Nonsteroidal Anti-inflammatory Drug-activated Gene (NAG-1/GDF15) Expression Is Increased by the Histone Deacetylase Inhibitor Trichostatin A*

Nonsteroidal anti-inflammatory drug-activated gene (NAG-1) is a putative tumor suppressor whose expression can be increased by drug treatment. Glioblastoma is the most common central nervous system tumor, is associated with high morbidity and mortality, and responds poorly to surgical, chemical, and radiation therapy. The histone deacetylase inhibitors are under current consideration as therapeutic agents in treating glioblastoma. We investigated whether trichostatin A (TSA) would alter the expression of NAG-1 in glioblastoma cells. The DNA demethylating agent 5-aza-dC did not increase NAG-1 expression, but TSA up-regulated NAG-1 expression and acted synergistically with 5-aza-dC to induce NAG-1 expression. TSA indirectly increases NAG-1 promoter activity and increases NAG-1 mRNA and protein expression in the T98G human glioblastoma cell line. TSA also increases the expression of transcription factors Sp-1 and Egr-1. Small interfering RNA experiments link NAG-1 expression to apoptosis induced by TSA. Reporter gene assays, specific inhibition by small interfering RNA transfections, and chromatin immunoprecipitation assays indicate that Egr-1 and Sp-1 mediate TSA-induced NAG-1 expression. TSA also increases the stability of NAG-1 mRNA. TSA-induced NAG-1 expression involves multiple mechanisms at the transcriptional and post-transcriptional levels.

Glioblastoma multiform is the most malignant human primary brain tumor in adults and represents the most malignant stage of astrocytoma progression, has a median survival time of 12 months and a 5-year survival rate of Ͻ3%. The survival rate has not significantly changed despite advances in anti-tumor therapy (1,2). Malignant gliomas aggressively infiltrate into normal adjacent brain tissue, are incurable by surgery alone, and are relatively resistant to radiotherapy and most forms of chemotherapy.
Nonsteroidal anti-inflammatory drug-activated gene (NAG-1) 2 is a divergent member of the transforming growth factor-␤ superfamily that was identified in this laboratory by a PCRbased subtractive hybridization from an indomethacin-induced library obtained from human colorectal cells. The cDNA has been cloned by six different groups (as known as MIC-1, PDF, GDF-15, PLAB, and PTGFB). The previous investigations on the regulation of NAG-1 expression revealed complex mechanisms that can be modulated by a number of drugs and chemicals: cyclooxygenase inhibitors (3), dietary agents (4 -6), PPAR agonist (7)(8)(9), and anti-cancer drugs (10). Moreover, NAG-1 is induced by several nonsteroidal anti-inflammatory drugs and other drugs known to have anti-tumorigenic and pro-apoptotic activities (11,12). In the brain, NAG-1 is expressed in epithelial cells of the choroid plexus and secreted into the cerebral spinal fluid and is reported to function as both a neurotrophic and a neuroprotective factor for midbrain dopaminergic neurons in vivo and in vitro (13,14).
Recently, we generated NAG-1 transgenic mice (Cre/NAG-1 Tg/Lox ), which express human NAG-1 in all tissues (15). The anti-tumorigenic activity of the Cre/NAG-1 Tg/Lox mice was evaluated with a known colon carcinogen, azoxymethane. Cre/ NAG-1 Tg/Lox mice had smaller numbers of pre-neoplastic polyps in their colon compared with littermate controls. Furthermore, Apc Min/ϩ mice mated to the Cre/NAG-1 Tg/Lox mice showed a 42% reduction in small intestine polyp numbers and a 40% reduction in tumor load (15). These findings suggest that NAG-1 plays a role in the development of intestinal polyps in mice and colorectal cancer in humans. These data and other findings reported in the literature suggest that NAG-1 may act as a tumor suppressor, but the association between NAG-1 and brain tumor development has been poorly investigated.
