Interplay between HDAC3 and WDR5 is essential for
hypoxia-induced epithelial-mesenchymal transition
Min-Zu Wu,1,9 Ya-Ping Tsai,1,4,9 Muh-Hwa Yang,2,3,5 Chi-Hung Huang,4 Shyue-Yih Chang,6 Cheng-Chi Chang,7 Shu-Chun Teng,8 and Kou-Juey Wu1, 3,
1Institutes of Biochemistry & Molecular Biology, 2Clinical Medicine, and 3Genome Research Center, National Yang-Ming University, Taipei 112; 4Taiwan Advance Biopharm (TABP), Inc., Xizhi city, New Taipei City 221; 5Division of Hematology-Oncology, Departments of Medicine and 6Otolaryngology, Taipei Veterans General Hospital, Taipei 112; 7Graduate Institute of Oral Biology, School of Dentistry, and 8Graduate Institute of Microbiology, College of Medicine, National Taiwan University, Taipei 100, Taiwan
9: These authors contributed equally to the work. Correspondence should be addressed to:
Kou-Juey Wu, Institute of Biochemistry and Molecular Biology, National Yang-Ming University, No.155, Li-Nong St., Sec.2, Peitou, Taipei 112, Taiwan; Email: [email protected]; Tel: 886-228267328; Fax: 886-228264843
Summary
Epithelial-mesenchymal transition (EMT) is important for organ development, metastasis, cancer stemness, and organ fibrosis. Molecular mechanisms to coordinately regulate hypoxia-induced EMT remain elusive. Here we show that HIF-1-induced histone deacetylase 3 (HDAC3) is essential for
hypoxia-induced EMT and metastatic phenotypes. Change of specific chromatin states is associated with hypoxia-induced EMT. Under hypoxia, HDAC3 interacts with hypoxia-induced WDR5, recruits the histone methyltransferase (HMT) complex to increase histone H3 lysine 4 (H3K4)-specific HMT activity, and activate mesenchymal gene expression. HDAC3 also serves as an essential co-repressor to repress epithelial gene expression. Knockdown of WDR5 abolishes mesenchymal gene activation but not epithelial gene repression during hypoxia. These results indicate that hypoxia induces different chromatin modifiers to coordinately regulate EMT through distinct mechanisms.
Highlights
Activation of HDAC3 by HIF-1 is required for hypoxia-induced EMT and metastasis
Change of chromatin states is associated with hypoxia-induced EMT
HDAC3 recruits hypoxia-induced WDR5 to increase histone methyltransferase activity
Introduction
Epithelial-mesenchymal transition (EMT) is fundamental for embryonic development and important for organ formation and differentiation (Thiery et al., 2009; Yang and Weinberg, 2008). Pathological processes which utilize EMT include tumor metastasis, maintenance of cancer stemness, and fibrosis of different organs such as lung, kidney, liver, etc., which constitute a broad spectrum of human diseases (Guarino et al., 2009; Higgins et al., 2008; Thiery et al., 2009; Yang and Weinberg, 2008). In cancer metastasis, different EMT regulators such as Snail, Twist1, SIP1, etc. are shown to mediate EMT and metastasis under the signaling of hypoxia/HIF-1, Wnt, Notch, TGF-, etc (Peinado et al., 2007; Thiery et al., 2009; Yang and Weinberg, 2008; Yang and Wu, 2008). However, the mechanisms to coordinately repress epithelial genes and activate mesenchymal genes during EMT are largely unknown.
Although the role of different EMT regulators to mediate EMT is well demonstrated, it is unknown whether a chromatin modifier is required to co-ordinate different EMT regulators to mediate EMT. Recent results showed that lysine demethylase LSD1-mediated decrease in dimethylated H3K4 was required for Snail-induced repression of epithelial genes (Lin et al., 2010). However, no specific change of chromatin states (i.e. change of histone marks) was assigned to the promoters of EMT marker genes during hypoxia-induced EMT. In addition to serving as
co-repressors (Lane and Chabner, 2009), histone deacetylase (HDAC) complexes were shown to associate with 5’ transcribed regions of active genes to prevent cryptic transcription or promote efficient elongation by RNA polymerase II in Saccharomyces cerevisiae (Kim et al., 2009; Li et al., 2007b). Different HDACs were also shown to associate with 5’ transcribed regions of transcriptionally active genes to reset chromatin for subsequent rounds of transcription in mammalian cells (Wang et al., 2009). However, the active role of HDACs in gene activation has not been demonstrated. WDR5 is a WD40 repeat protein which is essential for histone H3 lysine 4 (H3K4) methylation and vertebrate development (Wysocka et al., 2005). WDR5 is a core component of the human MLL and SET1 (hCOMPASS) histone H3K4 methyltransferase complexes (Trievel and Shilatifard, 2009). Although WDR5 is involved in different physiological functions (Ang et al., 2011; Brown et al., 2005; Gori et al., 2009; Wang et al., 2010; Zhu et al., 2008b), the role of WDR5 related to specific gene activation and whether WDR5 is regulated by extracellular signaling or stress remain largely unknown.
In this study, we searched for the chromatin modifiers which are regulated by hypoxia to coordinately mediate the transition of EMT marker genes during EMT. We also identified the specific change of histone marks associated with hypoxia-induced EMT. The mechanisms to coordinately regulate EMT marker genes during EMT and
Results
Direct activation of HDAC3 by HIF-1
A bioinformatic approach was used to identify novel HIF-1 downstream targets. Given that a HIF-1 ancillary sequence (HAS: CACG(T/C)) was shown to couple with a HIF-1 binding sequence (HBS: ACGTG) in the promoters of certain HIF-1 targets (Kimura et al., 2001), we used the HBS+HAS sequence to search for genes whose promoters contain such a sequence. Chromatin modifiers were our priority candidates since they may represent genes providing epigenetic control of EMT. HDAC3 is the candidate gene identified through this approach (Figure S1A). We verified that HDAC3 is a bona fide HIF-1 target gene using cell lines under hypoxia, overexpressing HIF-1, or knockdown of HIF-1 by quantitative real-time PCR and Western blot analysis (Figures 1A and 1B, and Figure S1B). The FADU and MCF-7 cell lines were used due to their low endogenous HIF-1 levels and low metastatic activity (epithelial in nature). The H1299 cell line was used due to its constitutively active HIF-1 and higher HDAC3 levels (mesenchymal in nature). Repression of endogenous HIF-1 by siRNA in either FADU or MCF-7 clones under hypoxia decreased the RNA and protein levels of HDAC3 to pre-hypoxic levels (Figure 1C and Figure S1C), demonstrating that HIF-1 is a major regulator of HDAC3
expression.
