuppressing -induced expression pathways. We and other investigators have shown that through suppressing E2F activation at the E2F-responsive AP-1 regulates cellular proliferation, differentiation, site (À726/À719). AP-1 blockade also led to reduced , oncogene-induced transformation and cancer expression of and , but not , at the cell invasion (McDonnell et al., 1990; Szabo et al., 1991; mRNA and protein levels. Chromatin immunoprecipita- Brown et al., 1993). Proliferation of breast cells requires tion and supershift assays demonstrated that AP-1 signals from growth factors such as estrogen, epidermal blockade caused decreased binding of E2F1 protein to growth factor (EGF), transforming growth factor a the E2F site in the . We also found that (TGFa), heregulin and insulin-like growth factors Tam67 suppressed the expression of the E2F1 dimerizing (IGFs), and these factors activate AP-1 signaling. partner, DP1 and E2F-upregulated cell cycle genes Therefore, blockade of AP-1 complex may arrest multi- (cyclins E, A, B and D3) and enhanced the expression of ple growth signals important for breast cell proliferation E2F-downregulated cell cycle genes (cyclins G2 and I). and transformation. Reduced expression of other E2F-regulated genes was also We have previously shown that AP-1 blockade seen with AP-1 blockade and E2F suppression. Thus, the induced by expression of a specific AP-1 inhibitor (a AP-1 factor regulates the expression of and E2F cJun dominant-negative mutant, Tam67) suppressed the (the latter in turn regulates E2F-downstream genes), growth of breast cancer cells induced by many growth leading to cell cycle progression and breast cancer cell factors such as estrogen, EGF, heregulin and IGF-1 proliferation. (Liu et al., 2002). We also demonstrated that AP-1 Oncogene (2008) 27, 366–377; doi:10.1038/sj.onc.1210643; blockade by Tam67 arrested cell cycle at the published online 16 July 2007 (Liu et al., 2004). In addition, G1 cyclins were down- regulated at the mRNA and protein levels when breast Keywords: AP-1 factor; proliferation arrest; cyclin D1; E2F cancer cells were synchronized at the G2/M phase and factors; expression regulation; breast cancer cells released from synchronization by serum stimulation (Liu et al., 2004). However, the explicit mechanism by which AP-1 blockade regulates the cell cycle remains unclear. Potential consensus AP-1 binding sites or AP-1-like binding sites have been found in the promoter of the cyclin D1 gene (Herber et al., 1994). In the present study, we measured cyclin D1 expression at the mRNA and protein levels in MCF-7 cells synchronized at G0/G1 Correspondence: Dr PH Brown, Breast Center, Baylor College of phases, and determined that the consensus AP-1 binding Medicine, One Baylor Plaza, MS600, Houston, TX 77030, USA. E-mail: [email protected] site is responsible for maintaining basal cyclin D1 Received 18 August 2006; revised 22 May 2007; accepted 23 May 2007; promoter activity. We further demonstrated that an published online 16 July 2007 E2F site present in the cyclin D1 promoter was required AP-1 blockade downregulates cyclins and E2F1 Q Shen et al 367 for the Tam67-induced suppression of cyclin D1 We next mutated the E2F sites at À726/À719 and expression. We then demonstrated that AP-1 and E2F À139/À131 (Figure 2b). Mutation of the E2F site at factors bind to AP-1 and E2F binding sites in the cyclin À139/À131 resulted in a slight decrease in promoter D1 promoter. We also determined that E2F1 and E2F2, activity, while the Tam67-induced suppression remained and their dimerizing partner DP1 were downregulated the same. Mutation of the E2F site at À726/À719 by Tam67, as were many E2F downstream genes. Thus, caused loss of Tam67-induced suppression of promoter the AP-1 transcription factor regulates breast cancer cell activity (Figure 2b). As shown in Figure 2c, the growth via multiple mechanisms, including regulation of fragment containing the E2F site at À726/À719 was cyclin D1, E2F factors and their target genes. These ligated back to the shortest construct used that does not findings support that the AP-1 transcription factor is a have the E2F site at À726/À719, and the resulting critical regulator of breast cell proliferation, making this promoter construct restored the repression by Tam67. factor a potential target for the treatment and preven- These results confirm the importance of the E2F site at tion of breast cancer. À726/À719 in the AP-1 regulation of cyclin D1. Our studies demonstrate that AP-1 regulates cyclin D1 expression by at least two mechanisms: activation of basal cyclin D1 promoter activity via the AP-1 site and Results further by increasing E2F activity. Thus both the AP-1 and E2F factors play important roles in regulating AP-1 blockade by Tam67 downregulates cyclin D1 at the cyclin D1 expression in breast cancer cells. mRNA and protein levels in MCF-7 cells We have previously shown that AP-1 blockade by Tam67 reduced cyclins D1, D2, D3 and at the Interaction of the AP-1 inhibitor, Tam67, with the mRNA and protein level in MCF-7 cells synchronized at promoter elements of cyclin D1 gene et al G2/M phase (Liu ., 2004). Here, we examined the We next performed chromatin immunoprecipitation effect of Tam67 on cyclin D1 expression in MCF-7 cells (ChIP) assays to demonstrate in vivo interaction between synchronized at G /G phase and released from 0 1 Tam67 and the AP-1 or E2F site within the cyclin D1 synchronization by serum stimulation. We confirmed promoter. Using cJun and cFos antibodies as positive that Tam67 expression suppressed cyclin D1 mRNA by controls, we found that cJun and cFos bound to the more than 50% in these cells (P 0.0004) (Figures 1a ¼ DNA fragments containing the AP-1 sites of either a and b). Tam67 reduced the expression of cyclin D1 positive control gene (collagenase/MMP-1) or the cyclin mRNA, while control cells still show a cyclic pattern D1 gene (Figure 3a). As shown using an anti-flag (Figure 1b). Tam67 also reduced the protein level of antibody in these ChIP experiments, flag-tagged Tam67 cyclin D1 (Figures 1c and d). Note that the maintained was found to bind the AP-1 sites of the control MMP-1 cyclic pattern of cyclin D1 protein expression still occurs and cyclin D1 promoters in doxycycline-untreated, thus suggesting that other mechanisms such as proteolysis Tam67-induced MCF-7 cells, but not in control cells still contribute to the degradation of cyclin D1. (Figure 3a). This suggests that Tam67 complex does bind to the AP-1 site in the cyclin D1 promoter, Tam67 suppresses cyclin D1 promoter activity supporting the conclusion from Figure 2 that AP-1 Since Tam67 suppressed cyclin D1 mRNA expression, maintains basal cyclin D1 promoter function. The ChIP we investigated whether the transcription factors that assays also show that the E2F1 protein, but not cJun or bind at the AP-1 site regulate cyclin D1 transcription. cFos, bound to the E2F site in the cyclin D1 promoter. Potential binding sites for AP-1, E2F, nuclear factor-kB To further verify the in vivo binding results, we (NF-kB) and Sp1 factors were found in the promoter performed in vitro electrophoretic mobility shift assay region of À2963 to þ 361 of the cyclin D1 promoter (EMSA) and supershift assays to determine direct or (Herber et al., 1994). To determine the cis-elements indirect interaction of Tam67 and E2F1 to responsible for the Tam67-induced inhibition, we respective binding sites. We found that Tam67 is examined the promoter activity for cyclin D1 gene in capable of binding the AP-1 site in the cyclin D1 the presence and absence of Tam67 expression. We promoter (Figure 3b, lane 7), consistent with our results found that deletion of the AP-1 site at –934/À928 of reporter assays showing suppression of the basal reduced cyclin D1 promoter activity (in the absence of cyclin D1 promoter activity (Figure 2). Tam67 did not Tam67) by approximately 40% (Figure 2a). Deletion of bind the E2F element of the cyclin D1 promoter at the NF-kB site at À840/À831 did not alter the promoter À726/À719 (Figure 3b, lanes 13 and 14). As expected, activity significantly. However, induction of Tam67 still E2F1 protein does bind the E2F site (Figure 3b, lanes 15 suppressed the promoter activity in constructs lacking and 16). We also observed reduced binding of E2F1 the AP-1 and NF-kB sites. Only when the E2F site was protein to the E2F site in Tam67-expressing cells deleted did we observe loss of Tam67-induced suppres- (Figure 3b, lanes 10and 14). This result suggests that sion of promoter activity. This result suggests that AP-1 Tam67 either affects E2F expression or affects E2F transcription factors bind to the AP-1 site to regulate DNA binding activity. Tam67 did not bind the NF-kB basal cyclin D1 expression and that Tam67 suppresses and Sp1 sites (Figure 3c, lanes 6 and 7, and lanes 13 and cyclin D1 expression in part by affecting the proteins 14, respectively). Taken together, the ChIP and super- that bind the E2F site in the cyclin D1 promoter. shift data suggest that the AP-1 inhibitor, Tam67,

