[CANCER RESEARCH 64, 3465–3473, May 15, 2004]
Microarray Analysis of Epigenetic Silencing of Gene Expression in the KAS-6/1
Multiple Myeloma Cell Line
Celine Pompeia,1 David R. Hodge,1 Christoph Plass,3 Yue-Zhong Wu,3 Victor E. Marquez,2 James A. Kelley,2 and
William L. Farrar1
1
Laboratory of Molecular Immunoregulation, and 2Laboratory of Medicinal Chemistry, National Cancer Institute-Frederick Cancer Research and Development Center, Frederick,
Maryland, and 3Division of Human Cancer Genetics, Department of Molecular Virology, Immunology and Medical Genetics, Ohio State University, Columbus, Ohio
ABSTRACT
The epigenetic control of gene transcription in cancer has been the
theme of many recent studies and therapeutic approaches. Carcinogenesis
is frequently associated with hypermethylation and consequent downregulation of genes that prevent cancer, e.g., those that control cell proliferation and apoptosis. We used the demethylating drug zebularine to
induce changes in DNA methylation, then examined patterns of gene
expression using cDNA array analysis and Restriction Landmark
Genomic Scanning followed by RNase protection assay and reverse transcription-PCR to confirm the results. Microarray studies revealed that
many genes were epigenetically regulated by methylation. We concluded
that methylation decreased the expression of, or silenced, several genes,
contributing to the growth and survival of multiple myeloma cells. For
example, a number of genes (BAD, BAK, BIK, and BAX) involved in
apoptosis were found to be suppressed by methylation. Sequenced methylation-regulated DNA fragments identified by Restriction Landmark
Genomic Scanning were found to contain CpG islands, and some corresponded to promoters of genes that were regulated by methylation. We
also observed that after the removal of the demethylating drug, the
addition of interleukin 6 restored CpG methylation and re-established
previously silenced gene patterns, thus implicating a novel role of interleukin 6 in processes regulating epigenetic gene repression and carcinogenesis.
INTRODUCTION
The control and cure of cancer is an important public health issue
because cancer is a significant contributor to patient mortality. Because a majority of patients suffer serious side effects or develop
resistance to traditional forms of cancer treatment, the development of
drugs as an adjunct to, or replacement for, current chemotherapy and
radiotherapy protocols is of paramount importance. One of the great
difficulties in such a quest is the vast variability inherent to malignant
cells, thus requiring drugs that are targeted to the common features
found in neoplastic disease. Recently, attention has focused on some
types of neoplastic growth that are thought to evolve from chronic
inflammatory states (1–3). In these cases, cancer cells are able to
proliferate by using inflammatory signals as growth signals and to
evade cell death by inactivating apoptotic pathways triggered in some
cases by immune surveillance and are themselves capable of promoting and inducing angiogenesis.
Many tumor cells respond to, or become, autocrine producers of the
inflammatory cytokine interleukin 6 (IL-6), using it for growth and
survival. This phenomenon is interesting because, in normal cells,
Received 12/18/03; revised 2/0/04; accepted 3/2/04.
Grant support: This work was supported by the Federal funds of the National Cancer
Institute, NIH, under contract N01-CO-5600 and the Department of Health and Human
Services, by contract with Science Applications International Corporation-Frederick. C.
Plass is a Leukemia and Lymphoma Society Scholar and his work is sponsored by NIH
Grant CA93548.
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.
Requests for reprints: William L. Farrar, Laboratory of Molecular Immunoregulation, National Cancer Institute-Frederick Cancer Research and Development Center, 1050
Boyles Street, Building 560, Room 31-68, Frederick, MD 21702. Phone: (301) 846-6867;
Fax: (301) 846-7042; E-mail: [email protected].
