Hypoxic Repression of STAT1 and Its Downstream Genes by a Pvhl/HIF-1 Target DEC1/STRA13

Home , MyoD, STAT1

Oncogene (2007) 26, 802–812 & 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.00 www.nature.com/onc ORIGINAL ARTICLE Hypoxic repression of STAT1 and its downstream genes by a pVHL/HIF-1 target DEC1/STRA13

SV Ivanov1,2, K Salnikow2, AV Ivanova3, L Bai4 and MI Lerman2

1Basic Research Program, SAIC-Frederick, Inc., Frederick, MD, USA; 2National Cancer Institute at Frederick, National Institutes of Health, Frederick, MD, USA; 3Department of Cardiothoracic Surgery, NYU Medical Center, New York, NY, USA and 4Department of Internal Medicine, University of Michigan, Ann Arbor, MI, USA

DEC1/STRA13 is a bHLH type transcriptional regulator transcriptional factor STRA13/DEC1 is one of them involved with immune regulation, hypoxia response and (Ivanov et al., 1998, 2001), with an as yet unclear role in carcinogenesis. We recently demonstrated that STRA13 the hypoxic response and poorly investigated down- interacts with STAT3 in the transcriptional activation of stream targets.STRA13 is expressed in differentiating STAT-dependent promoters. Here, we pursue STRA13 embryonic and adult tissues (Boudjelal et al., 1997; Shen involvement in the JAK/STAT pathway by studying its et al., 1997), responds to a broad range of growth stimuli role in STAT1 expression. First, we showed that VHL (Shen et al., 2001; Zawel et al., 2002), and serves as a key deficiency or HIF-1 activation resulted in the repression of regulator of lymphoid cell development and activation endogenous STAT1 mediated by STRA13. We then (Sun et al., 2001; Seimiya et al., 2004). Transcriptional characterized the STAT1 proximal promoter to assess repression activity of STRA13 can be established its response to STRA13 by transient coexpression in a through direct DNA binding to the E-box (St-Pierre luciferase reporter assay. Using sequential truncation and et al., 2002; Zawel et al., 2002; Azmi et al., 2003) via its site-directed mutagenesis of the STAT1 promoter with basic region, or through indirect involvement in STRA13 deletion constructs, we showed that the STRA13 transcriptional modulation via interaction with multiple C-terminal trans-activation domain, which is known to transcription factors, such as USF (Dhar and Taneja, bind HDAC1, mostly determines the repressive activity. 2001), TBP and TFIIB (Boudjelal et al., 1997), MyoD Involvement of HDAC activity in STAT1 regulation (Azmi et al., 2003), and MSP58 (Ivanova et al., 2005), was validated by TSA inhibition and chromatin immuno- or with the HDAC complex components (Sun and precipitation (ChIP) assay. Thus, we demonstrate that Taneja, 2000).Recently, our data on the interaction of STRA13-mediated repression of STAT1 transcription STRA13 with STAT3 (Ivanova et al., 2004) implicated utilizes an HDAC1-dependent mechanism. Furthermore, STRA13 in the JAK/STAT signal transduction path- we show that targets of unphosphorylated STAT1, such as way that plays a central role in regulation of the antigen presenting genes and CASP1, are also repressed immune system (Bach et al., 1997; Shuai and Liu, 2003). by hypoxia possibly through the same STRA13-mediated STAT1 and STAT3 are that two principal signal trans- mechanism. Thus, the newly discovered link between HIF- ducers in this pathway that serve disparate physiological 1 and STAT1 reveals a previously unknown role of roles through transcriptional activation of different STRA13 in hypoxia and carcinogenesis. targets (Qing and Stark, 2004).In carcinogenesis, these Oncogene (2007) 26, 802–812. doi:10.1038/sj.onc.1209842; proteins have opposite effects on cell growth and published online 31 July 2006 survival; while STAT3 is an oncogene (Bromberg et al., 1999), STAT1 possesses tumor suppressor proper- Keywords: VHL; HIF-1; hypoxia; STRA13; STAT1; ties (Kaplan et al., 1998; Bromberg and Darnell, 2000). CASP1 STAT1’s major functions, executed through the IFN- g-induced Y701 phosphorylation, are inhibition of proliferation, activation of antimicrobial and antiviral responses, and induction of antitumor immune defense Introduction (Shuai et al., 1996; Kaplan et al., 1998; Stark et al., 1998).Recent data suggest that unphosphorylated Hypoxic stress or VHL deficiency stabilizes HIF-1 and STAT1 may also play an important role in the activates multiple VHL/HIF1 transcriptional targets regulation of constitutive expression of certain genes (Maxwell et al., 1999; Maxwell, 2004). The bHLH (Kumar et al., 1997; Chatterjee-Kishore et al., 2000; Meyer et al., 2002). In this study, we identify STAT1 as a transcrip- Correspondence: Dr SV Ivanov, Basic Research Program, SAIC- tional target of STRA13, investigate the molecular Frederick, Inc., B. 560, P.O. Box B, N/A, Frederick, MD 21702, USA. mechanisms of STRA13 activity on the STAT1 E-mail: [email protected] Received 8 November 2005; revised 13 April 2006; accepted 22 May promoter, and assess the downstream effects of this 2006; published online 31 July 2006 activity. Hypoxic repression of STAT1 and its targets SV Ivanov et al 803 Results 786-0 cells were stimulated with IFN-g, and endogenous STAT1 protein levels assessed in treated and nontreated Introduction of VHL in clear cell carcinoma cells cells by Western blot.Both VHL-deficient and VHL- stimulates STAT1 expression positive 786-0 cells responded to the IFN-g treatment by Our previous study revealed functional interaction a dramatic transient upregulation of both Y701- between a VHL/HIF-1 target STRA13 and STAT3 phosphorylated STAT1 species, but produced no rapid (Ivanova et al., 2004). To find out if STRA13 can be change in the total pool of STAT1 proteins (compare associated with other constituents of the JAK/STAT 15 min of IFN-g treatment with point 0 on Figure 2). pathway, we took advantage of cell lines with different While prolonged IFN-g treatment caused upregulation VHL status.The original clear cell carcinoma cell lines of both STRA13 and STAT1, stimulation was more UM-RC-6 and 786-0 contain frameshift mutations in pronounced in the STRA13 case (compare points 0 and the remaining VHL allele that make aberrant proteins 18 h).VHL-positive 786-0 cells showed an increase in incapable of targeting and degrading HIF-a, leading to the basal (unphosporylated) STAT1 proteins as com- hypoxia-like conditions.To create a tool for identifica- pared to VHL-deficient cells (compare first two lanes, tion of hypoxia-associated genes, these original cell lines top row, on Figure 2).VHL status, however, had no were transfected with wild-type (wt) or mutant VHL effect on Y701 phosphorylation, since rows 2 and 3 from transgenes (Ivanov et al., 1998). Mammalian cells top showed the same amount of protein in VHL-positive produce two differentially spliced isoforms of STAT1 and negative cells.Interestingly, while the level of Y701- mRNA that correspond to two distinct proteins, a(p91) phosphorylation in both cases significantly declined and b(p84) (Schindler et al., 1992). 786-0 cells constitu- after 18 h, total STAT1 protein amount was markedly tively transfected with the wt VHL gene showed a B2.5- upregulated suggesting a long-lasting stimulating effect fold upregulation in endogenous levels of both STAT1 of IFN-g on STAT1 transcription.Full activation of mRNA isoforms when compared with the original STAT1a by IFN-g requires phosphorylation on S727 VHL-deficient cell line or mutant VHL-transfected cells (Wen et al., 1995). 786-0 cells showed constitutive (Figure 1a).The stimulating effect of VHL on STAT1 S727 phosphorylation that was not affected by IFN-g expression was reproduced on the protein level in both stimulation or VHL status (data not shown).These 786-0 and UM-RC-6 (Figure 1b).No variation was results combined with our data shown in Figure 1 observed between 786-0 and UM-RC- 6 in the levels of implied that VHL stimulated constitutive production of STAT1 or STRA13 in spite of the fact that these cell STAT1 mRNA and unphosphorylated protein, but had lines produce different HIF-a subunits (UM-RC-6 no impact on STAT1 phosphorylation. expresses HIF-1a, while 786-0 – HIF-2a (Martin-Puig et al., 2004)). Our analysis of the VHL effect on STAT3 (Figure 1b) and other STATs’ expression (data not Inverse correlation between STAT1 and STRA13 shown) demonstrated that only STAT1 expression is expression and regulation of STAT1 transcription by modulated by pVHL. HIF-1a and STRA13 The fact that pVHL stimulates STAT1 expression prompted us to investigate the molecular mechanism VHL has no effect on STAT1 phosphorylation Phosphorylation of STAT1 is critical for the IFN-a,-b, and -g pathways (Darnell, 1998).To provide more insight into the VHL effect on STAT1, we next sought to determine if STAT1 phosphorylation is VHL-dependent.

