Biochemical Journal

Research article

PLU1 histone demethylase decreases the expression of KAT5 and enhances the invasive activity of the cells

Masakazu Yoshida , Akihiko Ishimura , Minoru Terashima , Zanabazar Enkhbaatar , Naohito Nozaki , Kenji Satou , Takeshi Suzuki

Abstract

PLU1 is a candidate oncogene that encodes H3K4 (Lys4 of histone H3) demethylase. In the present study, we found that ectopic expression of PLU1 enhanced the invasive potential of the weakly invasive cells dependent on its demethylase activity. PLU1 was shown to repress the expression of the KAT5 gene through its H3K4 demethylation on the promoter. The regulation of KAT5 by PLU1 was suggested to be responsible for PLU1-induced cell invasion. First, knockdown of KAT5 similarly increased the invasive potential of the cells. Secondly, knockdown of PLU1 in the highly invasive cancer cells increased KAT5 expression and reduced the invasive activity. Thirdly, simultaneous knockdown of KAT5 partially relieved the suppression of cell invasion imposed by PLU1 knockdown. Finally, we found that CD82, which was transcriptionally regulated by KAT5, might be a candidate effector of cell invasion promoted by PLU1. The present study demonstrated a functional contribution of PLU1 overexpression with concomitant epigenetic dysregulation in cancer progression.

  • cancer progression
  • cell invasion
  • histone demethylase
  • Jumonji C domain (JmjC domain)
  • transcription

INTRODUCTION

Post-transcriptional modifications of histones including phosphorylation, ubiquitination, acetylation and methylation play important roles in regulating chromatin dynamics and gene expression [1]. Methylation occurs on both lysine and arginine residues, and four lysine residues on the tail of histone H3 (Lys4, Lys9, Lys27 and Lys36) have been shown to be mono-, di- or tri-methylated. Histone lysine methylation has been linked to transcriptional activation and repression, depending on the specific residues that become methylated and the state of methylation [24]. For example, H3K4me2 [dimethylation of H3K4 (Lys4 of histone H3)] or H3K4me3 (trimethylation of H3K4) is associated with gene activation, whereas H3K9me3 or H3K27me3 (trimethylation of Lys9 or Lys27 of histone H3 respectively) is linked to transcriptional repression [5,6]. The discovery of a large number of site-specific histone methyltransferases and demethylases indicates the dynamic regulation of histone methylation [1,4]. The misregulation of these enzyme activities has been shown to result in developmental defects and the pathogenesis of human disease. Especially, overexpression, amplification and mutations of several enzymes have been reported in many types of cancer [7].

In order to find novel genes involved in cancer development, we have utilized retroviral insertional mutagenesis in mice. We have identified many candidate cancer genes including the genes encoding histone methyltransferases and demethylases by high-throughput retroviral tagging ([8,9], and A. Ishimura, M. Terashima, K. Akagi and T. Suzuki, unpublished work). This result provided us with the opportunity to explore the functions of these enzymes in oncogenesis. Previously, we reported that Jmjd2c (Jumonji-domain-containing protein 2c) histone demethylase, one of the candidate oncogenes, up-regulated the expression of Mdm2 (murine double minute 2) oncogene, resulting in the reduction of p53 tumour suppressor in the cells [10]. We also showed that Utx demethylase, one of the candidate tumour suppressor genes, enhanced the expression of Rb (retinoblastoma) and Rbl2 (Rb-like 2) genes that play important roles in cell proliferation [11].

PLU1, also named JARID1B or KDM5B, is one of the four JARID family members and can demethylate H3K4me1 (monomethylation of H3K4), H3K4me2 and H3K4me3 [1215]. PLU1 was initially isolated as a transcript overexpressed in human breast-cancer cell lines and primary breast carcinomas [16,17]. It has been also reported to be up-regulated in many types of tumours such as prostate, bladder and lung cancers [15,18]. Since PLU1 acts as a transcriptional regulator by its ability to change histone H3K4 methylation, several cancer-associated genes regulated by PLU1 have been identified. They include BRCA1 (breast-cancer susceptibility gene 1) tumour suppressor gene that is repressed by PLU1, and E2F1 and E2F2 that are up-regulated by PLU1. The regulation of these genes by PLU1 was proposed to contribute to cell cycle progression and tumorigenesis [12,18,19]. In addition to the role in tumour initiation, PLU1 was suggested to be associated with malignant progression of prostate cancer [15]. However, the mechanism by which PLU1 contributes to tumour progression remains unclear.

In the present study, we demonstrated that overexpression of PLU1 histone demethylase increased the invasive activity of the cells. To elucidate the function of PLU1 in cancer progression, we have searched for the downstream target genes regulated by PLU1. We discovered that the KAT5/Tip60 gene was down-regulated by PLU1 through changes in histone H3 methylation on the promoter. Mechanistic investigations suggested that KAT5 down-regulation mediated the PLU1-induced cell invasion. We also found that CD82/KAI1 might be a candidate downstream effector for the invasive activity regulated by PLU1 via KAT5.