Oncogene AP-1 blockade downregulates cyclins and E2F1 Q Shen et al 368 a Hrs released from synch. 0 hr 6 hr 12 hr 18 hr 24 hr 36 hr 48 hr Dox + - + - + - + - + - + - + - Cyclin D1

L32 RPAs

Dox (+) Dox (-) b 1.00 p = 0.0004

0.75

0.50

0.25

Relative Cyclin D1 mRNA Level 0.00 06 1218243648 Hours Released from Synchronization at G0/G1 Phase

c Hrs released from synch. 0 hr 6 hr 12 hr 18 hr 24 hr 36 hr 48 hr

Dox + - + - + - + - + - + - + - Cyclin D1

β-actin Westerns

d 350 Dox (+) Dox (-) p = 0.0045 300

250

200

150

100

50

Relative Cyclin D1 Protein Level 0 0 hr 6 hr 12 hr 18 hr 24 hr 36 hr 48 hr Hours Releasing from Synchronization at G0/G1 Phase Figure 1 AP-1 blockade suppresses cyclin D1 expression. MCF-7 Tet-Off Tam67 cells were cultured in the presence/absence of doxycycline for 5 days and then synchronized by 48 h serum starvation, and cells were then plated at the G0/G1 phase. Total RNA was prepared from cells released from synchronization. (a) The cyclin D1 mRNA level was measured by RNase protection assay (RPA). L32 served as loading control. (b) Data represent relative cyclin D1 mRNA levels from three independent RPAs. (c) Total cell lysates were prepared from MCF-7 Tet-Off Tam67 cells and cyclin D1 protein was determined by western blot. b-Actin levels served as loading control. (d) Relative cyclin D1 protein levels based on three independent experiments.

