Genomic analyses reveal global functional alterations that promote tumor growth and novel tumor suppressor genes in natural killer-cell malignancies.

Natural killer (NK)-cell malignancies are among the most aggressive lymphoid neoplasms with very poor prognosis. We performed array comparative genomic hybridization analysis on a number of NK cell lines and primary tumors to gain better understanding of the pathogenesis and tumor biology of these malignancies. We also obtained transcriptional profiles of genes residing in these regions and compared them with normal and activated NK cells. Only 30-50% of the genes residing in the gained or deleted regions showed corresponding increased or decreased expression. However, many of the upregulated genes in regions of gain are functionally important for the proliferation and growth of the neoplastic population. Genes downregulated in regions of loss included many transcription factors or repressors, tumor suppressors or negative regulators of the cell cycle. The minimal common region of deletion in 6q21 included three known genes (PRDM1, ATG5 and AIM1) showing generally low expression. Mutations resulting in truncated PRDM1 and changes in conserved amino-acid sequences of AIM1 were detected. Highly methylated CpG islands 5' of PRDM1 and AIM1 correlated with low expression of the transcripts. Reversal of methylation by Decitabine induced expression of PRDM1 and cell death. In conclusion, we have shown a general tumor-promoting effect of genetic alterations and have identified PRDM1 as the most likely target gene in del6q21. ATG5, an essential gene for autophagy and AIM1, a gene implicated in melanoma, may also participate in the functional abnormalities.

Inhibition of (cytosine C5)-methyltransferase by oligonucleotides containing flexible (cyclopentane) and conformationally constrained (bicyclo[3.1.0]hexane) abasic sites.

Pseudorotationally locked sugar analogues based on bicyclo[3.1.0]-hexane templates were placed in DNA duplexes as abasic target sites in the M. HhaI recognition sequence. The binding affinity of the enzyme increases when the abasic site is constrained to the South conformation and decreases when it is constrained to the North conformation. A structural understanding of these differences is provided.

Synthesis of oligonucleotide inhibitors of DNA (Cytosine-C5) methyltransferase containing 5-azacytosine residues at specific sites.

The incorporation of 5-azacytosine residues into DNA causes potent inhibition of DNA (Cytosine-C5) methyltransferases. The synthesis of oligodeoxyribonucleotides incorporating single or multiple 5-aza-2'-deoxycytidine residues at precise sites was undertaken to generate an array of sequences containing the reactive 5-azacytosine base as specific target sites for enzymatic methylation. Preparation of these modified oligonucleotides requires the use of 2-(p-nitrophenyl)ethyloxycarbonyl (NPEOC) groups for the protection of exocyclic amino functions. These groups are removed under mild conditions, thus avoiding conventional protocols that are detrimental to the integrity of the 5-azacytosine ring.

Single-site methylation within the p53 promoter region reduces gene expression in a reporter gene construct: possible in vivo relevance during tumorigenesis.

It is not known whether transcriptional suppression by de novo methylation occurs within the promoter region of the p53 gene during multistage tumorigenesis. To address this question, in vivo alterations in the CpG methylation within the rat p53 promoter region were evaluated in control, preneoplastic, and tumor tissue during tumor progression using the folate/methyl-deficient model of hepatocarcinogenesis. Alterations in CpG methylation were found to be site-specific and to vary depending on the stage of carcinogenesis. To further explore the effect of site-specific methylation on p53 promoter activity, reporter gene constructs were prepared containing specifically methylated sites within the p53 promoter region, and the transcriptional activity in cultured mammalian cells was determined in a transient transfection assay. Relative to the unmethylated construct as a positive control, single-site methylation at nucleotide (nt) -450, which occurs 216 nt upstream from the 85-nt minimal promoter region, suppressed promoter activity by 85%. In contrast, single-site methylation at nt -179, which occurs within the minimal essential promoter region, suppressed activity by only 20%. The p53 promoter constructs containing the singly methylated CpG site at nt -450 were then reevaluated for processive changes in methylation status 48 h after transfection, during maximum suppression of promoter activity. Restriction analysis with methylation-sensitive enzymes revealed that de novo methylation had occurred after transfection at previously unmethylated sites. These findings suggest that nt -450 may constitute a critical site for initiation of de novo methylation and processive spreading of methylation associated with transcriptional inactivation of the p53 gene. Furthermore, the results suggest a possible alternative mechanism for the silencing of the p53 gene in tumors that do not have p53 mutations.

Mechanism of inhibition of DNA (cytosine C5)-methyltransferases by oligodeoxyribonucleotides containing 5,6-dihydro-5-azacytosine.

