Histone deacetylase inhibitors (HDI) have been reported to inhibit the growth and survival of cancer cells while leaving normal cells untouched. However, the mechanisms underlying this selective cell death are poorly understood. Gene expression analysis revealed that HDI treatment induced up-regulation of p21WAF1/Cip1 and down-regulation of ErbB2 in cancer cells but not normal cells. Overexpression of p21WAF1/Cip1 and/or silencing of ErbB2 enhanced cancer cell growth inhibition, suggesting that HDI-induced up-regulation/down-regulation of these genes play critical roles in HDI-induced growth inhibition of cancer cells. Most importantly, we found that the gene silencing factor methyl CpG–binding domain protein 3 (MBD3) was not only released from cancer-selective promoter of the HDI up-regulated p21WAF1/Cip1 gene but also recruited to that of the HDI-down-regulated ErbB2 gene. Furthermore, silencing of MBD3 by small interfering RNA abrogated the HDI-induced gene regulation and growth inhibition in lung cancer but not in normal cells. Together, our results support the critical potential of MBD3 in HDI-induced cancer-selective cell death via cancer differential gene expression. (Cancer Res 2005; 65(24): 11400-10)

The accumulation of genetic abnormalities is a major hallmark of cancer, but recent work has indicated that epigenetic change is also an important molecular determinant and may be a critical step in carcinogenesis (13). During cancer development, genes are transcriptionally silenced not only by mutations but also by DNA hypermethylation and histone modifications, which are important early events (16).

Of the epigenetic modifications, histone acetylation/deacetylation is a well-studied paradigm. Histone acetyltransferases and histone deacetylases (HDAC) determine the histone acetylation status and thus regulate gene expression, growth, development, and differentiation in both normal and cancer cells (79). HDACs form dynamic complexes containing diverse required and optional proteins, allowing great functional flexibility for silencing and chromatin remodeling (1014). HDACs interact with DNA methyltransferases (DNMT) and methyl CpG–binding domain (MBD) proteins, which are associated with CpG island methylation, another epigenetic modification involved in transcriptional repression and heterochromatin (1518).

Five mammalian members of the MBD family have been identified: MeCP2 and MBD1-MBD4 (19). All MBD family members recruit various HDAC-containing repressor complexes, which lead to silencing by generating repressive chromatin structures at and near the relevant binding sites (2026). MBD protein binding is determined by DNA methylation status and sequence context. MBD1, MBD2, MBD4, and MeCP2 bind methylated DNA, whereas MBD3 is unable to selectively recognize and bind methylated DNA (2628). Thus, MBD binding is determined by its ability to distinguish methylation status and recognize sequence context.

Frequent co-occurrence of aberrant DNA methylation and histone deacetylation in cancers suggests that these processes are functionally linked in carcinogenesis (2931). As these epigenetic changes can be reversed by DNA or histone modification, HDAC-specific small-molecule inhibitors (HDI) have shown anticancer activity (3241) and have been found to differentially affect growth, differentiation, and death of cancer and normal cells with little toxicity to normal cells (3441). However, the mechanism(s) by which HDACs selectively inhibit cancer cell growth and/or promote cancer cell death is not yet well understood.

Here, we investigated the mechanisms of HDI-induced growth inhibition and death in lung cancer cells. We identified MBD3 as playing an important role in mediating the HDI-induced gene regulations associated with cancer-selective cell death. In lung cancer and normal cells, the distribution of MBD3 in the cell and genome was distinctly regulated by HDIs and was associated with HDI target genes. Silencing of MBD3 abolished the ability of HDIs to induce cancer-specific gene expression and death, whereas altered expressions of the HDI-regulated p21WAF1/Cip1 and ErbB2 genes induce cancer-specific growth inhibition. Taken together, these findings indicate that MBD3 imparts HDI-induced selectivity in cancer cells via differential transcriptional regulation.

Cell culture and treatments. All lung cell lines, except for normal human bronchial epithelial (NHBE) cells (BioWhittaker, Inc., San Diego, CA), were obtained from American Type Culture Collection (Rockville, MD) and grown according to the manufacturer's instructions. Cells were treated with the indicated concentrations of HDIs, trichostatin A, scriptaid, and sodium butyrate (Sigma Chemical Co, St. Louis, MO).

3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay. The growth inhibition of lung cells was measured as described (42).

Microarray analysis. The GeneChip (HG-U133, Affymetrix, Inc., Santa Clara, CA) arrays were done and analyzed as described (43).

Reverse transcription-PCR analysis. Reverse transcription-PCR (RT-PCR) was done as described (4244) using primers as follows: glyceraldehyde-3-phosphate dehydrogenase (GAPDH)-forward (5′-GTCAACGGAATTTGGTCTGTATT-3′) and GAPDH-reverse (5′-AGTCTTCTGGGTGGCAGTGAT-3′), p21WAF1/Cip1-forward (5′-CCTCCTCGGCCCAGTGGAC-3′) and p21WAF1/Cip1-reverse (5′-CCGTTTTCGACCCTGAGAG-3′), E2F1-forward (5′-CTTGGCCGGGGCCCCTGCGG-3′) and E2F1-reverse (5′-TGTGGGCCGGGGCGCCTGGG-3′), cyclin D1-forward (5′-CCTCTTGTGCCACAGATG-3′) and cyclin D1-reverse (5′-GATGTCCACGTCCCGCAC-3′), cyclin B1-forward (5′-CCTGAGCCAGAACCTGAGCC-3′) and cyclin B1-reverse (5′-AGTCACCAATTTCTGGAGGG-3′), 14-3-3σ-forward (5′-GTGACGACAAGAAGCGCATCA-3′) and 14-3-3σ-reverse (5′-GTCGGCCGTCCACAGTGTCAG-3′), ErbB2-forward (5′-TGGCTGCAAGAAGATCTTTG-3′) and ErbB2-reverse (5′-TGCAGTTGACACACTGGGTG-3′), E-cadherin-forward (5′-CTCAAGCTCGCGGATAACCAG-3′) and E-cadherin-reverse (5′-AGGCCCCTGTGCAGCTGGCTC-3′), TS-forward (5′-CCTCTGATGGCGCTGCCTCC-3′) and TS-reverse (5′-GGATGCGGATTGTACCCTTC-3′), H2AX-forward (5′-ATGTCGGGCCGCGGCAAG-3′) and H2AX-reverse (5′-GTACTCCTGGGAGGCCTG-3′), and MBD3-forward (5′-AAGCCCGACCTGAACACGGCG-3′) and MBD3-reverse (5′-CTAGACGTGCTCCATCTCCGG-3′).

Plasmid constructs. To generate mammalian expression vectors for p21WAF1/Cip1 and ErbB2, full-length cDNAs were produced by RT-PCR from RNA of 293T cells using primers as follows: p21WAF1/Cip1 (5′-CCGGAATTCATGTCAGAACCGGCTGGG-3′ and 5′-CCGCTCGAGGGGCTTCCTCTTGGAGAAG-3′) and ErbB2 (5′-CTCAAGCTTCCATGGAGCTGGCGGCCTTG-3′ and 5′-GCCCTCGAGTCACACTGGCACGTCCAGAC-3′). The fragments of p21WAF1/Cip1 and ErbB2 were inserted into Flag-tagged pcDNA3.1 (Invitrogen, Carlsbad, CA) at the EcoRI and XhoI sites and pcDNA3 (Invitrogen) at the HindIII and XhoI sites, respectively. Double-stranded small interfering RNAs (siRNA) for ErbB2 and MBD3 were generated using pSUPER vector (42). The siRNA primers designed for targets, AAATTCCAGTGGCCATCAA and AGACGGCGTCCATCTTCAA, according to ErbB2 and MBD3 mRNA sequences, respectively, are as follows: ErbB2 (5′-GATCCCCAAATTCCAGTGGCCATCAATTCAAGAGATTGATGGCCACTGGAATTTTTTTTGGAAA-3′ and 5′-AGCTTTTCCAAAAAAAATTCCAGTGGCCATCAATCTCTTGAATTGATGGCCACTGGAATTTGGG-3′) and MBD3 (5′-GATCCCCAGACGGCGTCCATCTTCAATTCAAGAGATTGAAGATGGACGCCGTCTTTTTGGAAA-3′ and 5′-AGCTTTTCCAAAAAAGACGGCGTCCATCTTCAATCTCTTGAATTGAAGATGGACGCCGTCTGGG-3′).

Immunoblotting and immunofluorescence staining. Immunoblotting and immunostaining were done as described previously (42). Flag-p21WAF1/Cip1, ErbB2, and MBD3 were detected with anti-Flag (Sigma), anti-ErbB2 (NeoMarkers, Fremont, CA) and anti-MBD3 (Santa Cruz Biotechnology, Santa Cruz, CA) antibodies, respectively.