suppresses cyclin D1 expression by directly attenuating Tam67 expression leads to reduced E2F1 and E2F2, but cyclin D1 promoter activity through blockade at the not E2F4 mRNA levels as assessed by RNase protection AP-1 site, and indirectly reducing binding of E2F1 to assays (RPAs) (P ¼ 0.01 for E2F1 and E2F2) (Figures the E2F site in the cyclin D1 promoter. 4a–d). The E2F dimerizing partner, DP1, was also suppressed at the mRNA level (Figures 4a and e) (P ¼ 0.0004), while DP2 was not (data not shown). AP-1 blockade suppresses E2F1, E2F2 and DP1, but not Protein levels for E2F1, E2F2 and DP1 were reduced E2F4 expression upon AP-1 blockade, while the expression of E2F4 was We next determined whether reduced binding of E2F1 upregulated (Figures 4f and i). Thus, AP-1 blockade by to the E2F site was due to decreased E2F1 protein Tam67 differentially regulates the expression of E2F1, expression in AP-1 blocked MCF-7 cells. We found that E2F2, E2F4 and their dimerizing partner DP1, resulting

Oncogene AP-1 blockade downregulates cyclins and E2F1 Q Shen et al 369 +1 a AP-1NFκB E2F AP-2 E2F SP-1

-934/-928 -726/-719

Required for cyclin D1 No Tam67 repression by Tam67 Tam67 induced

0241 3 Relative Luciferase Activity

b

X

NoTam67 X Tam67 induced

X

039 6 12 15 Relative Luciferase Activity

c

Addition No Tam67 Tam67 induced

010155 Relative Luciferase Activity

Figure 2 Promoter activity of cyclin D1 gene. MCF-7 Tet-Off Tam67 cells were cultured in the presence/absence of doxycycline (Dox) for 5 days and co-transfected with the deletion constructs of cyclin D1 promoter and dual luciferase/renilla pRL-TK vector. Luciferase activity was measured and normalized with the renilla activity. (a) Deletion of the activating protein-1 (AP-1) site at À934/À928 partially reduced promoter activity, while deletion of the E2F site at À726/À719 resulted in the loss of Tam67-induced inhibition of promoter activity. (b) Promoter activity after mutation of the E2F sites in the cyclin D1 promoter. (c) Addition of the fragment containing the E2F site at À726/À719 back to the D1D-543pXP2 vector. Rescue of the E2F site at À726/À719 restored Tam67-induced repression of the cyclin D1 promoter activity. in modulated expression of E2F-dependent genes (DeGregori et al., 1995; Botz et al., 1996). We found (including cyclin D1). that Tam67 suppressed cyclin E mRNA at all time points examined, showing a 30–60% suppression (P ¼ 0.0005) (Figures 5a and b). Protein levels for cyclin E2F target genes are regulated upon AP-1 blockade E were reduced as well (Figures 5c and d). These results We next examined the expression of E2F-downstream suggest that the reduction of E2F1 activity induced by genes. Cyclin E is a direct target gene of E2F1 Tam67 leads to reduced expression of cyclin E.

Oncogene AP-1 blockade downregulates cyclins and E2F1 Q Shen et al 370 +1 a Cyclin D1 gene AP-1 E2F promoter -934/-928 -726/-719

MMP-1 gene Cyclin D1 gene Cyclin D1 gene Inputs AP-1 site ChIP AP-1 site ChIP E2F site ChIP AP-1 blockade No Yes No Yes No Yes No Yes

cJun

Antibody cFos Pulldown Anti-Flag ChIPs

E2F1

AP-1 site E2F site

b AP-1 site probe +++++++ ------E2F site probe ------+++++++++ Dox-treated NE (control) - +++--- - +-+++ - - - AP-1-blocked NE - - + --+ + -+----+++ 100X cold probe -+-- - + - -+---- + -- Anti-Flag Ab -+- - - - + -+----- + - Anti-E2F1 Ab ------+------+ Lanes 123 4 5 6 7 8 9 10 11 12 13 14 15 16 Loading wells

AP-1/DNA complex E2F/DNA

EMSA/ complex

Supershifts Unbound labeled probe

NF-kBsite Sp1 site

c NF-kB site probe +++++++ ------Sp1 site probe ------+ -+++++++ + Dox-treated NE (control) ----+++ - +----+++ AP-1-blocked NE -+---++ --+ - + -- + + 100X cold probe --+---+ - --+ +-- -- Anti-Flag Ab ---+---+ --- -+- + - Anti-Sp1 Ab ------+---+ Lanes 1234567 8910111213141516 Loading wells Sp1/DNA NF-kB/DNA complex complex EMSA/