Figure 1 Stimulation effect of VHL on STAT1 expression in renal clear cell carcinoma cell lines.( a) STAT1 endogenous mRNA expression in three 786-0 cell lines with different VHL status grown to late confluent stage.26S rRNA is shown to assure equal loading. (b) Expression of the STAT1, STAT3, and STRA13 proteins in renal Figure 2 The effect of VHL on the STAT1 protein level is clear cell carcinoma cell lines.UM-RC-6 and 786-0 renal clear cell independent of IFN-g.786-0 cells were stimulated with IFN- g carcinoma cell lines with or without constitutively transfected VHL (1000 units/ml) for the indicated time intervals.Levels of phos- were grown to a late confluent stage (4–6 days of incubation) and phorylated STAT1 proteins were assessed with Y701-STAT1 VHL effect was assessed using antibodies against endogenous proteins. specific antibodies.

Oncogene Hypoxic repression of STAT1 and its targets SV Ivanov et al 804 of this effect.We previously demonstrated that pVHL downregulates STRA13 expression on the mRNA level (Ivanova et al., 2001). In this study, using recently developed antibodies against endogenous STRA13, we observed a correlation between upregulation of STAT1 and downregulation of the STRA13 protein in VHL- positive renal clear carcinoma cells, as compared to VHL-deficient cells (Figures 1b and 2).In the course of IFN-g treatment, however, this negative correlation was not as pronounced, probably due to the fact that these gene products are tightly regulated by multiple factors. To examine if such correlation is seen on the mRNA level, four different cell lines were exposed to hypoxia and 0.5 mM NiSO4, used as a hypoxia mimetic.Hypoxic conditions produced elevated levels of STRA13 and decreased levels of STAT1 transcripts.Overall, the correlation coefficient between STRA13 and STAT1 band intensities in Figure 3a was – 0.8. Normal (1HAEo(À) and MRC-5) and cancer (U-87, U-251, LN-229) cell lines showed similar response suggesting that this effect is consistent and not cancer-specific.The observed hypoxic repression of STAT1 inversely associated with STRA13 expression prompted us to investigate whether STAT1 is a HIF-1 target regulated via STRA13.To test if STRA13 effect is HIF-1 dependent, we used MEFs with targeted HIF-1a (Salnikow et al., 2004a) and showed that HIF-1- deficient cells were not able to repress STAT1 at hypoxic Figure 3 STRA13 mediates VHL/hypoxia effect on STAT1 conditions (Figure 3b).This experiment unequivocally transcription.( a) Inverse correlation between STRA13 and STAT1 in transcriptional response to hypoxic conditions.Responsiveness proved that STAT1 is a HIF-1 target.Finally, we used of four cell lines to hypoxia and nickel was tested by RT–PCR on HaCat keratinocytes with inactivated STRA13 gene STRA13 and STAT1 transcripts.CA9 was used as a positive (Zawel et al., 2002) to further corroborate STRA13 control for hypoxia (Ivanov et al., 1998). RNA input was 50 ng per repressive activity on endogenous STAT1 (Figure 3c). RT–PCR reaction.The ribosomal protein gene RPLP0 used as a The ability of HaCat cells to downregulate STAT1 at loading control.Cycle numbers were optimized to avoid over- production and are shown for STRA13 and CA9 on the right.Cycle hypoxic conditions was significantly compromised in number for STAT1 was 32 for lanes 1–9 and 26 for lanes 10–12, for STRA13-deficient (compare lane 1 with lanes 4, 6 CA9 – 32 for lanes 1–3, 20 for lanes 4–9, and 26 for lanes 10–12. and 8), but not STRA13-proficient cells (compare lane 1 Cells were incubated with 0.5 mM NiSO4 or in hypoxia for 20 h. with lanes 3, 5 and 7).Moreover, STAT1 transcription (b) Targeted inactivation of HIF-1a in MIFs negates hypoxic was upregulated in STRA13-deficient HaCat cells in repression of STAT1.Mouse embryonic fibroblasts with or without HIF-1a were incubated for 20 h with 0.5 mM NiSO4,0.2mM desfer- lane 2 suggesting that STRA13 may as well repress rioxamine.( c) Targeted inactivation of the STRA13 gene in HaCat STAT1 at normoxic conditions.Thus, it is conceivable cells abolishes repressive effect of hypoxia on STAT1.HaCat that not only HIF-1, but other STRA13-inducing STRA13-proficient or deficient cells were incubated as above. factors, such as TGF-b or other cytokines, may have a similar repressive effect on STAT1 transcription. STAT1 transcription initiation site.Using in silico analysis, we identified in this area a potential CCAAT- Characterization of the STAT1 proximal promoter and its and ISRE-containing promoter (pos.90226–90852, STRA13-dependent activity reverse, GenBank accession # AC067945).The ISRE To test if STRA13 is capable of STAT1 suppression on sequence (IFN-stimulated response element with con- the transcriptional level, we used phSTAT1 reporter sensus AGTTTCNNTTTCNY) is a common element of constructs described earlier (Bai and Merchant, 2003) to IFN-responsive promoters that can be bound and trans- identify and characterize the STAT1 proximal promo- activated by the IFN-inducible STAT1/STAT2/IRF-9 ter.Transcriptional activity of these constructs in 293T complex, as well as other IRFs (Stark et al., 1998). In cells, with or without forced STRA13 expression, was our attempt to further delineate the STRA13-dependent evaluated using a luciferase reporter system.Out of region, we created two deletion mutants of the 1035 bp three constructs, only phSTAT1-1856 showed a sub- fragment, phSTAT1b and phSTAT1c (Figure 4a and c). stantial transcriptional activity and appeared to be As promoter activity and STRA13-dependent repression STRA13-dependent, producing a Bfourfold repression on both of them were very similar to that of the when co-transfected with STRA13 expression construct phSTAT1-1856 construct (data not shown), our further (Figure 4a and b).This observation prompted us to studies were performed on phSTAT1c (365 bp, pos. focus on the 1035 bp DNA fragment 50-adjacent to the 90287–90651 on AC067945).This DNA sequence is