Chromatin immunoprecipitation assay. The chromatin immunoprecipitation assay with antibodies against acetylated histones 3 and 4 (Upstate Biotechnology, Lake Placid, NY), HDAC1 and HDAC2 (Zymed, San Francisco, CA), DNMT1, and MBD1-MBD3 (Santa Cruz Biotechnology) was done as described previously (44). The immunoprecipitate was amplified using promoter-specific primers as follows: p21WAF1/Cip1 (5′-CCTTGCCTGCCAGAGGGG-3′ and 5′-CAGCTGCTCACACCTCAG-3′), ErbB2 (5′-CCCGGACTCCGGGGGAGG-3′ and 5′-CCCGGGGGGCTCCCCTGG-3′), and E-cadherin (5′-GCAGTGAGCTGTGATCGCACC-3′ and 5′-GAGAGGGGGTGCGTGGCTGC-3′).

Trichostatin A affects gene expression differently in lung cancer and normal cells. Based on the previous reports that HDIs exhibited cancer-selective cytotoxicity (3441), we first tested cytotoxicity of HDIs, including trichostatin A, scriptaid, and sodium butyrate, on NHBE cells, bronchial epithelial cells immortalized by SV40 (BEAS-2B), and several cancer cells, alveolar epithelial adenocarcinoma A549, epithelial-like sub-bronchial gland adenocarcinoma CALU-3, pleural effusion adenocarcinoma H460, adenocarcinoma H23, bronchioalveolar carcinoma H358, and non–small cell lung cancer H157 and H226. HDIs inhibited cell growth highly by ∼80% to 90% in all the tested lung cancer cells but only by ∼20% in NHBE cells and ∼40% in nontransformed BEAS-2B cells (data not shown), suggesting that HDIs inhibit cell growth more effectively in cancerous lung cells than in normal lung cells. We next examined another normal human lung fibroblast WI38 and its derivative SV40-transformed VA13 cells. HDIs inhibited the growth of VA13 cells (∼90%) but had little effect on WI38 cell growth (∼20%; data not shown), indicating that HDIs efficiently induce growth inhibition and cell death of transformed (VA13) but not normal (WI38) lung cells. Collectively, these results support the notion that HDIs inhibit cell growth more effectively in lung cancer cells than in primary normal (NHBE and WI38) and nontransformed (BEAS-2B) lung cells, consistent with the previous reports (3441).

The differential growth inhibition and death of cancer and normal cells treated with HDIs prompted us to examine possible mechanisms underlying this effect. HDAC activity plays an important role in regulating gene expression. Thus, we examined the possibility that HDIs could differentially modulate transcriptional profiles of genes in lung cancer and normal cells leading to the observed cancer-specific cytotoxicity. To identify potential genes mediating HDI-induced differential cytotoxicity in lung cancer and normal cells, we analyzed the expression profiles of trichostatin A–regulated genes in lung cancer (A549) and normal (NHBE) cells by performing microarray in cells treated with or without trichostatin A. We first identified trichostatin A–responsive genes that were up-regulated or down-regulated in A549 cells but were not altered or differentially regulated in NHBE cells (data not shown) and selected differentially expressed genes as being possibly associated with the cancer-selective mechanism of trichostatin A–induced growth inhibition and chromatin remodeling (Table 1) and then confirmed our microarray data by RT-PCR (Fig. 1A). Consistent with our microarray data (Table 1), trichostatin A efficiently induced the expression of the gene encoding p21WAF1/Cip1, a cyclin-dependent kinase inhibitor leading to G1 cell cycle arrest, in A549 and H460 cancer cells but not in NHBE cells (Fig. 1A). In contrast, RT-PCR confirmed our microarray data by showing that the ErbB2 gene, a clinically significant kinase implicated in cell growth (45), was down-regulated by trichostatin A in cancer cells but not in NHBE cells (Fig. 1A). Interestingly, expression of the E-cadherin gene was decreased by trichostatin A in A549 cells but not in either H460 or NHBE cells (Fig. 1A).

Table 1.

NHBE and A549 cells present different gene expression profiles in response to trichostatin A treatment

NHBE-trichostatin A vs NHBE-none
A549-trichostatin A vs A549-none
SymbolDescriptionFCSymbolDescriptionFC
Up-regulation   Up-regulation   
IL11RA Interleukin-11 receptor, α 17.1 AKR1B10 Aldo-keto reductase family 1, member B10 (aldose reductase) kynureninase 21.85 
SLC23A1 Nucleobase transporter-like 1 protein 16 CCK Cholecystokinin 16 
VGF VGF nerve growth factor inducible 16 PTPN12 Protein tyrosine phosphatase, nonreceptor type 12 15.78 
LMCD1 LIM and cysteine-rich domains 1 14.9 S100A9 S100 calcium-binding protein A9 (calgranulin B) 15.67 
STC1 Stanniocalcin 1 13.9 TFPI2 Tissue factor pathway inhibitor 2 13.92 
MBD2 Methyl CpG–binding domain protein 2 13.9 KYNU Kynureninase (l-kynurenine hydrolase) 10.55 
p55 Phosphoinositide 3-kinase, regulatory subunit, polypeptide 3 13 MAPK8 Mitogen-activated protein kinase 8 9.64 
NILR Novel interleukin receptor 12.1 ACTN4 Actinin, α4 8.69 
THBS2 Thrombospondin 2 11.3 HSPA2 Heat shock 70-kDa protein 2 6.68 
H3FK Histone H3 family, member K 10.6 TGFBR2 Transforming growth factor, β receptor II 5.93 
FLJ22408 Hypothetical protein FLJ22408 9.85 CDC42BPB CDC42-binding protein kinase β (DMPK-like) 
CRLF1 Cytokine receptor-like factor 1 5.66 EPS8 Epidermal growth factor receptor pathway substrate 8 3.68 
SBF1 Nuclear dual-specificity phosphatase BNIP3L BCL2/adenovirus E1B 19-kDa interacting protein 3-like 3.55 
MBD4 Methyl CpG–binding domain protein 4 3.48 PIGPC1 p53-induced protein PIGPC1 3.09 
ZNF277 Zinc finger protein 277 2.64 p21 Cyclin-dependent kinase inhibitor 1A (Cip1) 
TCFL5 Transcription factor-like 5 (basic helix-loop-helix) 2.64 NCOR2 Nuclear receptor corepressor 2 2.9 
CCNA1 Cyclin A1 2.46 PTP4A2 Protein tyrosine phosphatase type IVA, member 2 2.56 
PCTP Phosphatidylcholine transfer protein 2.3 STC1 Stanniocalcin 1 2.36 
KBRAS2 IκB-interacting Ras-like protein 2 2.14 TTRAP TRAF and tumor necrosis factor receptor-associated protein 2.31 
cPLA2 Phosphatidylcholine 2-acylhydrolase 2.14 YWHAZ 14-3-3 protein 
SURF1 Surfeit 1 2.14    
PAWR PRKC, apoptosis, WT1, regulator 2.14    
SIRT2 Sirtuin (silent mating type information regulation 2, Saccharomyces cerevisiae, homologue) 2    
      