The promoter of NAG-1 has several CpG islands, but no investigation has been done to epigenetically silence the protein by methylation or histone modification. Histone acetylation plays a key role in transcriptional activation, whereas deacetylation of histones correlates with transcriptional repression and silencing of genes involved in cell-cycle regulation, differentiation, and development, as well as human cancer (16,17). Various histone deacetylase (HDAC) inhibitors, such as trichostatin A, sodium butyrate (18), suberoylanilide hydroxamic acid (19), trapoxin (20), depudecin (21), FR901228 (22), oxanflatine (23), and MS27275 (24) promote accumulation of acetylated histones in the nucleus, arrest cell growth, and reverse neoplastic characteristics in cultured cells via the expression of a specific preprogrammed set of genes. During cancer therapy, TSA and 5-aza-dC, the demethylation drug, act synergistically for reactivation of silenced genes (25)(26)(27).
HDAC inhibitors are being investigated as potential therapeutic agents in the treatment of glioblastoma. We suspected that HDAC inhibitors may alter NAG-1 expression because the transcription factor Sp-1 has been shown to interact with HDAC1 (28) and NAG-1 expression is regulated by p53, Egr-1, and Sp-1 (29,30). However, whether histone acetylation modification is involved in NAG-1 transcriptional regulation has not been elucidated. The goal of the present study was to determine whether NAG-1 expression in glioblastoma cells is altered by HDAC inhibitors.
Western Blot Analysis-Total cell lysates were isolated in RIPA buffer with Complete Mini protease inhibitor mixture tablets from Roche (Indianapolis, IN), sonicated briefly, and quantitated by BCA assay (Pierce). Forty micrograms of total protein per lane were separated by SDS-PAGE 4 -12% BisTris gel (Invitrogen) and transferred onto a nitrocellulose membrane (Invitrogen). The blots were blocked for 1 h in 5% skim milk in Tris-buffered saline containing 0.05% Tween 20 (TBS-T) (Sigma) and probed overnight at 4°C in 5% skim milk in TBS-T with each primary antibody. After washing with TBS-T, the blots were incubated with the appropriate horseradish peroxidase-conjugated secondary antibody for 1 h at room temperature in 3% skim milk in TBS-T, washed several times in TBS-T, and detected by the Amersham Biosciences ECL plus Western blot detection system (GE Healthcare).
Isolation and Reverse Transcription-Total RNAs were isolated with the RNeasy MINI kit (Qiagen, Valencia, CA) according to the manufacturer's protocol. 1 g of RNA was treated with 1 unit of DNase I Amplification Grade (Invitrogen) at room temperature for 15 min to remove genomic DNA and followed by inactivation by DNase I with 2.5 mM EDTA (Invitrogen) at 65°C for 5 min. Reverse transcription (RT) was performed with SuperScript II Reverse Transcriptase (Invitrogen) according to the manufacturer's protocol. The cDNA was stored at Ϫ20°C until use.
Constructions of Plasmids-Several NAG-1 promoter clones linked to luciferase, the Egr-1 expression vector, and the Sp-1 expression vector were described previously (11).
Luciferase Reporter Assays-T98G cells were seeded in 12-well plates at 15 ϫ 10 4 cells/well in Dulbecco's modified Eagle's medium and grown to 50 -60% confluence. The plasmid mixtures, containing 1 g of NAG-1 promoter luciferase construct and 0.025 g of pRL-null (Promega), were transfected using FuGENE 6 Transfection Reagent (Roche) according to the manufacturer's protocol. The co-transfection experiment was carried out using plasmid mixtures containing 0.5 g of NAG-1 Ϫ133/ϩ41 promoter luciferase construct, 0.5 g of expression plasmid (Egr-1 or Sp-1), and 0.025 g of pRL-null. The pcDNA3.1 empty vector (Invitrogen) was used as a negative control for the expression plasmid. After 24 h transfection, the cells were treated with 500 ng/ml TSA or Control (0.1% Me 2 SO) for 24 h, then harvested in 1ϫ luciferase lysis buffer (Promega), and luciferase activity was measured and normalized with the values of pRL-null luciferase activity using a dual luciferase assay kit (Promega).
