PRIP Promotes Tumor Formation through Enhancing Serum-responsive Factor-mediated FOS Expression*

PRIP (peroxisome proliferator-activator receptor interacting protein) is a nuclear receptor coactivator required for mammary gland development. To understand the function of PRIP in breast tumorigenesis, we established a mammary tumor cell line with the PRIPLoxp/Loxp genotype. By knocking out the PRIP gene in the tumor cell line, we demonstrated that PRIP deficiency led to inhibited tumor formation without affecting tumor cell proliferation. The PRIP deficiency was associated with decreased cell invasion and migration capabilities. We found that PRIP deficiency substantially reduced FOS gene expression. A chromatin immunoprecipitation assay revealed that PRIP was recruited to the FOS promoter. In addition, we demonstrated that PRIP also directly up-regulated the FOS gene expression in human breast cancer cells. Promoter analysis showed that PRIP acted through serum-responsive factor to regulate FOS gene expression. Finally, by re-expressing the FOS gene, we confirmed that the inhibited tumor formation of PRIP-deficient tumor cells was due to reduced expression of the FOS gene.

We and others demonstrated that PRIP null mutation is embryonic lethal (15)(16)(17). To analyze the role of PRIP in mammary gland development, we generated mouse models with a null mutation of PRIP in mammary glands (18). Mammary glands with PRIP deficiency showed decreased ductal side branching and alveolar proliferation with relatively differentiated function intact. The ductal branching of mammary gland in response to estrogen treatment was attenuated in PRIP mutant glands. These studies demonstrated that PRIP plays critical roles in mammary gland development. In addition, the PRIP gene is amplified and overexpressed in a proportion of breast cancers (5), indicating that it may be involved in breast cancer development.
FOS is an immediate early gene belonging to a class of genes that are rapidly activated in response to intracellular signaling cascades (19). FOS and other members of the Fos family dimerize with Jun proteins to form AP-1 transcription factor complex (20). AP-1-regulated genes include critical modulators for invasion and metastasis; proliferation, differentiation, and survival; and angiogenesis (20). However, different cells often have a different set of genes controlled by AP-1, which leads to a different biological consequence due to increased AP-1 activity (20).
SRF is the transcription factor recognizing the serum-responsive element often found in the immediate early genes, including FOS (21). In addition to its role in cell proliferation, recent studies have revealed that SRF is a critical transcription factor regulating actin cytoskeleton and contractile genes (22).
To understand the function of PRIP in breast tumorigenesis, we established a mammary tumor cell line with the PRIP Loxp/Loxp genotype. By knocking out the PRIP gene in the tumor cell line, we demonstrated that PRIP promoted tumor formation by enhancing the SRF-mediated FOS gene expression.

Isolation of Mammary Epithelial Cells and Retroviral
Infection-The inguinal glands were collected from PRIP Loxp/Loxp mice. Mammary epithelial cells were separated from stromal cells by collagenase digestion and Percoll gradient centrifugation (23). 5 ϫ 10 6 Phoenix-Eco packaging cells were transfected with 15 g of Pbabe-H-rasV12 (provided by Bob Weinberg, MIT) through the calcium phosphate precipitation method.
Forty-eight hours after the transfection, viral supernatant was added to 5 ϫ 10 5 mammary epithelial cells with 8 g/ml Polybrene. After 5 h of incubation, the medium with retrovirus was replaced with fresh medium.
Knock-out of the PRIP Gene-Mammary tumor cells were infected with retroviruses expressing Cre-ERtam or control viruses. The cells were selected with 1 g/ml puromycin. The cells were then treated with tamoxifen (1 M) for 2 days.
Cell Proliferation Assay-Cells were seeded into 6-well plates at a density of 1 ϫ 10 4 cells/well with fresh DMEM plus 10% FBS. Viable cells were counted after trypan blue staining once a day for 4 days.