Down-regulation   Down-regulation   
PHAP1 Putative human HLA class II associated protein I 17.1 H2AFX H2A histone family, member X 124.5 
IL7R Interleukin-7 receptor 11.3 TYMS Thymidylate synthetase 25.99 
FGF5 Fibroblast growth factor 5 10.6 CDC6 CDC6 cell division cycle 6 homologue 22.9 
G0S2 Putative lymphocyte G0G1 switch gene 9.85 CENPF Centromere protein F, 350/400 kDa (mitosin) 22.78 
STK6 Serine/threonine kinase 6 9.19 PRC1 Protein regulator of cytokinesis 1 18.37 
SNRPF Small nuclear ribonucleoprotein polypeptide F 7.46 CKS1B CDC28 protein kinase regulatory subunit 1B 15.77 
CCNA2 Cyclin A2 7.46 SIVA CD27-binding (Siva) protein 11.39 
CUL4A Cullin 4A 6.96 BAD BCL2-antagonist of cell death 9.18 
NASP Nuclear autoantigenic sperm protein (histone-binding) interleukin-7 receptor 6.5 DNAJC9 DnaJ (Hsp40) homologue, subfamily C, member 9 7.83 
TGF-α Transforming growth factor-α 5.66 CKS1B CDC28 protein kinase regulatory subunit 1B 6.19 
PGAR PPAR(γ) angiopoietin-related protein 4.59 CDK2 Cyclin-dependent kinase 2 5.65 
TYMS Thymidylate synthetase 4.59 ERBB2 v-erb-b2 avian erythroblastic leukemia viral oncogene homologue 2 4.59 
ANX3 1,2-Cyclic-inositol phosphate phosphodiesterase 3.73 E2F1 E2F transcription factor 1 
INHBC Inhibin, β C 3.48 Cyclin B1 Cyclin B1 
Cyclin B1 Cyclin B1 3.25 CDK4 Cyclin-dependent kinase 4 3.5 
NORPEG Novel retinal pigment epithelial gene 2.83 XRCC5 X-ray repair complementing defective repair in Chinese hamster cells 5 2.82 
TFDP1 Transcription factor Dp-1 2.64 SAP18 Sin3-associated polypeptide, 18 kDa 2.69 
HSP90 Heat shock protein 1, α, 90 kDa 2.3 HSBP1 Heat shock factor–binding protein 1 2.09 
CHC1 Chromosome condensation 1 2.14 TXNIP Thioredoxin-interacting protein 2.04 
H2AFZ H2A histone family, member Z 1.62 MGC32043 Hypothetical protein MGC32043 
NHBE-trichostatin A vs NHBE-none
A549-trichostatin A vs A549-none
SymbolDescriptionFCSymbolDescriptionFC
Up-regulation   Up-regulation   
IL11RA Interleukin-11 receptor, α 17.1 AKR1B10 Aldo-keto reductase family 1, member B10 (aldose reductase) kynureninase 21.85 
SLC23A1 Nucleobase transporter-like 1 protein 16 CCK Cholecystokinin 16 
VGF VGF nerve growth factor inducible 16 PTPN12 Protein tyrosine phosphatase, nonreceptor type 12 15.78 
LMCD1 LIM and cysteine-rich domains 1 14.9 S100A9 S100 calcium-binding protein A9 (calgranulin B) 15.67 
STC1 Stanniocalcin 1 13.9 TFPI2 Tissue factor pathway inhibitor 2 13.92 
MBD2 Methyl CpG–binding domain protein 2 13.9 KYNU Kynureninase (l-kynurenine hydrolase) 10.55 
p55 Phosphoinositide 3-kinase, regulatory subunit, polypeptide 3 13 MAPK8 Mitogen-activated protein kinase 8 9.64 
NILR Novel interleukin receptor 12.1 ACTN4 Actinin, α4 8.69 
THBS2 Thrombospondin 2 11.3 HSPA2 Heat shock 70-kDa protein 2 6.68 
H3FK Histone H3 family, member K 10.6 TGFBR2 Transforming growth factor, β receptor II 5.93 
FLJ22408 Hypothetical protein FLJ22408 9.85 CDC42BPB CDC42-binding protein kinase β (DMPK-like) 
CRLF1 Cytokine receptor-like factor 1 5.66 EPS8 Epidermal growth factor receptor pathway substrate 8 3.68 
SBF1 Nuclear dual-specificity phosphatase BNIP3L BCL2/adenovirus E1B 19-kDa interacting protein 3-like 3.55 
MBD4 Methyl CpG–binding domain protein 4 3.48 PIGPC1 p53-induced protein PIGPC1 3.09 
ZNF277 Zinc finger protein 277 2.64 p21 Cyclin-dependent kinase inhibitor 1A (Cip1) 
TCFL5 Transcription factor-like 5 (basic helix-loop-helix) 2.64 NCOR2 Nuclear receptor corepressor 2 2.9 
CCNA1 Cyclin A1 2.46 PTP4A2 Protein tyrosine phosphatase type IVA, member 2 2.56 
PCTP Phosphatidylcholine transfer protein 2.3 STC1 Stanniocalcin 1 2.36 
KBRAS2 IκB-interacting Ras-like protein 2 2.14 TTRAP TRAF and tumor necrosis factor receptor-associated protein 2.31 
cPLA2 Phosphatidylcholine 2-acylhydrolase 2.14 YWHAZ 14-3-3 protein 
SURF1 Surfeit 1 2.14    
PAWR PRKC, apoptosis, WT1, regulator 2.14    
SIRT2 Sirtuin (silent mating type information regulation 2, Saccharomyces cerevisiae, homologue) 2    
      
Down-regulation   Down-regulation   
PHAP1 Putative human HLA class II associated protein I 17.1 H2AFX H2A histone family, member X 124.5 
IL7R Interleukin-7 receptor 11.3 TYMS Thymidylate synthetase 25.99 
FGF5 Fibroblast growth factor 5 10.6 CDC6 CDC6 cell division cycle 6 homologue 22.9 
G0S2 Putative lymphocyte G0G1 switch gene 9.85 CENPF Centromere protein F, 350/400 kDa (mitosin) 22.78 
STK6 Serine/threonine kinase 6 9.19 PRC1 Protein regulator of cytokinesis 1 18.37 
SNRPF Small nuclear ribonucleoprotein polypeptide F 7.46 CKS1B CDC28 protein kinase regulatory subunit 1B 15.77 
CCNA2 Cyclin A2 7.46 SIVA CD27-binding (Siva) protein 11.39 
CUL4A Cullin 4A 6.96 BAD BCL2-antagonist of cell death 9.18 
NASP Nuclear autoantigenic sperm protein (histone-binding) interleukin-7 receptor 6.5 DNAJC9 DnaJ (Hsp40) homologue, subfamily C, member 9 7.83 
TGF-α Transforming growth factor-α 5.66 CKS1B CDC28 protein kinase regulatory subunit 1B 6.19 
PGAR PPAR(γ) angiopoietin-related protein 4.59 CDK2 Cyclin-dependent kinase 2 5.65 
TYMS Thymidylate synthetase 4.59 ERBB2 v-erb-b2 avian erythroblastic leukemia viral oncogene homologue 2 4.59 
ANX3 1,2-Cyclic-inositol phosphate phosphodiesterase 3.73 E2F1 E2F transcription factor 1 
INHBC Inhibin, β C 3.48 Cyclin B1 Cyclin B1 
Cyclin B1 Cyclin B1 3.25 CDK4 Cyclin-dependent kinase 4 3.5 
NORPEG Novel retinal pigment epithelial gene 2.83 XRCC5 X-ray repair complementing defective repair in Chinese hamster cells 5 2.82 
TFDP1 Transcription factor Dp-1 2.64 SAP18 Sin3-associated polypeptide, 18 kDa 2.69 
HSP90 Heat shock protein 1, α, 90 kDa 2.3 HSBP1 Heat shock factor–binding protein 1 2.09 
CHC1 Chromosome condensation 1 2.14 TXNIP Thioredoxin-interacting protein 2.04 
H2AFZ H2A histone family, member Z 1.62 MGC32043 Hypothetical protein MGC32043 

NOTE: Oligonucleotide microarray analysis was used to distinguish trichostatin A–dependent genes in both cell types and identify genes significantly overexpressed or underexpressed in A549 cells compared with NHBE cells following trichostatin A treatment.

Figure 1.

Trichostatin A–induced transcriptional modulation in lung cells. A, RT-PCR-based validation of selected oligonucleotide microarray results in A549 and NHBE cells. RT-PCR analysis was done in normal (NHBE) and cancer cells (A549 and H460) treated with trichostatin A (TSA; 0.5 μmol/L) for 24 hours. The specific primer pairs for each gene are described in Materials and Methods and the results are normalized against GAPDH expression. B, expression of p21WAF1/Cip1 and ErbB2 was analyzed by RT-PCR in the trichostatin A–treated lung normal WI-38 and its derivative transformed VA13 cells, lung immortalized BEAS-2B, and CALU-3, H23, H358, H157, and H226 cancer cells.

Figure 1.

Trichostatin A–induced transcriptional modulation in lung cells. A, RT-PCR-based validation of selected oligonucleotide microarray results in A549 and NHBE cells. RT-PCR analysis was done in normal (NHBE) and cancer cells (A549 and H460) treated with trichostatin A (TSA; 0.5 μmol/L) for 24 hours. The specific primer pairs for each gene are described in Materials and Methods and the results are normalized against GAPDH expression. B, expression of p21WAF1/Cip1 and ErbB2 was analyzed by RT-PCR in the trichostatin A–treated lung normal WI-38 and its derivative transformed VA13 cells, lung immortalized BEAS-2B, and CALU-3, H23, H358, H157, and H226 cancer cells.

Close modal

We next examined differential expression of p21WAF1/Cip1 and ErbB2 genes in additional lung cancer and normal cells. Our RT-PCR results revealed that although the expression of genes encoding p21WAF1/Cip1 and ErbB2 was not altered in normal lung WI38 cells the expression of p21WAF1/Cip1 gene was increased and the expression of ErbB2 gene was reduced by trichostatin A in its transformed VA13 cells (Fig. 1B). Trichostatin A also induced the expression of p21WAF1/Cip1 gene and reduced the expression of ErbB2 gene in other CALU-3, H23, H358, and H157 lung cancer cells (Fig. 1B). Interestingly, the expression of p21WAF1/Cip1 gene was not affected, but the expression of ErbB2 was reduced by trichostatin A in BEAS-2B cells (Fig. 1B), indicating that trichostatin A–induced transcriptional regulation profiles in BEAS-2B cells differ between normal (expression of p21WAF1/Cip1 gene) and cancer (expression of ErbB2 gene) cells. Together, these results suggest that trichostatin A differentially modulates gene expression in lung cancer and normal cells and may target distinct genes for transcriptional regulation.