Short Interfering RNA (siRNA) Transfection-The NAG-1 siRNA (M-019875-01), the Sp-1 siRNA (M-026959-00), Egr-1 siRNA (M-006526-01), and control siRNA (D-001206-13) were purchased from Dharmacon (Lafayette, CO). T98G cells were grown to 50% confluence in antibiotic-free Dulbecco's modified Eagle's medium and transfected with each siRNA at 50 nM using Lipofectamine TM 2000 reagent (Invitrogen) according to the manufacturer's instructions. After incubating for 4 h, the cells were washed and changed to the complete media and recovered overnight. The cells were subsequently treated by 500 ng/ml TSA or Control (0.1% Me 2 SO) for 24 h, and the effect of NAG-1 expres-sion by Sp-1, or Egr-1 knock-down was investigated with Western blot analysis and real-time PCR. The effect of NAG-1, Egr-1, and Sp-1 siRNA were optimized with Western blot analysis.
Flow Cytometry-T98G cells that transfected with control or NAG-1 siRNA were seeded in 6-well plates at 8 ϫ 10 4 cells/well in Dulbecco's modified Eagle's medium and subsequently treated by 500 ng/ml TSA or control (0.1% Me 2 SO) for 72 h. Cells were trypsinized and harvested by centrifugation. Pellets were washed with phosphate-buffered saline, resuspended in phosphate-buffered saline containing 10 mg/ml propidium iodide (Roche) and 0.5 mg/ml RNase (Roche), and stored overnight at 4°C. Stained cells were analyzed using a BD Biosciences FACSort (Franklin Lakes, NJ) as per the manufacturer's instructions. The fraction of sub-G 1 cells in the population was determined using CellQuest 3.1 software (BD Biosciences Flow Cytometry System).
Statistical Analysis-All statistical differences between experimental groups were evaluated by the two-tailed unpaired Student's t test.

DNA Methylation, Combination Treatment of TSA, and 5-Aza-dC-
The NAG-1 promoter contains CpG islands and thus the expression of NAG-1 could be regulated by methylation. We examined the effect of the demethylation agent 5-aza-dC and the HDAC inhibitor TSA on the expression in glioblastoma cells, T98G. Cells were treated with vehicle or the chemicals for 48 h and protein and RNA were isolated. As shown in Fig. 1, treatment with 5-aza-dC did not increase NAG-1 protein or RNA expression, but TSA up-regulated both protein and RNA expression and acted synergistically with 5-aza-dC to induce NAG-1 expression. These findings suggest a role for histone deacetylation in the regulation of NAG-1 expression in brain cancer cells.
HDAC Inhibitor, TSA Effects on NAG-1 Expression in T98G Cells-The induction of NAG-1 expression by TSA was both dose-and time-dependent (Fig. 2). The T98G cells were treated with several concentrations of TSA ranging from 10 to 600 ng/ml for 12 h and then the cells were harvested for protein or RNA analysis. An increase in NAG-1 expression at both the protein and mRNA levels was observed at concentrations as low as 20 ng/ml (Fig. 2, A and B). Cells were then treated with 500 ng/ml TSA for 0 to 48 h. The highest expression of NAG-1 mRNA was observed at 12 h and highest NAG-1 protein expression was observed at 18 h (Fig. 2, C and D). To further confirm the dependence on histone deacetylase inhibition, the effect of another HDAC inhibitor, sodium butyrate, was used. The induction of NAG-1 by sodium butyrate was both timeand dose-dependent (data not shown). Both HDAC inhibitors TSA and sodium butyrate induced NAG-1 expression in T98G cells and lend support to the hypothesis that the induction of NAG-1 is dependent on the inhibition of histone deacetylation.