IL-6 functions mainly to mediate and augment inflammatory responses. IL-6 has been associated with the etiology of, and as a
promoter of, several forms of cancer, including hepatic and prostate
cancers and multiple myeloma (4 –7). DNA methylation is associated
with several changes in chromatin structure, including the regulation
of histone methylation and acetylation and the recruitment of proteins
to the methylated sites. This usually leads to the obstruction of the
promoter, which hinders gene transcription, and to subsequent gene
silencing (8). Our laboratory has reported that IL-6 exerted a regulatory influence on DNA methylation via the transcriptional activation
of FLI-1, a transcription factor that up-regulated the expression of
DNA methyltransferase-1 (DNMT-1; Refs. 9, 10). We have also
observed that IL-6 contributes to the epigenetic silencing of important
cell cycle and tumor suppressor genes.4 Therefore, it is conceivable
that some malignant tumors may develop from chronic inflammatory
activity based on the influence IL-6 exerts on the epigenetic control of
genes, such as tumor suppressors and modulators of differentiation
and apoptosis.
Many of the traits exhibited by cancer cells can be attributed to
epigenetic changes in gene expression induced by methylation of
DNA. Numerous tumor suppressor genes and proapoptotic genes, for
example, are down-regulated or silenced in cancer cells because of
promoter hypermethylation (11, 12). For this reason, some chemotherapy protocols have included the use of drugs that inhibit DNA
methylation. Among such drugs are several analogs of deoxycytidine,
the targeted nucleoside for methylation, such as 5-azacytidine
and 5-aza-2⬘-deoxycytidine. Additionally, 5-aza-2⬘-deoxycitidine has
been used extensively to study the reactivation of silenced genes in
numerous cell types. Although these drugs have shown some efficient
antitumor activity (reviewed in Ref. 13), they are highly unstable in
vivo and toxic to normal cells. Recently, a new drug, originally
developed as a cytidine deaminase inhibitor, zebularine, has been
shown to inhibit DNA methylation and is a promising candidate for
cancer therapy because of its lower toxicity and increased stability
under physiological conditions (14, 15).
To assess the importance of epigenetic regulation of gene expression in cancer cells and to explore the role of IL-6 as an inducer of
gene methylation and silencing, we used zebularine to demethylate
DNA in the IL-6-responsive multiple myeloma cell line KAS-6/1.
After treatment with zebularine, cells were washed and allowed to
grow in medium supplemented with 10 ng/ml IL-6 to assess the
consequences of gene remethylation. The overall effect of zebularine
on KAS-6/1 cell DNA methylation, and the silencing effect of IL-6,
were evaluated by cDNA microarrays and Restriction Landmark
Genomic Scanning (RLGS) analysis. Our data indicate that vital genes
were subject to epigenetic regulation via DNA methylation and suggest that this particular tumor type may have developed, in part, as a
result of the epigenetic silencing of differentiation modulators, tumor
suppressors, and apoptotic mediators. Furthermore, because IL-6 was
needed to restore the original DNA methylation levels present before
4
D. R. Hodge, B. Peng, C. Pompeia, S. D. Fox, V. E. Marquez, J. A. Kelley, and W. L.
Farrar. IL-6 regulates methylation of tumor suppressor gene promoters, manuscript in
preparation.
3465
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 2004 American Association for Cancer
Research.
EPIGENETIC TRANSCRIPTION SILENCING IN MYELOMA CELLS
drug treatment, the results obtained confirmed the important role of
IL-6 in methylation and subsequent down-regulation or silencing of
gene expression.
Standard statistical techniques were used and later validated using an experimental design called the “Latin square,” which is developed using naturally
absent transcripts spiked at known concentrations. This test allows the evaluation of the significance of values given the complex background of the
microarrays.
