Cancer-Type Regulation of MIG-6 Expression by Inhibitors of Methylation and Histone Deacetylation

Epigenetic silencing is one of the mechanisms leading to inactivation of a tumor suppressor gene, either by DNA methylation or histone modification in a promoter regulatory region. Mitogen inducible gene 6 (MIG-6), mainly known as a negative feedback inhibitor of the epidermal growth factor receptor (EGFR) family, is a tumor suppressor gene that is associated with many human cancers. To determine if MIG-6 is inactivated by epigenetic alteration, we identified a group of human lung cancer and melanoma cell lines in which its expression is either low or undetectable and studied the effects of methylation and of histone deacetylation on its expression. The DNA methyltransferase (DNMT) inhibitor 5-aza-2′-deoxycytidine (5-aza-dC) induced MIG-6 expression in melanoma cell lines but little in lung cancer lines. By contrast, the histone deacetylase (HDAC) inhibitor trichostatin A (TSA) induced MIG-6 expression in lung cancer lines but had little effect in melanoma lines. However, the MIG-6 promoter itself did not appear to be directly affected by either methylation or histone deacetylation, indicating an indirect regulatory mechanism. Luciferase reporter assays revealed that a short segment of exon 1 in the MIG-6 gene is responsible for TSA response in the lung cancer cells; thus, the MIG-6 gene can be epigenetically silenced through an indirect mechanism without having a physical alteration in its promoter. Furthermore, our data also suggest that MIG-6 gene expression is differentially regulated in lung cancer and melanoma.


Introduction
Mitogen inducible gene 6 (MIG-6) (also known as gene 33, ERRFI1, or RALT) is an immediate early response gene that is expressed in various tissues and plays a critical role in many pathophysiological states [1]. Its expression can be induced by a broad spectrum of growth factors, hormones, or stress stimuli, and it is associated with various chronic conditions [1,2]. Studies in mice have revealed that Mig-6 is required for skin morphogenesis and lung development and that it plays an important role in maintaining joint homeostasis [3,4,5].
As a cytoplasmic scaffolding adaptor, MIG-6 has several important protein-protein interaction motifs that may mediate interaction with signaling molecules downstream of receptor tyrosine kinases (RTKs) [2]. One of the most prominent roles of MIG-6 in regulating signal transduction comes from its ability to directly interact with epidermal growth factor receptor (EGFR) and other ErbB family members, inhibiting their phosphorylation and downstream signaling in a negative feedback fashion [6,7,8,9]. MIG-6 can be induced by hepatocyte growth factor (HGF) and functions as a negative feedback regulator of HGF-MET signaling [10,11], indicating that it has broad role as a signal checkpoint for modulating activated RTK pathways in a timely manner.
The evidence that MIG-6 is a tumor suppressor gene is compelling. It is located in chromosome 1p36, a locus that frequently has loss of heterozygosity in several human cancers including lung cancer [12,13,14], melanoma [15], and breast cancer [16]. Indeed, down-regulation or loss of MIG-6 expression has been reported in cancers and is often associated with poor prognosis [3,11,16,17,18,19,20,21,22,23]. MIG-6 down-regulation in non-small cell lung cancer (NSCLC) is associated with increased EGFR signaling and poorly differentiated cancer [21], while loss of its expression in ErbB2-amplified breast carcinoma renders the cancer cells more resistant to Herceptin, the neutralizing antibody against ErbB2 [16]. In glioblastoma, MIG-6 is identified as a single gene within the most commonly deleted region at the 1p36. 23 locus, and its expression is down-regulated in 34% of glioblastoma samples [19]. While MIG-6 down-regulation is reported in a high percentage of papillary thyroid cancers [22], high MIG-6 expression correlates with longer survival and is associated with favorable surgical outcomes for those patients [24]. Decreased MIG-6 expression has also been reported in skin cancer, endometrial cancer, and hepatocellular carcinomas [3,20,23]. Moreover, even though such events are rare, three mutations in the MIG-6 gene have been identified in human lung cancer and one in neuroblastoma [11,18]. Further evidence supporting MIG-6 as a tumor suppressor gene arose from mouse studies; Mig-6deficient mice are prone to develop epithelial hyperplasia or tumors in organs including the lung, skin, uterus, gallbladder, and bile duct [3,11,20].
Epigenetic alteration, one of the most well-known mechanisms leading to inactivation of a tumor suppressor gene [25], can result from DNA methylation or histone deacetylation in the gene's promoter [25]. Given that down-regulation of MIG-6 is frequently observed in many human cancers, we asked whether MIG-6 expression was affected by DNA methylation and histone deacetylation. Here, we show that the MIG-6 promoter itself is neither hypermethylated nor affected by histone deacetylation. However, its expression is induced by the DNA methyltransferase (DNMT) inhibitor 5-aza-29-deoxycytidine (5-aza-dC) in melanoma cell lines and by the histone deacetylase (HDAC) inhibitor trichostatin A (TSA) in lung cancer lines. By dissecting its promoter regulatory region using a luciferase reporter assay, we identified a minimal TSA-response element in exon 1 of MIG-6 that is essential for its induction by TSA in lung cancer cells.