EXPERIMENTAL

Plasmids and antibodies

Mouse PLU1 cDNA was tagged with a FLAG-His6-tag as described previously [10], and then cloned into pcDNA5/FRT/TO plasmid for establishing a doxycycline-inducible cell line using Flp-In T-Rex-293 cells (Invitrogen) or into the pDON-5 Neo plasmid (Takara) to produce retroviruses. The catalytically inactive PLU1 H499Y mutant [12] was constructed using a PCR-based mutagenesis kit (Takara). The shRNA (small hairpin RNA)-expressing retrovirus vectors were constructed by cloning the synthetic oligonucleotides into AgeI and EcoRI sites of pLKO.1-puro plasmid (Sigma–Aldrich). The sense strand sequences of the oligonucleotides were as follows: PLU1 shRNA#1, 5′-CCGGCCTGAGGAAGAGGAGTATCTTCTCGAGAAGATACTCCTCTTCCTCAGGTTTTT-3′; PLU1 shRNA#2, 5′-CCGGCGAGATGGAATTAACAGTCTTCTCGAGAAGACTGTTAATTCCATCTCGTTTTT-3′; KAT5 shRNA#1, 5′-CCGGCGTCCATTACATTGACTTCAACTCGAGTTGAAGTCAATGTAATGGACGTTTTT-3′; KAT5 shRNA#2, 5′-CCGGCCTCAATCTCATCAACTACTACTCGAGTAGTAGTTGATGAGATTGAGGTTTTT-3′; Control shRNA, 5′-CCGGCAACAAGATGAAGAGCACCAACTCGAGTTGGTGCTCTTCATCTTGTTGTTTTT-3′.

The sequence of control shRNA is the same as that of MISSION Non-Target shRNA (Sigma–Aldrich) that does not target human and mouse genes, making it useful as a negative control in experiments. Anti-FLAG (M2, Sigma), anti-KAT5 (ab23886, Abcam), anti-PLU1 (A301–813A, Bethyl) and anti-β-tublin (H-235, Santa Cruz Biotechnology) antibodies were used for Western blot analysis.

Cell culture and transfection

MCF10A, an immortalized breast cell line, MDA-MB-231 and BT549, breast-cancer cell lines, and DU145, a prostate cancer cell line, were maintained in DMEM (Dulbecco's modified Eagle's medium) with 10% FBS (fetal bovine serum), 2 mM glutamine and penicillin/streptomycin (Sigma–Aldrich) at 37 °C in 5% CO2. MEFs (mouse embryonic fibroblasts) were maintained as described previously [10]. To produce ecotropic retroviruses for the infection of mouse cells, the retrovirus vectors were transfected into Plat-E packaging cells [10] using FuGENE HD reagent (Roche). The retroviral stock was added into the medium with 6 μg/ml polybrene (Sigma–Aldrich) for infection. To produce amphotropic retroviruses for the infection of human cells, the pDON-5- or pLKO.1-puro-derived plasmids were co-transfected into HEK-293T cells [HEK (human embryonic kidney) -293 cells expressing the large T-antigen of simian virus 40] with pCMVdeltaR8.2 and pcDNA3.1-VSVG plasmids (kindly provided by Dr A. Hirao, Cancer Research Institute, Kanazawa University, Kanazawa, Japan) or MISSION Lentiviral Packaging Mix (Sigma–Aldrich). At 48 h after transfection, the supernatants were collected and used for infection. The retrovirus-infected cells were selected in the medium containing 500 ng/ml puromycin (nacalai) or 400 μg/ml G418 (nacalai) for 2–5 days, and used for the experiments. Each Stealth siRNA (small interfering RNA) for mouse PLU1 (Invitrogen) was introduced into MEFs using Lipofectamine™ RNAiMAX (Invitrogen) using a reverse transfection procedure as described by the manufacturer.

Quantitative PCR

Total RNA was extracted with TRIzol® (Invitrogen) using a standard method and transcribed into cDNA using SuperScript Vilo cDNA synthesis kit (Invitrogen). Quantitative RT–PCR (reverse transcription–PCR) analysis was performed with FastStart Universal SYBR Green Master (Roche) using a 7900HT Fast Real-Time PCR System (Applied Biosystems). PCR data were normalized with respect to control mouse β-actin or human GAPDH (glyceraldehyde-3-phosphate dehydrogenase) and are presented as the fold changes in mRNA expression. The averages from at least three independent experiments are shown with the S.D. Primers used for the quantitative PCR have been described previously [10] and are as follows: Mouse PLU1, 5′-AGAGGCTGAATGAGCTGGAG3′ and 5′-TGGCAATTTTGGTCCATTTT-3′; Human PLU1, 5′AGCTCCAGAACTCTTTGTGTCC-3′ and 5′-GCACACTGATTAGTTCGGTAAACA-3′; Mouse KAT5, 5′-ACCACCGCTCAACGAAAC-3′ and 5′-CTACGGGCTGACCCATTCT-3′; Human KAT5, 5′-TTGACCAAGTGTGACCTACGA-3′ and 5′-CACAGGTTCTGGGAATAACTCTT-3′; Mouse CD82, 5′-CCAATGCCACCAGTAGCC-3′ and 5′-CAGCCACAGCACTTGACCT-3′; Human CD82, 5′-TGCTGCTGTGTGGACGAC-3′ and 5′-CTTCCTTCCACGAAACCAGT-3′; Human GAPDH, 5′-CGAGATCCCTCCAAAATCAA-3′ and 5′-GTCTTCTGGGTGGCAGTGAT-3′; KAT5-Promoter region a, 5′-AAGGAGCGGACTTGTCAAC-3′ and 5′-CTGCTTCTGCCTCCACTG-3′; KAT5-Promoter region b, 5′-CTGGGGAGTCAGTAGACCTG-3′ and 5′-GGAAAAATCTTGCTAAGCCC-3′; KAT5-Promoter region c, 5′-CCCTCCACTGCCACTTCC-3′ and 5′-GGCCTATAGCGCACTCAGC-3′.

ChIP (chromatin immunoprecipitation) assays

ChIP experiments were performed as described previously [20]. MEFs were infected with the retroviruses or transfected with the siRNAs 3 days before harvesting for ChIP with mouse or rabbit antibodies (anti-H3K4me3, anti-H3K4me2, anti-H3K4me1, antiPLU1 and normal IgG), bound to Dynabeads M-280 sheep anti-mouse IgG or Dynabeads Protein G (Invitrogen). We examined the size of the chromatin fragments generated during the ChIP procedure by agarose gel electrophoresis (Supplementary Figure S1 at http://www.BiochemJ.org/bj/437/bj4370555add.htm). The size ranged from approximately 100 to 1000 bp with a median of 300 bp, which was suitable for our analysis for the promoter region of KAT5. The enrichment of the specific amplified region was analysed by quantitative PCR and the averages from at least three experiments are presented with the S.D.