A key step in the predicted mechanism of enzymatic transfer of methyl groups from S-adenosyl-l-methionine (AdoMet) to cytosine residues in DNA is the transient formation of a dihydrocytosine intermediate covalently linked to cysteine in the active site of a DNA (cytosine C5)-methyltransferase (DNA C5-MTase). Crystallographic analysis of complexes formed by HhaI methyltransferase (M.HhaI), AdoMet and a target oligodeoxyribonucleotide containing 5-fluorocytosine confirmed the existence of this dihydrocytosine intermediate. Based on the premise that 5,6-dihydro-5-azacytosine (DZCyt), a cytosine analog with an sp3-hybridized carbon (CH2) at position 6 and an NH group at position 5, could mimic the non-aromatic character of the cytosine ring in this transition state, we synthesized a series of synthetic substrates for DNA C5-MTase containing DZCyt. Substitution of DZCyt for target cytosines in C-G dinucleotides of single-stranded or double-stranded oligodeoxyribonucleotide substrates led to complete inhibition of methylation by murine DNA C5-MTase. Substitution of DZCyt for the target cytosine in G-C-G-C sites in double-stranded oligodeoxyribonucleotides had a similar effect on methylation by M. HhaI. Oligodeoxyribonucleotides containing DZCyt formed a tight but reversible complex with M.HhaI, and were consistently more potent as inhibitors of DNA methylation than oligodeoxyribonucleotides identical in sequence containing 5-fluorocytosine. Crystallographic analysis of a ternary complex involving M.HhaI, S-adenosyl-l-homocysteine and a double-stranded 13-mer oligodeoxyribonucleotide containing DZCyt at the target position showed that the analog is flipped out of the DNA helix in the same manner as cytosine, 5-methylcytosine, and 5-fluorocytosine. However, no formation of a covalent bond was detected between the sulfur atom of the catalytic site nucleophile, cysteine 81, and the pyrimidine C6 carbon. These results indicate that DZCyt can occupy the active site of M.HhaI as a transition state mimic and, because of the high degree of affinity of its interaction with the enzyme, it can act as a potent inhibitor of methylation.

Transcriptional activation of functional endogenous estrogen receptor gene expression in MCF10AT cells: a model for early breast cancer.

Utilizing the MCF10AT xenograft model for progression of human proliferative breast disease, we detected expression of the endogenous estrogen receptor (ER) gene only in MCF10AneoT and cells of the MCF10AT system, all of which stably express a transfected mutated T24 Ha-ras gene. ER transcripts were undetectable in the parental MCF10A cells and in MCF10A cells transfected with normal c-Ha-ras or vector. ER transcripts expressed in MCF10AT cells contain a normal full-length ER coding region and direct synthesis of a normally sized ER protein. The protein is functional based on its ability to mediate estradiol (E2)-induced increases of transcription from both endogenous and exogenous E2-regulated genes. Transcriptional activation of the endogenous ER gene does not appear to be related to a change in methylation status of the gene since a diagnostic CpG site in exon 1 that is methylated in ER-negative breast tumors and completely unmethylated in ER-positive breast tumors is hypomethylated to the same extent in ER-negative MCF10A cells and ER-positive MCF10AT cells. E2 increased both the number and size of soft-agar colonies formed by MCF10AT3c cells, a line from a third generation MCF10AT xenograft lesion. This suggests that xenograft passage has selected for growth regulatory pathways that are E2-responsive and that identification of these pathways and their role in progression will aid in determining how E2 acts to increase risk of breast cancer.

Characterization of the estrogen receptor transfected MCF10A breast cell line 139B6.

There has been increasing evidence which suggests that abnormal expression of the estrogen receptor (ER) protein in nonmalignant breast tissue may be important in the carcinogenic process. To examine the effects of ER expression in immortalized nonmalignant mammary epithelial cells, an expression vector containing human ER cDNA was transfected into the ER negative human breast cells, MCF10A. Characterization of a clone stably expressing ER, 139B6, provided evidence for the regulated synthesis of a functional ER capable of binding estradiol-17 beta (E2) and undergoing processing. Expression of the ER gene did not enable E2 to stimulate endogenous genes [progesterone receptor (PgR), pS2, cathepsin D and TGF alpha] which normally respond to estrogens in breast cancer cells. The ER in 139B6 cells was, however, capable of inducing expression of an ERE-regulated reporter gene, indicating its ability to interact with transcriptional machinery. Furthermore, cultures in log growth displayed a slight increase in doubling time in the presence of E2. These results indicate that ER expression alone is not sufficient to induce a transformed phenotype. Thus, the 139B6 cell line should provide a new model for determining what additional changes lead to increased growth potential in response to E2 and for exploring how E2 itself may help bring about changes leading to progression of preneoplastic breast epithelial cells.