Results
MIG-6 expression is differentially regulated by 5-aza-dC in melanoma cell lines and TSA in lung cancer cell lines To determine whether MIG-6 expression was affected by epigenetic alteration, we first identified human cancer cell lines in which its promoter is likely affected by methylation or histone deacetylation. As shown in Figure 1, we found four human NSCLC cell lines (A427, H226, H522, and H596) and five melanoma cell lines (M14, MALME-3M, SK-2, SK-MEL-28, and UACC-257) in which MIG-6 protein was either low or undetectable. We then treated these cell lines with or without 5-aza-dC, TSA, or a combination of both inhibitors.
To our surprise, we found that TSA treatment significantly increased the amount of MIG-6 protein in the lung cancer cell lines, but not in the melanoma lines ( Figure 2A). In contrast, 5aza-dC treatment significantly increased the MIG-6 protein in the melanoma cell lines, but not in the NSCLC lung cancer lines ( Figure 2B). To determine if the increase of MIG-6 protein was regulated at transcriptional level, we performed RT-PCR analysis. As shown in Figure 3, and consistent with protein expression, MIG-6 mRNA expression increased with TSA treatment only in the four lung cancer cell lines, and it increased with 5-aza-dC treatment only in the five melanoma lines. These data strongly suggest that the induction of MIG-6 expression by 5-aza-dC or TSA is regulated at the transcriptional level and is differentially regulated in the lung cancer and melanoma cells.
The MIG-6 promoter is neither hypermethylated nor directly affected by histone deacetylation Given that MIG-6 expression was induced by 5-aza-dC in the melanoma lines, we asked if its promoter was hypermethylated in those cells. We extracted genomic DNA from both lung cancer and melanoma cell lines and examined DNA methylation in the 596-bp MIG-6 promoter regulatory region, which contains abundant CpG sites ( Figure 4). To our surprise, the lung cancer cell lines ( Figure 4A) and the melanoma cell lines ( Figure 4B) were similar in having very few methylated CpG sites in the MIG-6 promoter regulatory region, indicating that induction of MIG-6 by 5-aza-dC in melanoma was independent of DNA methylation in its promoter. These results were confirmed by direct sequencing of the PCR products amplified from bisulfite-treated DNAs (data not shown).
Similarly, we asked if the MIG-6 promoter was influenced by histone deacetylation. By chromatin immunoprecipitation assay (ChIP), we found that TSA treatment did not increase the binding of acetyl-histone H3 to the MIG-6 promoter in the lung cancer lines or in the melanoma lines ( Figure 5), indicating that the MIG-6 promoter was not directly affected by histone deacetylation either.
MIG-6 transcription is indirectly regulated by a factor(s) that is affected by methylation or histone deacetylation Because the above data suggest that MIG-6 induction is not directly regulated, we looked for a secondary mechanism, with the inhibitors inducing expression of a transcription factor(s) or cofactor(s) that in turn regulates MIG-6 expression. Thus, we examined the responses of the MIG-6 promoter regulatory region to the inhibitors via luciferase reporter assay. A MIG-6 promoter reporter plasmid was constructed by inserting a 1.383-kb genomic DNA fragment (consisting of the MIG-6 promoter, its upstream regulatory region, and the downstream exon 1 and part of intron 1) in front of a luciferase reporter gene. Testing the reporter in both lung cancer and melanoma cell lines, we found that TSA significantly enhanced MIG-6 promoter activity in lung cancer cells but showed no such effect in melanoma cells ( Figure 6). This data was consistent with our prior western blot and RT-PCR analyses. 5-aza-dC, however, appeared to have no effect on reporter activity in either the melanoma or lung cancer lines ( Figure 6). These data indicate that while the TSA-responsive element is within the 1.383-kb region of MIG-6, the 5-aza-dCresponsive element is likely outside this region.