Up-regulation of p21WAF1/Cip1 and down-regulation of ErbB2 play roles in trichostatin A–induced cancer cell growth inhibition. To study whether differential trichostatin A–induced gene regulation was functionally linked to cancer-selective cell death, we evaluated the roles of p21WAF1/Cip1 up-regulation and ErbB2 down-regulation in growth inhibition of A549 cells. We first tested whether transcriptional modulation was required for HDI-induced cancer selective death. Cotreatment with trichostatin A and the transcription blocker, actinomycin D, significantly alleviated the death of cancer cells in comparison with cultures treated with trichostatin A alone or actinomycin D alone (data not shown). This indicates that transcription is required for HDI-induced cancer cell death.

We next examined cancer cell growth inhibition following up-regulation of p21WAF1/Cip1 and down-regulation of ErbB2 (Fig. 2A), which occurred in the trichostatin A–treated cancer cells (Table 1; Fig. 1). Overexpression of exogenous p21WAF1/Cip1 inhibited growth of A549 cells by ∼60% (Fig. 2B,, top), indicating that induction of p21WAF1/Cip1 expression can inhibit cancer cell growth in the absence of trichostatin A treatment. The growth inhibition of A549 cells 3 days after p21WAF1/Cip1 overexpression (Fig. 2B,, top) was similar to that seen in A549 cells treated with 0.5 μmol/L trichostatin A for 2 days (compare with Fig. 2B,, bottom). We then evaluated the effect of ErbB2 down-regulation on A549 cell growth inhibition (Fig. 2B). Knockdown of ErbB2 expression by siRNA enhanced the growth inhibition of A549 cells within 3 days in the absence of trichostatin A (Fig. 2B , top), consistent with the results of p21WAF1/Cip1 up-regulation. Taken together, these results indicate that up-regulation of p21WAF1/Cip1 or down-regulation of ErbB2 can inhibit the growth of A549 cells, providing support for our hypothesis that these trichostatin A–induced gene up-regulation/down-regulations may mediate the cancer-selective growth inhibition.

Figure 2.

Effects of p21WAF1/Cip1 overexpression or ErbB2 silencing on cancer cell growth inhibition. A, overexpression of exogenous p21WAF1/Cip1 and knockdown of ErbB2 by siRNA. A549 cells were transfected with pcDNA-Flag-p21WAF1/Cip1 (or empty pcDNA-Flag) and/or pSUPER-si-ErbB2. Forty-eight hours after transfection, the exogenously overexpressed Flag-p21WAF1/Cip1 (Flag-p21) and silenced ErbB2 (si-ErbB2) were measured by immunoblotting with anti-Flag and anti-ErbB2 antibodies, respectively. The level of β-actin protein was detected as a loading control. B, effect of the overexpressed p21WAF1/Cip1 and/or silenced ErbB2 on cancer cell growth inhibition. Twenty-four hours after transfection with Flag-p21WAF1/Cip1 and/or si-ErbB2, growth inhibition was measured by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay of A549 cells treated without (top) or with 0.2 μmol/L (middle) or 0.5 μmol/L (bottom) trichostatin A for the indicated times. Columns, mean of four independent experiments, each in triplicate; bars, SD. C, role of p21WAF1/Cip1 and effect of ErbB2 silencing on trichostatin A–induced growth inhibition. Silencing of ErbB2 expression in HCT116 cells expressing p21WAF1/Cip1 (p21-WT) or not (p21-KO) was detected by immunoblotting. Immunoblots were probed with anti-p21WAF1/Cip1 and anti-ErbB2 antibodies, respectively, to examine the expression of p21WAF1/Cip1 and ErbB2 in HCT116-p21-WT and HCT116-p21-KO cells transfected with empty pSUPER or si-ErbB2. β-Actin was detected as a loading control (top). In the presence of 0.5 μmol/L trichostatin A, growth inhibition was evaluated by MTT assay in HCT116-p21WAF1/Cip1-KO (HCT116-KO; bottom) cells versus wild-type HCT116 (HCT116-WT; middle) cells. In addition, growth inhibition was evaluated in wild-type HCT116 and HCT116-p21WAF1/Cip1-KO cells expressing empty pSUPER or si-ErbB2 following treatment with trichostatin A (0.5 μmol/L) for 3 days. Columns, mean of three experimental determinations; bars, SD. D, abrogation of trichostatin A–induced growth inhibition by exogenously introduced ErbB2 (exo-ErbB2). Overexpression of ErbB2 was measured by immunoblotting and growth inhibition was assayed by MTT analysis. Experiments in (C) were repeated.

Figure 2.

Effects of p21WAF1/Cip1 overexpression or ErbB2 silencing on cancer cell growth inhibition. A, overexpression of exogenous p21WAF1/Cip1 and knockdown of ErbB2 by siRNA. A549 cells were transfected with pcDNA-Flag-p21WAF1/Cip1 (or empty pcDNA-Flag) and/or pSUPER-si-ErbB2. Forty-eight hours after transfection, the exogenously overexpressed Flag-p21WAF1/Cip1 (Flag-p21) and silenced ErbB2 (si-ErbB2) were measured by immunoblotting with anti-Flag and anti-ErbB2 antibodies, respectively. The level of β-actin protein was detected as a loading control. B, effect of the overexpressed p21WAF1/Cip1 and/or silenced ErbB2 on cancer cell growth inhibition. Twenty-four hours after transfection with Flag-p21WAF1/Cip1 and/or si-ErbB2, growth inhibition was measured by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay of A549 cells treated without (top) or with 0.2 μmol/L (middle) or 0.5 μmol/L (bottom) trichostatin A for the indicated times. Columns, mean of four independent experiments, each in triplicate; bars, SD. C, role of p21WAF1/Cip1 and effect of ErbB2 silencing on trichostatin A–induced growth inhibition. Silencing of ErbB2 expression in HCT116 cells expressing p21WAF1/Cip1 (p21-WT) or not (p21-KO) was detected by immunoblotting. Immunoblots were probed with anti-p21WAF1/Cip1 and anti-ErbB2 antibodies, respectively, to examine the expression of p21WAF1/Cip1 and ErbB2 in HCT116-p21-WT and HCT116-p21-KO cells transfected with empty pSUPER or si-ErbB2. β-Actin was detected as a loading control (top). In the presence of 0.5 μmol/L trichostatin A, growth inhibition was evaluated by MTT assay in HCT116-p21WAF1/Cip1-KO (HCT116-KO; bottom) cells versus wild-type HCT116 (HCT116-WT; middle) cells. In addition, growth inhibition was evaluated in wild-type HCT116 and HCT116-p21WAF1/Cip1-KO cells expressing empty pSUPER or si-ErbB2 following treatment with trichostatin A (0.5 μmol/L) for 3 days. Columns, mean of three experimental determinations; bars, SD. D, abrogation of trichostatin A–induced growth inhibition by exogenously introduced ErbB2 (exo-ErbB2). Overexpression of ErbB2 was measured by immunoblotting and growth inhibition was assayed by MTT analysis. Experiments in (C) were repeated.

Close modal

As either p21WAF1/Cip1 up-regulation (∼55%) or ErbB2 down-regulation (∼52%) inhibited cell growth to levels lower than those induced by 0.2 μmol/L (∼68%; Fig. 2B,, top and middle) or 0.5 μmol/L trichostatin A (∼80%; Fig. 2B,, top and bottom), we next tested these effects in combination. The growth inhibition effect of p21WAF1/Cip1 up-regulation (∼84%) or ErbB2 down-regulation (∼83%) in the presence of 0.2 or 0.5 μmol/L trichostatin A was slightly higher than that of 0.2 μmol/L trichostatin A alone (∼68%) and similar to that of 0.5 μmol/L trichostatin A alone (∼80%; Fig. 2B), indicating that the combinatory effect of the gene up-regulation/down-regulation with trichostatin A seems to be relatively small compared with those induced by trichostatin A alone. These observations suggest that p21WAF1/Cip1 up-regulation and ErbB2 down-regulation induced by trichostatin A may be sufficient and potent for growth inhibition of A549 cells; consequently, either exogenous expression of p21WAF1/Cip1 or silencing of ErbB2 may not enhance the growth inhibition. In addition, A549 cells expressing both the p21WAF1/Cip1 expression vector and the ErbB2-targeting siRNA showed growth inhibitions similar to those following expression of either alone without regard to trichostatin A treatment (Fig. 2B). These results indicate that either up-regulation of p21WAF1/Cip1 expression or down-regulation of ErbB2 alone may play a role in growth inhibition of A549 cells.

