Distinct effects on gene expression of chemical and genetic manipulation of the cancer epigenome revealed by a multimodality approach.

We tested the hypothesis that the effects on gene expression of altered DNA methylation by 5-aza-2'-deoxycytidine (5-aza-CdR) and genetic (DNMT knockout) manipulation of DNA are similar, and distinct from Trichostatin A (TSA)-induced chromatin decondensation. Surprisingly, the effects of 5-aza-CdR were more similar to those of TSA than to DNMT1, DNMT3B, or double DNMT somatic cell knockout. Furthermore, the effects of 5-aza-CdR were similar at one and five days exposure, suggesting active demethylation or direct influence of both drugs on the stability of methylation and/or chromatin marks. Agents that induce gene activation through hypomethylation may have unintended consequences, since nearly as many genes were downregulated as upregulated after demethylation. In addition, a 75 kb cluster of metallothionein genes was coordinately regulated.

DNMT1 as a molecular target in a multimodality-resistant phenotype in tumor cells.

We have previously shown that hydrogen peroxide-resistant permanent (OC-14) cells are resistant to the cytotoxicity of several exogenous oxidative and anticancer agents including H(2)O(2), etoposide, and cisplatin; and we refer to this process as an oxidative multimodality-resistant phenotype (MMRP). Furthermore, OC-14 cells contain increased activator protein 1 activity, and inhibition of activator protein 1 reversed the MMRP. In this study, we show that permanent Rat-1 cell lines genetically altered to overexpress c-Fos also displayed a similar MMRP to H(2)O(2), etoposide, and cisplatin as OC-14 cells. Gene expression analysis of the OC-14 cells and c-Fos-overexpressing cells showed increased DNMT1 expression. Where OC-14 and c-Fos-overexpressing cells were exposed to 5-aza-2'-deoxycytidine, which inhibits DNMT activity, a significant but incomplete reversal of the MMRP was observed. Thus, it seems logical to suggest that DNMT1 might be at least one target in the MMRP. Rat-1 cells genetically altered to overexpress DNMT1 were also shown to be resistant to the cytotoxicity of H(2)O(2), etoposide, and cisplatin. Finally, somatic HCT116 knockout cells that do not express either DNMT1 (DNMT1(-/-)) or DNMT3B (DNMT3B(-/-)) were shown to be more sensitive to the cytotoxicity of H(2)O(2), etoposide, and cisplatin compared with control HCT116 cells. This work is the first example of a role for the epigenome in tumor cell resistance to the cytotoxicity of exogenous oxidative (H(2)O(2)) or systemic (etoposide and cisplatin) agents and highlights a potential role for DNMT1 as a potential molecular target in cancer therapy.

Thioredoxin reductase as a potential molecular target for anticancer agents that induce oxidative stress.

Redox-sensitive signaling factors regulate multiple cellular processes, including proliferation, cell cycle, and prosurvival signaling cascades, suggesting their potential as molecular targets for anticancer agents. It is logical to set constraints that a molecular target should meet at least one of the following criteria: (1) inhibition of prosurvival signaling pathways; (2) inhibition of cell cycle progression; or (3) enhancement of the cytotoxic effects of anticancer agents. Therefore, we hypothesized that thioredoxin reductase 1 (TR), a component of several redox-regulated pathways, might represent a potential molecular target candidate in response to agents that induce oxidative stress. To address this issue, permanent cell lines overexpressing either the wild-type (pCXN2-myc-TR-wt) or a Cys-Ser mutant (pCXN2-myc-mTR) TR gene were used, as were parental HeLa cells treated with 1-methyl-1-propyl-2-imidazolyl disulfide (IV-2), a pharmacologic inhibitor of TR. Cells were exposed to the oxidative stressors, H2O2 and ionizing radiation (IR), and analyzed for changes in signal transduction, cell cycle, and cytotoxicity. Analysis of HeLa cells overexpressing the pCXN2-myc-TR-wt gene showed increased basal activity of nuclear factor kappaB (NFkappaB) and activator protein (AP-1), whereas HeLa cells expressing a pCXN2-myc-mTR gene and HeLa cells treated with IV-2 were unable to induce NFkappaB or AP-1 activity following H2O2 or IR exposure. Fluorescence-activated cell sorting analysis showed a marked accumulation of pCXN2-myc-mTR cells in the late G1 phase, whereas pCXN2-myc-TR-wt cells showed a decreased G1 subpopulation. Chemical inhibition of TR with IV-2 also completely inhibited cellular proliferation at concentrations between 10 and 25 micromol/L, resulting in a G1 phase cell cycle arrest consistent with the results from cells expressing the pCXN2-myc-mTR gene. Following exposure to H2O2 and IR, pCXN2-myc-mTR- and IV-2-treated cells were significantly more sensitive to oxidative stress-induced cytotoxicity as measured by clonogenic survival assays. Finally, IV-2-treated cells showed increased tumor cell death when treated with H2O2 and IR. These results identify TR as a potential target to enhance the cytotoxic effects of agents that induce oxidative stress, including IR.