A small segment of exon 1 in MIG-6 is essential for TSA response in lung cancer We performed a series of deletion analyses in the 1.383-kb MIG-6 promoter regulatory region to determine the minimal region required for induction by TSA in lung cancer cells. Deletion from the 59-terminus to the proximal region of the MIG-6 promoter resulted in a decrease of the basal promoter activity, while the response to TSA was essentially retained ( Figure 7A). Deletion from the 39-terminus to the transcriptional initiation site of MIG-6 resulted in complete loss of response to TSA ( Figure 7B), indicating that the TSA response element was downstream of the MIG-6 promoter. Further deletion analyses revealed a small segment in the exon 1 starting from the transcriptional initiation site that was essential for TSA responsiveness ( Figure 7C). As summarized in Figure 7D, the minimal TSA response element is within the first 50 nucleotides of exon 1, with the distal 20nucleotide segment showing the highest activity.
We speculated that there exists a critical transcription factor binding motif in the minimal TSA response element. We performed mutation analyses of the 50-nucleotide segment to pinpoint potential transcription factor binding motif(s) ( Figure 8A). Compared with the wild-type P(-76/+50) reporter, mutation in the m4 and m5 elements resulted in a significant decrease of reporter activity in response to TSA, while mutation in other elements had lesser effect ( Figure 8B). This result agrees with the deletion analyses, as the m4 and m5 elements are within the distal 20nucleotide segment that, when deleted, resulted in a steep drop-off in TSA response. We generated another mutant reporter m11 in which half of the sequences in both m4 and m5 were mutated ( Figure 8A). We found that the m11 mutant had a much greater reduction in TSA response ( Figure 8C), indicating that those sequences are essential for the binding of a yet to be identified transcription factor which regulates MIG-6 gene expression induced by TSA in the lung cancer.   Other genes differentially regulated by 5-aza-dC and TSA in lung cancer and melanoma cells We next asked whether there were other genes differentially regulated by 5-aza-dC and TSA in lung cancer and melanoma cells. We performed DNA microarray analyses on samples derived from A427 lung cancer and M14 melanoma cells treated with 5aza-dC and/or TSA. Figure 9A shows the genes displaying an expression pattern similar to that of MIG-6 (which is indicated as ERRFI1 in the heat-map) in response to either 5-aza-dC or TSA treatment ( Figure 9A). Another group of genes appeared to be down-regulated, the opposite of MIG-6 expression ( Figure 9B).
Among the up-regulated genes were those coding for transcription factors such as EGR1 and STAT1, the MIG-6-inducible gene HBEGF, and genes coding for histone proteins ( Figure 9A). Even though those genes were differentially expressed in A427 and M14 cells ( Figure 9A), further analyses revealed that EGR1 (but not several other genes we examined) displayed an expression pattern similar to that of MIG-6 across the four lung cancer cell lines and five melanoma lines ( Figure 10). Thus, MIG-6 was not the only gene differentially regulated in the lung cancer and melanoma cells. Perhaps there are tissue-specific factors (either transcription factors or transcription factor co-activators/co-repressors) that respond differently to 5-aza-dC and TSA, leading to differential induction of MIG-6 and EGR1 in lung cancer and melanoma cells.