Oncogene Hypoxic repression of STAT1 and its targets SV Ivanov et al 805

Figure 4 Identification of the STAT1 proximal promoter as a target of STRA13 repressive activity.( a) Construction of 50-truncated derivatives of phSTAT1-1856.Predicted proximal promoter and the transcription initiation start are shown by hatched box and arrow. E-box1, E-box 2, and ISRE elements are indicated.STRA13 effect as determined by co-transfection in a Luc-reporter assay in 293 cells is shown on the right.( b) Identification of the STAT1 proximal promoter, contained within phSTAT1-1856, as a STRA13 target.293T cells were transfected with indicated plasmids for 48 h.( c) Putative protein-binding motifs identified inside the STAT1 minimal promoter, phSTAT1c.E-boxes, c-Myb, CAAT and ISRE (boxed) motifs are shown with introduced mutations above.Arrow indicates the position where the sequence was split to produce the 50-truncated derivative phSTAT1e.The transcription initiation site is shown as defined by (Wong et al., 2002). The 30-proximal sequence conserved between mouse and human is shown in bold. located just upstream of the transcription initiation site significant loss of basal transcriptional activity and contains several protein binding motifs that may be (Figure 5a), we considered phSTAT1c the STAT1 potentially important for STRA13 repressive activity minimal promoter.Notably, phSTAT1e transcriptional (Figure 4c). activity remained to be STRA13-dependent (Figures 4a and 5a).Thus, serial truncation of the STAT1 promoter in the 50–30 direction from 1035 to 180 bp failed to STRA13 represses the STAT1 promoter independently confine STRA13 repressive activity to any specific of E-boxes, c-Myb, or ISRE motifs segment of DNA. As truncation of phSTAT1c to its 30-half (construct To test if STRA13 repressive activity can be linked to phSTAT1e, marked with arrow in Figure 4c) caused a any protein-binding motifs identified in the minimal

Oncogene Hypoxic repression of STAT1 and its targets SV Ivanov et al 806 analysis of putative STRA13-binding sites, suggested involvement of STRA13 in the interaction with basal transcriptional machinery rather than direct binding of the repressor to promoter DNA.

The C-terminal transactivation domain is critically important for the STRA13 repressive activity For mapping STRA13 domains involved in transcriptional repression of the STAT1 promoter, we used the STRA13 deletion constructs described earlier (Ivanova et al., 2004). Co-expression of construct 68–412 (lacking the basic domain), 123–412 (devoid of both the basic and HLH domains), or 1–299 (devoid of the C-terminal transactivation domain) with phSTAT1-1856 Luc- reporter showed that elimination of the C-terminus virtually abrogated STRA13 suppression activity.Dele- tion of the basic or bHLH domains reduced this activity only partially (Btwofold) (Figure 5b).These results highlight the critical importance of the STRA13 C-terminus for the transcriptional repression.As the STRA13 C-terminal region was previously associated with protein–protein interaction (Sun and Taneja, 2000; Ivanova et al., 2004, 2005), the obtained data suggested a cooperation between STRA13 and other transcriptional regulator(s) on the STAT1 promoter.

Involvement of HDAC activity in the regulation of the Figure 5 STRA13 repressive activity is associated mostly with its STAT1 minimal promoter C-terminal transactivation domain.( a) Transcriptional and As our experiments showed the critical importance of STRA13-mediated repressive activities of phSTAT1c modifica- tions.Activities of phSTAT1c (C), phSTAT1e (E) and variants the STRA13 C-terminal protein-binding domain for the with mutations introduced in the c-Myb (mMyb), E-box 1 (mE1) STAT1 promoter repression and a limited involvement and ISRE (mISRE) motifs are compared.All mutations are shown of the STRA13 DNA-binding domain, we searched for in Figure 4c.phSTAT1e represents the 3 0-half of phSTAT1c STRA13-interacting proteins that could contribute to its (see Figure 4c for details).( b) Identification of the STRA13 protein repressive activity.STRA13 was previously associated domains involved in STAT1 promoter repression.Deletion constructs of STRA13 were coexpressed with phSTAT1-1856 in with HDAC1, which interacts with the STRA13 293 cells.Effects of the C-terminal truncated (1–299), basic-less C-terminal domain (Sun and Taneja, 2000).Also, (68–412) or bHLH-less (123–412) constructs are shown.STRA13 induction of interferon-stimulated genes requires his- co-transfected with ‘‘empty’’ Luc-vector served as a negative tone deacetylase activity (Sun and Taneja, 2000; Chang control (last column). et al., 2004). These observations prompted us to examine the effect of trichostatin A (TSA), a specific promoter sequence, we used site-directed mutagenesis histone deacetylase inhibitor, on the STAT1 minimal (Figure 4c).Mutations introduced in E1 and c-Myb promoter in a co-transfection assay.The addition of produced a limited effect on basal transcription and no TSA downregulated the minimal STAT1 promoter B effect on the STRA13 repressive activity (Figure 5a). 2.5- to threefold and abrogated the STRA13 repres- Mutation of the ISRE element caused a dramatic sive effect as assessed by luciferase activity (Figure 6).A reduction (approximately fivefold) of basal phSTAT1c similar effect was obtained on the full-length promoter transcriptional activity implicating this STAT-binding phSTAT1-1856 (data not shown).These observations motif in the regulation of promoter activity.However, suggest involvement of HDAC activity in both STAT1 the mutated ISRE did not change the STRA13 basal transcription regulation and STRA13-dependent repressive capacity (Bfourfold repression remained), repression. ruling out direct interaction between STRA13 and ISRE.Finally, introduction of mutations in the remain- Recruitment of HDAC1 to the STAT1 promoter is ing E2-box (pos.194–199 in Figure 4c) also did not suppressed by hypoxic conditions result in any change in the STRA13 repressive effect To further pursue HDAC involvement in the hypoxic (data not shown). regulation of STAT1, we explored HDAC1 binding to STRA13/BHLHB2 repressive activity is known to the STAT1 promoter using chromatin immunoprecipi- utilize either binding to the E-box or interaction with tation assay.As shown in Figure 7, treatment of other transcription factors in different signaling con- 1HAEo(À) cells with 0.5 mM NiSO4 significantly texts.In our study, two independent lines of evidence, reduced binding of HDAC1 to the STAT1 promoter serial truncation of the promoter and mutational supporting HDAC1 involvement in STAT1 hypoxic