To determine whether HDAC3 is directly regulated by HIF-1, the function of a putative hypoxia response element (HRE) in the proximal promoter of the HDAC3 gene was tested (Figure 1D and Table S1). A 3.5 to 4 fold increase in the HDAC3 promoter activity was observed after hypoxia or transient transfection with either wild type HIF-1 or HIF-1(∆ODD) vector (a constitutively active mutant with deletion of the oxygen degradation domain) (Huang et al., 1998). A further increase in the promoter activity (~6.5 fold) was observed in cells undergoing hypoxia and overexpressing HIF-1. The inactive HIF-1 mutant (HIF-1(LCLL)) (Huang et al., 1998) failed to activate the HDAC3 promoter and site-directed mutagenesis of the putative HRE in the HDAC3 promoter prevented activation under hypoxia or HIF-1 overexpression (Figure 1E). Increased HIF-1 binding was observed after incubation of nuclear extracts from hypoxic cells with the HRE-containing oligonucleotide from the HDAC3 promoter using electrophoretic mobility shift assays (EMSAs), and a supershifted band was detected after adding either an anti-HIF-1 or anti-HIF-1-specific antibody to the nuclear extracts of hypoxic cells (Figure S1D). Competition of HIF-1 binding by unlabelled oligonucleotide containing HRE abolished the HIF-1-shifted band, and probes containing mutated HRE were not HIF-1-shifted by HIF-1 (Figures S1D, S1E). Chromatin immunoprecipitation (ChIP) assays showed that HIF-1 bound
to the HDAC3 promoter containing the HRE (253 bp) in hypoxic FADU or MCF-7 cells but not in normoxic FADU or MCF-7 cells (Figure 1F). The binding was also detected in the FADU-HIF1(∆ODD) sample but not in the FADU-cDNA3 sample; whereas knockdown of HIF-1 in H1299 cells attenuated the binding of HIF-1 to the HRE (H1299-control vs. H1299-HIF-1-si; Figure S1F, upper panels). Control experiments showed that HIF-1 bound to the HRE in the VEGF promoter (262 bp) (Figure 1F, lanes 5 & 6 and Figure S1F, lower panels). These results demonstrate that HIF-1 activates HDAC3 expression directly by binding to the HRE in the HDAC3 promoter.
HDAC3 is essential for hypoxia/HIF-1 induced EMT and metastasis
To determine whether HDAC3 is essential in HIF-1- and hypoxia-mediated EMT and metastasis, siRNA-mediated repression of HDAC3 was performed in FADU-HIF1(∆ODD) clones. Repression of HDAC3 in FADU-HIF1(∆ODD)-HDAC3-si clones caused the shift in expression of mesenchymal markers (vimentin, N-cadherin) to epithelial markers (E-cadherin, plakoglobin), compared with the control FADU-HIF1(∆ODD) clones (Figure 2A and Figure S2A). Induction of metastatic phenotypes caused by HIF-1(∆ODD) overexpression was abolished by repression of HDAC3 as shown by migration, invasion, tail vein injections and
orthotopic implantation assays (Figures 2B and 2C, and Figures S2B-S2D, S2P). Similar results were observed in H1299 clones with HDAC3 repression (Figures 2D-2F, and Figures S2E-S2G). HDAC3 overexpression also induced EMT and metastatic phenotypes in FADU clones (Figures S2H-S2P). The enzymatic activity of HDAC3 was required for the induction of EMT since overexpression of an inactive HDAC3 mutant (Y298F) did not induce EMT (Figure S2Q). An in vitro HDAC3 deacetylase assay (Yang et al., 2002) further showed that wild type HDAC3, but not the HDAC3(Y298F) mutant, deacetylated the H3K4Ac peptide (Figure S2R).
Knockdown of endogenous HDAC3 caused a complete or significant loss of EMT and inhibition of migration and invasion activity in FADU or MCF-7 clones under hypoxia even in the presence of Snail and Twist1, suggesting the critical role of HDAC3 to facilitate the regulation of EMT by different EMT regulators (Figures 3A, 3B for Western analysis, Figures 3C, 3D and Figures S3A, S3B for migration/invasion activity, and Figures S3C, S3D for mRNA analysis of EMT markers). The morphology of different clones undergoing EMT was shown and the status of epithelial or mesenchymal morphology responded to HDAC3 modulation (Figures S3E, S3F). Knockdown of endogenous HDAC3 increased apoptosis and decreased cell proliferation/growth in FADU or MCF-7 clones using flow cytometry, TUNEL, alamarBlue staining, and nude mice xenograft assays (Figures S3G-S3N),
which were consistent with the reported role of HDAC3 in apoptosis and cell proliferation (Narita et al., 2005; Spurling et al., 2008; Trivedi et al., 2008; Zhu et al., 2008a). Taken together, these results demonstrate the essential and specific role of HDAC3 in the induction of EMT and metastatic phenotypes caused by hypoxia or HIF-1 overexpression.
Specific histone marks are associated with hypoxia-induced EMT
HDAC3 deacetylates histones with a preference for different acetylated histone marks in vitro (Johnson et al., 2002; Karagianni and Wong, 2007). It also deacetylates centromeric acetylated H3K4 (H3K4Ac) (Eot-Houllier et al., 2008). Nevertheless, the physiological significance of H3K4Ac in gene transcription is largely unknown (Garcia et al., 2007). To examine the effects of HDAC3 activation on the changes of different histone marks, Western blot analysis was performed. A significant decrease in H3K4Ac and increase in methylated H3K4 (H3K4me2/me3) levels was observed in FADU or MCF-7 clones under hypoxia (Figure 4A). Knockdown of HDAC3 abolished the changes of H3K4Ac and H3K4me2/me3 levels under hypoxia (Figure 4A). In contrast, there was no change in trimethylated H3K9 (H3K9me3), acetylated H3K9 (H3K9Ac), or trimethylated H3K27 (H3K27me3) levels (Figure 4A). The decrease in acetylated H4K5 (H4K5Ac), H4K12 (H4K12Ac), and H2AK5
(H2AK5Ac) levels induced by hypoxia was not modulated by HDAC3 since this decrease remained unchanged in HDAC3 knockdown clones (Figure 4A). These results suggest a correlation between HDAC3 expression and deacetylation of H3K4Ac/methylation of H3K4. The same histone mark changes were observed using acid-extracted histone proteins or titration experiments (Figures S4A, S4B). Hypoxia also increased the activity of HDAC3 to deacetylate H3K4Ac in vitro (Figure 4B).