Discussion
MIG-6, a tumor suppressor gene, has been found downregulated in many human cancers. To determine if downregulation of MIG-6 expression was affected by epigenetic modification in its promoter, we treated lung cancer and melanoma cell lines with inhibitors of methylation and histone deacetylation and then determined how those inhibitors influenced MIG-6 expression. Intriguingly, we found that DNMT inhibitor 5aza-dC specifically induced MIG-6 expression in melanoma cells but not in lung cancer cells, while the HDAC inhibitor TSA induced the reverse pattern (Figure 2 and 3). Despite both inductions being regulated at transcriptional level, we were surprised to find that the MIG-6 promoter was neither hypermethylated nor directly affected by histone deacetylation (Figure 4 and 5), indicating that an indirect mechanism might be responsible for differential induction. In fact, 5-aza-dC has also been reported to induce the expression of several other genes whose promoters are not directly affected by methylation in leukemia cells [26], suggesting that 5-aza-dC might have a broader influence on regulating gene expression via a methylation-independent manner.
Many DNMT inhibitors and HDAC inhibitors are currently in clinical trials for their anti-cancer properties [27,28,29]. Even though most of these epigenetic drugs are still in early development and the prospects for them to be used clinically for cancer treatment remain to be evaluated, that evaluation will depend on our understanding of how they work and what outcomes might be expected. 5-Aza-dC and TSA are viewed as potent and specific inhibitors for methylation and histone deacetylation, respectively [27,28,30], and they have been widely used for investigating epigenetic alteration of many tumor suppressor genes. These inhibitors usually cause global changes in gene expression by remodeling chromatin via directly converting methylated DNA to unmethylated DNA or unacetylated histones to the acetylated state, thereby allowing easy access of the transcription machinery to gene promoters. However, some inhibitors might be doing more, and their anti-cancer properties could be much more complicated. For instance, many non-histone cellular proteins such as transcription factors are also substrates of HDAC, and their transcriptional activities could be affected by the HDAC inhibitor TSA as well [29].
Most tumor suppressor genes are epigenetically silenced by either DNA methylation and/or histone deacetylation in their promoters [25]. To our knowledge, there is no report showing that the expression of such genes can be differentially regulated by inhibitors of methylation or histone deacetylation in a cancerspecific fashion without having epigenetic modifications in the promoter. The regulation of MIG-6 by these inhibitors, as we show here, unveils a novel mechanism by which a tumor suppressor gene can be epigenetically silenced in an indirect and tissuespecific manner. Our luciferase reporter assay results indicated that the regulation of MIG-6 expression in melanoma and in lung cancer was most likely mediated by different factors. We have identified a minimal TSA response element in exon 1 of MIG-6 proximal to its promoter (Figure 7 and 8), while location of the 5aza-dC response element is still uncertain ( Figure 6).
We speculate that the TSA response element in the MIG-6 gene is most likely regulated by a factor whose expression is affected by histone deacetylation in its promoter or whose protein activity is directly regulated by acetylation/deacetylation ( Figure 11A). This factor would be activated in lung cancer cells upon TSA treatment, but not in melanoma cells. Within the minimal TSAresponse element that we identified in MIG-6 gene exon 1 (Figure 7), there are putative DNA binding sequences for the transcription factor activator protein-2 (TFAP2), which has five family members and binds to the consensus sequence 59-GCCNNNGGC-39 [31,32]. When the putative TFAP2 binding sites were mutated, we observed a significant drop in TSAresponsiveness (Figure 8), indicating that those sequences are crucial for TSA-mediated regulation. It will be interesting to see if TFAP2 or other factor(s) binds to those sequences and regulates MIG-6 gene expression. As for 5-aza-dC, its response element is likely outside the tested 1.383-kb MIG-6 promoter regulatory region ( Figure 6); that is, it is either directly affected by methylation in its DNA sequences or is indirectly mediated by another transcriptional regulator whose promoter is modified by methylation in melanoma cells ( Figure 11B). Extensive studies will be required to determine what those factors are and how they control MIG-6 expression. Cancer-type regulation of gene expression by inhibitors of methylation and histone deacetylation is not unique to MIG-6. Other genes such as EGR1 [33] are also differentially regulated in lung cancer and melanoma cells by those inhibitors (see Figures 9 and 10). It remains to be determined whether-like the MIG-6 promoter-the EGR1 promoter is neither hypermethylated nor affected by histone deacetylation in those cells. If these character- istics are the same in the two promoters, it will be interesting to see if they are regulated by same factor(s) or via different mechanisms.
We report here that MIG-6 expression is differentially regulated by inhibitors of methylation and histone deacetylation in lung cancer and melanoma cells without physical epigenetic alterations in its promoter. MIG-6 (and possibly EGR1) may serve as valuable biomarkers for determining the sensitivity/suitability of a cancer type for treatment with DNMT and/or HDAC inhibitors in the clinic.