Supershifts Unbound labeled probe

Figure 3 In vivo and in vitro interaction of the cyclin D1 promoter elements and activating protein-1 (AP-1), E2F, nuclear factor-kB (NF-kB) and Sp1 factors. (a) Chromatin immunoprecipitation (ChIP) analysis of a known AP-1 regulated gene, MMP-1, at an AP-1 site contained within MMP-1 promoter region À242/À3 (positive control), and the cyclin D1 promoter at the AP-1 (À934/À928) and E2F binding sites (À726/À719). Antibodies used were cJun, cFos, flag-tagged Tam67 and E2F1. (b) Electrophoretic mobility shift assay (EMSA) and supershift was performed using the ds oligonucleotide probes containing AP-1 binding site (À934/À928 in cyclin D1 promoter) and E2F1 binding site (À726/À719 in cyclin D1 promoter) in the presence of nuclear extract (NE) from doxycycline (Dox)- treated normal MCF-7 cells or Dox-untreated, Tam67-induced cells. Specific shifted DNA complexes are shown with a bracket and supershifted complex are shown with an arrow to the right of each panel. (c) EMSA and supershift was performed using the NF-kB binding site (À838/À830in cyclin D1 promoter) and Sp-1 binding site ds oligonucleotide probes ( À113/À102 in cyclin D1 promoter) in the presence of NE from Dox( þ ) or Dox(À) cells. One representative autoradiograph from three independent experiments is shown for b and c.

Oncogene AP-1 blockade downregulates cyclins and E2F1 Q Shen et al 371 We also discovered that the expression of other and cathepsin D genes are directly regulated by E2F1 cyclins regulated by E2F was also altered by expression (DeGregori et al., 1995; Lavia and Jansen-Durr, of Tam67. and (both are E2F- 1999). We found reduced expression of these genes upregulated genes) were suppressed upon Tam67- in the AP-1-blocked MCF-7 cells (Figure 7a), while induced AP-1 blockade (Figure 6a), while cyclins G2 Bcl-2 and RARg genes were upregulated in the and I, both E2F-downregulated genes, were upregulated presence of AP-1 blockade, as shown by quantitative in the presence of AP-1 blockade (Figure 6b). In reverse-transcriptase–PCR (QRT–PCR) (Figure 7b). addition, we identified that cyclins C, G1 and H were As a control, the TOB1 gene did not show signi- not regulated in the presence of Tam67 (Figure 6c). ficant changes in mRNA levels (Figure 7c). Thus, Results from these studies support an important role of we have demonstrated that AP-1 blockade by E2F1 in mediating the effects of AP-1 blockade induced Tam67 causes downregulation of cyclins, in addition by Tam67. to genes involved in DNA replication and repair We next examined the expression of other down- (for example, PCNA). In addition, the E2F downstream stream target genes of E2F. The b-, PCNA target genes were also regulated upon AP-1 blockade.

a Hrs released from synch. 0 hr 6 hr 12 hr18 hr 24 hr 36 hr 48 hr

Dox + -+ - + - + - + - + - + -

E2F1

E2F2

E2F4 RPAs

DP1

L32

b c 0.200 0.12 E2F1 Dox (+) Dox (-) E2F2 Dox (+) Dox (-) 0.175 p = 0.011 p = 0.0092 0.10 0.150 0.08 0.125 0.100 0.06 0.075 0.04 0.050 0.02 0.025 Relative E2F2 mRNA Level Relative E2F1 mRNA Level 0.000 0.00 06 1218243648 06 1218243648 Hours Released from Synchronization at G0/G1 Phase Hours Released from Synchronization at G0/G1 Phase

de 0.200 1.2 E2F4 Dox (+) Dox (-) DP1 Dox (+) Dox (-) 0.175 p = 0.3692 p = 0.0004 1.0 0.150 0.8 0.125 0.100 0.6 0.075 0.4 0.050 0.2 0.025 Relative DP1 mRNA Level Relative E2F4 mRNA Level 0.000 0.0 06 1218243648 06 1218243648 Hours Released from Synchronization at G0/G1 Phase Hours Released from Synchronization at G0/G1 Phase Figure 4 AP-1 blockade regulates E2F and DP1 expression. MCF-7 Tet-Off Tam67 cells were cultured as in Figure 1. (a) The E2F1, E2F2, E2F4 and DP1 mRNA levels were measured by RNase protection assay (RPA). A representative autoradiograph is shown. (b–e) Relative mRNA levels of E2F1, E2F2, E2F4 and DP1 based on three separate experiments. (f–j) MCF-7 Tet-Off Tam67 cells were treated same as (a–e). E2F1, E2F2, E2F4 and DP1 proteins were determined by western blot. Data shown are the average of three independent experiments.