Basement membrane matrix invasion assay

Cell invasion assays were carried out using modified Boyden chambers consisting of Transwell membrane filter inserts (Corning Costar) in 24-well tissue culture plates. The upper surfaces of the Transwell membranes were coated with 1 mg/ml Matrigel™ matrix (Becton Dickinson) overnight at 4 °C. Serum-starved cells (2×105) suspended in 100 μl of DMEM containing 1 mg/ml BSA and 0.5% FBS were cultured in each Transwell chamber and allowed to invade towards the underside of the membrane for 16 h (MDA-MB-231 and DU145 cells) or 36 h (MCF10A cells). The lower chamber contained DMEM with 10% FBS and 5 μg/ml fibronectin as the chemoattractants. Non-invading cells were removed by wiping the upper side of the membrane, and the invaded cells were fixed and stained with Crystal Violet. The number of invaded cells was counted under a light microscope from at least five fields and three experiments. The average numbers from the three experiments are presented with the S.D. P values were calculated between control and the samples using Student's t test.

RESULTS

Overexpression of PLU1 promoted the invasive activity of the cells

It has been reported that the expression of PLU1 is up-regulated in breast cancer cell lines and primary metastatic breast cancers but is very low in normal tissues [16,17]. To clarify the role of PLU1 in the malignant progression of cancer, we tried to examine whether ectopic expression of PLU1 could affect the invasive potential of the cells. We used a benign immortalized breast cell line, MCF10A, because it has very low invasive potential [21]. MCF10A cells were infected with the amphotropic retroviruses expressing either without insert or with FLAG-tagged wild-type PLU1 or catalytically inactive PLU1 mutant H499Y [12]. The infected cells were selected in the medium containing 400 μg/ml G418, and the cell lysates were subjected to Western blot analysis. Exogenously introduced wild-type and the mutant PLU1 proteins were detected with anti-FLAG antibody (Figure 1A, upper panel). The blot with the anti-PLU1 antibody showed that the ectopic expression of PLU1 protein was approx. 5-fold higher than its endogenous level (Figure 1A, middle panel). A modified Boyden chamber assay was used to determine whether MCF10A cells undergo invasion on ectopic expression of PLU1. PLU1 was reported to be involved in cell proliferation of cancer cells [12,18]. To rule out the effect of PLU1 on cell growth, the infected cells were serum starved to suppress growth and then subjected to invasion assays. As shown in Figures 1(B) and 1(C), cell invasion through membranes coated with Matrigel™ was increased in the cells expressing wild-type PLU1 compared with the control cells. Furthermore, no significant change was observed after infection with the virus expressing the inactive H499Y mutant. These results indicated that the enzyme activity of PLU1 was required for the enhanced invasive activity of the cells. There were no significant differences in total cell numbers of each infected cell, which were determined by counting the cells on both sides of the membrane after incubation (results not shown). Therefore we concluded that ectopic expression of PLU1 enhanced the invasive potential of the weakly invasive cells, dependent on its demethylase activity.

Figure 1 Overexpression of PLU1 increased the invasive activity of the cells

(A) Overexpression of PLU1 proteins in MCF10A cells infected with the retroviruses. MCF10A cells were infected with the control retrovirus or the retrovirus expressing FLAG-tagged wild-type (WT) PLU1 or the H499Y mutant (Mut). PLU1 proteins were detected by Western blot (WB) analysis with an anti-FLAG antibody (upper panel) or an anti-PLU1 antibody (middle panel). The asterisk shows a non-specific band that reacts with the anti-PLU1 antibody. As a control, an anti-β-tublin antibody was used to show that equal amounts of proteins were loaded on to the gel (lower panel). The molecular mass in kDa is indicated on the left-hand side. (B) Ectopic expression of PLU1 induced cell invasion. Cell invasion assays were performed using the retrovirus-infected MCF10A cells as described in the Experimental section. Representative fields of invaded and stained cells are shown. (C) PLU1 enhanced the invasive activity of the cells. Cells that invaded through the basement membrane to the lower surface of the membrane were fixed, stained and counted. *P<0.001 compared with control.

KAT5/Tip60 was one of the target genes regulated by PLU1

PLU1 is one of the histone H3K4 demethylases and has been shown to be involved in the transcriptional regulation of several genes [12,18,19]. To uncover the molecular function of PLU1 in tumour initiation and progression, we have searched for the downstream target genes regulated by PLU1. We have established an inducible cell line expressing PLU1, and the cDNAs derived from the cells were applied to the massive transcriptional start analysis [22]. This is a method to collect positional information of transcriptional start sites together with the digital data of the expression levels of transcripts by using a massively parallel sequencing technology [22]. From this digital expression profile, we identified some candidate genes whose expression changes were predicted (M. Yoshida, A. Ishimura, M. Terashima, Y. Suzuki, S. Sugano, K. Satou and T. Suzuki, unpublished work). Among them, we focused on the KAT5/Tip60 gene whose expression was potentially down-regulated by PLU1 (2.5-fold down-regulation from the digital expression profile; P<0.0001), since KAT5 was considered a tumour suppressor gene and reported to be more frequently down-regulated in advanced carcinoma [2325].