Quantitative ChIP (qChIP) assays were performed to examine the levels of different histone marks on the promoters of two epithelial genes (E-cadherin, plakoglobin) and two mesenchymal genes (N-cadherin, vimentin) in FADU or MCF-7 clones under hypoxia. The results showed that HDAC3 bound to the E-cadherin and plakoglobin promoters in hypoxic clones and it correlated with decreased H3K4Ac, increased H3K4me2, and increased H3K27me3 levels on these promoters (Figure 5, and Figure S5A), which indicated the existence of bivalent domains on epithelial gene promoters (Bernstein et al., 2006; Jiang et al., 2011). Decreased H3K4Ac/increased H3K4me2 levels on the mesenchymal gene (N-cadherin, vimentin) promoters were accompanied by decreased H3K27me3 levels (Figure 5, and Figure S5A) (Li et al., 2007a; Martin and Zhang, 2005). Knockdown of HDAC3 abolished the changes of H3K4Ac, H3K4me2, and H3K27me3 levels on the promoters of EMT marker genes induced by hypoxia (Figure 5, and Figure S5A), compared to no change of these
histone marks in non-EMT (HSP90, VEGF) genes (Figure S5B). However, hypoxia increased the H3K9me2/me3 levels and EZH2 binding selectively on the E-cadherin promoter but not on the N-cadherin promoter (Figures 5B, 5C, and Figures S5C, S5D); whereas hypoxia did not change the protein levels of the components of EZH2 complex (Figure S5E). Knockdown of HDAC3 decreased the H3K9me2/me3 levels and EZH2 binding on the E-cadherin promoter under hypoxia (Figures 5B, 5C). It is intriguing that HDAC3 occupied the promoters of the activated mesenchymal genes (Figure 5A, and Figure S5A) as the functions of most HDAC family members are linked to gene repression (Lane and Chabner, 2009). The regulation of H3K4Ac levels by HDAC3 under hypoxia were further confirmed in the promoters of other epithelial (claudin 1, claudin 10) or mesenchymal (VE-cadherin, fibronectin) marker genes (Figure S5F) (Creighton et al., 2010; Labelle et al., 2008). In contrast, overexpression of a HDAC3 mutant (Y298F) did not regulate the different histone marks of EMT marker genes (E-cadheirn, N-cadherin, CLDN1, CDH5) even with the binding of the mutant HDAC3 to the EMT marker gene promoters (Figure S5G). Finally, examination of tiling of the promoter (-10 kb, -5kb upstream) and exonic regions of E-cadherin and N-cadherin genes was performed. The change of H3K4Ac mark was observed throughout the promoter (-10 kb, -5kb upstream) and exonic regions of E-cadherin and N-cadherin genes; whereas the changes of other histone
marks mainly occurred around the proximal promoter region (Figures S5H, S5I). These results show that specific histone marks are changed in response to HDAC3 modulation under hypoxia in either the expression levels (Western blot) or the levels of binding to EMT marker gene promoters (qChIP).
HDAC3 recruits WDR5/HMT complex to increase H3K4-specific histone methyltransferase activity
To investigate the mechanism of global increase in H3K4me2/me3 levels during hypoxia-induced EMT, we tested whether the protein levels of the common components of the histone methyltransferase (HMT) complexes related to gene activation are induced by hypoxia (Trievel and Shilatifard, 2009; Wysocka et al., 2005). Western blot analysis showed that hypoxia induced the protein levels of WDR5 (a WD40 repeat protein), but not other common components of the HMT complexes (Ash2L, RbBP5) in two cell lines (Figure 6A). To test the link between HDAC3 and WDR5 under hypoxia, co-immunoprecipitation assays showed that WDR5 interacted with HDAC3 using different antibodies to pull down endogenous proteins or proteins overexpressed in 293T cells (Figure 6B, and Figures S6A-S6E). To confirm the recruitment of WDR5 by HDAC3 on the N-cadherin and vimentin promoters during hypoxia-induced EMT, sequential qChIP assays using the
anti-HDAC3 antibody followed by the anti-WDR5 antibody were performed. WDR5 was pulled down in the anti-HDAC3 ChIP-immunoprecipitants from hypoxic cells, but not from the HDAC3-knockdown clones under hypoxia (Figure 6C, and Figures S6F, S6G). qChIP experiments using the anti-WDR5 antibody in control vs. HDAC3 knockdown clones (normoxia or hypoxia) showed the absence of WDR5 binding to the mesenchymal gene promoters in HDAC3 knockdown clones (Figure S6H). Co-immunoprecipitation assays showed that anti-Twist1 antibody pulled down HDAC3 and WDR5 (Figure 6D) and the anti-WDR5 antibody pulled down Twist1 (Figure 6E). GST pull down assay showed a direct interaction between Twist1 and HDAC3 (Figure S6I). Sequential qChIP assay showed a co-occupancy of Twist1 and WDR5 on the N-cadherin promoter (Figure 6F). Finally, a HMT activity assay was performed using the anti-HDAC3 antibody-immunoprecipitants to test the ability of HDAC3 to increase HMT activity under hypoxia (McKinnell et al., 2008). These results showed that histone H3 tail methyltransferase activity increased in anti-HDAC3-immunoprecipitants from two cell lines under hypoxia, compared with the immunoprecipitants from these cell lines under normoxia or control IgG-immunoprecipitants (Figure 6G). In addition, the increased HMT activity methylated unmodified H3K4, H3K4me1 (monomethylation), and H3K4me2 (dimethylation) peptides, but not the control H3R4 peptide, demonstrating the increased
H3K4-specific HMT activity in the anti-HDAC3-immunoprecipitants under hypoxia (Figure 6G). These results suggest that HDAC3 recruits the WDR5/HMT complex to increase H3K4-specific HMT activity in hypoxic cells.
Co-occupancy of HDAC3 and Snail on the epithelial gene promoters and interaction between Snail and HDAC3
To test the mechanism of HDAC3-mediated repression of epithelial genes during EMT, we examined whether HDAC3 could interact with an EMT regulator. The EMT regulator Snail was chosen since Snail mediates E-cadherin repression by recruiting the Sin3A/HDAC1/HDAC2 complex and drosophila Ebi mediates Snail-dependent transcription repression through HDAC3 (Peinado et al., 2004; Qi et al., 2008). Sequential qChIP assays using an anti-HDAC3 antibody followed by an anti-Snail antibody showed a co-occupancy of HDAC3 and Snai1 on the E-cadherin and plakoglobin promoters in H1299 cells (Figure S6J). Furthermore, qChIP assays using the anti-HDAC3 antibody showed a significant decrease in HDAC3 binding to the E-cadherin and plakoglobin promoters in H1299-Snail-si clones (Figure S6K). Sequential qChIP assays using an anti-Snail antibody followed by an anti-HDAC3 antibody showed a significant decrease in HDAC3 binding to the E-cadherin and plakoglobin promoters in H1299-Snail-si clones (Figure S6L).
Co-immunoprecipitation experiments and GST pull down assay showed the direct interaction between Snail and HDAC3 (Figures S6M-S6O). Finally, reporter gene assays showed a cooperative repression of an E-cadherin promoter driven reporter construct by Snail and HDAC3 compared to Snail alone, whereas this repression was abolished when all the E-box sites were mutated (Figures S6P, S6Q). Since Twist1 interacted with HDAC3 (Figure S6I), reporter gene assay was also performed and the results showed a cooperative repression of an E-cadherin promoter driven reporter construct by Twist1 and HDAC3 compared to Twist1 alone (Figure S6R). These results indicate that HDAC3 could interact with different EMT regulators to mediate E-cadherin repression.