Oncogene AP-1 blockade downregulates cyclins and E2F1 Q Shen et al 372 f Hrs released from synch. 0 hr 6 hr 12 hr18 hr 24 hr 36 hr 48 hr

Dox + - + - + - + - + - + - +-

E2F1

E2F2

E2F4 Westerns

DP1

β-actin

gh 550 E2F1 Dox (+) Dox (-) 250 E2F2 Dox (+) Dox (-) 500 p = 0.0389 p = 0.0719 450 200 400 350 150 300 250 200 100 150 100 50

50 Relative E2F2 Protein Level Relative E2F1 Protein Level 0 0 0 hr 6 hr 12 hr 18 hr 24 hr 36 hr 48 hr 0 hr 6 hr 12 hr 18 hr 24 hr 36 hr 48 hr Hours Released from Synchronization at G0/G1 Phase Hours Released from Synchronization at G0/G1 Phase

i j 250 E2F4 Dox (+) Dox (-) 400 DP1 Dox (+) Dox (-) p = 0.0034 p = 0.0118 200 300 150 200 100

100 50 Relative DP1 Protein Level Relative E2F4 Protein Level 0 0 0 hr 6 hr 12 hr 18 hr 24 hr 36 hr 48 hr 0 hr 6 hr 12 hr 18 hr 24 hr 36 hr 48 hr Hours Released from Synchronization at G0/G1 Phase Hours Released from Synchronization at G0/G1 Phase Figure 4 Continued.

These results support a role for AP-1 in regulating cell downregulation (shown schematically in Figure 8). cycle at several checkpoints. First, AP-1 blockade suppresses basal cyclin D1 activity which is controlled by the consensus AP-1 binding site within the cyclin D1 promoter. Second, suppression of cyclin D1 expression also occurs via an indirect Discussion mechanism by downregulation of the E2F1 and E2F2 transcriptional factors, leading to reduced binding of In these studies, we investigated the molecular mechan- E2F protein to the E2F binding site within the cyclin D1 isms by which the AP-1 inhibitor, Tam67, suppresses promoter. Third, AP-1 blockade suppresses the expres- breast cancer cell growth. We have shown that AP-1 sion of E2F dimerizing partner, DP1, thus reducing blockade suppresses cyclin D1 at the mRNA and the availability of the E2F/DP1 complex and modulat- protein levels, and have defined the mechanisms of this ing the expression of E2F1-targeted genes. Finally,

Oncogene AP-1 blockade downregulates cyclins and E2F1 Q Shen et al 373 a Hrs released from synch. 0 hr 6 hr 12 hr18 hr 24 hr 36 hr 48 hr Dox + - + - + - + - + - + - + - Cyclin E

L32 RPAs

Dox (+) Dox (-) 7 b p = 0.0005 6 5 4 3 2 1

Relative Cyclin E mRNA Level 0 06 1218243648 Hours Released from Synchronization at G0/G1 Phase

c Hrs released from synch. 0 hr 6 hr 12 hr 18 hr 24 hr 36 hr 48 hr

Dox + - + - + - + - + - + - + -

Cyclin E

β-actin Westerns

d 400 Dox (+) Dox (-) p = 0.0043 300

200

100

Relative Cyclin E Protein Level 0 0 hr 6 hr 12 hr 18 hr 24 hr 36 hr 48 hr Hours Releasing from Synchronization at G0/G1 Phase

Figure 5 Cyclin E, a direct E2F1-targeted gene, was suppressed in serum starvation-synchronized MCF-7 cells. MCF-7 Tet-Off Tam67 cells were cultured same as in Figure 1. (a) The cyclin E mRNA level was measured by RNase protection assay (RPA). (b) Data shown are the average cyclin E mRNA of three independent RPAs. (c) Protein levels of cyclin E were determined by western blot. (d) Data shown are the average levels of cyclin E protein from three independent experiments. suppression of the expression of E2F1-targeted genes regulates cyclin D1. Our results herein demonstrate further enhances the growth-suppressive effect of several mechanisms that may coexist to allow modula- Tam67 (Figure 8). tion of cell cycle regulators such as cyclin D1 and cyclin The role of cyclin D1 in cancer cell growth has been E. Thus, mechanisms by which AP-1 regulates G1 well documented and reviewed (Donnellan and Chetty, cyclins are complicated. Interaction of the AP-1 factor 1998). Signals from growth factors and mitogens and the AP-1 binding site at the cyclin D1 promoter was eventually activate AP-1, which in turn activates cyclin shown here to directly contribute to the basal promoter D1. Cyclin D1 regulation by AP-1 has also been activity of the cyclin D1 gene. An important additional reported in other breast cancer cells, such as T47D finding in these studies was that the attenuation of the and SKBR3 cells (Shah et al., 2002; Bu et al., 2005), and E2F1, E2F2 and DP1 expression by Tam67 caused a in silica-induced transformed embryonic lung fibroblasts negative regulation of cyclin D1 expression through the (Shen et al., 2006a). However, these previous studies E2F site in cyclin D1 promoter. Thus, both the basal have not fully described the mechanism by which AP-1 and activated cyclin D1 transcriptional activities are

Oncogene AP-1 blockade downregulates cyclins and E2F1 Q Shen et al 374 a E2F-upregulated cell cycle genes

Hrs released from synch. 0 hr 6 hr 12 hr18 hr 24 hr 36 hr 48 hr

Dox + - + - + - + - + - + - + - Cyclin A

Cyclin A1

Cyclin B RPAs

Cyclin D3

b E2F-downregulated cell cycle genes

Cyclin G2

Cyclin I RPAs

c Non-AP-1-regulated cell cycle genes

Cyclin C

Cyclin G1

Cyclin H RPAs

L32

Figure 6 Regulation of E2F-targeted and non-AP-1-regulated cell cycle modulators. MCF-7 Tet-Off Tam67 cells were cultured same as in Figure 1. RNase protection assays (RPAs) were performed to determine the mRNA levels. (a) Downregulation of cyclins A, A1, B and D3 in the presence of Tam67. This panel shows E2F-upregulated genes. (b) Upregulation of cyclins G1 and cyclin I in the presence of Tam67. This panel shows E2F-downregulated genes. (c) A panel of cyclins that were not regulated by Tam67.