Geldanamycin and 17-allylamino-17-demethoxygeldanamycin potentiate the in vitro and in vivo radiation response of cervical tumor cells via the heat shock protein 90-mediated intracellular signaling and cytotoxicity.

Ansamycin antibiotics inhibit function of the heat shock protein (HSP) 90, causing selective degradation of several intracellular proteins regulating such processes as proliferation, cell cycle regulation, and prosurvival signaling cascades. HSP90 has been identified previously as a molecular target for anticancer agents, including ionizing radiation (IR). Therefore, we hypothesized that the ansamycin geldanamycin and its 17-allylamino-17-demethoxy analog (17-AAG), which inhibit HSP90, would enhance tumor cell susceptibility to the cytotoxicity of IR. Treatment of two human cervical carcinoma cell lines (HeLa and SiHa) with geldanamycin and 17-AAG resulted in cytotoxicity and, when combined with IR, enhanced the radiation response, each effect with a temporal range from 6 to 48 h after drug exposure. In addition, mouse in vivo models using 17-AAG at clinically achievable concentrations yielded results that paralleled the in vitro radiosensitization studies of both single and fractioned courses of irradiation. The increase in IR-induced cell death appears to be attributable to a combination of both programmed and nonprogrammed cell death. We also measured total levels of several prosurvival and apoptotic signaling proteins. Akt1, extracellular signal-regulated kinase-1, Glut-1, HER-2/neu, Lyn, cAMP-dependent protein kinase, Raf-1, and vascular endothelial growth factor expression were down-regulated in 17-AAG-treated cells, identifying these factors as molecular markers and potential therapeutic targets. Finally, a series of immortalized and human papillomavirus-transformed cell lines were used to demonstrate that the radiosensitizing effects of 17-AAG were limited to transformed cells, suggesting a possible differential cytotoxic effect. This work shows that altered HSP90 function induces significant tumor cytotoxicity and radiosensitization, suggesting a potential therapeutic utility.

Simultaneous inhibition of hsp 90 and the proteasome promotes protein ubiquitination, causes endoplasmic reticulum-derived cytosolic vacuolization, and enhances antitumor activity.