Direct activation of WDR5 by HIF-1and HIF-2and the role of WDR5 and
HIF-2 in the regulation of hypoxia-induced EMT phenotypes
Since WDR5 expression was induced by hypoxia (Figure 6A), we tested whether it is activated by HIF-1 or HIF-2. The results showed that WDR5 was activated by HIF-1 or HIF-2at the mRNA and protein levels in FADU clones (Figure 7A and Figure S7A). Transient transfections and ChIP experiments showed that both HIF-1 and HIF-2 directly bound to the HRE located in the proximal promoter of the WDR5 gene to activate its expression (Figure 7B, Figures S7B, S7C,
and Table S2). Western blot analysis showed that WDR5 was only partially activated by hypoxia in HIF-1 or HIF-2 knockdown clones (Figures S7D, S7E). We further tested the role of WDR5 on hypoxia-induced EMT phenotypes. In hypoxic WDR5 knockdown clones there was preservation of epithelial transition but mesenchymal activation was abolished (Figure 7C, and Figure S7F), suggesting that WDR5 is a major regulator of hypoxia-induced mesenchymal gene activation. Knockdown of WDR5 decreased the migration/invasion activity of the clones under hypoxia (Figure 7D, and Figures S7G-S7H). Finally, knockdown of HIF-2 abolished the EMT phenotypes and decreased the migration/invasion activity of the clones under hypoxia (Figures S7I-S7L). These results show the direct activation of WDR5 by both HIF-1 and HIF-2 and the critical role of WDR5 and HIF-2 in the regulation of hypoxia-induced EMT and metastatic phenotypes.
Co-expression of HIF-1 WDR5, and HDAC3 correlates with metastasis and
predicts a worse prognosis of head and neck squamous cell carcinoma (HNSCC) patients
Tumors with increased HIF-1 activity are more likely to develop metastasis and correlate with poor survival (Gupta and Massague, 2006; Harris, 2002; Semenza, 2002). To investigate whether HDAC3 activation by HIF-1 indeed occurs in human
cancers and to evaluate the prognostic significance of HIF-1/HDAC3/WDR5 co-expression, tissue-microarray immunohistochemistry analysis of HIF-1, HDAC3 and WDR5 expression was performed in 88 sets of HNSCC samples (Table S3.1, samples from a representative case was shown in Figure S7M). Tumors with increased HIF-1 expression significantly correlated with HDAC3 overexpression (P=0.030, Table S4), and the expression level of HDAC3 was also associated with WDR5 (P=0.001, Table S4). Prognostic prediction analysis showed that co-expression of HIF-1/HDAC3/WDR5 had a significantly shorter metastasis-free period and a significantly worse outcome than did non-co-expression cases (P=0.003; Figures S7N, S7O). The prognostic effect of HIF-1/HDAC3/WDR5 co-expression was independent of other prognostic markers (advanced T stage, N stage) (P=0.029; Table S5.1). A different series of oral cancer patients were analyzed using qRT-PCR to categorize patients according to the mRNA levels. The results also showed the prognostic value of HIF-1/HDAC3/WDR5 co-expression to predict the worse outcome of oral cancer patients (P=0.024; Figure S7P). Collectively, the correlation analysis indicates that activation of HDAC3 by HIF-1 and association of WDR5 with HDAC3 indeed occur in HNSCC samples, and survival analysis supports the prognostic value of co-expression of HIF-1/HDAC3/WDR5 in HNSCC cases.
Discussion
Our results demonstrate that different chromatin modifiers are regulated by hypoxia to coordinately regulate EMT gene expression. WDR5, one of the core subunits of MLL and SET1 (hCOMPASS) histone methyltransferase complexes, is regulated by hypoxia, suggesting that these core subunits may be regulated by different environmental cues or extracellular signaling. It is intriguing that HDAC3-containing HMT complex had increased H3K4-specific HMT activity under hypoxia, suggesting that participation of HDAC3 in the HMT complex enhanced the HMT activity which transformed the repressor role of HDAC3 into the activator role on the mesenchymal gene promoters under hypoxia. This result points to the ability of chromatin modifiers to switch between different roles under different environmental conditions through interaction with different complexes. The dual role of HDAC3 and the novel role of WDR5 are thereby revealed. It will be interesting to test whether the transition of EMT marker genes is regulated through similar mechanisms under other signaling pathways such as Wnt, Notch, TGF-, etc (Peinado et al., 2007; Thiery et al., 2009; Yang and Weinberg, 2008).
Different functions of HDAC3 such as regulation of circadian metabolic physiology, S phase progression/DNA damage control, sister chromatid cohesion, and
liver metabolic transcription were recently delineated (Alenghat et al., 2008; Bhaskara et al., 2008; Eot-Houllier et al., 2008; Knutson et al, 2008) in addition to its co-repressor role (Karagianni and Wong, 2007; Lane and Chabner, 2009). The role of H3K4Ac mark is poorly addressed regarding its physiological significance (Garcia et al, 2007). Recently, a chromodomain switch mediated by H3K4Ac was shown to regulate heterochromatin assembly in fission yeast (Xhemalce and Kouzarides, 2010). Our results demonstrate the role of hypoxia-induced HDAC3 to deacetylate H3K4Ac and regulate EMT marker gene expression. Since histone acetylation may neutralize the charge on histone and/or facilitate the binding of chromatin remodeling factors, deacetylation of H3K4Ac by HDAC3 may induce a less accessible chromatin state. To compensate for this effect, HDAC3 would interact with the HMT complex to increase H3K4 methylation levels and promote gene activation. The increased H3K4 methylation levels on mesenchymal gene promoters could activate mesenchymal gene expression during EMT; whereas the increased H3K4 methylation and H3K27me3 levels on epithelial gene promoters may serve as bivalent domains to allow for later mesenchymal-epithelial transition (MET). Similar bivalent domains (i.e. co-existence of H3K27me3 and H3K4me3) were observed in the developmental genes of ES cells (Bernstein et al., 2006; Jiang et al., 2011). It is proposed that bivalent domains would silence such genes while keeping them poised for activation. Due to the constant
transition between epithelial and mesenchymal status during EMT/MET, bivalent domains could facilitate the smooth transition of EMT marker genes. Finally, it appears that H3K27me3 levels could faithfully reflect the repression/activation status induced by hypoxia (Martin and Zhang 2005). The mechanism of signal relaying from H3K4Ac deacetylation/H3K4 methylation induced by HDAC3 to selective regulation of H3K27 trimethylation on different promoters remains to be determined (a model is depicted in Figure 7E). Investigation of a possible crosstalk (direct or indirect) between HDAC3 and other chromatin modifier complexes may further reveal the unknown mechanism. Recent results showed that HDAC3 is essential for the maintenance of chromatin structure and genomic instability (Bhaskara et al., 2010), whereas our results discovered the role of HDAC3 induction in the regulation of hypoxia-induced EMT.
Finally, our results support the hypothesis that chromatin modifications mediated by chromatin modifiers are essential for EMT to proceed in addition to the transcription role of numerous EMT regulators. We propose that the interplay between HDAC3 and WDR5 coordinately regulates hypoxia-induced EMT. The information obtained in the present study will provide insightful guidance for future development of therapeutic agents against hypoxic tumors, hypoxia-induced metastasis, cancer stemness, and hypoxia-mediated organ fibrosis (Guarino et al.,
2009; Higgins et al., 2008; Thiery et al., 2009; Yang and Weinberg, 2008), thereby contributing to personalized management of cancer patients.