To further support the importance of p21WAF1/Cip1 up-regulation and/or ErbB2 down-regulation in trichostatin A–induced cancer-selective growth inhibition, cell growth was measured in p21WAF1/Cip1-knockout (HCT116-p21WAF1/Cip1-KO) and p21WAF1/Cip1 stably expressing HCT116 cells (HCT116-p21WAF1/Cip1-WT) with or without silencing of the ErbB2 gene (si-ErbB2; Fig. 2C). Silencing of ErbB2 induced growth inhibition of HCT116-p21WAF1/Cip1-WT cells (33%) by >2-fold that of HCT116-p21WAF1/Cip1-KO cells (12%) without trichostatin A. In addition, trichostatin A efficiently induced growth inhibition of HCT116-p21WAF1/Cip1-WT cells by >2-fold that observed in HCT116-p21WAF1/Cip1-KO cells (75% versus 33%, respectively; Fig. 2C), indicating that p21WAF1/Cip1 is required for efficient trichostatin A–induced inhibition of cancer cell growth. We next examined whether constitutive expression of ErbB2 could block trichostatin A–mediated growth inhibition. Following overexpression of exogenously introduced ErbB2, the growth inhibition effect of trichostatin A was alleviated in both HCT116-p21WAF1/Cip1-WT (68-35%) and HCT116-p21WAF1/Cip1-KO cells (30-13%; Fig. 2D), indicating that cells overexpressing ErbB2 do not exhibit efficient trichostatin A–induced growth inhibition. Taken together, these findings suggest that trichostatin A–induced up-regulation of p21WAF1/Cip1 and down-regulation of ErbB2 may be important determinants for HDI-induced lung cancer cell cytotoxicity.

Release of MBD3 from the p21WAF1/Cip1 promoter and its recruitment to the ErbB2 promoter in trichostatin A–treatedlung cancer cells. To investigate the molecular mechanism underlying HDI-mediated differential transcription in lung cancer and normal cells, we examined the promoters of the differentially expressed p21WAF1/Cip1 (Fig. 3A), ErbB2 (Fig. 3B), and E-cadherin (Fig. 3C) genes. Because HDIs mainly act on repression/derepression rather than activation/inactivation of transcriptional regulation, we used a chromatin immunoprecipitation assay to study the recruitment of components in the HDAC complex, which is generally associated with repression/derepression processes. Trichostatin A treatment increased the acetylation status of histone 3 (data not shown) and histone 4 (Fig. 3A,, Ac-H4) on the p21WAF1/Cip1 promoter in both normal and cancer cells, suggesting that the induction of histone acetylation by trichostatin A may result in opening of nearby chromatin in both normal and cancerous lung cells. In contrast, the association of HDAC2 with the p21WAF1/Cip1 promoter increased following trichostatin A treatment in NHBE and WI38 cells but decreased in cancer cells (Fig. 3A). In the presence of trichostatin A, the association of DNMT1 with the p21WAF1/Cip1 promoter increased in normal cells, whereas its association was absent or differed in cancer cell lines. Additionally, chromatin immunoprecipitation analyses with anti-MBD1-MBD3 antibodies revealed that these transcriptional repressor proteins were differentially associated with the p21WAF1/Cip1 promoter in cancer and normal lung cells before and after trichostatin A treatment (Fig. 3A). Following trichostatin A treatment, the association of MBD1 and MBD3 with the p21WAF1/Cip1 promoter decreased in cancer cells and increased in NHBE and WI38 cells (Fig. 3A), indicating that the composition of the transcriptional repressor complex located on the p21WAF1/Cip1 promoter differed in normal and cancer cells. The recruitment of MBD2 to the p21WAF1/Cip1 promoter was observed in normal NHBE and WI38 cells and was not affected by trichostatin A. In addition, the recruitment of MBD2 to the p21WAF1/Cip1 promoter was absent in A549, H460, and H358 cancer cells or was decreased by trichostatin A in CALU-3 and H23 cells (Fig. 3A). These results seem to suggest that trichostatin A–induced up-regulation of p21WAF1/Cip1 in lung cancer cells may be associated with the cancer-specific dissociations of MBD1 and MBD3 at the p21WAF1/Cip1 promoter.

Figure 3.

Recruitment of MBD1-MBD3 proteins. A to C, chromatin immunoprecipitation assays at the p21WAF1/Cip1 (A), ErbB2 (B), and E-cadherin (C) promoters in NHBE, A549, H460, H358, WI38, VA13, CALU-3, or H23 cells. D, association of MBD3 with the p21WAF1/Cip1 and ErbB2 promoters in trichostatin A–treated immortalized BEAS-2B and cancer (H226 and H157) cells. Chromatin immunoprecipitation assays were done following treatment without (none) or with 0.5 μmol/L trichostatin A for 24 hours. Anti-acetylated histone H4 (Ac-H4), anti-HDAC2, anti-DNMT1, and anti-MBD antibodies were used to pull-down acetylated histone, HDAC2, DNMT1, and MBD (MBD1-MBD3) proteins, respectively. The negative control was immunoprecipitation without antibody (no Ab) and the positive control consisted of 0.1% of the total chromatin sample before immunoprecipitation (Before IP). Pull-down chromatin pools were PCR amplified with specific primers for the p21WAF1/Cip1 (p21 promoter; A and D), ErbB2 (ErbB2 promoter; B and D), and E-cadherin (E-cadherin promoter; C) promoters. The primer sets were designed to amplify regions containing conserved GC boxes in each promoter are listed in Materials and Methods.

Figure 3.

Recruitment of MBD1-MBD3 proteins. A to C, chromatin immunoprecipitation assays at the p21WAF1/Cip1 (A), ErbB2 (B), and E-cadherin (C) promoters in NHBE, A549, H460, H358, WI38, VA13, CALU-3, or H23 cells. D, association of MBD3 with the p21WAF1/Cip1 and ErbB2 promoters in trichostatin A–treated immortalized BEAS-2B and cancer (H226 and H157) cells. Chromatin immunoprecipitation assays were done following treatment without (none) or with 0.5 μmol/L trichostatin A for 24 hours. Anti-acetylated histone H4 (Ac-H4), anti-HDAC2, anti-DNMT1, and anti-MBD antibodies were used to pull-down acetylated histone, HDAC2, DNMT1, and MBD (MBD1-MBD3) proteins, respectively. The negative control was immunoprecipitation without antibody (no Ab) and the positive control consisted of 0.1% of the total chromatin sample before immunoprecipitation (Before IP). Pull-down chromatin pools were PCR amplified with specific primers for the p21WAF1/Cip1 (p21 promoter; A and D), ErbB2 (ErbB2 promoter; B and D), and E-cadherin (E-cadherin promoter; C) promoters. The primer sets were designed to amplify regions containing conserved GC boxes in each promoter are listed in Materials and Methods.

Close modal

Similar analyses of the ErbB2 promoter revealed that trichostatin A had little effect on histone acetylation in normal NHBE and WI38 cells but induced histone acetylation of this promoter in cancer cell lines, except for CALU-3 and H23 cells (Fig. 3B). Analysis of histone acetylation and transcriptional regulator association/dissociation at the ErbB2 promoter revealed no apparent difference in NHBE and WI38 cells with and without trichostatin A treatment (Fig. 3B), whereas the cancer cell lines showed significantly enhanced recruitment of MBD3 following trichostatin A treatment. We observed slight or negligible changes in the associations of HDAC2, MBD1, and MBD2 with the ErbB2 promoter in trichostatin A–treated cancer cells (Fig. 3B) and noted a significant release of DNMT1 from the ErbB2 promoter following trichostatin A treatment in cancer cells (Fig. 3B). Collectively, these results suggest that trichostatin A–induced down-regulation of ErbB2 in lung cancer cells may be implicated with the cancer-specific recruitment of MBD3 at the ErbB2 promoter.

In contrast to the cancer-selective regulation of p21WAF1/Cip1 and ErbB2 following trichostatin A treatment, transcription of the E-cadherin gene decreased slightly in A549 cells, increased in H460 cells, and remained unchanged in NHBE cells (Fig. 1A). Our analyses revealed slight increases in histone acetylation and increased association of MBD1-MBD3 and DNMT1 with the E-cadherin promoter in trichostatin A–treated NHBE cells (Fig. 3C). However, HDAC2 was not associated with the E-cadherin promoter of NHBE cells (Fig. 3C), perhaps accounting for the minimal change in E-cadherin transcription in these cells under conditions that would seem to favor transcription. In trichostatin A–treated A549 cells, we observed recruitment of MBD1, MBD2, DNMT1, and HDAC2 to the E-cadherin promoter but no recruitment of MBD3 (Fig. 3C), perhaps accounting for the slight repression of E-cadherin expression observed in these cells. In trichostatin A–treated H460 cells, we observed decreased association of HDAC2 and MBD3, increased association of MBD2, and no change in recruitment of MBD1 (Fig. 3C). These differences in E-cadherin promoter recruitment and histone acetylation patterns between A549 and H460 cells likely account for the noted differences in the transcriptional expression of E-cadherin in these two cancer cell types (Fig. 1A). Taken together, the chromatin immunoprecipitation data from the E-cadherin promoter suggest that (a) increased association of MBD1 and MBD2 with the promoter has no effect on transcription, but recruitment of HDAC2 along with MBD1 and MBD2 can induce transcriptional repression, and (b) release of HDAC2 and MBD3 from the promoter can induce transcriptional up-regulation.