TSA Effect on NAG-1 Expression in Other Glioblastoma Cells-The basal level expression and induction by TSA of NAG-1 in several gliomas and normal human astrocyte cells were measured by mRNA and protein analysis. The basal expression of NAG-1 at the mRNA and the protein levels were very low in all cells except in A172 and the low grade cells, Hs683, SW1088, and normal human astrocyte NHA (Fig. 3, A and B). An increase in protein expression of NAG-1 was most notable in T98G and SW1088 cells after TSA treatment. Treatment with TSA increased the NAG-1 mRNA expression (Fig. 3C) in T98G, U87, and SW1088 cells. For all subsequent experiments we used T98G cells because of the large -fold increase in NAG-1 expression after TSA treatment.  NOVEMBER 28, 2008 • VOLUME 283 • NUMBER 48 NAG-1 Induced by TSA Is Associated with Apoptosis-TSA is known to induce apoptosis and increased expression of NAG-1 has been linked to proapoptotic activity. Apoptosis induced by TSA was measured with cleaved PARP, an indicator of apoptosis, and flow cytometry. Cells were treated with TSA, and PARP cleavage and NAG-1 were measured. The TSA-dependent induction of PARP cleavage in T98G cells was both dose-and time-dependent (Fig. 4, A and B). TSA also induced PARP cleavage in U87, A172, and Hs683 cells but not in U118, U138 NHA, and SW1008 cells (data not shown). To investigate whether there is a link between the expression of NAG-1 and PARP cleavage, we used siRNA to inhibit TSA induced NAG-1 expression (Fig. 4C). T98G cells were transfected with control and NAG-1 siRNAs and then treated with TSA. Suppression of NAG-1 expression by siRNA and attenuation of cleaved PARP induced by TSA was observed.

TSA Induces NAG-1 Expression
Next we measured apoptosis induced by TSA with flow cytometry. The increase in the sub-G 1 population by TSA was dose-and time-dependent (data not shown). To investigate whether there is a link between the expression of NAG-1 and the increase of sub-G 1 population, T98G cells were transfected with control and NAG-1 siRNAs, and then treated with TSA. After a 72-h treatment, cells were harvested and analyzed by flow cytometry. Suppression of NAG-1 expression by siRNA significantly decreased the TSA-increased sub-G 1 population. These results support the linkage of TSA-induced apoptosis to increased expression of NAG-1, suggesting that the expression of NAG-1 may mediate, in part, TSA-induced apoptosis.
TSA Indirectly Induces NAG-1 Expression in T98G Cells-To investigate if TSA directly activate NAG-1 expression, or indirectly act by increasing the expression of specific transcriptional factors, cells were pretreated with cycloheximide (CHX) to inhibit de novo protein synthesis, and then incubated with TSA. The CHX pretreatment inhibited both NAG-1 protein and mRNA expressions (Fig. 5). These results indicate that TSA does not directly induce NAG-1 expression and requires new protein synthesis.
Expression of Other Proteins After TSA Treatment-Because the induction of NAG-1 expression by TSA is dependent on de novo protein synthesis, expression of other proteins induced by TSA was investigated. Previously, our laboratory demonstrated that HDAC inhibitors induce COX-1 expression in normal human astrocyte cells and others have reported a suppression of COX-2 induction (31, 32). Thus, we investigated the effect of  TSA on the expression of COX-1 and COX-2 in T98G cells (Fig. 6). TSA did not alter COX-1 and COX-2 expression in T98G cells. Because TSA increases the expression of other genes via Egr-1/Sp-1 transcription factors, the effect of TSA on the expression of Egr-1 and Sp1 was examined. TSA-induced Egr-1 and NAG-1 expression at the mRNA and protein levels in a dosedependent manner. Furthermore, both mRNA and protein expression of Egr-1 was induced by TSA at earlier times than observed, respectively, for NAG-1 protein and mRNA expression (Fig. 7). The peak protein expression for Egr-1 and NAG-1 occurred at 12 and 18 h, respectively, whereas peak mRNA expression for Egr-1 and NAG-1 occurred at 9 and 12 h. In contrast, Sp-1 protein expression was induced at 3 h and then decreased. Increases in histone H3 and H4 acetylation were observed within 3 h after treatment and is observed at times earlier than the increase in Egr-1 or NAG-1 expression. We proposed based on these findings that the induction of NAG-1 is associated with Sp-1 and Egr-1 expression.
Analysis of the Promoter Activity of Promoter Regions of the NAG-1 Gene-Luciferase NAG-1 promoter constructs comprised of different sequences of the promoter were transfected into T98G cells and then treated with TSA (Fig. 8, A and  B). As an internal control, the plasmid pRL-null was used to determine the transfection efficiency. A large increase in luciferase activity was observed after TSA treatment for all NAG-1 promoter constructs, including the construct with the first 133 bp, a region containing the Sp-1/ Egr-1 binding site. Thus, the TSA response element in the NAG-1 promoter is located in the first 133-base region and possibly implicates the Sp-1/Egr-1 binding site.