Soft Agar Colony Formation Assay-Tumor cells (2 ϫ 10 3 ) were suspended in a 1-ml mixture containing 0.35% (w/v) agar and DMEM supplemented with 10% FBS and layered on a 1-ml base of 0.7% (w/v) agar in DMEM with 10% FBS in a 6-well plate. Colonies were counted and photographed 2 weeks after plating. Each soft agar assay was performed in triplicate.
In Vivo Tumorigenicity Assay-1 ϫ 10 6 mammary epithelial cells were subcutaneously injected into 8-week-old BALB/c athymic nude mice. The maximal tumor diameter was determined by caliber measurements once a week up to the 4th week.
Migration and Invasion Assays-For migration assays, transwell chamber membranes (Costar) were coated with 15 g/ml collagen I. 1 ϫ 10 5 cells suspended in DMEM/bovine serum albumin were added to the upper chamber, and DMEM containing 5% FBS was added to the lower chamber. After incubating for 14 h, nonmigrating cells were removed from the upper chamber with a cotton swab. Cells that had migrated to the lower surface of the membrane were stained with crystal violet. Migration was quantified by counting cells/mm 2 using bright field optics. For invasion assays, chambers (8-m pore size; Costar) were coated with Matrigel diluted 1:2 with ␣-minimum Eagle's medium. 1 ϫ 10 5 cells were added to the chamber and allowed to invade for 14 h. Cells that had invaded to the lower surface of the filter were stained and counted. Each assay was repeated three times.
Microarray Hybridization-The tumor cells were used for the preparation of total RNA by the TRIzol (Invitrogen) method. Total RNA was used to make cRNA for hybridization with mouse genome (Affymetrix) as instructed. Microarray data analysis was performed with Affymetrix microarray software.
siRNA Transfection-siRNA smart pool oligonucleotides targeting human PRIP and control siRNAs were purchased from Dharmacon. Human breast cancer BT-474 cells were purchased from ATCC and maintained in DMEM containing 10% heat-inactivated FBS. The cells were transfected with control siRNA or PRIP siRNA using Lipofectamine 2000 reagent (Invitrogen) and cultured in phenol red-free minimum Eagle's medium containing 10% charcoal-stripped FBS. After 24 h, total RNA was then isolated using TRIzol reagent.
Chromatin Immunoprecipitation-Mammary tumor cells were maintained in phenol red-free modified Eagle's medium supplemented with 10% charcoal-dextran-stripped fetal bovine serum. Cells were cross-linked with 1% formaldehyde for 10 min at room temperature. After cells were collected, chromatin immunoprecipitation was performed as previously described (24), using antibodies against PRIP, SRF, or trimethylated H3K4 (Abcam, Cambridge, MA). The primer pair F1/R1 for the mouse FOS promoter is 5Ј-CGTCAATCCCTCCCTCC-TTT-3Ј and 5Ј-AGGATTTCGGAGATGGTCCC-3Ј. Sequences for control primer pair F2/R2 upstream of the mouse FOS promoter are 5Ј-CTTCTCTGCACTGATTTGGG-3Ј and 5Ј-GGTCATTGTCCAGCAATCTG-3Ј. Sequences for the primer F1/R1 for the human FOS promoter are 5Ј-TCAATCCCTCC-CCCCTTACA-3Ј and 5Ј-TCTAAACGTCACGGGCTCAA-3Ј. Sequences for the primer F2/R2 located upstream of the human FOS promoter are 5Ј-ACTGCTACCCTGTAAGC-TCG-3Ј and 5Ј-CTTAACTCTTTCCTTCTGGC-3Ј. In all cases, PCR was performed with a serial dilution of input and various cycles (29 -35 cycles) to ensure that amplification was maintained in the linear range. Each assay was repeated three times. Representative results were presented.
Immunoprecipitation-Nuclear extracts from Ras-induced tumor cells (100 g) were immunoprecipitated by anti-SRF (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) or control IgG in the buffer (20 mM Tris-HCl, pH 7.9, 180 mM KCl, 0.2 mM EDTA, 0.05% Nonidet P-40, 0.5 mM phenylmethylsulfonyl fluoride, and 1 mM dithiothreitol). After extensive washing using the same buffer, the bound proteins were subjected to Western blot analysis using anti-PRIP.