Oncogene Hypoxic repression of STAT1 and its targets SV Ivanov et al 807 repression.The observed effect could not be explained VHL target regulated via STRA13.We then tested if simply by HDAC1 expression downregulation at similar effects could be observed on a hypoxic back- hypoxic conditions, since incubation with Ni produced ground.Indeed, incubation of minimally transformed no change in HDAC1 level but dramatically upregulated human lung epithelial 1HAEo(À) cells with 0.5 mM STRA13 expression (Western blot data not shown).The NiSO4 caused a B2–4-fold downregulation of STAT1 incubation with Ni had no visible effect on the amount and virtually abrogated constitutive LMP2 expression, of acetylated H3, p300, and RNA pol II, indicating that while the STRA13 level was B2.4-fold increased the HDAC1 substrate and cooperating transcription (Figure 8b).These data further substantiate the physio- factors remain to be identified. logical importance of hypoxic signaling to STAT1 and LMP2.Since HIF-1 stimulates both STRA13 (Miyazaki Repression of LMP2 and other ISRE-containing genes by et al., 2002) and represses STAT1 (Figure 3c), we then VHL deficiency or hypoxia sought to verify HIF-1 involvement in LMP2 repression. After we demonstrated involvement of STRA13 and We took advantage of the fact that HIF-1a is unstable in HDAC in the regulation of the STAT1 promoter, our the presence of ascorbate, a cofactor of prolyl hydro- next goal was to study if other IFN-stimulated xylases (Salnikow et al., 2004a). Addition of 50 mM promoters can be repressed by hypoxic conditions. ascorbate to NiSO4-containing media partially restored One well-known IFN-g/STAT1 target is the LMP2 gene, expression of both the LMP2 and STAT1 genes which encodes a subunit of the 20S proteasome, (Figure 8c).Thus, our experiments confirm that hypoxia implicated in antigen presentation via MHC class I and VHL are both capable of LMP2 transcriptional molecules (Chatterjee-Kishore et al., 2000). We studied modulation, and that this effect is most likely mediated the effect of VHL on LMP2 expression in 786-0 cells. by STRA13 and STAT1 in a HIF1-dependent manner. Both STAT1 and LMP2 showed a VHL-dependent Seeking other IFN-associated genes that can be modulation of constitutive expression in nontreated cells repressed by hypoxia, we applied RT–PCR to total (Figure 8a) and were approximately twofold upregu- RNA samples isolated from VHL-positive or negative lated.In contrast, pVHL downregulated STRA13 B1.4- cells with or without IFN-g stimulation.As a result, we fold in agreement with the hypothesis that LMP2 is a expanded the list of VHL targets with MHC class I associated genes, such as HLA-A, -B, and -C, TAP2, and LMP7 (Figure 8d).As in the STAT1 case, promoters of HLA class I genes are controlled via the conserved ISRE motif (van den Elsen et al., 2004). Transcriptional activation from ISRE requires cooperation of STAT1, STAT2 and IRF-9 (Horvath, 2000). Remarkably, we showed that 786-0 cells expressed IRF-9 in a VHL-dependent manner providing further support for the involvement of ISRE in hypoxia- associated regulation of constitutively expressed STAT1 targets.Expression of IRFs 3 and 4 in 786-0 was not detectable, while IRF-1 was expressed but showed no association with VHL (Figure 8e and data not shown). The MHC class II genes are another set of immunity- associated genes induced by IFN-g and tightly regulated at the transcriptional level (Glimcher and Kara, 1992). VHL-deficient 786-0 cells showed a low-level constitu- Figure 6 Association of STRA13 repressive activity with HDAC. Effect of TSA on the STAT1 proximal promoter activity in 293T tive expression of HLA-DP and HLA-DMA, while cells as assessed by coexpression in a Luc-reporter assay.Where HLA-DR expression was not detectable (Figure 8e). indicated, cells were incubated with 200 ng/ml TSA for 18 h. Introduction of wt VHL stimulated basal expression of

Figure 7 Incubation with Ni prevents the recruitment of HDAC1 to the STAT1 promoter.1HAEo( À) cells were incubated with or without 0.5 mM NiSO4 for 20 h.Chromatin fragments were immunoprecipitated with antibodies against acetylated histone 3, HDAC1, p300, RNA polymerase II, or mock antibodies and the STAT1 promoter region (À152 to À402) was ampliﬁed by PCR.We were unable to detect the STAT1 promoter when mock IgG was used.

Oncogene Hypoxic repression of STAT1 and its targets SV Ivanov et al 808

Figure 8 Expression analysis of STAT1 targets via Northern (a–c) and RT-PCR (d–g), (a) Effect of VHL on LMP2 expression in 786- 0 cells.Folds of suppression or induction are indicated with negative or positive numbers, respectively.( b) Hypoxic suppression of STAT1 targets in 1HAEo(À) cells.To mimic hypoxic conditions, cells were incubated with 0.5m M NiSO4 for 20 h.The highly inducible HIF-1 target NDRG1/Cap43 was used as a control.( c) Transcriptional regulation of LMP2 is ascorbate-dependent. (d) Effects of VHL and IFN-g on expression of MHC class I – associated genes as determined by RT–PCR.( e) Effects of VHL and IFN-g on expression of MHC class II – associated genes and the transcriptional factor IRF-9.Expression of house-keeping genes b-actin and USP11 was used as a control.( f) CASP1 expression in 786-0 cells is VHL-dependent.CASP1, 2 and 3 transcript levels were assessed by RT–PCR.( g) CASP1 hypoxic repression is STRA13-dependent.Hypoxic response of HaCat cells with or without STRA13 was studied by RT–PCR.

all three genes putting them on the list of VHL targets. required more than 1 h, suggesting long-time transcrip- Lack of CIITA transcripts in uninduced 786-0 cells tional effects (Figure 8e and data not shown). ruled out involvement of this MHC II genes’ master In addition to antigen presentation genes, CASP1/ regulator in the hypoxic downregulation of constitu- ICE, CASP2/ICH1, and CASP3/CPP32 were also tively expressed MHC II genes.It is worth mentioning characterized as STAT1 constitutive targets (Kumar that upregulation of most of the studied genes by IFN-g et al., 1997; Chatterjee-Kishore et al., 2000; Meyer et al.,