A detection P was calculated for each probe set by the signal difference
MATERIALS AND METHODS
from perfect-match probes and mismatch probes, rendering a discrimination
Materials. KAS-6/1 cells (16) were kindly provided by Dr. Dianne Jelinek, score that was used in a one-sided Wilcoxon’s signed-rank test. The signal
of the Mayo Clinic, Rochester, MN. Zebularine was synthesized in Dr. Victor from each probe set was a quantitative value that represented the relative
E. Marquez’s Laboratory of Medicinal Chemistry (National Cancer Institute, expression level and was calculated using the one-step Tukey’s biweight
NIH, Frederick, MD; Ref. 14). Affymetrix DNA microarrays were obtained estimate. The comparison between chips was done by two algorithms, one that
from Affymetrix (Santa Clara, CA). Fetal bovine serum was from Gemini generated a “change P” (using Wilcoxon’s signed-rank test) and the other for
Bio-Products (Woodland, CA). RPMI 1640, ampicillin/streptomycin solution, the quantitative estimate of the gene expression, which was associated with a
and L-glutamine were from Cellgro (Mediatech, Herndon, VA), IL-6 was signal log ratio (calculated with Tukey’s biweight method). Before chips were
obtained from PeproTech (Rock Hill, NJ). Agarose, Taq polymerase, and compared, scaling and normalization were also carried out using several
dNTPs were from Invitrogen (Carlsbad, CA). DNA molecular weight standard methods (not detailed by Affymetrix).
(100 bp) was from New England BioLabs (Beverly, MA), and other reagents
The microarray assay was carried out with two sets of chips, U133A and
were obtained from Sigma (Saint Louis, MO), unless otherwise stated in the U133B, with a total of 45,000 probe sets representing more than 39,000
methods.
transcripts, derived from ⬃33,000 well-substantiated human genes (based on
Cell Treatment. The human multiple myeloma cell line KAS-6/1 was the genome Build of April 2001). The set design uses sequences selected from
grown in RPMI 1640 supplemented with 10% fetal bovine serum, 10 ng/ml GenBank, dbEST, and RefSeq. The U133A set was generated from cDNA data
IL-6, 50 IU/ml ampicillin, 50 ␮g/ml streptomycin, 2 mM L-glutamine in a sequences previously represented on the Human Genome U95Av2 Array,
humidified incubator, at an atmosphere of 5% CO2, at 37°C. Control cells whereas the U133B was generated primarily from EST clusters. Each array has
received no treatment; zebularine-treated cells received 200 ␮M drug hereafter ⬃22,000 “probe sets,” each containing 11 DNA oligonucleotides of 25 bases
referred to as zebularine for a period of 48 h; “remethylated” cells were treated corresponding to different regions of the 3⬘ end of a given transcript. In some
with zebularine as described, washed 3 times with fresh medium, diluted
cases, several probe sets exist for the same gene, so that different splice forms
10-fold in fresh medium containing 10 ng/ml IL-6, and allowed to grow for 1
or transcripts with different polyadenylation sites can be assessed. Three sets
week.
of chips were analyzed for each group of cells: two with probes generated from
RLGS. RLGS was performed as described in Hatada et al. (17). Cell the same RNA (experiments 1 and 2) and one with probes generated from
genomic DNA was extracted, and sheared ends were blocked with nucleotide RNA obtained in a different experiment (experiment 3). To ascertain the
analogs (␣-S-dGTP, ␣-S-dCTP, dideoxyadenosine triphosphate, and dideoxy- accuracy of our first result (experiments 1 and 2), we repeated the microarray
thymidine 5⬘-triphosphate) in the presence of DNA polymerase I. Fragments analysis de novo with total RNAs from the KAS 6/1 cells treated as described
were then cut with NotI, which is specific for nonmethylated strands, and cut previously. Results obtained from this confirmatory work is presented in
ends were radiolabeled with [␣-32P]dGTP. This method ensured that highly experiment 3.
methylated regions of DNA were not radiolabeled. The resulting fragments
Semiquantitative RT-PCR. Total RNA was used as a template for the
were further cut with EcoRV and subjected to electrophoresis on a 0.8%
synthesis of cDNA using the “1st Strand cDNA Synthesis kit for RT-PCR
agarose tube gel. The DNAs were cut in situ with HinfI and the tube gel was
(AMV),” according to the manufacturer’s protocol (Roche Applied Science,
placed on the top of a 5% polyacrylamide gel for a second dimension elecIndianapolis, IN). The RNA (1 ␮g/20 ␮l) was added to a mix containing 10
trophoresis of the DNA. The gels were dried and exposed to an X-ray film.