To confirm whether PLU1 regulates the expression of KAT5 in normal cells, we used MEFs infected with the ecotropic retrovirus expressing PLU1. MEFs were infected with the control retrovirus or the retrovirus expressing FLAG-tagged wild-type PLU1 or its mutant, and the cell lysates were prepared. Both wild-type and the mutant PLU1 proteins were detected by Western blot analysis with an anti-FLAG antibody (Figure 2A). Then we examined the expression level of KAT5 genes by quantitative RT–PCR. KAT5 transcripts were decreased relative to the control with the expression of wild-type PLU1, but not with the mutant (Figure 2B). This observation was consistent with the results from the digital expression profiling. We also detected the reduction of KAT5 protein induced by wild-type PLU1, but not by the mutant, by Western blot analysis (Figure 2C). These results indicated that PLU1 repressed the expression of the KAT5 gene product in normal cells, which was dependent on its demethylase activity.

Figure 2 PLU1 decreased the expression of the KAT5 gene in normal cells

(A) Overexpression of PLU1 proteins in MEFs infected with the retroviruses. FLAG-tagged wild-type PLU1 or the H499Y mutant proteins were detected by Western blot (WB) analysis with an anti-FLAG antibody (upper panel). An anti-β-tublin antibody was used as a control (lower panel). (B) PLU1 reduced the expression of KAT5 transcripts through its demethylase activity. The cDNA derived from each retrovirus-infected MEFs was used for quantitative PCR detecting KAT5 expression. Relative amounts of transcripts are presented as the fold changes with respect to the control. The averages from at least three experiments are presented with the S.D. (C) PLU1 decreased the expression of KAT5 protein. KAT5 proteins in the cell lysates of the retrovirus-infected MEFs were detected by Western blot analysis with an anti-KAT5 antibody (upper panel). The molecular mass in kDa is indicated on the left-hand side. Mut, mutant; WT, wild-type.

PLU1 induced the conversion of histone H3 methylation on the promoter of the KAT5 gene

Since PLU1 has the demethylase activity of histone H3 [1215], we examined the methylated status of histone H3 on the KAT5 gene promoter with or without overexpression of PLU1. ChIP was performed using the antibodies that specifically recognize methylated lysine residues of histone H3 [20]. MEFs were infected with the control retrovirus or the retrovirus expressing wild-type PLU1 or its mutant, and the cell lysates were prepared. Following immunoprecipitation, several sets of primers positioned upstream from or spanning the transcription start site of the mouse KAT5 gene were used in quantitative PCR analysis (Figure 3A).

Figure 3 PLU1 regulated histone H3 methylation on the promoter of the KAT5 gene

(A) Schematic representation of the promoter region of the KAT5 gene. The boxes shown on the scheme indicate the first and second exons and the dark area corresponds to the coding region. The arrow points to the transcription initiation site. The regions covered by the primer sets used for ChIP assays are shown as a–c. (B and C) ChIP analysis of H3K4me3 (B) and PLU1 (C) on the KAT5 promoter regions. The cross-linked chromatins from MEFs infected with the control retrovirus or the retrovirus expressing wild-type PLU1 or the mutant were immunoprecipitated with the specific antibodies. The occupancies of methylated histones or PLU1 on the regions were analysed by quantitative PCR and presented as the relative fold of enrichment. The means from at least three experiments are presented with the S.D. Mut, mutant; WT, wild-type.

We first analysed the transcriptionally active H3K4me3 status because PLU1 can demethylate H3K4. On regions a and b of the KAT5 promoter, H3K4me3 was dramatically decreased by the expression of wild-type PLU1, but not by the mutant (Figure 3B). On region c, however, the level of H3K4me3 was not affected by PLU1 expression (Figure 3B), indicating that PLU1 induced H3K4me3 demethylation on the specific regions of the KAT5 gene promoter. For the status of H3K4me2, similar results were obtained, but the level of H3K4me1 was not significantly decreased by PLU1 expression (results not shown). These results agreed with the previous reports suggesting that PLU1 preferentially regulated the levels of H3K4me3 and/or H3K4me2 at the target genes, although it has the ability to catalyse all three H3K4 methylated states [12,13].

These ChIP experiments raised the possibility that PLU1 directly associated with the regions of the KAT5 gene promoter to change the methylated status of histone H3. ChIP assays using an anti-PLU1 antibody clearly demonstrated the increased occupancies of wild-type PLU1 and the mutant on regions a and b, but not on region c of the KAT5 promoter (Figure 3C). These results suggested that PLU1 which was recruited on the specific regions of KAT5 promoter converted the chromatin structure to a more transcriptionally repressed state by demethylating H3K4. Although the association of the PLU1 H499Y mutant was observed on the promoter (Figure 3C), the decrease of H3K4me3 was not detected (Figure 3B), which was consistent with the inability of KAT5 repression by the mutant (Figure 2B).

On the other hand, there were no significant changes of H3K4 methylation and PLU1 occupancies on the unrelated Actg1 promoter by PLU1 expression (results not shown). Therefore the PLU1 recruitment and the changes of histone H3 methylation were specific to the KAT5 gene promoter, which was closely correlated with the specificity of transcriptional regulation by PLU1.

Knockdown of PLU1 recovered the expression level of the target gene, KAT5

To extend our understanding for the regulation of KAT5 expression by PLU1, we examined the effects of PLU1 knockdown in MEFs by introducing its siRNAs. Quantitative RT–PCR indicated that PLU1 transcripts were significantly reduced with mouse PLU1 siRNA#1 and siRNA#2, but not with the control siRNA (Figure 4A). We also confirmed the reduction of PLU1 at the protein level in the siRNA-treated cells by Western blot analysis with the anti-PLU1 antibody (Figure 4B). Then we analysed the expression levels of KAT5 in the PLU1-knockdown cells. As shown in Figure 4(C), a significant increase in KAT5 transcripts was observed upon PLU1 knockdown with siRNA#1 and siRNA#2, but not with the control siRNA, suggesting that there is a PLU1-dependent mechanism of regulating KAT5 expression. We also detected the increase of KAT5 proteins in the PLU1-knockdown cells (Figure 4D). Next we tried to perform the rescue experiment. The PLU1 siRNA-treated MEFs were infected with the ecotropic retrovirus expressing PLU1. As shown in Figure 4(E), the expression of PLU1 protein was recovered when the siRNA-treated cells were infected with PLU1-expressing virus. After the recovery of PLU1 expression, the expression of KAT5 protein was decreased (Figure 4E). This result indicated that the increased expression of KAT5 induced by PLU1 siRNA was cancelled when PLU1 expression was rescued in the cells. These results strongly suggested the involvement of endogenous PLU1 in the transcriptional regulation of KAT5 expression. Thus we concluded that the KAT5 gene was a physiological downstream target of PLU1 in normal cells.