To further explore whether release/recruitment of MBD3 was correlated with transcriptional up-regulation/down-regulation by trichostatin A in cancer cells, we evaluated the association of MBD1-MBD3 with p21WAF1/Cip1 and ErbB2 promoters in other lung cancer cell lines (H226 and H157). The association of MBD1 and MBD2 at both p21WAF1/Cip1 and ErbB2 promoters was not altered by trichostatin A in H226 and H157 cells (Fig. 3D). Together with the chromatin immunoprecipitation results from other cancer cell lines (Fig. 3A and B), these observations suggest the association of MBD1 and MBD2 at p21WAF1/Cip1 and ErbB2 promoters differed in cancer cell lines. In contrast, MBD3 was released from the p21WAF1/Cip1 promoter and it was recruited to the ErbB2 promoter by trichostatin A in these lung cancer cells. Together, the release/recruitment of MBD3 at two trichostatin A–responsive promoters (Fig. 3A,, B, and D) was in good agreement with up-regulation/down-regulation of these genes (Fig. 1A and B).

Additionally, the association of MBD1 with the p21WAF1/Cip1 and ErbB2 promoters was slightly increased, but the association of MBD2 was not altered at both promoters by trichostatin A in the immortalized BEAS-2B cells (Fig. 3D). The association of MBD3 increased at both p21WAF1/Cip1 and ErbB2 promoters by trichostatin A in BEAS-2B cells (Fig. 3D), suggesting that the association of MBD3 with these promoters differs between normal (recruitment of MBD3 to the p21WAF1/Cip1 promoter) and cancer (the recruitment of MBD3 to the ErbB2 promoter) lung cells (Fig. 3A and B). These results further support the possibility that the trichostatin A–induced transcriptional modulation of p21WAF1/Cip1 and ErbB2 genes (Fig. 1B; no change in p21WAF1/Cip1 expression and down-regulation of ErbB2 expression by trichostatin A in BEAS-2B cells) may be related with the association of MBD3 at these promoters (Fig. 3D; the recruitment of MBD3 to p21WAF1/Cip1 and ErbB2 promoters by trichostatin A). As the results of our p21WAF1/Cip1, ErbB2, and E-cadherin promoter analyses in lung normal, immortalized, and cancer cells indicate that release and recruitment of MBD3 are associated with transcriptional up-regulation and down-regulation, respectively, we propose that release and recruitment of MBD3 at the promoters is implicated in transcriptional derepression and repression.

MBD3 mediates trichostatin A–induced cancer-selective gene expression and cytotoxicity. Because MBD3 dissociates from the cancer-selective, up-regulated p21WAF1/Cip1 promoter and associates with the cancer-selective down-regulated ErbB2 promoter in response to trichostatin A (Fig. 3), we evaluated the effect of MBD3 silencing on trichostatin A–induced gene regulation and lung cancer cell death. Silencing of MBD3 by siRNA abrogated the trichostatin A–induced ErbB2 down-regulation, although it did not affect the p21WAF1/Cip1 up-regulation by trichostatin A (Fig. 4A). Additionally, the knockdown of MBD3 blocked growth inhibition in response to trichostatin A (Fig. 4B), suggesting that MBD3 is required for HDI-induced cytotoxicity of lung cancer cells. In A549 cells expressing the MBD3-specific siRNA, treatment with 0.5 μmol/L trichostatin A for 3 days induced only 45% growth inhibition compared with ∼80% growth inhibition in cells not expressing the MBD3-siRNA (Fig. 4B,, top). In H460 cells expressing the MBD3-siRNA, trichostatin A induced only 30% growth inhibition versus 80% in nonsilenced cells (Fig. 4B , bottom). Collectively, these results indicate that MBD3 is required for efficient trichostatin A–induced gene regulation and cell death in lung cancer cells.

Figure 4.

Abrogation of trichostatin A–induced gene regulation and cytotoxicity by silencing of MBD3 in lung cancer but not normal cells. A, transcriptional modulation of p21WAF1/Cip1 and ErbB2 genes by trichostatin A was abrogated by silencing of MBD3 expression in lung cancer cells. Following treating A549 cells with 0.5 μmol/L trichostatin A for 1 day, 24 hours after transfection with an empty pSUPER (si-vector) or MBD3 siRNA (si-MBD3), RT-PCR analysis was carried out to evaluate knockdown of MBD3 and the expression levels of p21WAF1/Cip1 and ErbB2 mRNA. B, effect of silencing of MBD3 on trichostatin A–induced cancer cell growth inhibition. Immunoblots for MBD3 in A549 (top) and H460 (bottom) cells transfected with empty pSUPER (mock) or MBD3 siRNA for 48 hours. Silencing of MBD3 was confirmed by immunoblotting with an anti-MBD3 antibody. Growth inhibition was evaluated by treating A549 (top) and H460 (bottom) cells expressing empty pSUPER or MBD3 siRNA vectors without (white columns) or with 0.2 μmol/L (striped columns) or 0.5 μmol/L (black columns) trichostatin A for 3 days before MTT assay. Columns, mean absorbance of three independent experiments, each in triplicate; bars, SD. **, P < 0.01, percentage of relative growth compared with no treatment (none). C, silencing of MBD3 expression had no effect on the transcriptional modulation of p21WAF1/Cip1 and ErbB2 genes by trichostatin A in lung normal (WI38) cells. Experiments in (A) were repeated with mRNA of WI38 (normal) and VA13 (cancer) cells treated with 0.5 μmol/L trichostatin A for 24 hours 1 day after transfection with an empty pSUPER or MBD3 siRNA. D, effect of silencing of MBD3 on the trichostatin A–induced growth inhibition of normal cells compared with cancer cells. Experiments in (B) were repeated for evaluating growth inhibition by trichostatin A treatment.

Figure 4.

Abrogation of trichostatin A–induced gene regulation and cytotoxicity by silencing of MBD3 in lung cancer but not normal cells. A, transcriptional modulation of p21WAF1/Cip1 and ErbB2 genes by trichostatin A was abrogated by silencing of MBD3 expression in lung cancer cells. Following treating A549 cells with 0.5 μmol/L trichostatin A for 1 day, 24 hours after transfection with an empty pSUPER (si-vector) or MBD3 siRNA (si-MBD3), RT-PCR analysis was carried out to evaluate knockdown of MBD3 and the expression levels of p21WAF1/Cip1 and ErbB2 mRNA. B, effect of silencing of MBD3 on trichostatin A–induced cancer cell growth inhibition. Immunoblots for MBD3 in A549 (top) and H460 (bottom) cells transfected with empty pSUPER (mock) or MBD3 siRNA for 48 hours. Silencing of MBD3 was confirmed by immunoblotting with an anti-MBD3 antibody. Growth inhibition was evaluated by treating A549 (top) and H460 (bottom) cells expressing empty pSUPER or MBD3 siRNA vectors without (white columns) or with 0.2 μmol/L (striped columns) or 0.5 μmol/L (black columns) trichostatin A for 3 days before MTT assay. Columns, mean absorbance of three independent experiments, each in triplicate; bars, SD. **, P < 0.01, percentage of relative growth compared with no treatment (none). C, silencing of MBD3 expression had no effect on the transcriptional modulation of p21WAF1/Cip1 and ErbB2 genes by trichostatin A in lung normal (WI38) cells. Experiments in (A) were repeated with mRNA of WI38 (normal) and VA13 (cancer) cells treated with 0.5 μmol/L trichostatin A for 24 hours 1 day after transfection with an empty pSUPER or MBD3 siRNA. D, effect of silencing of MBD3 on the trichostatin A–induced growth inhibition of normal cells compared with cancer cells. Experiments in (B) were repeated for evaluating growth inhibition by trichostatin A treatment.

Close modal

To determine whether MBD3 silencing in normal lung cells may affect the HDI-induced responses, we next assessed the expression of p21WAF1/Cip1 and ErbB2 genes and growth inhibition with or without trichostatin A following silencing MBD3 in normal (WI38) and its transformed derivative (VA13) lung cells. The expression of p21WAF1/Cip1 and ErbB2 genes was not affected by the MBD3 knockdown in WI38 cells without regard to trichostatin A treatment (Fig. 4C,, left). In VA13 cells expressing MBD3-siRNA, the ErbB2 down-regulation by trichostatin A (Fig. 1B) was not observed, whereas the p21WAF1/Cip1 expression (Fig. 1B) was not affected (Fig. 4C,, right), consistent with the results from other cancer cells (Fig. 4A). These results suggest that silencing of MBD3 by siRNA abrogate the trichostatin A–induced ErbB2 down-regulation in cancer cells but may not affect its expression in normal cells, further supporting the cancer-specific role of MBD3 in trichostatin A–mediated differential transcriptional regulation in normal and cancer cells. Additionally, silencing of MBD3 had little effect on the growth of trichostatin A–unresponsive normal WI38 cells before and after trichostatin A treatment (Fig. 4D,, top), whereas silencing of MBD3 in trichostatin A–responsive VA13 cells inhibited the trichostatin A–induced cytotoxicity by about half (from 48% to 18% by 0.2 μmol/L trichostatin A and from 82% to 38% by 0.5 μmol/L trichostatin A, following expressing the MBD3-siRNA; Fig. 4D , bottom), supporting that MBD3 is necessary for cancer-selective cytotoxicity induced by trichostatin A in lung cancer cells.