mM Tris (pH 8.3), 50 mM KCl, 5 mM MgCl2, 1 mM dATP, 1 mM dTTP, 1 mM
Differences observed in the intensity of spots indicate changes in the degree of
dCTP, 1 mM dGTP, 80 ng/ml oligo-p(dT)15 primer, 2.5 units/␮l RNase
DNA methylation between samples. RLGS fragments were cloned from the
inhibitor, and 1 unit/␮l AMV reverse transcriptase. After primer annealing at
two-dimensional gel using the strategy described by Smiraglia et al. (18).
room temperature, the cDNA synthesis proceeded at 42°C, and the final
Sequences were aligned using BLAST to DNA databases such as GenBank
reaction was stopped by enzyme inactivation at 99°C for 5 min. The final
and BLAT, for alignment to the University of California-Southern California
product was further diluted with 30 ␮l water.
human genome to map their positions and identify putative gene promoter
The PCR was carried out with 20 mM Tris (pH 8.4), 50 mM KCl, 0.2 mM
regions.
dATP,
0.2 mM dTTP, 0.2 mM dCTP, 0.2 mM dGTP, 1.5 mM MgCl2 (unless
RNA Extraction. RNA was extracted from cells using the TRIzol reagent
(Invitrogen), as recommended by the manufacturer. Briefly, cells were lysed in otherwise stated), 0.2 ␮M each primer, 1 unit AmpliTaqDNA polymerase
a solution of phenol and guanidine isothiocyanate, and on the addition of (Invitrogen), and 40 ␮l/ml cDNA template. After an initial step for enzyme
chloroform, the aqueous phase, containing the RNA, was isolated from the activation (2 min, 96°C), cycles consisting of 30 s denaturing (96°C), 30 s
organic phase containing solvents, denatured proteins, and DNA. The RNA annealing, and 30 s extension (72°C) were performed. Primers, annealing
was then precipitated with isopropanol, washed with 75% ethanol, dried, and temperature, and cycle number are indicated in Table 1. The PCR conditions
dissolved in water. RNA quantity and quality were assessed by agarose gel were optimized to obtain data from the exponential phase of the reaction, when
electrophoresis (1% gel), spectrophotometry (wavelength ratio 260 and 280 relative levels of expression could be assessed. The housekeeping gene glycnm). The RNAs used for microarray analysis were also evaluated by high- eraldehyde 3-phosphate (GAPDH) was used to normalize data. After PCR, the
DNA products were run on a 2% Tris-acetate-EDTA (0.5⫻ TAE), ethidium
pressure liquid chromatography by Expression Analysis (Durham, NC).
Affymetrix Microarray Analysis. Samples of RNA from control (un- bromide-stained agarose gel. The molecular weight of the PCR products was
treated), zebularine-treated, and remethylated cells were processed by Expres- estimated by comparison with the 100-bp molecular weight standard. The
sion Analysis (Durham, NC). The steps involved in Affymetrix microarray fluorescence of each band was quantified using LabWorks Analysis Software,
assays were as follows: generation of a double-stranded cDNA using the RNA Version 3.0.02.00, UVP (Upland, CA).
RNA Protection Assay. RNase protection assay was carried out with the
as a template; in vitro transcription for the synthesis of an antisense complementary biotin-labeled cRNA; hybridization to the microarray chip; stringency “Multi-Probe RNase Protection Assay System,” human apoptosis-related set
washing of the chip; and scanning of the bound fluorescence tags on the chip. (hAPO-2c; BD Biosciences, Franklin Lakes, NJ), according to the manufac33
The data obtained was then processed by a series of programs to normalize the turer’s instructions. The kit contains DNA templates for [␣- P]-labeled RNA
antisense
synthesis.
The
probe
mixture
was
added
to
the
sample RNA and
data, statistically determine whether a given transcript was present or not, and
whether there were significant differences in expression between the different allowed to hybridize. Nonannealing RNA was then digested with RNases and
duplicates (experiments 1 and 2) or groups (control, zebularine-treated, and protected RNA fragments were resolved in a 6% polyacrylamide-urea gel. The
remethylated). The Affymetrix statistical algorithms were run using the Af- gel was dried and exposed to X-ray film. Internal standards [L32 and glycerfymetrix Microarray Suite version 5.0, a proprietary group of programs. aldehyde 3-phosphate (GAPDH)] were used to normalize for sample loading.