Figure 4 PLU1 is involved in the expression of KAT5 in normal cells

Using MEFs transfected with the PLU1 siRNAs, the relative amounts of RNA and protein were analysed, and ChIP assays were performed. (A) Quantitative PCR analysis of PLU1 transcripts after knockdown by its siRNAs (PLU1 si#1 and si#2). (B) The amount of PLU1 protein was decreased by its knockdown. PLU1 proteins were detected by Western blot (WB) analysis with an anti-PLU1 antibody (upper panel). The asterisk shows a non-specific band that reacts with the antibody. (C) The expression of KAT5 transcripts was increased by PLU1 knockdown. A primer set that detects KAT5 was used for quantitative PCR. (D) The expression of KAT5 protein was up-regulated after PLU1 knockdown detected by Western blot. (E) The increase of KAT5 expression by PLU1 siRNA was cancelled when PLU1 expression was rescued by the retrovirus. Western blot analysis was performed with an anti-PLU1 antibody (upper panel), an anti-KAT5 antibody (middle panel) and an anti-β-tublin antibody (lower panel). (F and G) The effects of PLU1 knockdown on PLU1 recruitment (F) and H3K4me3 (G) were analysed on region a of the KAT5 gene promoter by ChIP assays. The molecular mass in kDa is indicated on the left-hand side.

Next we investigated the PLU1 recruitment and the histone methylation on the KAT5 promoter in the PLU1-knockdown cells. ChIP experiments demonstrated that the level of PLU1 recruitment was significantly reduced by the PLU1 siRNAs on representative region a of the KAT5 promoter (Figure 4F). This result clearly indicated the recruitment of endogenous PLU1 protein on the promoter region of the KAT5 gene. The increase of active H3K4me3 marks was detected on the KAT5 promoter induced by PLU1 knockdown (Figure 4G). This increase of H3K4me3 was consistent with the observed elevated expression of KAT5 transcripts (Figure 4C). These changes of histone H3 methylation were specific to the KAT5 promoter, and were not observed on the Actg1 promoter (results not shown), suggesting that transcription from the KAT5 promoter was specifically regulated by PLU1-mediated H3K4 demethylation in normal cells.

KAT5 repression by PLU1 was implicated in the invasive activity of the cells

In order to elucidate the molecular mechanism of PLU1 in cell invasion, we examined whether KAT5 repression by PLU1 might be implicated in the invasive activity. First, we analysed the expression of KAT5 in MCF10A cells infected with the retrovirus expressing PLU1. KAT5 transcripts were decreased relative to the control with the expression of wild-type PLU1, but not with the mutant (Figure 5A). This result indicated that PLU1 could repress the expression of the KAT5 gene in human immortalized cells as well as in mouse primary cells. To analyse the effect of KAT5 down-regulation in cell invasion, we used KAT5 shRNA-expressing amphotropic retroviruses. We designed two different shRNAs for KAT5 knockdown (KAT5 sh#1 and sh#2) in order to avoid the off-target effects. MCF10A cells were infected with the control retrovirus or the retrovirus expressing each KAT5 shRNA, and the infected cells were selected with the medium containing 500 ng/ml puromycin. Quantitative RT–PCR indicated that KAT5 transcripts were significantly reduced with the infection of both KAT5 shRNA-expressing retroviruses, but not with the control retrovirus (Figure 5B). We also confirmed the reduction of KAT5 at the protein level in the shRNA-introduced cells by Western blot analysis with an anti-KAT5 antibody (Figure 5C). Then we examined the effect of KAT5 knockdown on the invasive potentials of MCF10A cells. Figure 5(D) showed that both KAT5 shRNAs led to the enhanced invasive activity of MCF10A cells. We also confirmed that KAT5 knockdown did not affect cell growth (results not shown). These observations suggested that KAT5 repression by PLU1 might play a critical role in PLU1-enhanced invasive potential of the weakly invasive cells.

Figure 5 KAT5 repression by PLU1 was implicated in the invasive activity of the cells

(A) The expression of KAT5 was reduced by PLU1 overexpression in MCF10A cells. MCF10A cells were infected with the control retrovirus or the retrovirus expressing wild-type PLU1 or the mutant. The expression of KAT5 mRNAs was detected by quantitative RT–PCR. (B) Quantitative PCR analysis of KAT5 after knockdown by its shRNAs. MCF10A cells were infected with the control retrovirus or the retrovirus expressing each KAT5 shRNA (KAT5 sh#1 and sh#2). The KAT5 expression was detected by quantitative RT–PCR. (C) The amount of KAT5 protein was reduced by its shRNAs. KAT5 proteins were detected by Western blot (WB) analysis with an anti-KAT5 antibody (upper panel). The molecular mass in kDa is indicated on the left-hand side. (D) KAT5 knockdown enhanced the invasive potential of MCF10A cells. Invasion assays were carried out as described in the Experimental section. *P<0.001 and **P<0.005 compared with control.