Finally, we used an immunofluorescence staining to examine the cellular localization of MBD3 in normal and cancer cells with and without trichostatin A treatment. In NHBE cells before trichostatin A treatment, MBD3 was located primarily in the cytoplasm, with some MBD3 proteins localizing to the nucleus (Fig. 5I). Following trichostatin A treatment, the majority of the MBD3 proteins were localized in the nucleus (Fig. 5J), suggesting that trichostatin A can induce nuclear localization of MBD3 in NHBE cells. In contrast, the majority of MBD3 proteins were located in the nucleus of BEAS-2B (Fig. 5K) and A549 and H460 cells (Fig. 5M and O) before trichostatin A treatment and then formed micronuclei and/or dispersed throughout the cells in response to trichostatin A in A549 and H460 cells (Fig. 5N and P) but not NHBE and BEAS-2B cells (Fig. 5J and L). These observations suggest that the cellular location of MBD3 differs between normal and cancer lung cells and may form the basis for their differential responses to trichostatin A. Overall, these findings indicate that the subcellular localization of MBD3 together with its localization on the specific promoters and the trichostatin A–induced transcriptional regulation may be important in the determination of cellular susceptibility to trichostatin A.

Figure 5.

Subcellular localization of MBD3 in NHBE, BEAS-2B, A549, and H460 cells. The cells were left untreated (none) or treated with trichostatin A (0.5 μmol/L) for 24 hours and stained with anti-MBD3 antibody (green; I-P) and 4′,6-diamidino-2-phenylindole for nuclei (blue; A-H). Q to X, overlaps of these images. Bar, 10 μm.

Figure 5.

Subcellular localization of MBD3 in NHBE, BEAS-2B, A549, and H460 cells. The cells were left untreated (none) or treated with trichostatin A (0.5 μmol/L) for 24 hours and stained with anti-MBD3 antibody (green; I-P) and 4′,6-diamidino-2-phenylindole for nuclei (blue; A-H). Q to X, overlaps of these images. Bar, 10 μm.

Close modal

Cancer cells exhibit genomic instability due to multiple genetic defects, so strategies targeting single genetic abnormalities are therapeutically ineffective. However, accumulating evidence has shown that cancer cells also suffer from epigenetic alterations that can be pivotal to carcinogenesis via inactivation of tumor suppressors and/or activation of oncogenes (16). Previous studies have identified drugs capable of targeting epigenetic regulation; some of these show cancer selectivity (lower cytotoxicity against normal cells), indicating that they may be promising candidates for anticancer therapeutics (3441). One subset of these drugs, the HDIs, induces histone acetylation, which can lead to reactivation of transcriptionally silenced genes or inactivation of transcriptionally hyperactive genes (2941). Here, we showed that HDIs exhibited cancer-selective cytotoxicity (Fig. 4D; data not shown) and cancer-specific transcriptional modulation (Table 1; Fig. 1).

Microarray analysis was used to identify differentially expressed genes that seem to be related with trichostatin A treatment of cancer cells, including tumor suppressors and oncogene, such as regulators of cell cycle, growth, and apoptosis (Table 1; Fig. 1). Thus, our data indicate that HDIs fulfill their cancer-selective roles through both transcriptional activation and silencing.

Our mechanistic analysis of cancer-selective growth inhibition by trichostatin A revealed that the growth inhibition effects were associated with reduced proliferation and apoptosis and occurred in a transcription-dependent manner (data not shown; Fig. 2). Trichostatin A–induced cancer-selective cell death was associated with up-regulation of p21WAF1/Cip1 and down-regulation of ErbB2 (Fig. 2), and overexpression of p21WAF1/Cip1 or silencing of ErbB2 inhibited cancer cell growth without regard to trichostatin A (Fig. 2). Therefore, our results indicate that the selective antitumor activities of HDIs seem to require transcriptional modulation of cancer-specific gene expressions, including up-regulation of p21WAF1/Cip1 and down-regulation of ErbB2.

Importantly, chromatin immunoprecipitation analyses of proteins released from or recruited to the p21WAF1/Cip1 and ErbB2 gene promoters revealed that trichostatin A treatment resulted in dissociation/association of MBD1-MBD3 at the activated/inactivated promoters, respectively (Fig. 3). Treatment of cancer cells with trichostatin A caused release of MBD3 from the p21WAF1/Cip1 promoter, whereas treatment of normal cells caused recruitment of MBD3 to this promoter, suggesting a reverse correlation between MBD3 binding and p21WAF1/Cip1 mRNA expression (Fig. 1). Consistent with this, we observed another reverse correlation between MBD3 association with ErbB2 promoter (Fig. 3) and reduced ErbB2 mRNA expression (Fig. 1) in cancer cells. On the other hand, both trichostatin A–induced up-regulation/down-regulation and protein recruitment differed between two cancer cell lines at the E-cadherin promoter (Fig. 3C), indicating that trichostatin A responses can differ in terms of MBD recruitment and transcriptional activation not only between cancer and normal cells but also between cancer types. However, our observations of MBD3 association/dissociation in trichostatin A–treated cells seem to indicate that MBD3 dissociates from the up-regulated promoter and associates with the down-regulated promoter following trichostatin A treatment. Here, our findings provide evidence that MBD3 may be another facet of the complex machinery influencing transcriptional status.

Unexpectedly, HDIs could induce histone acetylation at the promoters without regard to transcriptional expression. Repression of ErbB2 and E-cadherin expression in the trichostatin A–treated cancer cells was accompanied by increased histone acetylation (Fig. 3), suggesting that the open chromatin structure induced by HDAC inhibition may facilitate the access of transcriptional activators and repressors equally well; thus, binding of repressor (e.g., MBD3 and HDACs) to promoter may result in transcriptional repression. Thus, the nature of the complex formed at the promoter can be determined by the global chromatin context and a wide range of factors, such as histone modification profile, DNA methylation pattern, or even growth and differentiation stages.

The striking trichostatin A–induced recruitment of MBD3 to the ErbB2 promoter and its release from the p21WAF1/Cip1 promoter in cancer cells led us to speculate about the role of MBD3 in HDI-induced differential gene expression and cancer-selective death. Previous reports indicated that MBD1 and MBD2 preferentially bind to methylated DNA, whereas MBD3 preferentially binds unmethylated DNA (2628). Notably, HDIs modulate histone acetylation/deacetylation but not DNA methylation. Indeed, we found that silencing of MBD3 abrogated HDI-induced transcriptional reprogramming and growth inhibition in HDI-treated cancer cells (Fig. 4A and B) but not in normal cells (Fig. 4C and D). We further showed that MBD3, but not MBD1 and MBD2, was relocalized within cells in response to HDI treatment, and this localization differed between cancer and normal cells (Fig. 5). Thus, the relocation of MBD3 to the nucleus may facilitate MBD3 recruitment to the genome and allow MBD3 to function as a regulatory molecule. We additionally noted that MBD1 and MBD2 dissociated/associated differently in HDI-treated cancer and normal cells. Collectively, these results indicate that MBD3 likely plays a role in mediating the cancer-selective effects of HDI in transcriptional modulation and growth inhibition and that further work is warranted to examine whether silencing of MBD1 and MBD2 affects HDI-induced cancer-selective effects.

HDI-induced cancer-selective cytotoxicity is likely regulated by complex mechanisms. Our data indicate the importance of differential gene expression, protein localization, and histone acetylation/deacetylation. Moreover, HDIs are known to regulate additional proteins, such as α-tubulin and heat shock proteins, which may affect additional biochemical processes and physiologic phenomena. Nevertheless, our data indicate that the MBD proteins may play underlying roles in the nuclear/transcriptional effects of HDIs on cancer cells. Our observation that MBD3 silencing blocks HDI-induced cancer cell–specific gene regulation and selective cytotoxicity provides strong evidence for role of MBD3 in mediating cancer selectivity of HDIs. Thus, we have herein advanced our understanding of HDI-induced cancer-selective effects by exploring key molecular transcriptional changes in response to HDI treatment in lung cancer and normal cells. Our experiments revealed that MBD3 is differentially located in lung cancer and normal cells at both chromatin and cellular levels leading to differential transcription and cancer-selective cytotoxicity. In this way, our work provides new insights into the value of HDIs and their future potential in cancer treatment.