Experimental Procedures
Cell culture, oxygen deprivation, protein extraction, Western blot analysis, RNA extraction, quantitative real-time PCR, in vitro migration/invasion, tail vein metastatic assay, orthotopic implantation assay, and immunofluorescence staining
Protein extraction, Western blot, RNA extraction and qRT-PCR were performed following standard protocols as described (Yang et al., 2008). Boyden chambers were used in in vitro migration/invasion assay. NOD-SCID mice were used in tail vein injection or orthotopic implantation assays through injecting tumor cells into tail vein or local areas. All the details were described in the Supplemental Experimental Procedures section. The characteristics of the antibodies used were listed (Table S6). The sequences of primers used in the quantitative real-time PCR experiments were shown (Table S7).
Plasmids, stable transfection, transient transfection, luciferase assays, electrophoretic mobility shift assay (EMSA), and Chromatin immunoprecipitation (ChIP) assays
Co-transfections of different expression vectors with a reporter construct were performed in 293T cells under normoxia or hypoxia. Either no antibody, IgG control, anti-HIF-1 or anti-HIF-2 antibody was used in ChIP assays. All the details were
described in Supplemental Experimental Procedures (Yang et al., 2008).
Quantitative chromatin immunoprecipitation (qChIP) and sequential qChIP assays
For qChIP assay, DNA samples were quantified by the SYBR® Green assay using SYBR® Green PCR Master Mix (Applied Biosystems) with specific primer. Data were analyzed by the CT method and plotted as % input DNA. qChIP values were calculated by the following formula: % input recovery =[100/(input fold dilution/bound fold dilution)]×2(input CT － bound CT). In sequential ChIP experiment, cells were crosslinked by incubation at RT with 1% formaldehyde for 15 min and stopped the reaction by adding 1 M glycine to final concentration of 0.125 M. Fixed cells were harvested in 5 ml of SDS buffer (50mM Tris, pH 8.0, 0.5% SDS, 100 mM NaCl, 5 mM EDTA, and proteaseinhibitors). After centrifugation, cell pellets were suspended in 2 ml of IP buffer (100 mM Tris, pH 8.6, 0.3% SDS, 1.7% Triton X-100, 5 mM EDTA). After sonication, the lysates were incubated with protein A beads (50% protein A beads slurry) conjugating antibodies specific for different antibodies or IgG control. The percentage of IgG control pull down consistently reached below 0.02% of input. For re-ChIP reaction, DNA-protein complexes were eluted by incubation at 37℃ with 25 l 10 mM dithiothreitol (DTT) for 30 min. After centrifugation, supernatant was collected and diluted with IP buffer to total volume 500 l and
subjected to the ChIP procedure again. The primers used to scan each gene were shown in Table S7. The promoter regions used in Figure S5 & S6 were described in Table S7. The characterization of the anti-H3K4Ac antibody was described in Supplemental Experimental Procedures.
Co-immunoprecipitation, GST pull down assays, in vitro HDAC3 deacetylase and histone methyltransferase (HMT) activity assays
Co-immunoprecipitation and GST pull down assays were performed following standard procedures as described using different antibodies (Table S6) (Tsai et al., 2009). In vitro deacetylase assay was performed as described (Yang et al., 2002). Briefly, the 293T cells were transfected with CMV, HDAC3 or pFlag-HDAC3 (Y298F) plasmid. The cell lysates were immunoprecipitated with the anti-Flag antibody or control IgG and incubated with H3K4Ac peptide for reaction. The details were described in Supplemental Experimental Procedures. HMT activity assays were performed as described (McKinnell et al., 2008). Briefly, cell lysates were incubated with the anti-HDAC3 antibody or control IgG followed by reacting in a reaction mixture containing the unmodified peptide, peptide with K4 position modified (monomethylation or dimethylation), or peptide with replacement by arginine at the K4 position. The final products were detected by radiography after exposure to films for 3 weeks. The details were described in Supplemental
Experimental Procedures.
Study population, sample collection and tissue microarray construction Eighty-eight HNSCC patients who underwent treatment at Taipei Mackay Memorial Hospital and Taipei Veterans General Hospital between January 2001 and December 2004 were retrospectively analyzed. This study has been approved by the Institutional Review Board of Taipei Veterans General Hospital. The clinical characteristics of 88 HNSCC patients are illustrated (Table S3). Primary tumor samples and the corresponding non-cancerous matched tissue were obtained during surgery. A high-density tissue microarray (TMA) was constructed as described in Supplemental Experimental Procedures. For the 72 oral cancer patient samples used in qRT-PCR analysis, the samples were collected at National Taiwan University Hospital (NTUH) and the study was approved by the IRB of NTUH.
Immunohistochemstry (IHC), validation of antibodies and scoring
The sample processing and IHC procedure for determining the immunoreactivity of HIF-1HDAC3 and WDR5 were described in Supplemental Experimental Procedures (Weichert et al., 2008; Yang et al., 2008). For these markers, we defined the IHC result as a positive one only if the nuclear expression of the target protein was identified in 50% of tumor cells. All the antibodies used in IHC are listed in Table S6.
The Kaplan-Meier estimate was used for metastasis-free and overall survival analysis. The rest of the statistical analysis was performed as described in Supplemental Experimental Procedures.
References
Alenghat, T., Meyers, K., Mullican, S.E., Leitner, K., Adeniji-Adele, A., Avila, J., Bućan, M., Ahima, R.S., Kaestner, K.H., and Lazar, M.A. (2008) Nuclear receptor corepressor and histone deacetylase 3 govern circadian metabolic physiology. Nature 456, 997-1000.
Ang, Y.S., Tsai, S.Y., Lee, D.F., Monk, J., Su, J., Ratnakumar, K., Ding, J., Ge, Y., Darr, H., Chang, B. et al. (2011) Wdr5 Mediates Self-Renewal and Reprogramming via the Embryonic Stem Cell Core Transcriptional Network. Cell 145, 183-197.
Bernstein, B.E., Mikkelsen, T.S., Xie, X., Kamal, M., Huebert, D.J., Cuff, J., Fry, B., Meissner, A., Wernig, M., Plath K. et al. (2006) A bivalent chromatin structure marks key developmental genes in embryonic stem cells. Cell 125, 315-326. Bhaskara, S., Chyla, B.J., Amann, J.M., Knutson, S.K., Cortez, D., Sun, Z.W., and
Hiebert, S.W. (2008) Deletion of histone deacetylase 3 reveals critical roles in S phase progression and DNA damage control. Mol. Cell 30, 61-72.
Bhaskara, S., Knutson, S.K., Jiang, G., Chandrasekharan, M.B., Wilson, A.J., Zheng, S., Yenamandra, A., Locke, K., Yuan, J.L., Bonine-Summers, A.R. et al. (2010) Hdac3 is essential for the maintenance of chromatin structure and genome stability. Cancer Cell 18, 436-447.
M., and Schibler, U. (2005) PERIOD1-associated proteins modulate the negative limb of the mammalian circadian oscillator. Science 308, 693-696.
Cao, R., and Zhang, Y. (2004) The functions of E(Z)/EZH2-mediated methylation of lysine 27 in histone H3. Cur. Opin. Genet. Dev. 14, 155-64.