Electrophoretic Mobility Shift Assay-The electrophoretic mobility shift assay was performed according to standard protocols. Nuclear extracts were prepared from wild type tumor cells. The double-stranded SRE oligonucleotide (5Ј-AGGATGTCCATATTAGGACATCTG-3Ј) was end-labeled using T4 polynucleotide kinase and [␥-32 P]ATP. 5 g of nuclear extracts were incubated for 20 min at room temperature with 32 P-labeled SRE probe. DNA-protein complexes were resolved on a 6% TBE polyacrylamide gel. Gels were dried and autoradiographed at Ϫ80°C. For supershift experiments, nuclear extracts were incubated with anti-SRF or anti-PRIP in binding buffer for 45 min at room temperature before 32 P-labeled probe was added to the binding mixture.

Generation of Stable Cell Lines Expressing FOS and Tumor
Formation-The PRIP Ϫ/Ϫ cells were transfected with 10 g of pcDNA3-FLAG-Fos (Addgene) using Lipofectamine 2000. After selection with G418 (300 g/ml), the individual clones  were picked up, expanded, and examined for Fos expression. The PRIP ϩ/ϩ cells, PRIP Ϫ/Ϫ cells not expressing exogenous Fos, and PRIP Ϫ/Ϫ cells expressing exogenous Fos were injected subcutaneously into nude mice. The maximal tumor diameter was determined by caliber measurements once a week up to the 4th week.

PRIP Is Not Required for the Proliferation of Ras-induced
Mammary Tumor Cells-PRIP LoxP/LoxP mammary cells were isolated and infected with a retrovirus expressing an oncogenic form of the H-ras gene, H-rasV12. Two weeks later, individual clones were picked up and expanded. Many of the cells underwent senescence followed by apoptosis, whereas some of the cells proliferated. The process continued for about 4 months. Finally, a cell line emerged with vigorous growth. The cell line is estrogen receptor-negative and expresses keratin 18 but not smooth muscle actin, indicating that the cell line was derived from a luminal epithelial cell. The cells also formed colonies when cultured in soft agar. In addition, we found that the cells generated invasive tumors in nude mice.
To mutate the PRIP gene, the Ras-induced tumor cells were infected with a retrovirus expressing Cre-ERtam or a control retrovirus. After selection with puromycin, the cells were treated with tamoxifen for 2 days to activate the Cre recombinase and induce the knock-out of the PRIP gene. The deletion of the PRIP gene in mammary epithelial cells was demonstrated by PCR using primers specific for the deleted and LoxP-PRIP gene (Fig. 1A). RT-PCR showed that PRIP mRNA from tumor cells with deleted PRIP gene was absent (Fig. 1B). The successful deletion indicated that PRIP is not essential for the growth of Ras-induced mammary tumor cells. To find if PRIP deficiency affects the cell proliferation, PRIP Ϫ/Ϫ and PRIP ϩ/ϩ tumor cells were inoculated at equal densities, and their numbers were counted every day for 5 days. The PRIP Ϫ/Ϫ and wild type tumor cells showed a similar rate of proliferation (Fig. 1C).
Loss of PRIP Inhibits Tumor Formation-We subcutaneously injected PRIP Ϫ/Ϫ or PRIP LoxP/LoxP Ras-induced tumor cells into nude mice for tumor formation. Surprisingly, the tumor formed by PRIP-deficient tumor cells grew far more slowly than wild type tumor cells ( Fig. 2A). To determine the potential influence of PRIP on the anchorage-independent growth, we inoculated PRIP-deficient tumor cells and wild type tumor cells into soft agar. The PRIP-deficient tumor cells showed both reduced number and size of colonies compared with wild type tumor cells (Fig. 2B). Since PRIP expression had no significant effect on the growth rate of tumor cells on plastic, we tested if loss of PRIP affects the cell migration and invasion capability. PRIP-deficient cells and wild type cells were plated onto Matrigel-coated transwell chambers to measure invasion or onto collagen-coated transwell chambers to measure migration. The loss of PRIP significantly decreased migration and invasion by mammary tumor cells (Fig. 2C).