Grant support: National Cancer Center grant 0410040 and Korea Institute of Science and Technology Evaluation and Planning grant 2005-01112.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1
Jones PA, Baylin SB. The fundamental role of epigenetic events in cancer.
Nat Rev Genet
2002
;
3
:
415
–28.
2
Feinberg AP. The epigenetics of cancer etiology.
Semin Cancer Biol
2004
;
6
:
427
–32.
3
Feinberg AP, Tycko B. The history of cancer epigenetics.
Nat Rev Cancer
2004
;
4
:
143
–53.
4
Laird PW. Cancer epigenetics.
Hum Mol Genet
2005
;
14
Spec No1:
R65
–76.
5
Bird A. DNA methylation patterns and epigenetic memory.
Genes Dev
2002
;
16
:
6
–21.
6
Drummond DC, Noble CO, Kirpotin DB, et al. Clinical development of histone deacetylase inhibitors as anticancer agents.
Annu Rev Pharmacol Toxicol
2005
;
45
:
495
–528.
7
Yamagoe S, Kanno T, Kanno Y, et al. Interaction of histone acetylases and deacetylases in vivo.
Mol Cell Biol
2003
;
23
:
1025
–33.
8
Rountree MR, Bachman KE, Herman JG, Baylin SB. DNA methylation, chromatin inheritance, and cancer.
Oncogene
2001
;
20
:
3156
–65.
9
Ehrlich M. Expression of various genes is controlled by DNA methylation during mammalian development.
J Cell Biochem
2003
;
88
:
899
–910.
10
Narlikar GJ, Fan HY, Kingston RE. Cooperation between complexes that regulate chromatin structure and transcription.
Cell
2002
;
108
:
475
–87.
11
Struhl K. Histone acetylation and transcriptional regulatory mechanisms.
Genes Dev
1998
;
12
:
599
–606.
12
Milutinovic S, Zhuang Q, Szyf M. Proliferating cell nuclear antigen associates with histone deacetylase activity, integrating DNA replication and chromatin modification.
J Biol Chem
2002
;
277
:
20974
–8.
13
Verger A, Crossley M. Chromatin modifiers in transcription and DNA repair.
Cell Mol Life Sci
2004
;
61
:
2154
–62.
14
Smith MM. Centromeres and variant histones: what, where, when and why?
Curr Opin Cell Biol
2002
;
14
:
279
–85.
15
Bowen NJ, Fujita N, Kajita M, Wade PA. Mi-2/NuRD: multiple complexes for many purposes.
Biochim Biophys Acta
2004
;
1677
:
52
–7.
16
Tong JK, Hassig CA, Schnitzler GR, Kingston RE, Schreiber SL. Chromatin deacetylation by an ATP-dependent nucleosome remodelling complex.
Nature
1998
;
395
:
917
–21.
17
Zhang Y, LeRoy G, Seelig HP, Lane WS, Reinberg D. The dermatomyositis-specific autoantigen Mi2 is a component of a complex containing histone deacetylase and nucleosome remodeling activities.
Cell
1998
;
95
:
279
–89.
18
Rountree MR, Bachman KE, Baylin SB. DNMT1 binds HDAC2 and a new co-repressor, DMAP1, to form a complex at replication foci.
Nat Genet
2000
;
25
:
269
–77.
19
Hendrich B, Bird A. Identification and characterization of a family of mammalian methyl-CpG binding proteins.
Mol Cell Biol
1998
;
18
:
6538
–47.
20
Jiang CL, Jin SG, Pfeifer GP. MBD3L1 is a transcriptional repressor that interacts with methyl-CpG-binding protein 2 (MBD2) and components of the NuRD complex.
J Biol Chem
2004
;
279
:
52456
–64.
21
Jones PL, Veenstra GJ, Wade PA, et al. Methylated DNA and MeCP2 recruit histone deacetylase to repress transcription.
Nat Genet
1998
;
19
:
187
–91.
22
Hendrich B, Tweedie S. The methyl-CpG binding domain and the evolving role of DNA methylation in animals.
Trends Genet
2003
;
19
:
269
–77.
23
Roloff TC, Ropers HH, Nuber UA. Comparative study of methyl-CpG-binding domain proteins.
BMC Genomics
2003
;
4
:
1
.
24
Ng HH, Zhang Y, Hendrich B, et al. MBD2 is a transcriptional repressor belonging to the MeCP1 histone deacetylase complex.
Nat Genet
1999
;
23
:
58
–61.
25
Feng Q, Zhang Y. The MeCP1 complex represses transcription through preferential binding, remodeling, and deacetylating methylated nucleosomes.
Genes Dev
2001
;
15
:
827
–32.
26
Brackertz M, Boeke J, Zhang R, Renkawitz R. Two highly related p66 proteins comprise a new family of potent transcriptional repressors interacting with MBD2 and MBD3.
J Biol Chem
2002
;
277
:
40958
–66.
27
Saito M, Ishikawa F. The mCpG-binding domain of human MBD3 does not bind to mCpG but interacts with NuRD/Mi2 components HDAC1 and MTA2.
J Biol Chem
2002
;
277
:
35434
–9.
28
Fraga MF, Ballestar E, Montoya G, Taysavang P, Wade PA, Esteller M. The affinity of different MBD proteins for a specific methylated locus depends on their intrinsic binding properties.
Nucleic Acids Res
2003
;
31
:
1765
–74.
29
Patra SK, Patra A, Dahiya R. Histone deacetylase and DNA methyltransferase in human prostate cancer.
Biochem Biophys Res Commun
2001
;
287
:
705
–13.
30
Fahrner JA, Eguchi S, Herman JG, Baylin SB. Dependence of histone modifications and gene expression on DNA hypermethylation in cancer.
Cancer Res
2002
;
62
:
7213
–8.
31
Cameron EE, Bachman KE, Myohanen S, Herman JG, Baylin SB. Synergy of demethylation and histone deacetylase inhibition in the re-expression of genes silenced in cancer.
Nat Genet
1999
;
21
:
103
–7.
32
Strahl BD, Allis CD. Core histones are subjected to a variety of post-translational modifications, including acetylation and methylation.
Nature
2000
;
403
:
41
–5.
33
Goffin J, Eisenhauer E. DNA methyltransferase inhibitors—state of the art.
Ann Oncol
2002
;
13
:
1699
–716.
34
Ungerstedt JS, Sowa Y, Xu WS, et al. Role of thioredoxin in the response of normal and transformed cells to histone deacetylase inhibitors.
Proc Natl Acad Sci U S A
2005
;
102
:
673
–8.
35
Marks PA, Richon VM, Miller T, et al. Histone deacetylase inhibitors.
Adv Cancer Res
2004
;
91
:
137
–68.
36
Marks PA, Richon VM, Breslow R, Rifkind RA. Histone deacetylase inhibitors as new cancer drugs.
Curr Opin Oncol
2001
;
13
:
477
–83.
37
Jaboin J, Wild J, Hamidi H, et al. MS-27-275, an inhibitor of histone deacetylase, has marked in vitro and in vivo antitumor activity against pediatric solid tumors.
Cancer Res
2002
;
62
:
6108
–15.
38
Cheong JW, Chong SY, Kim JY, et al. Induction of apoptosis by apicidin, a histone deacetylase inhibitor, via the activation of mitochondria-dependent caspase cascades in human Bcr-Abl-positive leukemia cells.
Clin Cancer Res
2003
;
9
:
5018
–27.
39
Vanhaecke T, Papeleu P, Elaut G, Rogiers V. Trichostatin A-like hydroxamate histone deacetylase inhibitors as therapeutic agents: toxicological point of view.
Curr Med Chem
2004
;
11
:
1629
–43.
40
Atadja P, Gao L, Kwon P, et al. Selective growth inhibition of tumor cells by a novel histone deacetylase inhibitor, NVP-LAQ824.
Cancer Res
2004
;
64
:
689
–95.
41
Burgess A, Ruefli A, Beamish H, et al. Histone deacetylase inhibitors specifically kill nonproliferating tumour cells.
Oncogene
2004
;
23
:
6693
–701.
42
Jang ER, Lim SJ, Lee ES, et al. The histone deacetylase inhibitor trichostatin A sensitizes estrogen receptor α-negative breast cancer cells to tamoxifen.
Oncogene
2004
;
23
:
1724
–36.
43
Noh EJ, Lee JS. Functional interplay between modulation of histone deacetylase activity and its regulatory role in G2-M transition.
Biochem Biophys Res Commun
2003
;
310
:
267
–73.
44
Jang ER, Lee JH, Lim DS, et al. Analysis of ATM- and NBS-regulated gene expression patterns.
J Cancer Res Clin Oncol
2004
;
130
:
225
–34.
45
Cappuzzo F, Varella-Garcia M, Shigematsu H, et al. Increased HER2 gene copy number is associated with response to gefitinib therapy in epidermal growth factor receptor-positive non-small-cell lung cancer patients.
J Clin Oncol
2005
;
23
:
5007
–18.