The ansamycin antibiotic, geldanamycin, targets the hsp 90 protein chaperone and promotes ubiquitin-dependent proteasomal degradation of its numerous client proteins. Bortezomib is a specific and potent proteasome inhibitor. Both bortezomib and the geldanamycin analogue, 17-N-allylamino-17-demethoxy geldanamycin, are in separate clinical trials as new anticancer drugs. We hypothesized that destabilization of hsp 90 client proteins with geldanamycin, while blocking their degradation with bortezomib, would promote the accumulation of aggregated, ubiquitinated, and potentially cytotoxic proteins. Indeed, geldanamycin plus bortezomib inhibited MCF-7 tumor cell proliferation significantly more than either drug alone. Importantly, while control cells were unaffected, human papillomavirus E6 and E7 transformed fibroblasts were selectively sensitive to geldanamycin plus bortezomib. Geldanamycin alone slightly increased protein ubiquitination, but when geldanamycin was combined with bortezomib, protein ubiquitination was massively increased, beyond the amount stabilized by bortezomib alone. In geldanamycin plus bortezomib-treated cells, ubiquitinated proteins were mostly detergent insoluble, indicating that they were aggregated. Individually, both geldanamycin and bortezomib induced hsp 90, hsp 70, and GRP78 stress proteins, but the drug combination superinduced these chaperones and caused them to become detergent insoluble. Geldanamycin plus bortezomib also induced the formation of abundant, perinuclear vacuoles, which were neither lysosomes nor autophagosomes and did not contain engulfed cytosolic ubiquitin or hsp 70. Fluorescence marker experiments indicated that these vacuoles were endoplasmic reticulum derived and that their formation was prevented by cycloheximide, suggesting a role for protein synthesis in their genesis. These observations support a mechanism whereby the geldanamycin plus bortezomib combination simultaneously disrupts hsp 90 and proteasome function, promotes the accumulation of aggregated, ubiquitinated proteins, and results in enhanced antitumor activity.

Radiosensitizing and anti-proliferative effects of resveratrol in two human cervical tumor cell lines.

Resveratrol is a polyphenol isolated from the skins of grapes that has been shown to significantly alter the cellular physiology of tumor cells, as well as block the process of initiation and progression. At least one mechanism for the intracellular actions of resveratrol involves the suppression of prostaglandin (PG) biosynthesis. The involvement of PGs and other eicosanoids in the development of human cancer is well established. PGs are synthesized from arachidonic acid via the cyclooxygenase pathway and have multiple physiological and pathological functions. In addition, evidence has arisen suggesting that PGs may be implicated in the cytotoxic and/or cytoprotective response of tumor cells to ionizing radiation (IR). As such, we hypothesized that tumor cells may exhibit changes in the cellular response to IR following exposure to resveratrol, a naturally occurring compound that inhibits cyclooxygenase-1 (COX-1) activity. Thus, clonogenic cell survival assays were performed using irradiated HeLa and SiHa cells pretreated with resveratrol prior to IR exposure, and resulted in enhanced tumor cell killing by IR in a dose-dependent manner. Further analysis of COX-1 inhibition indicated that resveratrol pretreatment: (1), inhibited cell division as assayed by growth curves; and (2), induced an early S phase cell cycle checkpoint arrest, as demonstrated by fluorescence-activated cell sorting, as well as bromodeoxyuridine pulse-chase analysis. These results suggest that resveratrol alters both cell cycle progression and the cytotoxic response to IR in two cervical tumor cell lines.

The effect of 835.62 MHz FDMA or 847.74 MHz CDMA modulated radiofrequency radiation on the induction of micronuclei in C3H 10T(1/2) cells.

To determine if radiofrequency (RF) radiation induces the formation of micronuclei, C3H 10T(1/2) cells were exposed to 835.62 MHz frequency division multiple access (FDMA) or 847.74 MHz code division multiple access (CDMA) modulated RF radiation. After the exposure to RF radiation, the micronucleus assay was performed by the cytokinesis block method using cytochalasin B treatment. The micronuclei appearing after mitosis were scored in binucleated cells using acridine orange staining. The frequency of micronuclei was scored both as the percentage of binucleated cells with micronuclei and as the number of micronuclei per 100 binucleated cells. Treatment of cells with cytochalasin B at a concentration of 2 microg/ml for 22 h was found to yield the maximum number of binucleated cells in C3H 10T(1/2) cells. The method used for the micronucleus assay in the present study detected a highly significant dose response for both indices of micronucleus production in the dose range of 0.1-1.2 Gy and it was sensitive enough to detect a significant (P > 0.05) increase in micronuclei after doses of 0.3 Gy in exponentially growing cells and after 0.9 Gy in plateau-phase cells. Exponentially growing cells or plateau-phase cells were exposed to CDMA (3.2 or 4.8 W/kg) or FDMA (3.2 or 5.1 W/kg) RF radiation for 3, 8, 16 or 24 h. In three repeat experiments, no exposure condition was found by analysis of variance to result in a significant increase relative to sham-exposed cells either in the percentage of binucleated cells with micronuclei or in the number of micronuclei per 100 binucleated cells. In this study, data from cells exposed to different RF signals at two SARs were compared to a common sham-exposed sample. We used the Dunnett's test, which is specifically designed for this purpose, and found no significant exposure-related differences for either plateau-phase cells or exponentially growing cells. Thus the results of this study are not consistent with the possibility that these RF radiations induce micronuclei.