Relationship between estrogen structure and conformational changes in estrogen receptor/DNA complexes.

The effect of estrogen structure on the conformation of the complex formed with estrogen receptor (ER) and the consensus estrogen response element (EREc) has been examined with gel mobility shift assay. Proteins in MCF-7 cell extracts formed three distinct complexes with ERE. Only the slowest moving complex contained ER as indicated by binding with anti-ER antibodies H222 and D547. This ER-ERE complex displayed increased electrophoretic mobility when formed in the presence of estradiol (E2) and bound radiolabeled 16 alpha-iodoestradiol. The antiestrogen ICI 164,384 decreased the mobility of the ER-ERE complex and blocked the effect of E2. The results reported here indicate that the position and location of hydroxyl groups on the estratriene nucleus is an important factor in determining the mobility of ER-EREc (or a variant ERE) in gel shift assays. The ability of E2 analogs to cause conformational changes detectable as altered mobility was not directly related either to their binding affinity for ER or to their ability to activate E2 responsive genes. Although several dihydroxyestrogens (estradiol-16 alpha, 1- and 2-hydroxyestratrien-17 beta-ol) caused an increase in the mobility of the ER-EREc, other ligands (estradiol-17 alpha, 4-hydroxyestratriene-17 beta-ol, 3-hydroxy estratriene, estratrien-17 beta-ol and 5-androsten-3 beta, 17 beta-diol) with a capacity for activating at least some E2 responsive genes in MCF-7 cells had little or no effect. On the basis of these and previously published results, it can be concluded that specific structure features of estrogens are responsible for conformational changes of ER-ERE complexes detectable in gel-shift assays. Furthermore, the identified structural characteristics of the ligand which are required for gel-shift are not the same as those previously reported to be essential for stimulation of transcriptional activity of ER.

5-Methyl-2'-deoxycytidine in single-stranded DNA can act in cis to signal de novo DNA methylation.

Methylation of cytosine residues in DNA plays an important role in regulating gene expression during vertebrate embryonic development. Conversely, disruption of normal patterns of methylation is common in tumors and occurs early in progression of some human cancers. In vertebrates, it appears that the same DNA methyltransferase maintains preexisting patterns of methylation during DNA replication and carries out de novo methylation to create new methylation patterns. There are several indications that inherent signals in DNA structure can act in vivo to initiate or block de novo methylation in adjacent DNA regions. To identify sequences capable of enhancing de novo methylation of DNA in vitro, we designed a series of oligodeoxyribonucleotide substrates with substrate cytosine residues in different sequence contexts. We obtained evidence that some 5-methylcytosine residues in these single-stranded DNAs can stimulate de novo methylation of adjacent sites by murine DNA 5-cytosine methyltransferase as effectively as 5-methylcytosine residues in double-stranded DNA stimulate maintenance methylation. This suggests that double-stranded DNA may not be the primary natural substrate for de novo methylation and that looped single-stranded structures formed during the normal course of DNA replication or repair serve as "nucleation" sites for de novo methylation of adjacent DNA regions.

Moderate folate deficiency does not cause global hypomethylation of hepatic and colonic DNA or c-myc-specific hypomethylation of colonic DNA in rats.

Global and gene-specific DNA hypomethylation is considered to be an important early epigenetic event in several human neoplasms. A growing body of evidence has suggested that DNA methylation can be altered by dietary manipulation of methyl group donors. This study investigated whether moderate depletion of folate, a dietary component needed for the synthesis of methyl groups, would cause decreased hepatic and colonic S-adenosylmethionine concentrations, and thereby lead to global and/or protooncogene-specific DNA hypomethylation. Weanling rats were fed an amino acid-defined diet containing either 0 or 8 mg folate/kg diet for 15 or 24 wk. Significantly lower systemic, hepatic and colonic folate concentrations were observed in the moderately folate-depleted rats than in controls at both 15 and 24 wk (P < 0.005). Although hepatic S-adenosylmethionine was significantly lower in the moderately folate-depleted rats than in controls at the two time points (P < 0.03), colonic S-adenosylmethionine concentrations were not significantly different between the two groups at either time point. No significant differences between the folate-depleted and control animals could be detected with regard to global DNA methylation in the liver or colonic mucosa. Furthermore, c-myc protooncogene-specific DNA methylation in the colonic mucosa was not significantly different between these two groups of animals. These results indicate that moderate folate depletion does not cause a significant reduction in global DNA methylation in liver or colonic mucosa or in c-myc-specific colonic mucosal DNA methylation in this rat model.