Oncogene Hypoxic repression of STAT1 and its targets SV Ivanov et al 809 2002)).Our assessment of expression of these apoptosis- is not involved in the regulation of constitutively associated genes in 786-0 cells showed that CASP1, but expressed STAT1 targets.Instead, the hypoxic mecha- not CASP2 or 3, was considerably upregulated by the nism may recruit other STAT1-associated transcriptional WT VHL gene (Figure 8f).The hypothesis that this regulators, such as IRF9. effect is also mediated via STRA13 was tested on It was previously reported that renal cell carcinomas STRA13-deficient HaCat keratinocytes subjected to and other tumors downregulate expression of the MHC hypoxia (Figure 8g).This experiment showed that while I and II HLA genes (Gastl et al., 1996; Osborne et al., normal HaCat cells are capable of CASP1 hypoxic 1997; Atkins et al., 2004; Satoh et al., 2004; Zagzag repression (lanes 1 and 3), STRA13-deficient cells lost et al., 2005). As it is commonly believed that abnormali- such a capacity (lanes 2 and 4).Thus, STRA13 may ties in antigen presentation may help tumors evade the serve as a CASP1 repressor via downregulation of full-scale immune response (Seliger et al., 2000), the STAT1 transcription. molecular mechanisms of such deficiencies may provide a clue to novel therapeutic approaches.Studies on the HLA-DRA promoter in breast carcinoma cells showed Discussion that Rb protein introduction stimulated basal HLA-DR expression in a CIITA-independent manner (Osborne While tumor suppressor properties of STAT1 were et al., 1997). It is tempting to hypothesize, therefore, recognized a few years ago, they have been associated that the newly characterized link between VHL/HIF and mostly with IFN-g and phosphosylated STAT1 (Kaplan the regulation of HLA genes is a yet another example of et al., 1998; Bromberg and Darnell, 2000). However, the such a mechanism that predisposes tumors to a low-level recent finding that methylation of the STAT1 promoter constitutive immunogenicity and obstructs immune is critical for squamous carcinoma progression (Xi et al., reaction.The connection found between hypoxia and 2006) combined with the data on transcriptional activity CASP1 downregulation is also intriguing.To our of unphosphorylated STAT1 protein (Kumar et al., knowledge, CASP1 expression has never been reported 1997; Chatterjee-Kishore et al., 2000; Meyer et al., 2002) to be hypoxia-dependent.The new link identified underscore the importance of transcription in the between HIF-1/STRA13 and CASP1 may provide a regulation of STAT1 activity.Our previous studies on better insight into the regulation of apoptosis and the role of hypoxia-inducible STRA13 suggested a inflammation in hypoxic tumors. connection between STRA13 and the JAK/STAT path- In conclusion, our results can be summarized in a way.Here, for the first time, we demonstrate that scheme where STRA13 provides a previously unknown hypoxia has a negative impact on STAT1 transcription, link between hypoxia, STAT1, and targets of unpho- and that this effect is mediated by STRA13, which sphorylated STAT1 (Figure 9).It remains to be serves as a STAT1 repressor. determined, which of the vital STAT1 functions, Pursuing the molecular mechanisms of the STRA13 growth-suppressive, proapoptotic or proinflammatory, repressive effect, we isolated and characterized the can be affected through the VHL/HIF1 pathway.The STAT1 ISRE-type proximal promoter and implicated events triggered by hypoxia downstream unphosphory- histone deacetylase activity in its stimulation.These lated STAT1 require more research since ISRE-contain- findings are consistent with the recent reports on the ing promoters of IFN-inducible genes are known to be requirement of HDAC for stimulation of IFN-g regulated via transcriptional cooperation between inducible targets (Nusinzon and Horvath, 2003; Chang STATs, IRFs, CBP/p300, HDACs, NFkB, and other et al., 2004; Sakamoto et al., 2004) and with the data on factors (Shuai et al., 1996). The STRA13 role in these the functional interaction of STRA13 with HDAC1 multiple interactions and different physiological con- (Sun and Taneja, 2000).Suppression of HDAC1 texts must be defined. binding to the STAT1 promoter by NiSO4 (Figure 7), Besides hypoxia, STRA13 can be induced by a variety a well-established chemical inducer of hypoxia-respon- of other stimuli, such as TGF-b (Zawel et al., 2002), sive genes such as STRA13 (Figure 3a), suggests that cytokines (Ivanova et al., 2004), retinoic acid (Boudjelal sequestering HDAC1 by STRA13 from the STAT1 promoter is the most plausible mechanism for STAT1 hypoxic repression. While deficiency of phosphorylated STAT1 is known to cause suppression of both the adaptive and innate immune systems (Kaplan et al., 1998), the biological significance of transcriptional modulation of STAT1 remained poorly investigated.In our study, we suc- ceeded in linking VHL inactivation with the repression of STAT1-dependent antigen-presenting genes, such as MHC classes I and II genes, and CASP1.We show that events downstream from transcriptionally modulated Figure 9 Hypoxic repression of STAT1 transcription through STRA13 The negative effect of hypoxia on STAT1 expression is STAT1 are different from the ‘‘classic’’ IFN-g-respon- mediated through HIF-1 and STRA13.HDAC1 is involved in sive pathway because CIITA, the master coordinator of basal (normoxic) STAT1 transcription regulation and hypoxic IFN-g-dependent signaling (van den Elsen et al., 2004), STRA13-dependent repression of the STAT1 promoter.