TSA Response Element Is Located in Sp-1 and Egr-1 Binding Region of NAG-1 Promoter-The
TSA response element is located between positions Ϫ133 and ϩ41 in   the NAG-1 promoter. Three Sp-1 binding sites (Sp-1A, Sp-1B, and Sp-1C) are located in this region of the NAG-1 promoter. Sp-1B and Sp-1C sites overlap with a putative Egr-1 binding site, and play a pivotal role in COX inhibitor-induced NAG-1 expression (3). To further define the responsible element for TSA, 2 internal deletion constructs were transfected into T98G cells (Fig. 8, C and D). The pNAG133del.Sp-1A is a deletion in the Sp-1A site, whereas pNAG133del.Sp-1BC is a deletion in the Sp-1B and Sp-1C sites in the NAG-1 promoter region. After transfecting the deletion constructs into T98 cells, the cells were treated with TSA. Both deletion constructs showed significant reduction of luciferase activity as compared with each non-deletion construct. The luciferase activity of pNAG44, a shorter proximal construct without any Sp-1 binding site but containing the core promoter elements, was not increased by TSA. These findings indicate that the TSA response element is located in the region containing both the Sp-1 and Egr-1 binding sites.
TSA-induced NAG-1 Expression Required Sp-1 and Egr-1-We next determined whether overexpression of Sp-1 or Egr-1 would affect TSAinduced NAG-1 expression. Sp-1 and Egr-1 expression vectors were generated and cotransfected along with the pNAG133/LUC reporter vector into T98G cells (Fig. 9). TSA treatment of Sp-1-and Egr-1-transfected cells enhanced the luciferase activity relative to TSA-treated vector-transfected cells. Egr-1 expression enhanced the promoter activity higher than Sp-1 expression, thus implicating Egr-1 expression as critical for TSA-induced NAG-1 expression.
For further analysis, Sp-1 and Egr-1 siRNA were used to suppress Sp-1 and Egr-1 expression during the TSA-induced NAG-1 expression (Fig. 10). TSA-induced NAG-1 mRNA expression was reduced by 60% with Sp-1 siRNA and by 80% with Egr-1 siRNA. siRNA efficiencies to suppress respective protein expressions were confirmed by Western blotting.
ChIP assay was also used to determine whether Sp-1 protein, Egr-1 protein, and acetylated histones are present in the NAG-1 promoter region containing the putative Sp-1 and Egr-1 binding sites (Fig. 11). We designed 3 site-specific primers of the NAG-1 promoter; two are for random sites not containing the Sp-1/ Egr-1 binding site (non-binding site) and one is for the Sp-1 ABC site. It was impossible to investigate each Sp-1 binding site as these sites are located too close together for a ChIP assay to identify separately. The Sp-1 and Egr-1 proteins appear to bind   (LUC). B, each construct was cotransfected with pRL-null vector into T98G cells using FuGENE 6, and the cells were treated with control (white bars) or 500 ng/ml TSA (black bars) for 24 h in the absence of serum. Promoter activities were measured by luciferase activity. Transfection efficiency for luciferase activity was normalized to Renilla luciferase (pRL-null vector) activity. The y axis shows relative luciferase units (RLU; luciferase activity/Renilla units). The results are the mean Ϯ S.D. of three independent transfections. The mean is given above the bar. C, deletion effects on NAG-1 promoter activity. The constructs of NAG-1 promoter vectors with internal deletion mutations have been described previously. The internal deletion is shown by a gap in the bar. D, wild type (pNAG133), pNAG133 Sp1 site deletion, or wild type (pNAG44) reporters and pRL-null were co-transfected into T98G cells and followed as in B. The results are the mean Ϯ S.D. of three independent transfections. The mean is given above the bar. p values derived from Student's t test are: *, p Ͻ 0.01; **, p Ͻ 0.05. only at the Sp-1 ABC site after TSA treatment. This was about 3-fold higher than controls. In contrast, after TSA treatment, acetyl-histone H3 and H4 protein did not bind at the Sp-1 ABC site, but we did observe a decrease in binding at the non-binding site 1 and an increase at non-binding site 2. These findings indicate that TSA-induced NAG-1 expression is directly regulated by Egr-1 and Sp-1 binding to the Sp-1 ABC site of the NAG-1 promoter.