a 50 b-myb Dox (+) Dox (-) 20 PCNA Dox (+) Dox (-) 120 cath D Dox (+) Dox (-) 110 p = 0.073 p = 0.016 p = 0.0015 100 40 15 90 80 30 70 10 60 20 50 40 5 30 10 20 Relative b-myb mRNA Level

Relative PCNA mRNA Level 10 0 0 0 Relative cathepsin D mRNA Level 0 6 1218243648 0 6 1218243648 0 6 1218243648 Hours Released from Synchronization at G0/G1 Phase Hours Released from Synchronization at G0/G1 Phase Hours Released from Synchronization at G0/G1 Phase

TOB1 b Bcl-2 Dox (+) Dox (-) c 3.5 p = 0.0379 40 RARγ 150 p = 0.4135 Dox (+) Dox (-) Dox (+) Dox (-) 3.0 p = 0.0012 30 2.5 100

2.0 mRNA Level

ϑ 20 1.5 50 1.0 10 0.5 Relative Bcl-2 mRNA Level Relative TOB1 mRNA Level Relative RAR 0.0 0 0 0 6 1218243648 0 6 1218243648 0 6 1218243648 Hours Released from Synchronization at G0/G1 Phase Hours Released from Synchronization at G0/G1 Phase Hours Released from Synchronization at G0/G1 Phase Figure 7 Expression of E2F-downstream genes was regulated by AP-1 blockade. MCF-7 Tet-Off Tam67 cells were cultured same as in Figure 1. QRT–PCR was performed to determine the mRNA level for the indicated genes. (a) Downregulation of E2F-upregulated genes, b-myb, PCNA and cathepsin D, in the presence of Tam67. (b) Upregulation of E2F-downregulated genes, Bcl-2 and RARg,in the presence of Tam67. (c) Non-E2F-targeted gene: TOB1 in the presence of Tam67.

suppressed by AP-1 blockade with the AP-1 inhibitor, transcription factors has been shown to be involved in Tam67. many aspects of fundamental cell cycle control. E2F1-3 Our results also demonstrate that E2F1, E2F2 and binds to cell cycle-regulated gene promoters and activate DP1 proteins are regulated by AP-1. The E2F family of transcription of these genes. E2F1 is a proliferation

Oncogene AP-1 blockade downregulates cyclins and E2F1 Q Shen et al 375 binding sites for E2F factors have been identified in a Fos or Tam67 large number of genes that control cell cycle and DNA Jun synthesis, including cdk 2 and 4, cyclin A, D and E, DNA polymerase, ribonucleotide reductase and PCNA as reviewed (Yamasaki, 1998). It was also shown that E2F1 E2F1 and E2F2 transcription factors regulate their own E2F2 transcriptional activity by binding to the E2F site in the DP1 promoters of E2F1, E2F2, pRB and other cell cycle regulators (Yamasaki, 1998). We have previously showed that AP-1 blockade reduced E2F promoter Fos or Tam67 Jun activity (Liu et al., 2004), and now have demonstrated E2F DP that the E2F1 and E2F2 proteins were decreased, AP-1 E2F consistent with the decreased promoter activity of these -934/-928 -726/-719 genes.

Cyclin D1 Promoter The importance of the AP-1 factor in regulating breast cancer cell growth has been demonstrated by our