Human Cell Lines
The human lung cancer cell lines A427, NCI-H292, NCI-H2122, NCI-H596, and SK-MES-1 were obtained from American Type Culture Collection (Manassas, VA). EBC-1 was from the Health Science Research Resources Bank (Tokyo, Japan).

Western Blot Analysis
To prepare protein lysates, cells were treated with or without 5aza-dC (10 mM) for 3 d or TSA (1 mM) for 1 d in complete growth medium. For the combination of the two, cells were treated with 5aza-dC for 2 d, followed by TSA treatment for 1 d. Whole cell lysates were extracted in RIPA buffer (20 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% NP 40, 0.5% deoxycholate, 0.1% SDS, 1 mM EDTA, 50 mM NaF, 1 mM sodium orthovanadate, and Complete Proteinase Inhibitor Cocktail tablets), quantified using a DC Protein Assay Kit (Bio-Rad, Hercules, CA), and adjusted with equal volume of 26 Laemmli Sample Buffer (Sigma, Saint Louis, MO). The protein was run in a 10% Tris-glycine gel (Invitrogen, Grand Island, NY), transferred onto a PVDF membrane, and probed with anti-MIG-6 [11] or anti-b-actin antibody (Sigma).

Bisulfite DNA Sequencing
Genomic DNAs isolated from each cell line were subjected to sodium bisulfite conversion of unmethylated cytosine in DNA using EpiTect Bisulfite Kit (QIAGEN, Valencia, CA) according to the manufacturer's protocol. Bisulfite-treated DNAs were then used for nested PCR using following primers: 59-GGAGAG-GAAAAAATATATAATTTTGTT-39 (forward primer p1) and 59-TCTCCCTCCATCCCAAAAACTC-39 (reverse primer p5) for the first round of PCR, and for the second round, forward primer p1 and 59-TAACCCTCCCCACCCCCTCAAC-39 (reverse primer p4). The second round PCR products (596 bp) were then cloned into the TOPO TA Cloning vector (Invitrogen). For each cell line, 10 different colonies were picked and the inserts were sequenced. Direct sequencing of the second round PCR products was also performed. The MethPrimer program (www. urogene.org/methprimer) was used for analyzing CpG islands in the MIG-6 gene.

Chromatin Immunoprecipitation (ChIP) Assay
The ChIP assay was performed using a Chromatin Immunoprecipitation Assay Kit (Upstate, Lake Placid, NY) according to the manufacturer's instructions. Briefly, cells were treated with or without TSA (1 mM) for overnight, and then histones in the cells were cross-linked to DNA by 1% formaldehyde treatment for 10 min. After crosslinking, cells were collected, resuspended in SDS lysis buffer, and sonicated to shear the DNA. The lysates were centrifuged to clear debris, and the supernatants were collected and diluted with 10-fold ChIP Dilution Buffer containing protease inhibitor cocktail (Roche, Indianapolis, IN). The diluted supernatants were then incubated with either anti-acetyl histone H3 antibody (Upstate) or normal rabbit serum (as negative control) overnight at 4uC, followed by a 1 h incubation with Protein A agarose/salmon sperm DNA for immunoprecipitation. After extensive washing, the precipitated antibody/histone/DNA complex was eluted in freshly prepared elution buffer (1% SDS, 0.1 M NaHCO 3 ), and the histone-DNA crosslinks were reversed by heating at 65uC for 4 h. Precipitated DNAs were recovered by phenol/chloroform extraction and ethanol precipitation and then stored at 220uC until used for PCR. Primers used for PCR were as follows: for the MIG-6 gene promoter, 59-AGACGCCTCTCCGGGAGAC-39 (forward) and 59-ATAGGCCGCCGGGCCGTGA-39 (reverse); and for the GAPDH gene promoter, 59-TCGGTGCGTGCCCAGTT-GAACC-39 (forward) and 59-ATGCGGCTGACTGTCGAA-CAGG-39 (reverse).

Microarray Analysis
Total RNA was isolated from A427 and M14 cells treated with 5-aza-dC and/or TSA and from controls, as described above. Agilent 44K One-Color Microarrays (Agilent, Santa Clara, CA) were used for detection and analyses of gene expression.

Luciferase Reporter Assay
The day before transfection, 5610 4 cells were seeded in each well of a 96-well plate. The firefly luciferase reporter plasmid pGL3-luc or its derivative (100 ng) was co-transfected with pU6B-Renilla reporter (5 ng) using Lipofectamine 2000 (Invitrogen) overnight. The cells were then treated with 5-aza-dC (10 mM) for 2 d or TSA (1 mM) for 1 d. The luciferase reporter activities were assayed using the Dual-Luciferase Reporter Assay System (Promega) and measured using an EnVision 2104 Multilabel Reader (PerkinElmer, Waltham, MA). The firefly luciferase activities were normalized for analyses using Renilla luciferase activities. Figure 11. Working models for epigenetic regulation of MIG-6 expression by 5-aza-dC and TSA. (A) In lung cancer cells, TSA might indirectly regulate MIG-6 through two mechanisms. The first could be that TSA inhibits histone deacetylation in the promoter of gene X, resulting increased production of X protein which enhances MIG-6 gene expression by associating with the TSA-response element in exon 1. The second might be that direct acetylation of protein X influenced by TSA results in increased transcription of MIG-6. (B) In melanoma cells, the regulation of MIG-6 expression by 5-aza-dC might be direct or indirect. 5-Aza-dC might inhibit DNA methylation in the promoter of gene Y, leading to up-regulation of its product and thus indirectly enhancing MIG-6 expression; or it might directly inhibit DNA methylation outside the tested 1.383-kb MIG-6 promoter regulatory region, allowing easy access of a transcriptional co-activator to enhance MIG-6 expression. ''P'' indicates the promoter of each gene; ''E1'' indicates exon 1 of the MIG-6 gene. doi:10.1371/journal.pone.0038955.g011