Oncogene Hypoxic repression of STAT1 and its targets SV Ivanov et al 810 et al., 1997), insulin (Yamada et al., 2003) and light Washington Biotechnology (Washington, DC) and validated (Honma et al., 2002). It appears that in many of these in our previous study (Ivanova et al., 2005). STAT1 and seemingly disparate physiological contexts, STRA130s STAT3 antibodies were purchased from Santa Cruz Biotech- main function is relaying suppressive environmental nology (Santa Cruz, CA, USA).The pY701 STAT1 and pS727 signals to STAT1 and other ISRE-containing genes. STAT1 antibodies were produced by Cell Signaling Techno- logy (Beverly, MT, USA).For quantification of Northern blot Association of STRA13 activity with various stress data, we used the Packard Cyclone Storage Phosphor System factors, such as hypoxia (Wykoff et al., 2000; Ivanova (PerkinElmer, Shelton, CT, USA). et al., 2001; Miyazaki et al., 2002), light (Honma et al., 2002) and serum deprivation (Sun and Taneja, 2000; Li et al., 2002) fits well into the emerging role of the JAK/ RT–PCR analysis: optimization, loading control, and quantitative assessment STAT pathway in mediating the diverse cellular RT–PCR reactions were standardized using the GeneFisher responses to stress (Dudley et al., 2004). In our previous Primer Design web site (http://bibiserv.techfak.uni- study, we showed that STRA13 may cooperate with bielefeld.de), SuperScriptt One Step RT–PCR kit (Invitrogen, STAT3 in transcriptional regulation via the GAS, ISRE Carlsbad, CA, USA), and universal RT–PCR protocol: 501C– and STAT3 cis-elements (Ivanova et al., 2004). Our 30 min, 941C–2 min, 25–30 cycles of (941C–15 s, 561C–30 s, present data linking STRA13 with the negative regula- 701C–60 s), 701C–5 min.cDNA sequences for genes of interest tion of STAT1 and its targets shed more light on were extracted from the Source web site (http://source. STRA13’s newly discovered role in balancing the stanford.edu/cgi-bin/source/source), and primers were designed B transcriptional activities of STAT1 and STAT3. to produce 1 kb PCR products as follows: Human, STAT1: 50-gttttcatgacctcctgtcac-30,50-gtggactcctc catgttcatc-30; STRA13: 50-ccggattaacgagtgcatcg-30 50-aaagggg ggattggcttcag-30; CA9: 50-tggcttctcacattctccaag-30, cagggacaaa Materials and methods gaaggggatg-30; RPLP0: 50-tgatgcccagggaagacag-30,50-ctcctc cgactcttccttgg-30; LMP2, LMP-7: 50-tctctttggctcaggctagg-30, Cell lines, VHL and STRA13 expression constructs, and STAT1 50-cacgctcgccttcaagttc-30; TAP-2, AKTB: 50-gtcgtcgacaacggct reporter assay cc-30,50-tggccgtcaggcagctc-30; CAP43, HLA-A: 50-tcatcaacctct Renal clear cell carcinoma cell lines 786-0 and UM-RC-6 were catggcaag-30,50-aaaggattacatcgccctgaac-30; HLA-B: 50-ggggcg incubated to late confluent stage to ensure maximum effect of tgaagaaatcc-30,50-tcatctcagtgggctacg-30; HLA-C: 50-tgcagaaa VHL on its transcriptional targets.Mutant (mut 2) or gagatgccagag-30,50-gttcgtgcggttcgacag-30; IRF-1: 50-actcggatgc transgenic wild-type VHL (wtVHL) constructs used in these gcatgagac-30,50-atgccaggtcctgcttgc-30; IRF-9: 50-tgttgctgagccc study were described earlier (Ivanov et al., 1998). Human 293T tacaagg-30;50-gtgtgtgcgaggattttcag-30; CII2A: 50-gcatccgaagg embryonic kidney cell line was purchased from ATCC; catccttg-30; USP11: 50-cttcgaccccttctgctac-30,50-ggtccagcttct LN-229 (glioma) and U-251 (astrocytoma) were kindly tggttgc-30; CASP1: 50-gacaagggtgctgaacaagg-30,50-gcggaaaatt provided by Dr D Zagzag, and MRC-5 cells (normal tcctccacatc-30; CASP2: 50-gtgttgagcaatgtgcacttc-30,50-tcacttgt untransformed human fetal lung fibroblasts) were purchased caggctcttcaac-30; CASP3: 50-cagacagtggtgttgatgatg-30,50-tagatc from Coriell Cell Repositories, Camden, NJ, USA.Human caggggcattgtagc-30; lung minimally transformed epithelial 1HAEo(À) cells were Mouse, mSTRA13: 50-gcatgccgtagaagtgagag-30,50-actgcc described elsewhere (Salnikow et al., 2004b). MEF þ / þ tacctgcctatgc-30; mSTAT1: 50-aaaagctggaggagttggaac-30,50-tc and MEF HIF-1À/À were obtained from Dr R Johnson cacccatgtgaatgtgatg-30.Reliability and accuracy of RT–PCR- (University of California San Diego, CA, USA) and were based comparison was tested by measurements in triplicate described previously (Salnikow et al., 2000). HaCat cells with experiments.As a loading control, we used the gene encod- targeted STRA13 gene, described previously (Zawel et al., ing ribosomal protein RPLP0 recommended by (http:// 2002), were kindly provided by Dr B Vogelstein (Howard pathmicro.med.sc.edu/pcr/realtime-home.htm), AKTB, or Hughes Medical Institute and The Sidney Kimmel Compre- USP11.After separation of RT–PCR products in a conven- hensive Cancer Center, The Johns Hopkins Medical Institu- tional agarose gel, intensities of ethidium bromide stained tions).STRA13 deletion constructs were produced and verified DNA bands were measured using Kodak 1D 3.6.3 Network earlier (Ivanova et al., 2004). STAT1 reporter constructs Software.The correlation coefficient was calculated using the phSTAT1-1856, 1-888 and 1-300 were described in Bai and UCLA Department of Statistics website (http://calculators. Merchant (2003).Additional constructs phSTAT1b, stat.ucla.edu/correlation.php). phSTAT1c and phSTAT1e were generated by PCR on the phSTAT1-1856 template and then cloned into the KpnI/ Luc-reporter assay HindIII site of the pGL3-Basic vector. Thirty thousand 293T cells per well were plated into 96-well plates in 100 ml media.On the next day, cells were transfected Northern and Western blot analyses with 75 ng/well of the STAT1 promoter constructs and 75 ng/ Cell cultures were grown in DMEM or EMEM media.To well of the STRA13 expression plasmid, or empty vector and produce hypoxic conditions cells were incubated in a chamber 50 ng/well of pCMV-Renilla luciferase (RL-CMV), using the in the atmosphere containing 0.5% O2,5%CO2, and 94.5% transfection reagent effecten (Qiagen, Valencia, CA, USA). N2 at 371C for 20 h.To mimic hypoxic conditions, cells were After overnight incubation, plasmid expression was allowed incubated with 0.5 mM NiSO4 or 0.2 mM desferoxamine for 24–36 h.The reporter expression was detected using the for 20 h.RNA samples were extracted with RNeasy Mini Dual-Glo luciferase system (Promega, Madison, WI, USA) total RNA (Qiagen, Valencia, CA, USA) or FastTrack polyA and the MicroLumat LB 96V Luminometer (Bad Wildbad, RNA (Invitrogen, Carlsbad, CA, USA) kits.For Western blot, Germany).The data are presented as mean values 7s.d. of cells were lysed in RIPA buffer and proteins fractionated normalized relative luciferase units (RLU), which represent the as described earlier (Ivanova et al., 2004). STRA13 anti- ratio of luminescence produced by the STAT1 promoter-Luc bodies against the endogenous protein were ordered from plasmid to the luminescence of the RL-CMV reporter in the

Oncogene Hypoxic repression of STAT1 and its targets SV Ivanov et al 811 same well.Each condition has been tested in quadruplicate in (pos.90721–90704 on AC067945) and 5 0-GCGCTGGGG at least three independent experiments. TATTTCC-30 (pos.90449–90471 on AC067945).