SIRT3 is a mitochondria-localized tumor suppressor required for maintenance of mitochondrial integrity and metabolism during stress.

The sirtuin gene family (SIRT) is hypothesized to regulate the aging process and play a role in cellular repair. This work demonstrates that SIRT3(-/-) mouse embryonic fibroblasts (MEFs) exhibit abnormal mitochondrial physiology as well as increases in stress-induced superoxide levels and genomic instability. Expression of a single oncogene (Myc or Ras) in SIRT3(-/-) MEFs results in in vitro transformation and altered intracellular metabolism. Superoxide dismutase prevents transformation by a single oncogene in SIRT3(-/-) MEFs and reverses the tumor-permissive phenotype as well as stress-induced genomic instability. In addition, SIRT3(-/-) mice develop ER/PR-positive mammary tumors. Finally, human breast and other human cancer specimens exhibit reduced SIRT3 levels. These results identify SIRT3 as a genomically expressed, mitochondria-localized tumor suppressor.

BAT3 and SET1A form a complex with CTCFL/BORIS to modulate H3K4 histone dimethylation and gene expression.

Chromatin status is characterized in part by covalent posttranslational modifications of histones that regulate chromatin dynamics and direct gene expression. BORIS (brother of the regulator of imprinted sites) is an insulator DNA-binding protein that is thought to play a role in chromatin organization and gene expression. BORIS is a cancer-germ line gene; these are genes normally present in male germ cells (testis) that are also expressed in cancer cell lines as well as primary tumors. This work identifies SET1A, an H3K4 methyltransferase, and BAT3, a cochaperone recruiter, as binding partners for BORIS, and these proteins bind to the upstream promoter regions of two well-characterized procarcinogenic genes, Myc and BRCA1. RNA interference (RNAi) knockdown of BAT3, as well as SET1A, decreased Myc and BRCA1 gene expression but did not affect the binding properties of BORIS, but RNAi knockdown of BORIS prevented the assembly of BAT3 and SET1A at the Myc and BRCA1 promoters. Finally, chromatin analysis suggested that BORIS and BAT3 exert their effects on gene expression by recruiting proteins such as SET1A that are linked to changes in H3K4 dimethylation. Thus, we propose that BORIS acts as a platform upon which BAT3 and SET1A assemble and exert effects upon chromatin structure and gene expression.

CTCFL/BORIS is a methylation-independent DNA-binding protein that preferentially binds to the paternal H19 differentially methylated region.

The CTCF paralog BORIS (brother of the regulator of imprinted sites) is an insulator DNA-binding protein thought to play a role in chromatin organization and gene expression. Under normal physiologic conditions, BORIS is predominantly expressed during embryonic male germ cell development; however, it is also expressed in tumors and tumor cell lines and, as such, has been classified as a cancer-germline or cancer-testis gene. It has been suggested that BORIS may be a pro-proliferative factor, whereas CTCF favors antiproliferation. BORIS and CTCF share similar zinc finger DNA-binding domains and seem to bind to identical target sequences. Thus, one critical question is the mechanism governing the DNA-binding specificity of these two proteins when both are present in tumor cells. Chromatin immunoprecipitation (ChIP) in HCT116 cells and their hypermethylated variant showed that BORIS binds to methylated DNA sequences, whereas CTCF binds to unmethylated DNA. Electromobility shift assays, using both whole-cell extracts and in vitro translated CTCF and BORIS protein, and methylation-specific ChIP PCR showed that BORIS is a methylation-independent DNA-binding protein. Finally, experiments in murine hybrid cells containing either the maternal or paternal human chromosome 11 showed that BORIS preferentially binds to the methylated paternal H19 differentially methylated region, suggesting a mechanism in which the affinity of CTCF for the unmethylated maternal allele directs the DNA binding of BORIS toward the paternal allele.