NAG-1 mRNA Stability-Because NAG-1 protein induction was robust compared with mRNA induction by TSA, we suspected TSA could also alter the stability of NAG-1 mRNA and thereby increase NAG-1 protein expression. The half-lives of NAG-1 and Egr-1 mRNA were assessed by mRNA chase experiments using the transcription inhibitor actinomycin D in the presence or absence of TSA (Fig. 12). Egr-1 mRNA stability was not altered by TSA, but the half-life of NAG-1 mRNA increased in the presence of TSA by 2-fold compared with control. Thus TSA increases NAG-1 mRNA stability, which contributes to the increase in NAG-1 protein expression. This finding indicates that TSA-induced NAG-1 expression occurs at both the transcriptional and post-transcriptional levels.

DISCUSSION
The development of suitable drugs for treatment for glioblastoma, based on inhibition of histone acetylation is an area of intense interest. Histone deacetylase inhibitors, like the model compound TSA, are thought to exert their anticancer effects by altering histone acetylation, which results in alteration in the transcription of target genes or components of the transcriptional machinery. TSA induces cell differentiation, cell-cycle arrest, and apoptosis in cancer cells (16,17). Furthermore, HDAC inhibitors alter apoptotic signaling cascades, such as Rb (33), PTEN (34), tumor necrosis factor-␣ (35), p21 WAF/Cip1 (36 -38), and p53 (39 -41). NAG-1, a pro-apoptotic and anti-cancer gene, is induced by TSA in glioblastoma cells. TSA treatment of glioma cells used in this study increased apoptosis as measured by PARP cleavage and sub-G 1 population. Suppression of TSA-induced NAG-1 expression by siRNA attenuated both cleaved PARP expression and sub-G 1 population. These finding identify NAG-1 as a novel target for histone deacetylase inhibitors and appears to play an important role in mediating the apoptosis observed in response to TSA in these glioblastoma cells. This study provides a direct link between HDAC inhibitor, TSA, and the expression of a protein known to inhibit tumor growth and induce apoptosis.
Both transcriptional and post-transcriptional mechanisms are responsible for TSA-dependent induction of NAG-1. Increases in histone H3 and H4 acetylation were observed after treatment with TSA and similar results were observed with NaBT, also a HDAC inhibitor. We suspect that the effect of TSA on NAG-1 expression may be an indirect effect of changes in histone acetylation. CHX pretreatment abolished TSA-induced NAG-1 expression and thus is dependent on de novo protein synthesis. TSA activates NAG-1 gene expression as measured by luciferase promoter activity, mRNA, and protein expression levels. Based on analysis with NAG-1 promoter luciferase deletion constructs, we suspected that the Sp-1 and Egr-1 binding sites present between positions Ϫ133 and ϩ41 are involved in transcriptional activation. Increases in the protein expression of the transcription factors Sp-1 and Egr-1 were observed after TSA treatment and were observed at times earlier than the increase in NAG-1 expression, a result that indicates these proteins mediate TSA-induced NAG-1 expression.