Stimulation of 786-0 cells with INF-g Cells grown in DMEM media supplemented with bovine fetal Abbreviations serum were stimulated with INF-g (Upstate Biotechnology, Waltham, MA, USA) at the final cytokine concentration STAT1, signal transducer and activator 1; VHL and pVHL, 1000 m/ml. von Hippel Lindau gene and protein; HIF-1, hypoxia- inducible factor 1; GAS, IFN-g activating sequence; ISRE, STAT1 promoter analysis IFN-stimulated responsive element; TSA, trichostatin A; For promoter prediction and analysis, we used Genomatix MEF, mouse embryonic fibroblasts, HRE, hypoxia-responsive (http://www.genomatix.de/) and Motif (http://motif.gen- element; HDAC, histone deacetylase; ChIP, chromatin im- ome.jp) servers. Introduction of mutations into the putative munoprecipitation; DFX, desferioxamine. protein-binding motifs, identified in the promoter sequence, was made with Stratagene (La Jolla, CA, USA) Quick Change Acknowledgements Site-Directed Mutagenesis kit according to the manufacturer’s protocol. This project was funded in whole or in part with Federal funds by the Intramural Research Program of the NIH, NCI, and Chromatin immunoprecipitation assay Contract No.NO1-CO-12400.The content of the publication For immunoprecipitation studies on the STAT1 promoter, we does not necessarily reflect the views or policies of the used Active Motif ChIP-IT kit (Carlsbad, CA, USA) and Department of Health and Human Services, nor does mention RNA polymerase II antibodies according to the manufac- of trade names, commercial products, or organizations imply turer’s instructions.Antibodies against acetyl-H3, HDAC1 endorsement by the US Government. and p300 were purchased from Upstate Cell Signaling Authors thank Dr Bert Vogelstein for providing STRA13- Solutions (Lake Placid, NY, USA).PCR primers for the deficient HaCat cells, Robert Nalewaik for technical help with STAT1 promoter were 50-CCCTGGGTTTAGCAACAC-30 experimentation, and Laura Geil for editing the manuscript.

References

Atkins D, Ferrone S, Schmahl GE, Storkel S, Seliger B.(2004). Dhar M, Taneja R.(2001).Cross-regulatory interaction Down-regulation of HLA class I antigen processing mole- between Stra13 and USF results in functional antagonism. cules: an immune escape mechanism of renal cell carcinoma? Oncogene 20: 4750–4756. J Urol 171: 885–889. Dudley AC, Thomas D, Best J, Jenkins A.(2004).The STATs Azmi S, Sun H, Ozog A, Taneja R.(2003).mSharp-1/DEC2, a in cell stress-type responses. Cell Commun Signal 2:8. basic helix-loop-helix protein functions as a transcriptional Gastl G, Ebert T, Finstad CL, Sheinfeld J, Gomahr A, repressor of E box activity and Stra13 expression. J Biol Aulitzky W et al. (1996).Major histocompatibility complex Chem 278: 20098–20109. class I and class II expression in renal cell carcinoma and Bach EA, Aguet M, Schreiber RD.(1997).The IFN gamma modulation by interferon gamma. J Urol 155: 361–367. receptor: a paradigm for cytokine receptor signaling. Annu Glimcher LH, Kara CJ.(1992).Sequences and factors: a guide Rev Immunol 15: 563–591. to MHC class-II transcription. Annu Rev Immunol 10: 13–49. Bai L, Merchant JL.(2003).Transcription factor ZBP-89 is Honma S, Kawamoto T, Takagi Y, Fujimoto K, Sato F, required for STAT1 constitutive expression. Nucleic Acids Noshiro M et al. (2002).Dec1 and Dec2 are regulators of the Res 31: 7264–7270. mammalian molecular clock. Nature 419: 841–844. Boudjelal M, Taneja R, Matsubara S, Bouillet P, Dolle P, Horvath CM.(2000).STAT proteins and transcriptional Chambon P.(1997).Overexpression of Stra13, a novel responses to extracellular signals. Trends Biochem Sci 25: retinoic acid-inducible gene of the basic helix-loop-helix 496–502. family, inhibits mesodermal and promotes neuronal differen- Ivanov SV, Kuzmin I, Wei MH, Pack S, Geil L, Johnson BE tiation of P19 cells. Genes Dev 11: 2052–2065. et al. (1998).Down-regulation of transmembrane carbonic Bromberg J, Darnell Jr JE.(2000).The role of STATs in anhydrases in renal cell carcinoma cell lines by wild-type von transcriptional control and their impact on cellular function. Hippel-Lindau transgenes. Proc Natl Acad Sci USA 95: Oncogene 19: 2468–2473. 12596–12601. Bromberg JF, Wrzeszczynska MH, Devgan G, Zhao Y, Pestell Ivanova AV, Ivanov SV, Danilkovitch-Miagkova A, Lerman RG, Albanese C et al. (1999).Stat3 as an oncogene. Cell 98: MI.(2001).Regulation of STRA13 by the von Hippel- 295–303. Lindau tumor suppressor protein, hypoxia, and the UBC9/ Chang HM, Paulson M, Holko M, Rice CM, Williams BR, ubiquitin proteasome degradation pathway. J Biol Chem Marie I et al. (2004).Induction of interferon-stimulated gene 276: 15306–15315. expression and antiviral responses require protein deacety- Ivanova AV, Ivanov SV, Lerman ML.(2005).Association, lase activity. Proc Natl Acad Sci USA 101: 9578–9583. mutual stabilization, and transcriptional activity of the Chatterjee-Kishore M, Wright KL, Ting JP, Stark GR.(2000). STRA13 and MSP58 proteins. Cell Mol Life Sci 62: 471–484. How Stat1 mediates constitutive gene expression: a complex Ivanova AV, Ivanov SV, Zhang X, Ivanov VN, Timofeeva of unphosphorylated Stat1 and IRF1 supports transcription OA, Lerman MI.(2004).STRA13 interacts with STAT3 and of the LMP2 gene. EMBO J 19: 4111–4122. modulates transcription of STAT3-dependent targets. J Mol Darnell Jr JE.(1998).Studies of IFN-induced transcriptional Biol 340: 641–653. activation uncover the Jak-Stat pathway. J Interferon Kaplan DH, Shankaran V, Dighe AS, Stockert E, Aguet M, Cytokine Res 18: 549–554. Old LJ et al. (1998).Demonstration of an interferon