Knockdown of PLU1 decreased the invasive activity of the cancer cell lines

To expand our investigations, we examined the expression of PLU1 and KAT5 in the weakly invasive breast cell line MCF10A, and the highly invasive breast cancer cell lines MDA-MB-231 and BT549. Western blot analysis revealed that the expression level of PLU1 approx. 2-fold higher in MDA-MB-231 and BT549 cells than in MCF10A cells (Figure 6A, upper panel). Interestingly, the expression of KAT5 was lower in MDA-MB-231 and BT549 cells than in MCF10A cells (Figure 6A, middle panel), indicating an inverse correlation of KAT5 and PLU1 expression. This result suggested the important roles of endogenous PLU1 and KAT5 in the invasion potentials of the cells.

Figure 6 Knockdown of PLU1 restored the expression of KAT5 and decreased the invasion activity of the highly invasive cancer cells

(A) The expression of PLU1 and KAT5 proteins in the cell lines. Cell lysates were prepared from the indicated cell lines, and PLU1 and KAT5 proteins were detected by Western blot (WB) analysis with an anti-PLU1 antibody (upper panel) or an anti-KAT5 antibody (middle panel). The asterisk shows a non-specific band that reacts with the anti-PLU1 antibody. As a loading control, an anti-β-tublin antibody was used (lower panel). The molecular mass in kDa is indicated on the left-hand side. (B) Quantitative PCR analysis of PLU1 after knockdown by its shRNAs. MDA-MB-231 cells were infected with the control retrovirus or the retrovirus expressing each PLU1 shRNA (PLU sh#1 and sh#2). The expression of PLU1 mRNAs was detected by quantitative RT–PCR. (C) The expression of KAT5 transcripts was increased by shRNA-mediated PLU1 knockdown. A primer set that detects KAT5 was used for quantitative PCR. (D) PLU1 knockdown reduced the invasive potential of MDA-MB-231 cells. Invasion assays were performed using MDA-MB-231 cells infected with the control retrovirus or PLU1 shRNA-expressing retroviruses. *P<0.001 and **P<0.005 compared with control. (E) The increase of KAT5 expression upon PLU1 knockdown was cancelled by introducing KAT5 shRNA into MDA-MB-231 cells. A primer set that detects KAT5 was used for quantitative PCR. (F) The reduced invasive activity on PLU1 knockdown was partially recovered by KAT5 knockdown. *P<0.001 compared with control.

Then we tried to examine whether knockdown of endogenous PLU1 would affect the invasion activity of the highly invasive cancer cell lines. We used PLU1 shRNA-expressing retroviruses for stable knockdown and the MDA-MB-231 cancer cell line. MDA-MB-231 cells were infected with the control retrovirus or the retrovirus expressing each PLU1 shRNA (PLU1 sh#1 and sh#2). The expression of PLU1 mRNAs was detected by quantitative RT–PCR. PLU1 transcripts were significantly reduced with the infection of PLU1 shRNA-expressing retroviruses, but not with the control retrovirus in both cells (Figure 6B). Then we examined the expression levels of KAT5 in the PLU1-knockdown cells. As shown in Figure 6(C), a significant increase in KAT5 transcripts was observed upon PLU1 knockdown by the shRNAs, suggesting the involvement of endogenous PLU1 in the regulation of KAT5 expression in the cancer cell line. Next we investigated the effect of PLU1 knockdown on the invasive potentials of the cells. To rule out the effect of PLU1 knockdown on cell growth, the infected cells were serum starved and then subjected to invasion assays as shown in Figure 1. The knockdown of PLU1 expression markedly reduced the invasive activity of MDA-MB-231 cells (Figure 6D). Similar results were obtained when we examined the effects of PLU1 knockdown in another breast cancer cell line, BT549, and the highly invasive prostate cell line DU145 (results not shown). These results clearly indicated that endogenous PLU1 was responsible for the invasive potentials of the highly invasive cancer cell lines.

The recovery of KAT5 expression by PLU1 knockdown was suggested to contribute to the marked attenuation of the invasive potentials. Therefore we examined whether simultaneous knockdown of KAT5 could reverse the phenotypes induced by PLU1 knockdown. MDA-MB-231 cells were infected with two retroviruses expressing PLU1 shRNA and KAT5 shRNA. The increase of KAT5 expression on PLU1 knockdown was cancelled by introducing KAT5 shRNA (Figure 6E). As shown in Figure 6(F), the decreased invasive activity induced by PLU1 knockdown was partially recovered by KAT5 shRNA. These results strongly suggested that PLU1 affected the invasion potential of the cancer cells through its transcriptional regulation of the KAT5 gene.

CD82/KAI1 was suggested as a candidate mediator of cell invasion enhanced by PLU1

In addition to the tumour suppressor activity of KAT5/Tip60 [23], it has also been reported that KAT5 is involved in transcriptional activation of CD82/KAI1, a metastasis suppressor gene [26,27]. Thus we tried to explore the involvement of CD82 in the PLU1-mediated enhancement of cell invasion. We analysed the expression of CD82 in MCF10A cells infected with the retrovirus expressing PLU1 by quantitative RT–PCR. CD82 expression was decreased relative to the control with the expression of wild-type PLU1, but not with the mutant (Figure 7A). To examine whether CD82 down-regulation might be related to KAT5 repression by PLU1, we used KAT5 shRNA-expressing retroviruses. The expression of CD82 was analysed in MCF10A cells infected with the control retrovirus or the retrovirus expressing each KAT5 shRNA. Quantitative RT–PCR showed that CD82 transcripts were significantly reduced by the introduction of KAT5 shRNA (Figure 7B), indicating KAT5 function in transcriptional activation of CD82. We could confirm the previous result demonstrating that KAT5 was a member of the important transcriptional activator complex for the CD82 gene [26].