E2F-targeted Genes laboratory and other investigators. Clinically, AP-1 cyclin D1 Cyclin E, E2, A, A1, B, D3 expression is related to the estrogen (ER)- Cyclin G2, I b-myb, PCNA negative phenotype and is a marker of poor prognosis, and AP-1 proteins are variably expressed in human breast cancer tissues (Chen et al., 1996). cJun, the predominant AP-1 protein and the major dimerizing partner, is associated with low ER expression and tamoxifen resistance (Johnston et al., 1999; Gee et al., 2000). Importantly, cJun regulates the growth of breast Cell cycle progression cancer cell lines. Its overexpression in MCF-7 breast & proliferation cancer cells produces a tumorigenic, invasive and hormone-resistant phenotype (Smith et al., 1999). Taken Figure 8 Schematic diagram of the mechanisms by which activating protein-1 (AP-1) regulates cyclin D and E2F factor to control cell together, these studies suggest that AP-1 plays a critical proliferation. role in regulating breast cell proliferation and may have an important role in the development of breast cancers, including ER-negative breast cancer and tamoxifen- marker of breast cells, 6.3% of cells seen in ductal resistant breast cancers. carcinoma in situ were E2F1 positive, 15.3% in invasive We have previously demonstrated that Tam67 blocks breast cancer while only 1.9% in normal breast tissue the cell cycle at the G0/G1 phase by reducing phospho- (Zhang et al., 2000). Also, E2F1 was expressed in 44% Rb levels and downregulating cyclin D, cyclin E, CDK2/4 of node-positive breast cancers and therefore serves as a and . We have now further characterized the effect of prognostic marker (Han et al., 2003). Therefore, based AP-1 blockade and shown that Tam67 suppresses the upon the results of our studies, it may be possible to cell cycle by downregulating cyclin D1 and transcription control breast cancer cell growth by modulating E2F factors E2F1 and E2F2, and their dimerization partner factors. E2F1-3 are generally considered as activators of DP1. These results suggest that targeting the AP-1 E2F-dependent gene expression, while E2F4 usually transcription factor results in cell cycle blockade, and functions as a repressor (Dimova and Dyson, 2005). Our that such targeted therapy may be useful for breast results are consistent with this notion, as we have shown cancer treatment. that E2F1 and E2F2 were significantly suppressed by Tam67-induced AP-1 blockade while E2F4 was upre- gulated. This is in line with the literature demonstrating Materials and methods that E2F4 is a functional antagonist of E2F1, 2 and 3 proteins. Interestingly, the dimerizing partner of E2F Cell culture and transfection factors, DP1, but not DP2, was differentially down- MCF-7 Tet-Off Tam67 clones #62, #67 and vector clones #1 regulated by AP-1 blockade. This would have reduced and #3 have been described previously (Liu et al., 2002). the available level of DP1 protein and eventually the E2F/DP complexes. This also suggests a hypothesis that Cyclin D1 promoter constructs different dimerizing partners may preferentially form Cyclin D1 deletion constructs were obtained from Dr specific dimer complexes with different E2F family RMu¨ ller (Herber et al., 1994). The D1D-848pXP2 and D1D- members. From our previous findings and the results 742pXP2 plasmids were used to mutate the E2F sites at À726/À719 (TTTCGGGC to TTTCATTC) and À139/À131 presented herein, it is evident that AP-1 blockade by (TTTGGCGC to TTTGATTC), using a QuickChange XL Tam67 causes global inhibition of G1 cyclins D and E, Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA, as well as other cell cycle regulators in G1 phase. USA). A 220bp fragment was amplified with the primer set The E2F transcription factors are key regulators of used in ChIP assay containing the E2F site at À726/À719 and cell proliferation (Muller and Helin, 2000). In fact, cloned into the linearized D1D-543pXP2 plasmid, in front of

Oncogene AP-1 blockade downregulates cyclins and E2F1 Q Shen et al 376 the 50 cyclin D1 promoter sequence. The selected promoter ChIP assay constructs were used in transient transfection experiments. ChIP assays were performed as described (Shang et al., 2000). Antibodies used were: cJun/AP-1, cat# PC06, Oncogene; cFos, Western blot analysis cat# PC05, Oncogene; Anti-Flag M2, cat# 3165, Sigma- Western blotting was conducted as previously described (Liu Aldrich (St Louis, MO, USA); E2F1, cat# 39601, Active et al., 2004). The antibodies were: cyclin D1 (sc-8396), cyclin E Motif. (sc-247), E2F2 (sc-633), E2F4 (sc-866) and DP1 (sc-610), from Santa Cruz Biotech Inc. (Santa Cruz, CA, USA) and E2F1 (#36901, Active Motif, CA, USA). Statistical analyses The mRNA and protein assays were statistically analysed using paired t-tests with GraphPad Prism 4 software (San RNase protection assay Diego, CA, USA). RPAs were performed as previously described (Liu et al., 2004).

Quantitative reverse-transcriptase–PCR Abbreviations Quantitative reverse-transcriptase (QRT)–PCR assays were conducted as previously described (Shen et al., 2006b). AP-1, activating protein-1; ChIP, chromatin immunoprecipi- tation; Dox, doxycycline; EMSA, electrophoretic mobility Reporter luciferase assay shift assay; ER, ; NE, nuclear extract; RPA, Cyclin D1 promoter activity was measured using the Dual- RNase protection assay; Tam67, a cJun dominant-negative Luciferase Reporter Assay (Promega, Madison, WI, USA) as mutant. described previously (Liu et al., 2002).

EMSA and supershift assay Acknowledgements EMSA and supershift assays were performed as previously described (Shen and Singh, 2004). Probes used were 20 bp We thank Shirley Pennington for her assistance in preparing oligos representing the AP-1, NF-kB, Sp-1 and E2F sites in this manuscript. This work was supported by a pilot grant human cyclin D1 promoter, synthesized by IDT Inc. (Coralville, from the Dan L Duncan Cancer Center at Baylor College of IA, USA). Medicine.