To gain further insight into whether the induction of NAG-1 by TSA involves Sp-1 and Egr-1, we used different assays: NAG-1 promoter, siRNA, and ChIP. The NAG-1 promoter between positions Ϫ133 and ϩ41 contains three Sp-1 binding sites. Two of the sites, Sp-1B and Sp-1C, overlap with an Egr-1 binding site, and play a pivotal role in sulindac sulfite-and Troglitazone-induced NAG-1 expression. The Sp-1A, Sp-1B, and Sp-1C deletion constructs showed significant reduction in luciferase activity compared with non-deletion constructs, indicating that not only Sp-1B and Sp-1C, but also Sp-1A sites are the TSA response element. A ChIP assay showed that Sp-1 and Egr-1 proteins bind at the Sp-1 ABC site of NAG-1 promoter in TSA-treated cells. These results confirm other reports that show that HDAC inhibitor activation of the Sp-1 binding site (GC-rich sites) in many target genes, such as p21 WAF/Cip1 (38), COX-1 (31), DR-5 (35), hydroxymethylglutaryl-CoA synthase gene (42), PTEN (34), Mn-superoxide dismutase (43), and thrombospondin-1 (44). To investigate the function of Sp-1 and Egr-1 in NAG-1 expression, we examined the effect of Sp-1 and Egr-1 overexpression on NAG-1 promoter activity and the effect of specific siRNAs. Both Sp-1 and Egr-1 overexpression significantly increased the activity of the NAG-1 promoter, and both specific siRNAs for Sp-1 and Egr-1 reduced NAG-1 expression. Thus, both Sp-1 and Egr-1 have positive effects on the induction of NAG-1 expression by TSA. The COX inhibitor, sulindac sulfite, and the PPAR␥ ligand, troglitazone, induce NAG-1 expression in several human cells and are also dependent on Sp-1/Egr-1 binding sites in the NAG-1 promoter. However, this mechanism of NAG-1 transcription requires the MEK1-extracellular signal-regulated kinase 1 and 2 (ERK1/2) signaling pathway (9). For TSA-induced NAG-1 expression the ERK1/2 pathway is not necessary (data not shown) and therefore it is intriguing to speculate that TSA may indirectly up-regulate NAG-1 expression by altering histone acetylation.
In addition, TSA-induced NAG-1 protein expression was robust compared with mRNA expression, and thus we hypothesized other post-transcriptional mechanisms were at work. The half-life of NAG-1 mRNA increased more than 2-fold after TSA treatment. The TSA-mediated increase in NAG-1 mRNA stability could contribute to the increase in NAG-1 protein expression. Other agents that increase NAG-1 expression were reported to have changed the stability of NAG-1 mRNA. Retinoid 6-[3-(1-ademantyl)-4-hydroxyphenyl]-2-naphthalene carboxylic acid induces NAG-1 expression in human lung cancer cells by increasing the stability of NAG-1 mRNA dependent on de novo protein syntheses (7). The experimental anti-cancer drug, 5F-203 in human breast cells, and the PPAR␥ ligand, MCC-555 in human colorectal cells, increase NAG-1 expression by an ERK1/2-dependent increase in the stability of NAG-1 mRNA (10). The 3Ј-untranslated region of human NAG-1 mRNA contains three AU-rich elements that play a role in determining the stability of NAG-1 mRNA. TSA and other HDCA inhibitors are also reported to affect mRNA stability of some genes via induction of the mRNA destabilizing factor (45) and via alteration of subcellular localization of the RNA-binding protein, HuR (46).
Thus, we propose that TSA-induced NAG-1 expression occurs at both the transcriptional and post-transcriptional levels. TSA induces the expression of Egr-1 and Sp-1 transcription factors that subsequently bind to the SP-1/Egr-1 sites present in the Ϫ133 to ϩ41 bp of the NAG-1 promoter and increase the expression of NAG-1 mRNA and protein. In addition, TSA increases the stability of NAG-1 mRNA, which further contributes to the increase in protein expression. TSA increases the acetylation of histones and increases NAG-1 expression, presumably indirectly resulting in changes in chromatin structure, which leads to altered gene expression. This is a new and novel mechanism for the regulation of NAG-1.  Our present results establish for the first time that TSA-induced NAG-1 expression and apoptosis are linked mechanistically in human glioblastoma cells. These finding suggest that a role for NAG-1 in the response of glioblastoma to the histone deacetylase inhibitor such as TSA appears to play an important role in TSA-induced apoptosis. We observed in malignant glioma, glioblastoma multiform (except A172 cell line), less NAG-1 expression than in low grade gliomas and normal human astrocyte. Another study also shows low expression of NAG-1 in glioblastoma cell lines and less expression of NAG-1 in primary glioblastoma grade IV than other brain tumors (47). Thus, malignant transformation may silence the NAG-1 gene in malignant brain tumors.
In summary, our data suggest that NAG-1 induction by TSA is linked to TSA-induced apoptosis. This induction involves not only transcriptional regulation via Sp-1 and Egr-1, but also post-transcriptional regulation. This molecular study provides useful information for the development of optional combinational therapeutic strategies for malignant brain tumor involving HDAC inhibitors.