DNA methyltransferase 1 and 3B activate BAG-1 expression via recruitment of CTCFL/BORIS and modulation of promoter histone methylation.

In a previous genomic analysis, using somatic methyltransferase (DNMT) knockout cells, we showed that hypomethylation decreased the expression of as many genes as were observed to increase, suggesting a previously unknown mechanism for epigenetic regulation. To address this idea, the expression of the BAG family genes was used as a model. These genes were used because their expression was decreased in DNMT1(-/-), DNMT3B(-/-), and double knockout cells and increased in DNMT1-overexpressing and DNMT3B-overexpressing cells. Chromatin immunoprecipitation analysis of the BAG-1 promoter in DNMT1-overexpressing or DNMT3B-overexpressing cells showed a permissive dimethyl-H3-K4/dimethyl-H3-K9 chromatin status associated with DNA-binding of CTCFL/BORIS, as well as increased BAG-1 expression. In contrast, a nonpermissive dimethyl-H3-K4/dimethyl-H3-K9 chromatin status was associated with CTCF DNA-binding and decreased BAG-1 expression in the single and double DNMT knockout cells. BORIS short hairpin RNA knockdown decreased both promoter DNA-binding, as well as BAG-1 expression, and changed the dimethyl-H3-K4/dimethyl-H3-K9 ratio to that characteristic of a nonpermissive chromatin state. These results suggest that DNMT1 and DNMT3B regulate BAG-1 expression via insulator protein DNA-binding and chromatin dynamics by regulating histone dimethylation.

SIRT3 interacts with the daf-16 homolog FOXO3a in the mitochondria, as well as increases FOXO3a dependent gene expression.

Cellular longevity is a complex process relevant to age-related diseases including but not limited to chronic illness such as diabetes and metabolic syndromes. Two gene families have been shown to play a role in the genetic regulation of longevity; the Sirtuin and FOXO families. It is also established that nuclear Sirtuins interact with and under specific cellular conditions regulate the activity of FOXO gene family proteins. Thus, we hypothesize that a mitochondrial Sirtuin (SIRT3) might also interact with and regulate the activity of the FOXO proteins. To address this we used HCT116 cells overexpressing either wild-type or a catalytically inactive dominant negative SIRT3. For the first time we establish that FOXO3a is also a mitochondrial protein and forms a physical interaction with SIRT3 in mitochondria. Overexpression of a wild-type SIRT3 gene increase FOXO3a DNA-binding activity as well as FOXO3a dependent gene expression. Biochemical analysis of HCT116 cells over expressing the deacetylation mutant, as compared to wild-type SIRT3 gene, demonstrated an overall oxidized intracellular environment, as monitored by increase in intracellular superoxide and oxidized glutathione levels. As such, we propose that SIRT3 and FOXO3a comprise a potential mitochondrial signaling cascade response pathway.

The epigenome as a molecular marker and target.

Tumor cell proliferation, de-differentiation, and progression depend on a complex combination of altered cell cycle regulation, excessive growth factor pathway activation, and decreased apoptosis. The understanding of these complex mechanisms should lead to the identification of potential molecular markers, targets, and molecular profiles that should eventually expand and improve therapeutic intervention. It now appears clear that methylation plays a central role in transformation, both in vitro and in vivo. However, the exact targets and mechanism(s) are not yet fully understood. This is partly due to the significant number of genes altered by changes in intracellular methyltransferase activity and the chemical agents used to modulate gene expression. The complex nature of methylation's role in regulating gene expression suggests that in addition to investigating individual genes, researchers should develop more comprehensive methods to examine gene expression patterns and their predictive value as this will likely be necessary in the future. If methylation plays a role in transformation, then it seems logical that genes regulating intracellular methylation status may be used as molecular markers to profile tumors by any new methods currently being developed. Perhaps more noteworthy is that DNMT genes may be found to be novel molecular targets for new factor-specific anticancer agents. This idea will be addressed.