Creighton, C.J., Chang, J.C., and Rosen, J.M. (2010) Epithelial-mesenchymal transition (EMT) in tumor-initiating cells and its clinical implications in breast cancer. J. Mammary Gland Biol. Neoplasia. 15, 253-60.
Eot-Houllier, G., Fulcrand, G., Watanabe, Y., Magnaghi-Jaulin, L., and Jaulin, C. (2008) Histone deacetylase 3 is required for centromeric H3K4 deacetylation and sister chromatid cohesion. Genes & Dev. 22, 2639-2644.
Garcia, B.A., Hake, S.B., Diaz, R.L., Kauer, M., Morris, S.A., Recht, J., Shabanowitz, J., Mishra, N., Strahl, B.D., Allis, C.D. et al. (2007) Organismal differences in post-translational modifications in histones H3 and H4. J. Biol. Chem. 282, 7641-7655.
Gori, F., Zhu, E.D., and Demay, M.B. (2009) Perichondrial expression of Wdr5 regulates chondrocyte proliferation and differentiation. Dev. Biol. 329, 36-43. Guarino, M., Tosoni, A., and Nebuloni, M. (2009) Direct contribution of
epithelium to organ fibrosis: epithelial-mesenchymal transition. Hum. Pathol. 40, 1365-1376.
Cell 127, 679-695.
Harris, A.L. (2002) Hypoxia-a key regulatory factor in tumor growth. Nat. Rev. Cancer 2, 38-47.
Higgins, D.F., Kimura, K., Iwano, M., and Hasse, V.H. (2008) Hypoxia-inducible signaling in the development of fibrosis. Cell Cycle 7, 1128-1132.
Huang, L.E., Gu, J., Schau, M., and Bunn, H.F. (1998) Regulation of hypoxia-inducible factor 1alpha is mediated by an O2-dependent degradation domain via the ubiquitin-proteasome pathway. Proc. Natl. Acad. Sci. USA 95, 7987-7992. Jiang, H., Shukla, A., Wang, X., Chen, W.Y., Bernstein, B.E., and Roeder, R.G.
(2011) Role for Dpy-30 in ES cell-fate specification by regulation of H3K4 methylation within bivalent domains. Cell 144, 513-25.
Johnson, C.A., White, D.A., Lavender, J.S., O'Neill, L.P., and Turner, B.M. (2002) Human class I histone deacetylase complexes show enhanced catalytic activity in the presence of ATP and co-immunoprecipitate with the ATP-dependent chaperone protein Hsp70. J. Biol. Chem. 277, 9590-9597.
Karagianni, P., and Wong, J. (2007) HDAC3: taking the SMRT-N-CoRect road to repression. Oncogene 26, 5439-5449.
Kim, T., and Buratowski, S. (2009) Dimethylation of H3K4 by Set1 recruits the Set3 histone deacetylase complex to the 5’ transcribed regions. Cell 137, 259-272. Kimura, H., Weisz, A., Ogura, T., Hitomi, Y., Kurashima, Y., Hashimoto, K.,
D'Acquisto, F., Makuuchi, M., and Esumi, H. (2001) Identification of hypoxia-inducible factor 1 ancillary sequence and its function in vascular endothelial growth factor gene induction by hypoxia and nitric oxide. J. Biol. Chem. 276, 2292-2298.
Knutson, S.K., Chyla, B.J., Amann, J.M., Bhaskara, S., Huppert, S.S., and Hiebert, S.W. (2008) Liver-specific deletion of histone deacetylase 3 disrupts metabolic transcriptional networks. EMBO J. 27, 1017-1028.
Kwon, S.H., and Workman, J.L. (2008) The heterochromatin protein 1 (HP1) family: put away a bias toward HP1. Mol. Cells 26, 217-227.
Labelle, M., Schnittler, H.J., Aust, D.E., Friedrich, K., Baretton, G., Vestweber, D., and Breier, G. (2008) Vascular endothelial cadherin promotes breast cancer progression via transforming growth factor beta signaling. Cancer Res. 68, 1388-97.
Lane, A.A. and Chabner, B.A. (2009) Histone deacetylase inhibitors in cancer therapy. J. Clin. Oncol. 27, 5459-5468.
Li, B., Carey, M., and Workman, J.L. (2007a) The role of chromatin during transcription. Cell 128, 707-719.
Li, B., Gogol, M., Carey, M., Lee, D., Seidel, C., and Workman, J.L. (2007b) Combined action of PHD and chromo domains directs the Rpd3S HDAC to transcribed chromatin. Science 316, 1050-1054.
Lin, T., Ponn, A., Hu, X., Law, B.K., and Lu, J. (2010) Requirement of the histone demethylase LSD1 in Snai1-mediated transcriptional repression during epithelial-mesenchymal transition. Oncogene 29, 4896-904.
Martin, C., and Zhang, Y. (2005) The diverse functions of histone lysine methylation. Nat. Rev. Mol. Cell Biol. 6, 838-849.
McKinnell, I.W., Ishibashi, J., Le Grand, F., Punch, V.G., Addicks, G.C., Greenblatt, J.F., Dilworth, F.J., and Rudnicki, M,A. (2008) Pax7 activates myogenic genes by recruitment of a histone methyltransferase complex. Nature Cell Biol. 10, 77-84.
Narita, N., Fujieda, S., Tokuriki, M., Takahashi, N., Tsuzuki, H., Ohtsubo, T., and Matsumoto, H. (2005) Inhibition of histone deacetylase 3 stimulates apoptosis induced by heat shock under acidic conditions in human maxillary cancer. Oncogene 24, 7346-7354.
Peinado, H., Ballestar, E., Esteller, M., and Cano, A. (2004) Snail mediates E-cadherin repression by the recruitment of the Sin3A/histone deacetylase 1 (HDAC1)/HDAC2 complex. Mol. Cell. Biol. 24, 306-319.
Peinado, H., Olmedo, D., and Cano, A. (2007) Snail, ZEB, and bHLH factors in tumour progression: an alliance against the epithelial phenotype? Nat. Rev. Cancer 7, 415-428.
Qi, D., Bergman, M., Aihara, H., Nibu, Y., and Mannervik M. (2002) Drosophila Ebi mediates Snail-dependent transcriptional repression through HDAC3-induced histone deacetylation. EMBO J. 27, 898-909.
Semenza, G.L. (2002) HIF-1 and tumor progression: pathophysiology and therapeutics. Trends Mol. Med. 8 (4 Suppl), S62-67.
Spurling, C.C., Godman, C.A., Noonan, E.J., Rasmussen, T.P., Rosenberg, D.W., and Giardina, C. (2008) HDAC3 overexpression and colon cancer cell proliferation and differentiation. Mol. Carcinog. 47, 137-147.
Thiery, J.P., Acloque, H., Huang, R.Y., and Nieto, M.A. (2009) Epithelial-mesenchymal transitions in development and disease. Cell 139, 871-890.
Trievel, R., and Shilatifard, A. (2009) WDR5, a complexed protein. Nature Struc. Mol. Biol. 16, 678-680.