Oncogene Hypoxic repression of STAT1 and its targets SV Ivanov et al 812 gamma-dependent tumor surveillance system in immuno- and function in Clast5/Stra13/DEC1-transgenic mice. Eur J competent mice. Proc Natl Acad Sci USA 95: 7556–7561. Immunol 34: 1322–1332. Kumar A, Commane M, Flickinger TW, Horvath CM, Stark Seliger B, Maeurer MJ, Ferrone S.(2000).Antigen-processing GR.(1997).Defective TNF-alpha-induced apoptosis in machinery breakdown and tumor growth. Immunol Today STAT1-null cells due to low constitutive levels of caspases. 21: 455–464. Science 278: 1630–1632. Shen M, Kawamoto T, Teramoto M, Makihira S, Fujimoto K, Li Y, Zhang H, Xie M, Hu M, Ge S, Yang D et al. (2002). Yan W et al. (2001).Induction of basic helix-loop-helix Abundant expression of Dec1/stra13/sharp2 in colon carci- protein DEC1 (BHLHB2)/Stra13/Sharp2 in response to the noma: its antagonizing role in serum deprivation-induced cyclic adenosine monophosphate pathway. Eur J Cell Biol apoptosis and selective inhibition of procaspase activation. 80: 329–334. Biochem J 367: 413–422. Shen M, Kawamoto T, Yan W, Nakamasu K, Tamagami M, Martin-Puig S, Temes E, Olmos G, Jones DR, Aragones J, Koyano Y et al. (1997).Molecular characterization of the Landazuri MO.(2004).Role of iron (II)-2-oxoglutarate- novel basic helix-loop-helix protein DEC1 expressed in dependent dioxygenases in the generation of hypoxia- differentiated human embryo chondrocytes. Biochem induced phosphatidic acid through HIF-1/2 and von Biophys Res Commun 236: 294–298. Hippel-Lindau-independent mechanisms. J Biol Chem 279: Shuai K, Liao J, Song MM.(1996).Enhancement of 9504–9511. antiproliferative activity of gamma interferon by the specific Maxwell PH.(2004).HIF-1’s relationship to oxygen: simple inhibition of tyrosine dephosphorylation of Stat1. Mol Cell yet sophisticated. Cell Cycle 3: 156–159. Biol 16: 4932–4941. Maxwell PH, Wiesener MS, Chang GW, Clifford SC, Vaux Shuai K, Liu B.(2003).Regulation of JAK-STAT signalling in EC, Cockman ME et al. (1999).The tumour suppressor the immune system. Nat Rev Immunol 3: 900–911. protein VHL targets hypoxia-inducible factors for oxygen- St-Pierre B, Flock G, Zacksenhaus E, Egan SE.(2002).Stra13 dependent proteolysis. Nature 399: 271–275. homodimers repress transcription through class B E-box Meyer T, Begitt A, Lodige I, van Rossum M, Vinkemeier U. elements. J Biol Chem 277: 46544–46551. (2002).Constitutive and IFN-gamma-induced nuclear im- Stark GR, Kerr IM, Williams BR, Silverman RH, Schreiber port of STAT1 proceed through independent pathways. RD.(1998).How cells respond to interferons. Annu Rev EMBO J 21: 344–354. Biochem 67: 227–264. Miyazaki K, Kawamoto T, Tanimoto K, Nishiyama M, Sun H, Lu B, Li RQ, Flavell RA, Taneja R.(2001).Defective Honda H, Kato Y.(2002).Identification of functional T cell activation and autoimmune disorder in Stra13- hypoxia response elements in the promoter region of the deficient mice. Nat Immunol 2: 1040–1047. DEC1 and DEC2 genes. J Biol Chem 277: 47014–47021. Sun H, Taneja R.(2000).Stra13 expression is associated with Nusinzon I, Horvath CM.(2003).Interferon-stimulated growth arrest and represses transcription through histone transcription and innate antiviral immunity require deace- deacetylase (HDAC)-dependent and HDAC-independent tylase activity and histone deacetylase 1. Proc Natl Acad Sci mechanisms. Proc Natl Acad Sci USA 97: 4058–4063. USA 100: 14742–14747. van den Elsen PJ, Holling TM, Kuipers HF, van der Stoep N. Osborne A, Tschickardt M, Blanck G.(1997).Retinoblastoma (2004).Transcriptional regulation of antigen presentation. protein expression facilitates chromatin remodeling at the Curr Opin Immunol 16: 67–75. HLA-DRA promoter. Nucleic Acids Res 25: 5095–5102. Wen Z, Zhong Z, Darnell Jr JE.(1995).Maximal activation of Qing Y, Stark GR.(2004).Alternative activation of STAT1 transcription by Stat1 and Stat3 requires both tyrosine and and STAT3 in response to interferon-gamma. J Biol Chem serine phosphorylation. Cell 82: 241–250. 279: 41679–41685. Wong LH, Sim H, Chatterjee-Kishore M, Hatzinisiriou I, Sakamoto S, Potla R, Larner AC.(2004).Histone deacetylase Devenish RJ, Stark G et al. (2002).Isolation and charac- activity is required to recruit RNA polymerase II to the terization of a human STAT1 gene regulatory element. promoters of selected interferon-stimulated early response Inducibility by interferon (IFN) types I and II and role of genes. J Biol Chem 279: 40362–40367. IFN regulatory factor-1. J Biol Chem 277: 19408–19417. Salnikow K, Blagosklonny MV, Ryan H, Johnson R, Costa Wykoff CC, Pugh CW, Maxwell PH, Harris AL, Ratcliffe PJ. M.(2000).Carcinogenic nickel induces genes involved with (2000).Identification of novel hypoxia dependent and hypoxic stress. Cancer Res 60: 38–41. independent target genes of the von Hippel-Lindau (VHL) Salnikow K, Donald SP, Bruick RK, Zhitkovich A, Phang JM, tumour suppressor by mRNA differential expression profil- Kasprzak KS.(2004a).Depletion of intracellular ascorbate by ing. Oncogene 19: 6297–6305. the carcinogenic metals nickel and cobalt results in the Xi S, Dyer KF, Kimak M, Zhang Q, Gooding WE, Chaillet JR induction of hypoxic stress. J Biol Chem 279: 40337–40344. et al. (2006).Decreased STAT1 expression by promoter Salnikow K, Li X, Lippmann M.(2004b).Effect of nickel and methylation in squamous cell carcinogenesis. J Natl Cancer iron co-exposure on human lung cells. Toxicol Appl Inst 98: 181–189. Pharmacol 196: 258–265. Yamada K, Kawata H, Shou Z, Mizutani T, Noguchi T, Satoh A, Toyota M, Ikeda H, Morimoto Y, Akino K, Mita H Miyamoto K.(2003).Insulin induces the expression of the et al. (2004).Epigenetic inactivation of class II transactivator SHARP-2/Stra13/DEC1 gene via a phosphoinositide (CIITA) is associated with the absence of interferon-gamma- 3-kinase pathway. J Biol Chem 278: 30719–30724. induced HLA-DR expression in colorectal and gastric Zagzag D, Salnikow K, Chiriboga L, Yee H, Lan L, Ali MA cancer cells. Oncogene 23: 8876–8886. et al. (2005).Downregulation of major histocompatibility Schindler C, Fu XY, Improta T, Aebersold R, Darnell Jr JE. complex antigens in invading glioma cells: stealth invasion of (1992).Proteins of transcription factor ISGF-3: one gene the brain. Lab Invest 85: 328–341. encodes the 91-and 84-kDa ISGF-3 proteins that are activated Zawel L, Yu J, Torrance CJ, Markowitz S, Kinzler KW, by interferon alpha. Proc Natl Acad Sci USA 89: 7836–7839. Vogelstein B et al. (2002).DEC1 is a downstream target of Seimiya M, Wada A, Kawamura K, Sakamoto A, Ohkubo Y, TGF-beta with sequence-specific transcriptional repressor Okada S et al. (2004).Impaired lymphocyte development activities. Proc Natl Acad Sci USA 99: 2848–2853.

Oncogene