Altered Gene Expression by Knock-out of PRIP-In an effort to identify potential PRIP target genes that might reveal the mechanism by which PRIP regulates tumor formation, a microarray analysis was performed with wild type and PRIP Ϫ/Ϫ  Ras-induced tumor cells. We identified a total of 37 genes down-regulated with the loss of PRIP (Table 1) and a total of 35 genes up-regulated with the loss of PRIP ( Table 2). The expression of Csh1 (chorionic somatomammotropin hormone 1 (placental lactogen)) (25) was increased more than 40-fold. Also increased was the expression of Xist, which is involved in X chromosome inactivation (26,27). Among the down-regulated genes are six histone genes and trp53. Notably, the FOS and FOSB were markedly decreased with the loss of PRIP. Both FOS and FOSB genes are immediate responsive genes (20). The promoters of FOS and FOSB genes share similar regulatory elements. Although FOS and FOSB mRNA showed similar -fold decrease with PRIP deficiency, the level of FOS mRNA was about 20 times as that of FOSB. The microarray data showed that the FOS gene was one of most highly expressed transcription factors in this cell line.

PRIP Is Required for Expression of the FOS Gene-
The decreased expression of FOS and FOSB due to loss of PRIP was validated by real time RT-PCR (Fig. 3A). BT474 is breast cancer cell line with overexpression of PRIP due to the amplification of the PRIP gene. To find if PRIP is required for FOS and FOSB expression in human breast cancers, BT474 cells were trans-fected with siRNAs targeting PRIP, which reduced PRIP transcript at least 75% compared with controls (Fig. 3B). Real time PCR showed that FOS and FOSB expression was decreased with the suppression of PRIP expression (Fig. 3B). Similar to Ras-induced tumor cells, suppression of PRIP expression also inhibited the invasion (Fig. 3C) and migration (Fig. 3D) capability of BT474 cells.
Direct Recruitment of PRIP to the FOS Promoter-ChIP assays were performed to find if PRIP is directly recruited to the FOS gene promoter in vivo. Formaldehyde-cross-linked chromatin-protein complexes from wild type tumor cells or PRIP Ϫ/Ϫ tumor cells were immunoprecipitated using antibodies against PRIP and analyzed by semiquantitative PCR. We used two different PCR primer sets, which span two different regions of the FOS promoter. Primer set F1/R1 (Ϫ348 to Ϫ50 bp) spans the region containing the SRE site, c-Sis-inducible element site, FOS activator protein 1 site, and Cre site, whereas primer set F2/R2 (Ϫ2840 to Ϫ2580 bp) is located upstream of promoter containing no identified regulatory elements. We observed occupancy of PRIP at the FOS promoter containing regulatory elements and its absence upstream of promoter in wild type tumor cells (Fig. 4A). As a negative control, no ampli- fication of the FOS promoter was observed in PRIP Ϫ/Ϫ tumor cells. The PRIP is a component of a complex carrying histone methyltransferase activity. We performed chromatin immunoprecipitation to find if the recruitment of PRIP alters the methylated status of histone in the FOS promoter. We observed a marked increase of H3K4 trimethylation in the promoter of FOS in wild type tumor cells compared with PRIP Ϫ/Ϫ tumor cells (Fig. 4A).
We then examined if PRIP is recruited to the FOS gene promoter in human breast cancer cell line BT474. ChIP assays were performed with anti-PRIP. We observed amplification with the primer set covering the regulatory elements of the FOS gene and no amplification with the primer set located upstream of the FOS promoter (Fig. 4B), indicating that PRIP is also recruited to the FOS gene in human breast cancers. Similarly, we found that H3K4 trimethylation in the human FOS promoter was decreased in BT474 cells with suppressed expression of PRIP (Fig. 4B).