3466
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 2004 American Association for Cancer
Research.
EPIGENETIC TRANSCRIPTION SILENCING IN MYELOMA CELLS
Table 1 PCR primers and conditions for each of the analyzed genes
Gene
ALDH12 aldehyde dehydrogenase 8 family,
member A1
AML1 runt-related transcription factor 1;
acute myeloid leukemia 1
BIK BCL2-interacting killer
ICAM1/CD54 intercellular adhesion
molecule 1 (CD54), human rhinovirus
receptor
JTV1
PA2G4 proliferation associated 2G4
RASSF1 Ras association (RalGDS/AF-6)
domain family 1
RNF7 ring finger protein 7
SLC3A2 Solute carrier family 3 member 2
GAPDH glyceraldehyde 3-phosphate
Primers
5⬘-tggtgagcataggtgctctg-3⬘
5⬘-caccgtgggaagcataaagt-3⬘
5⬘-aatgacctcaggtttgtcgg-3⬘
5⬘-caatggatcccaggtattgg-3⬘
5⬘-tcttgatggagaccctcctg-3⬘
5⬘-agtgtggtgaaaccgtccat-3⬘
5⬘-ggctggagctgtttgagaac-3⬘
Annealing temperature (°C)
No. of cycles
Fragment size (bp)
55
35
164
60
34
400
49
36
310
49
36
222
62
33
294
55
23
313
60
36
250
62
32
286
50
37
432/336
55
32
319
5⬘-aggagtcgttgccataggtg-3⬘
5⬘-gtcgtgacctctgacggttt-3⬘
5⬘-ccatcaactgcagctttcaa-3⬘
5⬘-ggctgcagtggtggtagtagg-3⬘
5⬘-cctggtcgctcttcaaaggg-3⬘
5⬘-cacttccttttacctgccca-3⬘
5⬘-cttcaggacaaagctcaggg-3⬘
5⬘-ggaagacggagaggaaacct-3⬘
5⬘-gcagcgattgttctgtttca-3⬘
5⬘-ggttttctcacccagtgcat-3⬘
5⬘-tcttggaagctaaggcagga-3⬘
5⬘-aggtgaaggtcggagtcaacgg-3⬘
5⬘-cccagccttctccatggtggtg-3⬘
fragments indicate methylation control. Darker spots correspond to
unmethylated DNA and are found more frequently in zebularinetreated-derived DNA, as compared with control DNA (untreated) and
DNAs obtained after wash-out of zebularine from the cells, associated
with the culture of the cells for several days in the presence of IL-6
(“remethylated”). The patterns of the DNA fragments obtained from
control and remethylated cells were very similar, indicating that the
removal of zebularine and addition of IL-6 restores control DNA
methylation patterns.
Of a total of 1360 spots, 23 DNA fragments were darker in DNA
derived from cells treated with zebularine. These 23 well-resolved
spots were identified, and 7 were sequenced. The results are shown in
Table 2.
After alignment with public databases, all of the sequenced fragments were found to correspond to CpG islands. Such alignment,
particularly to the human DNA genome, enabled us to map the
sequenced fragments and, in some cases, associate them with putative
gene promoters. Five sequences aligned with known genes, one with
a predicted open reading frame and one with an expressed sequence
tag. The first known gene is the poly(rC)-binding protein 1 (PCBP1),
a multifunctional protein that binds RNA and is involved in translation and RNA stability. PCBP1 is known to participate in the increased stability of mRNAs containing internal ribosome entry segments. Next, the microtubule-associated protein 1 light chain 3 ␤
(MAP1LC3B) protein, which functions in both neurogenesis and
autophagy. Third, the estrogen receptor 1 (ESR1), that was found to
Fig. 1. Restriction Landmark Genomic Scanning (RLGS) two-dimensional gel film
X-rays. The genomic DNA from each sample was digested with NotI, radioactively
labeled, digested with EcoRI, run on an agarose tube gel; digested in situ with HinfI and
run on a 5% polyacrylamide gel. Regions with good resolution were selected to pick spots
with various intensity. Each spot received identification. Arrows, DNA fragments (spots)
that appear in zebularine-treated cells and are absent, or found at a lower intensity, in
DNA fragments from control and remethylated cells. Alphanumerical codes to the right
correspond to the names assigned to each of these regulated spots.