Figure 7 CD82 was found to be a candidate effector of PLU1-induced cell invasion

(A) The expression of CD82 was reduced by PLU1 overexpression in MCF10A cells. The expression of CD82 mRNAs in the retrovirus-infected MCF10A cells was detected by quantitative RT–PCR. (B) Quantitative PCR analysis of CD82 after knockdown of KAT5 by its shRNAs. The CD82 expression was detected by quantitative RT–PCR in the MCF10A cells expressing each KAT5 shRNA. (C) The expression of CD82 transcripts was up-regulated in MDA-MB-231 cells by the shRNA-mediated PLU1 knockdown. (D) The increase of CD82 transcripts by PLU1 knockdown was cancelled by introducing KAT5 shRNA. (E) The expression of CD82 transcripts in the cell lines. The CD82 expression was examined by quantitative RT–PCR in the indicated cell lines.

Next we investigated whether knockdown of endogenous PLU1 would affect the expression of CD82 in the cells. In MDA-MB-231 cells infected with PLU1 shRNA-expressing retroviruses, a significant increase in CD82 transcripts was observed by quantitative RT–PCR (Figure 7C). However, the increase of CD82 expression on PLU1 knockdown was cancelled by the simultaneous introduction of KAT5 shRNA (Figure 7D). These results indicated that the recovery of CD82 expression was dependent on the recovery of KAT5 expression in the PLU1-knockdown cells. Moreover, we examined the expression of endogenous CD82 in MCF10A, MDA-MB-231 and BT549 cells. Quantitative RT–PCR revealed that the expression level of CD82 was lower in MDA-MB-231 and BT549 cells than in MCF10A cells (Figure 7E). This result, together with the results shown in Figure 6(A), indicated that CD82 expression was inversely correlated with PLU1 expression, which was the same as KAT5 expression. Although the expressions of CD82 and KAT5 were shown to be inversely correlated with PLU1 expression, our observations suggested that CD82 was not directly regulated by PLU1, but through KAT5. Importantly, the expression levels of CD82 and KAT5 were also shown to be inversely correlated with the invasive potentials of MCF10A cells and MDA-MB-231 cells. Since CD82 was reported to be a metastasis suppressor gene, we proposed that CD82 might be a candidate effector of cell invasion controlled by PLU1-mediated KAT5 expression.

DISCUSSION

In the present study, we describe a novel function of PLU1 that promoted the invasive potential of the cells. For the mechanism, we found that PLU1 repressed the expression of the KAT5/Tip60 gene in both normal mouse cells and human immortalized cells. PLU1 was shown to associate with the promoter region of the KAT5 gene to induce the conversion of histone H3 methylation into a more transcriptionally repressive state through its H3K4 demethylase activity. A series of knockdown experiments revealed that the decrease of KAT5 expression was responsible for the PLU1-induced cell invasion. Furthermore, CD82/KAI1, which was transcriptionally regulated by KAT5, was suggested as a candidate effector of cell invasion promoted by PLU1.

PLU1 has a restricted expression pattern with low levels of expression in all normal tissues other than testis [17]. On the other hand, it is up-regulated in many types of malignant tumours including breast, prostate, bladder and lung cancer [15,16,18]. PLU1 was also demonstrated to promote cancer cell growth in vitro and in vivo, because knockdown of PLU1 reduced the cell proliferation of cancer cells [12,18]. In addition to the function of PLU1 in cell growth and tumorigenesis, a previous paper suggested the correlation of PLU1 expression and the malignant stage of prostate cancer [15]. In the present study, we demonstrated that overexpression of PLU1 enhanced the invasive potential of the weakly invasive cells and shRNA-mediated knockdown of PLU1 in the highly invasive cancer cells reduced cell invasion. The observed changes of invasive activities were not due to the PLU1 effect on cell growth, since the invasion assays were carried out using the serum-starved cells to rule out the effect of PLU1 on cell growth. Thus our experiments provided a novel important role of PLU1 that could contribute to cell invasion, a hallmark of cancer aggressiveness.

A growing body of evidence indicates that overexpression or mutations of histone methyltransferases and demethylases have been linked to the development of many human cancers [1,7]. However, the roles of these enzymes in the processes of cancer progression, such as invasion and metastasis, have just begun to emerge. Ezh2, a histone H3K27 methyltransferase, was shown to be overexpressed in metastatic prostate and breast cancer, and the expression levels were strongly associated with tumour progression [28,29]. Overexpression of Ezh2 in immortalized mammary epithelial cells was shown to promote cell invasion through transcriptional repression of the E-cadherin gene [29,30]. Recently, the expression of H3K9 methyltransferase G9a was also reported to correlate with poor prognosis of lung cancer [31]. The authors showed that G9a promoted cell invasion of lung cancer cells by repressing the transcription of the Ep-CAM gene. We have provided for the first time evidence that PLU1 H3K4 demethylase functions as a promoter of cell invasion. We observed that PLU1 did not affect the expression of the E-cadherin or Ep-CAM genes (results not shown) but repressed the KAT5 gene, suggesting different pathways for the enhancement of cell invasion. Currently, the possible mechanisms for PLU1 overexpression associated with malignant cancer progression remain elusive. For example, further studies using an experimental metastasis model will be required to extend our understanding.

Although PLU1 was initially thought to act as a transcriptional repressor by its ability to demethylate active H3K4 modification status, it was proved to function in both gene activation and repression depending on the specific genes analysed [12,18,19]. Several cancer-associated genes regulated by PLU1 have been identified so far. PLU1 was reported to positively regulate cell proliferation by repressing the expression of cell growth inhibitors such as BRCA1, CAV1 (caveolin 1) and HOXA5 (homeobox A5) [12]. PLU1 was also shown to up-regulate the expression of E2F1 and E2F2 transcription factors for cell cycle progression [18]. In the present paper, we describe the KAT5 gene as one of the novel target genes regulated by PLU1. We found that knockdown of KAT5 expression led to enhanced invasive activity of cells as did PLU1 overexpression. Furthermore, knockdown of KAT5 was shown to partially reverse the suppression of cell invasion due to PLU1 knockdown. Although we could not rule out the possibility that other target genes regulated by PLU1 might be involved in cell invasion, these findings strongly suggested that KAT5 was an important downstream mediator of PLU1-promoted cell invasion.