Trivedi, C.M., Lu, M.M., Wang, Q., and Epstein, J.A. (2008) Transgenic overexpression of Hdac3 in the heart produces increased postnatal cardiac myocyte proliferation but does not induce hypertrophy. J. Biol. Chem. 283, 26484-26489.
Tsai, Y.P., Yang, M.H., Huang, C.H., Chang, S.Y., Liu, C.J., Chen, P.M., Teng, S.C., and Wu, K.J. (2009) Interaction between HSP60 and -catenin promotes metastasis. Carcinogenesis 30, 1049-1057.
Wang, Y.Y., Liu, L.J., Zhong, B., Liu, T.T., Li, Y., Yang, Y., Ran, Y., Li, S., Tien, P., and Shu, H.B. (2010) WDR5 is essential for assembly of the VISA-associated signaling complex and virus-triggered IRF3 and NF-kappaB activation. Proc. Natl. Acad. Sci. USA 107, 815-820.
Wang, Z., Zang, C., Cui, K., Schones, D.E., Barski, A., Peng, W., and Zhao, K. (2009) Genome-wide mapping of HATs and HDACs reveals distinct functions in active and inactive genes. Cell 138, 1019-1031.
Weichert, W., Röske, A., Niesporek, S., Noske, A., Buckendahl, A.C., Dietel, M., Gekeler, V., Boehm, M., Beckers, T., and Denkert, C. (2008) Class I histone deacetylase expression has independent prognostic impact in human colorectal cancer: specific role of class I histone deacetylase in vitro and in vivo. Clin. Cancer Res. 14, 1669-1677.
Wysocka, J., Swigut, T., Milne, T.A., Dou, Y., Zhang, X., Burlingame, A.L., Roeder, R.G., Brivanlou, A.H., and Allis CD. (2005) WDR5 associates with histone H3 methylated at K4 and is essential for H3 K4 methylation and vertebrate development. Cell 121, 859-872.
Xhemalce, B., and Kouzarides, T. (2010) A chromodomain switch mediated by histone H3 lys 4 acetylation regulates heterochromatin assembly. Genes & Dev. 24, 647-652.
Yang, J., and Weinberg, R.A. (2008) Epithelial-mesenchymal transition: at the crossroads of development and tumor metastasis. Dev. Cell 14, 818-829.
Yang, M.H., and Wu, K.J. (2008) TWIST activation by hypoxia inducible factor-1 (HIF-1): implications in metastasis and development. Cell Cycle 7, 2090-2096. Yang, M.H., Wu, M.Z., Chiou, S.H., Chang, S.Y., Chen, P.M., Liu, C.J., Teng,
S.C., and Wu, K.J. (2008) Direct regulation of TWIST by HIF-1promotes metastasis. Nature Cell Biol. 10, 295-305.
Yang, W.M., Tsai, S.C., Wen, Y.D., Fejer, G., and Seto, E. (2002) Functional domains of histone deacetylase-3. J. Biol. Chem. 277, 9447-54.
Zhu, C.C., Bornemann, D.J., Zhitomirsky, D., Miller, E.L., O’Conner, M.B., and Simon, J.A. (2008a) Drosophila histone deacetylase-3 controls imaginal disc size through suppression of apoptosis. PLoS Genet. 4, e1000009.
Zhu, E.D., Demay, M.B., and Gori, F. (2008b) Wdr5 is essential for osteoblast differentiation. J. Biol. Chem. 283, 7361-7367.
Acknowledgments
We greatly appreciate Drs. Y. Zhang and L.J. Juan for suggestions and comments on the manuscript. We thank Drs. L.E. Huang and W.M. Yang for the gifts of pHA-HIF1/pHA-HIF1(∆ODD)/pHA-HIF1(LCLL) plasmids and HDAC3 cDNA plasmid, respectively. We appreciate Drs. T.Y. Chou and W.Y. Li of the Department of Pathology, Taipei Veterans General Hospital for providing expert opinions on pathology reading and IHC analysis. We are grateful to Dr. K.W. Chang of the Institute of Oral Biology, National Yang-Ming University for TMA construction. We thank Dr. J.J. Hung, C.Y. Hsieh, J.Y. Tang and Yang-Ming Genomic Center for technical assistance. The authors declare no competing financial interests. This work was supported in part by National Science Council & Frontier grant (NSC100-2321-B-010-011, NSC97-2320-B-010-029, NSC-97-2311-B-010-007) (K.J.W.), National Research Program for Genomic Medicine (NSC100-3112-B-010-003)(K.J.W.), Taipei Veterans General Hospital (VGH98-ER2-008)(M.H.Y.), (VGH98-ER2-009)(S.Y.C.), (VGH99-ER2-009) (K.J.W.), center of excellence for cancer research at Taipei Veterans General Hospital (DOH100-TD-C-111-007), a grant from Ministry of Education, Aim for the Top University Plan (100AC-T505)(K.J.W.), and National Health Research Institutes (NHRI-EX-100-9931BI)(K.J.W.).
Figure Legends
Figure 1. Direct activation of HDAC3 expression by HIF-1.
(A) Upper: fold change of mRNA levels of HIF-1, VEGF, HDAC3 and HDAC4 by real-time PCR analysis in FADU/MCF-7 cells under normoxia (N) or hypoxia (H); lower: Western blot analysis of HDAC3 expression. The different bar graph symbols represented different species of molecules, which applied to panels (A)-(C).
(B) Decreased expression of HIF-1/VEGF/HDAC3 mRNA levels and HDAC3 protein levels in H1299-HIF1-si clones. Knockdown of unrelated protein topoisomerase 3 top3-si) and transfection with an empty vector (H1299-cont.) were used as controls. 1 & 2 indicate the samples from two different clones. (C) Knockdown of endogenous HIF-1 abolishes the induction of HDAC3 (mRNA and protein levels). Control experiments to knockdown Top3 did not change the levels of HIF-1 or HDAC3 expression (Figure S1G). The asterisk (*) indicates statistical significance (P<0.05) between experimental and control clones.
(D) Schematic representation of the HDAC3 promoter region and the reporter constructs used in HIF-1 transfection experiments. The constructs contained wild type (pXP2-HDAC3-HRE) or mutated (pXP2-HDAC3-mut) HRE located -37 to -33 bp upstream of the transcription start site of HDAC3.
(E) Luciferase activity of pXP2-HDAC3-HRE or pXP2-HDAC3-mut after co-transfection of different expression constructs under normoxia or hypoxia. The
luciferase activity/-galactosidase of 293T cells co-transfected with pXP2-HDAC3/pcDNA3 control vector under normoxia was applied as the baseline control. The asterisk (*) indicates statistical significance (P<0.05) between experimental and control transfections. The protein levels of different expression vectors were shown (Figure S1H).
(F) Chromatin immunoprecipitation (ChIP) analysis. Chromatin was incubated with IgG control or anti-HIF-1 antibody. The 253-bp fragment contains the HRE; whereas the 202-bp fragment does not contain any HRE in the HDAC3 promoter (upper 4 lanes). The 262-bp fragment contains the HRE in the VEGF promoter (lower 2 lanes). Input, 2% of total input lysate.