PRIP Enhances FOS Promoter Activity through Serum Response Factor-To determine the mechanism by which PRIP regulates FOS expression, we analyzed whether PRIP regulates the promoter activity of the FOS gene. We transfected PRIP Ϫ/Ϫ or wild type cells with FOS promoter-luciferase construct. As shown in Fig. 5A, the promoter activity was dramatically decreased in PRIP Ϫ/Ϫ cells compared with wild type cells. The re-expression of PRIP increased the FOS promoter activity in a dose-dependent manner, suggesting that PRIP enhances FOS promoter activity (Fig. 5B).
The FOS promoter is under the control of several transcription factors, including signal transducers and activators of transcription, AP-1, Cre, and SRF (28). PRIP has been reported to promote the transcriptional activity of Cre, AP-1, and SRF (3,14). To find which factor PRIP acts on to regulate FOS gene expression, we transfected the reporter gene with the FOS promoter carrying different regulatory elements and PRIP expression vector into PRIP-deficient tumor cells. PRIP enhanced the FOS promoter activity in the absence of the c-Sis-inducible element, which interacts with signal transducers and activators of transcription (Fig. 5C). When the SRE was not included in the reporter gene, PRIP had minimal effect on the FOS promoter, which still contained Cre and the FOS activator protein 1 site, indicating that PRIP acts through transcription factor SRF to promote FOS gene expression. To confirm the role of SRE in mediating activation of the FOS gene by PRIP, the SRE in the Fos-Luc reporter was mutated. As expected, the mutation of SRE abolished the enhancement of FOS promoter activity by PRIP (Fig. 5D). To explore if PRIP directly interacts with SRF, we performed immunoprecipitation with anti-SRF or control serum. The precipitates were subjected to Western blot analysis with anti-PRIP. We found that anti-SRF rather than control serum pulled down PRIP (Fig. 5E), suggesting that SRF interacts with PRIP in intact cells. The interaction of PRIP with SRF on SRE was examined by gel shift assays (Fig.  5F). Two complexes were formed with nuclear extracts from Ras-induced tumor cells. Unlabeled SRE competitor abolished both bands, indicating that both complexes resulted from the specific binding of SRF to SRE. Anti-SRF caused the supershift for both complexes, confirming that both complexes contained SRF. The inclusion of anti-PRIP led to the supershift of the band with slower mobility, indicating that this complex was formed by PRIP, SRF, and SRE. Sequential ChIP was performed to address if PRIP and SRF form a complex on the FOS promoter in mammary tumor cells. The eluated chromatin precipitated by anti-SRF was subjected to a second immunoprecipitation with anti-PRIP or control IgG. There was specific enrichment of DNA corresponding to the FOS promoter compared with IgG controls (Fig. 5G). However, we did not observe any specific enrichment of DNA 2.5 kb upstream of the FOS promoter.
Re-expression of FOS Promotes the Tumor Formation-Given that the FOS gene is highly expressed in Ras-induced tumor cells and is substantially down-regulated with PRIP deficiency, we further tested if reduced FOS expression leads to attenuated tumor formation by PRIP Ϫ/Ϫ tumor cells. We re-expressed FOS in the PRIP Ϫ/Ϫ tumor cells (Fig. 6A). Two clones expressing FOS at levels similar to that from wild type tumor cells were inoculated into nude mice for tumor formation. The PRIP Ϫ/Ϫ cells expressing exogenous FOS showed tumor formation at a growth rate similar to that by wild type tumor cells (Fig.  6B), confirming that reduced FOS expression is responsible for the decreased tumor formation as a result of PRIP deficiency.

DISCUSSION
To understand the role of PRIP in mammary tumorigenesis, we created a mammary tumor cell line carrying a conditional flox PRIP gene. By deleting the PRIP gene, we found that PRIP is not essential for the proliferation of mammary tumor cells in vitro on plates. However, PRIP is important for in vivo invasive growth, since loss of PRIP decreases the tumor cell migration and invasion capability. Further studies found that PRIP is required for the expression of the FOS gene by interacting with the SRF factor and promoting its activity.