KAT5 is a histone acetyltransferase that is part of a conserved multi-subunit complex, NuA4 [32]. It is recruited by many transcription factors to their target promoters, where it is generally involved in transcriptional activation. KAT5 can function as a co-activator of the transcription factors that either promotes (e.g. Myc and E2F) or suppresses (e.g. Tp53) tumorigenesis. In addition, KAT5 has transcription-independent roles in DNA damage response signalling and repair, which is believed to constitute an important barrier against tumorigenesis [32]. Importantly, KAT5 has been reported to possess a haplo-insufficient tumour suppressor activity in both mouse and human cells [23]. In our present study, it remains unclear as to whether KAT5 would act as a tumour suppressor in the cells overexpressing PLU1. At least, it does not appear to be involved in the cell growth regulated by PLU1, because knockdown of KAT5 did not enhance the proliferation of the cells (results not shown). On the other hand, down-regulation of KAT5 was reported in various human malignancies, such as breast, head and neck, colon and lung cancers, and was more frequently observed in high-grade tumours [2325]. Thus KAT5 was suggested to be involved in the processes for malignant tumour progression. In the present study, we could propose one of the important roles of KAT5 in tumour progression by demonstrating that KAT5 repression by PLU1 enhanced the invasive activity of the cells.

We also found that CD82 was down-regulated in the cells overexpressing PLU1 and its down-regulation was mediated by KAT5. In the digital expression profiling, we could not detect the reduction of CD82 expression by PLU1, because the sequence tags for CD82 representing its expression level were originally very few in HEK-293 cells-derived PLU1-inducible cells (results not shown). However, our results revealed that CD82 down-regulation by KAT5 repression was closely correlated with the invasive potential of the cells. Therefore we proposed that CD82 might be a candidate effector of cell invasion promoted by PLU1. CD82/KAI1 is a member of the tetraspanin family of type III membrane proteins [33]. It is a metastasis suppressor gene for human malignancies including prostate, breast, pancreatic, colon and lung cancers [27,33,34]. The expression level of CD82 was inversely correlated with metastatic potential of cancer cells, indicating that CD82 may be a useful indicator of tumour progression [35]. There is little evidence for gene mutation, loss of heterozygosity, promoter mutation or hypermethylation to explain the loss of CD82 expression in clinical samples of metastatic cancers. Altered transcription remains a possible mechanisms for the loss of CD82 transcripts. This suggested mechanism agreed with our results showing that CD82 expression could be down-regulated by PLU1 through KAT5. In order to validate our proposed invasion cascade induced by PLU1, we will need further studies that uncover the correlation between the expression levels of PLU1, KAT5 and CD82 in various malignant stages of human cancers.

PLU1 is a unique enzyme that can demethylate all three methyl groups (mono-, di- and tri-methyl) in H3K4 in contrast with other histone demethylases [1215]. The paper by Ellinger et al. [36] reported that renal cell carcinoma patients with unfavourable prognostic parameters had significantly lower global levels of H3K4me1, H3K4me2 or H3K4me3. The underlying mechanism leading to altered global histone H3 methylation is poorly understood. It may be caused by the complex interplay of histone methyltransferases and demethylases. Although the expression of methyl-modifying enzymes are largely unknown in renal cell carcinoma patients, it is possible that PLU1 H3K4 demethylase plays an important role. Functional analysis of these enzymes will clarify how the aberrant histone modification patterns are critically involved in malignant progression of cancer.

We have discovered a novel role of PLU1 histone demethylase in cell invasion. PLU1 represses the downstream target gene KAT5/Tip60 to lead to the down-regulation of CD28/KAI1, a metastasis suppressor, thereby promoting the invasive potential of the cells. Our present study suggests the importance of future studies to explore the relevance of histone methyl-modifying enzymes to malignant progression of cancer.

AUTHOR CONTRIBUTION

Masakazu Yoshida, Akihiko Ishimura, Minoru Terashima and Zanabazar Enkhbaatar performed the experiments and analysed the data. Naohito Nozaki and Kenji Satou contributed vital reagents and analytical tools, and provided advice on experimental design. Takeshi Suzuki designed the research project and wrote the manuscript.

FUNDING

This work was supported in part by Grants-in-Aid for Scientific Research C [grant number 21590303 (to T.S.)] and Science Research for Priority Areas [grant number 21022018 (to T.S.)] from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

Acknowledgments

We thank Dr T. Takino, Dr Y. Endo and Dr H. Sato (Cancer Research Institute, Kanazawa University, Kanazawa, Japan) for providing the cell lines and the protocol for the cell invasion assay. We also thank Dr H. Kimura (Graduate School of Frontier Biosciences, Osaka University, Suita, Japan) for providing the information about the antibodies that specifically recognize methylated H3K4 residues.

Abbreviations: BRCA1, breast-cancer susceptibility gene 1; ChIP, chromatin immunoprecipitation; DMEM, Dulbecco's modified Eagle's medium; FBS, fetal bovine serum; MEF, mouse embryonic fibroblast; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; H3K4, Lys4 of histone H3; H3K4me1, monomethylation of H3K4; H3K4me2, dimethylation of H3K4; H3K4me3, trimethylation of H3K4; HEK, human embryonic kidney; RT–PCR, reverse transcription–PCR; shRNA, small hairpin RNA; siRNA, small interfering RNA

References

View Abstract