Error bars indicate standard deviations (s.d.) of duplicate mRNA levels by qRT-PCR (A-C) or triplicate luciferase activity (E).
Figure 2. Knockdown of HDAC3 in FADU-HIF1(∆ODD) or H1299 clones
reverts EMT and metastasis.
(A) Western blot analysis of HIF-1(∆ODD), HDAC3, epithelial and mesenchymal markers in FADU-HIF1(∆ODD)-HDAC3-si vs. FADU-HIF1(∆ODD) control clones. 1 & 2 indicate the samples from two different clones.
(B) Fold change of migration and invasion activity of cDNA3, HIF1(∆ODD) and HIF1(∆ODD)-HDAC3-si clones. The
FADU-HIF1(∆ODD)-1 clone was selected as the control group.
(C) In vivo metastatic ability of cDNA3, HIF1(∆ODD) and FADU-HIF1-HDAC3-si clones as assayed by tail vein injection (open bars) or orthotopic implantation (closed bars) methods. The FADU-HIF1(∆ODD)-1 clone was selected as the control group.
(D) EMT marker changes in control vs. HDAC3-si clones. H1299-top3-si clone was also used as a control.
(E) Fold change of migration and invasion activity of control vs. H1299-HDAC3-si clones. H1299-control clone was used as a control.
(F) In vivo metastatic ability of H1299 control and H1299-HDAC3-si clones as assayed by tail vein injection methods. H1299-control clone was used as a control. The asterisk (*) indicates statistical significance (P<0.05) between experimental and control clones.
Error bars indicate standard deviations (s.d.) of quadruple measurement of migration/invasion activity (B, E) or metastastic tumor nodules in mice experiments (n=6)(C, F).
Figure 3. Inhibition of the EMT phenotypes and migration/invasion activity by
HDAC3 knockdown in two cell lines under hypoxia.
FADU-HDAC3-si or MCF7-FADU-HDAC3-si clones under hypoxia even in the presence of Snail and Twist1. 1 & 2 indicate the samples from two different clones. N: normoxia; H: hypoxia.
(C) & (D) Inhibition of in vitro migration/invasion activity in FADU-HDAC3-si or MCF7-HDAC3-si clones vs. the FADU or MCF-7 control clone (under normoxia (N) or hypoxia (H)). The asterisk (*) indicates statistical significance (P<0.05) between experimental and control clones (i.e. FADU or MCF-7 control clones in A, B). For C and D panels, normoxic samples of each clone were used as controls vs. their hypoxic counterparts. Error bars indicate standard deviations (s.d.) of quadruple measurement of migration/invasion activity (C, D).
Figure 4. Specific histone marks are associated with hypoxia-induced EMT using Western blot analysis and in vitro deacetylase assay.
(A) Western blot analysis showed the decrease in H3K4Ac and increase in H3K4me2/me3 levels in hypoxic clones. Other histone marks are either not changed under hypoxia or not modulated by HDAC3. 1 & 2 indicate the samples from two different clones. Total histone 3, 4, or 2A levels were used as controls. N: normoxia; H: hypoxia.
(B) Increased activity of HDAC3 to deacetylate the H3K4Ac peptide in hypoxic extracts using in vitro deacetylase assay. Anti-HDAC3 antibody was used to pull
down the endogenous HDAC3 proteins and test its activity. WCE: whole cell extracts.
Figure 5. Specific histone marks are associated with hypoxia-induced EMT using qChIP analysis of promoters.
(A)qChIP results using different antibodies to measure the levels of different histone marks on the promoters of EMT marker genes. The covered regions for each gene were described in Table S7. The asterisk (*) indicates statistical significance (P<0.05) between experimental and control clones.
(B) & (C) qChIP results showed the increased H3K9me2/me3 levels and EZH2 binding on the E-cadherin promoter.
Error bars indicate standard deviations (s.d.) of duplicate qChIP values for each sample (A-C).
Figure 6. Induction of WDR5 by hypoxia, interaction between WDR5 and HDAC3, and increased HMT activity in hypoxic cells.
(A) Western blot analysis showed the increased WDR5 levels in hypoxic FADU and MCF-7 cells. N: normoxia; H: hypoxia.
(B) Co-immunoprecipitation assays showed that the anti-WDR5 antibody pulled down endogenous HDAC3 and other components of the HMT complex in FADU cells under hypoxia. IgG was used as a negative control.
(C) Sequential qChIP assays showed the recruitment of WDR5 by HDAC3 on the N-caherin/vimentin gene promoters -HDAC3, -WDR5: antibody against HDAC3 or WDR5. The asterisk (*) indicates statistical significance (P<0.05) between experimental and control clones.
(D) Co-immunoprecipitation assay showed that the antibody against Twist1 pulled down HDAC3 and WDR5. IgG was used as a negative control. WCE: whole cell extracts.
(E) Co-immunoprecipitation assay showed that the antibody against WDR5 pulled down Twist1. IgG was used as a negative control antibody. WCE: whole cell extracts. (F) Sequential qChIP assays showed that Twist1 recruited WDR5 on the N-cadherin promoter containing the E-box site in FADU or MCF-7 cells under hypoxia. IgG was used as a negative control. GAPDH promoter was used as an input control.
(G) HMT assays showed the increase in H3K4-specific HMT activity in two cell lines under hypoxia. Peptides used in the assay were described in Supplemental Experimental Procedures. -HDAC3: antibody against HDAC3. The assays were performed in the presence of DNase.
Error bars indicate standard deviations (s.d.) of duplicate ChIP values for each sample (C, F).
the regulation of hypoxia-induced EMT phenotypes, and a model to depict the role of chromatin modifiers in hypoxia-induced EMT.
(A) Western blot analysis showed that overexpression of HIF-1(ODD) or HIF-2 induced WDR5 expression.
(B) Transient transfection experiments identified a HRE (CGTG) located at the position -643 to -640 bp upstream of the transcription start site of the WDR5 proximal promoter. Mutation of this site abolished activation by HIF-1 or HIF-2. Open box represented the location of the HRE. The luciferase activity/-galactosidase of 293T cells co-transfected with pXP2-WDR5-900/pcDNA3 control vector under normoxia was applied as the baseline control. The asterisk (*) indicates statistical significance (P<0.05) between experimental and control transfections. The protein levels of different expression vectors were shown (Figure S7C).
(C) Knockdown of WDR5 in hypoxic clones abolished the activation of mesenchymal genes with the preservation of epithelial gene repression.
(D) Knockdown of WDR5 in FADU clones decreased the in vitro migration/invasion activity induced by hypoxia.
(E) A model of hypoxia-induced chromatin modifiers and the changes in associated histone marks which lead to hypoxia-induced EMT. Empty arrowheads indicate selective regulation of either epithelial or mesenchymal genes. Question mark
indicates unknown mechanism. Upward or downward dark arrowheads indicate increased or decreased H3K27me3 levels.
Error bars indicate standard deviations (s.d.) of triplicate luciferase activity (B) or quadruple measurement of migration/invasion activity (D).