The PRIP gene is frequently amplified and overexpressed in breast cancers. In addition to Ras-induced tumor cells, we found that PRIP is also required for FOS expression in human breast cancer BT474 cells. BT474 cells, which contain no Ras mutation but possess an oncogenic mutation on PIK3CA, carry an amplified PRIP gene with overexpression of PRIP (3). There-fore, PRIP promotes breast tumorigenesis at least partly by enhancing the FOS gene expression in human breast tumors with PRIP overexpression.
SRF also regulates other immediate response genes (20). Among these genes, we found that JUNB, EGR1, CYR61, and NURR1 are expressed in the Ras-induced tumor cells (data not shown). However, the expression of these genes is not affected due to the loss of PRIP expression. It appears that PRIP regulates gene expression through SRF in a gene-specific fashion. The SRF activity is affected by cofactors mainly including members of the ternary complex factor family of Ets domain proteins and the myocardin-related transcription factors (29). These two families of SRF cofactors are regulated by separate signaling pathways and control the SRF targets differentially. One possibility is that one of these collaborating transcription factors is required for PRIP action.
The FOS promoter is regulated by a number of transcription factors, including AP-1, signal transducers and activators of transcription, Cre, and SRF. PRIP has been shown to increase the transcriptional activity of AP-1, Cre, and SRF using minimal enhancer-containing reporters in a transient transfection assay. However, we found that PRIP acted through SRF to regulate FOS gene expression while having no effect on a truncated FOS promoter containing binding sites for AP-1 and Cre. The functions of PRIP on AP-1 and Cre, just like SRF, seem to be affected by other regulatory elements of target genes.
Notably, microarray studies revealed that the expression of CSH1, which is normally expressed in placenta, was highly increased in PRIP-deficient tumor cells. PRIP is a component of the complex ASCOM, which includes MLL3 or MLL4. MLL3 and MLL4 are involved in epigenetic regulation of genes. It is likely that loss of PRIP disrupts the epigenetic regulation of certain genes, which leads to aberrant gene expression. Microarray analysis also showed that the expression of six histone genes was decreased with the loss of PRIP. It is consistent with the notion that histone genes are regulated coordinately (30). Despite reduced expression of six histone genes, the proliferation of the tumor cells is not affected. PRIP could serve as a coactivator for certain transcriptional factors regulating histone expression.
The FOS gene plays an important role in tumorigenesis. Depending on the cell types, FOS may promote cell prolifer-  ). E, the direct interaction between PRIP and SRF in vivo. The nuclear extract from wild type tumor was prepared and precipitated with anti-SRF or control IgG. The precipitate was subject to Western blot analysis using anti-PRIP. F, PRIP and SRF form a complex on SRE. Gel shift was performed with 32 P-labeled SRE, nuclear extracts from Rasinduced tumor cells, and antibodies against SRF or PRIP. Lane 1, no nuclear extract; lane 2, nuclear extract plus anti-SRF; lane 3, nuclear extract; lane 4, nuclear extract plus anti-PRIP; lane 5, nuclear extract plus a 50-fold molar excess of unlabeled competitor SRE oligonucleotide. G, sequential ChIP assay demonstrating PRIP and SRF protein complexes over SRE in the FOS promoter. The first-step ChIP was performed with Ras-induced tumor cells and anti-SRF. The second step ChIP was carried out with the eluates of the initial ChIP, anti-PRIP, and IgG. PCR was performed using primer set F1/R1 and F2/R2, as described above.
ation, migration and invasion, and angiogenesis. In breast cancer, we found that the expression of Fos is critical for the invasive growth of Ras-induced tumor cells, and PRIP is critical for FOS gene expression. Given that increased FOS gene expression is commonly present in human cancers, including breast cancers, inhibiting PRIP activity could be a potential way to reduce FOS activity for the breast cancer treatment.