References

Angel P, Karin M. (1991). The role of Jun, Fos and the AP-1 complex Johnston SR, Lu B, Scott GK, Kushner PJ, Smith IE, in cell-proliferation and transformation. Biochim Biophys Acta 1072: Dowsett M et al. (1999). Increased activator protein-1 DNA 129–157. binding and c-Jun NH2-terminal kinase activity in human breast Botz J, Zerfass-Thome K, Spitkovsky D, Delius H, Vogt B, Eilers M tumors with acquired tamoxifen resistance. Clin Cancer Res 5: et al. (1996). Cell cycle regulation of the murine cyclin E gene 251–256. depends on an E2F binding site in the promoter. Mol Cell Biol 16: Lavia P, Jansen-Durr P. (1999). E2F target genes and cell-cycle 3401–3409. checkpoint control. Bioessays 21: 221–230. Brown PH, Alani R, Preis LH, Szabo E, Birrer MJ. (1993). Liu Y, Lu C, Shen Q, Munoz-Medellin D, Kim H, Brown PH. (2004). Suppression of oncogene-induced transformation by a deletion AP-1 blockade in breast cancer cells causes cell cycle arrest by mutant of c-jun. Oncogene 8: 877–886. suppressing G1 cyclin expression and reducing cyclin-dependent Bu X, Avraham HK, Li X, Lim B, Jiang S, Fu Y et al. (2005). Mayven kinase activity. Oncogene 23: 8238–8246. induces c-Jun expression and cyclin D1 activation in breast cancer Liu Y, Ludes-Meyers J, Zhang Y, Munoz-Medellin D, Kim HT, Lu C cells. Oncogene 24: 2398–2409. et al. (2002). Inhibition of AP-1 transcription factor causes blockade Chen TK, Smith LM, Gebhardt DK, Birrer MJ, Brown PH. (1996). of multiple signal transduction pathways and inhibits breast cancer Activation and inhibition of the AP-1 complex in human breast growth. Oncogene 21: 7680–7689. cancer cells. Mol Carcinog 15: 215–226. McDonnell SE, Kerr LD, Matrisian LM. (1990). Epidermal growth DeGregori J, Kowalik T, Nevins JR. (1995). Cellular targets for factor stimulation of stromelysin mRNA in rat fibroblasts requires activation by the E2F1 transcription factor include DNA synthesis- induction of proto-oncogenes c-fos and c-jun and activation of and G1/S-regulatory genes. Mol Cell Biol 15: 4215–4224. protein kinase C. Mol Cell Biol 10: 4284–4293. Dimova DK, Dyson NJ. (2005). The E2F transcriptional network: old Muller H, Helin K. (2000). The E2F transcription factors: acquaintances with new faces. Oncogene 24: 2810–2826. key regulators of cell proliferation. Biochim Biophys Acta 1470: Donnellan R, Chetty R. (1998). Cyclin D1 and human neoplasia. M1–M12. Mol Pathol 51: 1–7. Shah S, Pishvaian MJ, Easwaran V, Brown PH, Byers SW. (2002). The Gee JM, Barroso AF, Ellis IO, Robertson JF, Nicholson RI. (2000). role of cadherin, beta-catenin, and AP-1 in retinoid-regulated Biological and clinical associations of c-jun activation in human carcinoma cell differentiation and proliferation. J Biol Chem 277: breast cancer. Int J Cancer 89: 177–186. 25313–25322. Han S, Park K, Bae BN, Kim KH, Kim HJ, Kim YD et al. (2003). Shang Y, Hu X, DiRenzo J, Lazar MA, Brown M. (2000). Cofactor E2F1 expression is related with the poor survival of lymph node- dynamics and sufficiency in estrogen receptor-regulated transcrip- positive breast cancer patients treated with fluorouracil, doxorubicin tion. Cell 103: 843–852. and cyclophosphamide. Breast Cancer Res Treat 82: 11–16. Shen F, Fan X, Liu B, Jia X, Du H, You B et al. (2006a). Herber B, Truss M, Beato M, Muller R. (1994). Inducible Overexpression of cyclin D1-CDK4 in silica-induced transformed regulatory elements in the human cyclin D1 promoter. Oncogene cells is due to activation of ERKs, JNKs/AP-1 pathway. Toxicol 9: 2105–2107. Lett 160: 185–195.

Oncogene AP-1 blockade downregulates cyclins and E2F1 Q Shen et al 377 Shen Q, Singh P. (2004). Identification of a novel SP3 binding site in the Szabo E, Preis LH, Brown PH, Birrer MJ. (1991). The role of promoter of human IGFBP4 gene: role of SP3 and AP-1 in regulating jun and fos gene family members in 12-O-tetradecanoylphorbol-13- promoter activity in CaCo2 cells. Oncogene 23: 2454–2464. acetate induced hemopoietic differentiation. Cell Growth Differ 2: Shen Q, Zhang Y, Uray IP, Hill JL, Kim HT, Lu C et al. (2006b). The 475–482. AP-1 transcription factor regulates postnatal mammary gland Yamasaki L. (1998). Growth regulation by the E2F and DP development. Dev Biol 295: 589–603. transcription factor families. Results Probl Cell Differ 22: Smith LM, Wise SC, Hendricks DT, Sabichi AL, Bos T, Reddy P et al. 199–227. (1999). cJun overexpression in MCF-7 breast cancer cells produces a Zhang SY, Liu SC, Al-Saleem LF, Holloran D, Babb J, Guo X et al. tumorigenic, invasive and hormone resistant phenotype. Oncogene (2000). E2F-1: a proliferative marker of breast neoplasia. Cancer 18: 6063–6070. Epidemiol Biomarkers Prev 9: 395–401s.

Oncogene