Table 2 Overall results of Restriction Landmark Genomic Scanninga
After exposure of the two-dimensional gel to an X-ray film, the spot intensities and
patterns among the control, zebularine-treated, and remethylated samples were compared.
In all cases, the intensity of the control and the remethylated spots was the same.
DNA spot intensity in control, zebularine
and remethylated-derived samples
DNA spot identification
DNA fragments that are absent in control 3D3; 3A25 (EST), 2F55 (X03635, ESR1),
and remethylated samples, but appear
3E47, 3F9, 3F71 (morfar),b N3⫹,
N7⫹, N8⫹, 3F19 (CSDA)
in zebularine-treated samples
RESULTS
RLGS. The overall effect of zebularine on KAS-6/1 genomic DNA
methylation was evaluated using RLGS (Fig. 1; Table 2). This technique uses methylation-sensitive restriction enzymes with a high CG
and CpG island content to cleave genomic DNA at the CpG islands.
Cleaved sites are radioactively labeled and, after further digestion
with restriction enzymes, the fragments are sequentially resolved in
agarose and polyacrylamide gels, yielding a two-dimensional level of
resolution. As shown in Fig. 1, changes in the intensity of the
DNA fragments that are present in
control and remethylated samples, but
that appear with a higher intensity in
zebularine-treated samples
3D19 (X78137, PCBP1), 4C1 (EST), 5E2
(NM㛭022818, MAP1LC3B), 6E5, 3E54,
3F53 (AB007897, SETBP1), 5F3 or
5F4 (AF074924, NDST3), N1⫹, N2⫹,
N4⫹, N5⫹, N6⫹, N9⫹
a
DNA fragments that were sequenced are followed by parentheses containing the
GenBank accession number and gene symbol, when known.
b
morfar, predicted gene with unknown function; CSDA, cold shock domain protein A,
a transcriptional regulator that binds and represses the promoter of the granulocyte
macrophage colony-stimulating factor (GM-CSF) gene.
3467
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 2004 American Association for Cancer
Research.
EPIGENETIC TRANSCRIPTION SILENCING IN MYELOMA CELLS
Fig. 2. Gene regulation by zebularine (Array U188A, average of experiments 1 and 2).
Full diamond-shaped symbols, the relative intensity of zebularine/remethylated samples
(Y axis) and zebularine/control samples (X axis). The intensity of each sample corresponds
to an average from experiments 1 and 2 of the results of each probe set present in Array
U188A. The empty boxes, data considered significant between groups by the Affymetrix
statistical suite (P ⱕ 0.05). The gray lines, the cutoff value of 2 in each axis. Only probe
sets corresponding to white squares with a value within the upper right quadrant were used
for further analysis. These correspond to genes that were up-regulated at least 2-fold by
zebularine and then down-regulated at least 2-fold after remethylation.
be down-regulated by demethylation from its basal level in untreated
cells. Fourth, the SET binding protein 1 (SETBPI), binds to SET, the
translocation breakpoint-encoded protein found in acute undifferentiated leukemia, which participates in protein transport and transcriptional regulation. Fifth, the N-deacetylase/N-sulfotransferase (heparin
glucosaminyl) 3 (NDST3) enzyme, responsible for heparin and protein deacetylation and sulfation. This finding is interesting because of
the role heparin is believed to play in cell proliferation, metastasis,
angiogenesis, and cell adhesion (reviewed in Ref. 19).
Microarray. On verification that zebularine reduced DNA methylation levels in KAS-6/1 cells, we decided to perform a systematic
study on genes epigenetically regulated by zebularine using total cell
RNA as a probe in the expression cDNA microarray technique. This
approach has already been used to study methylation-regulated gene
transcription, e.g., 5-aza-deoxycytidine-treated human fibroblast and
bladder cancer-derived RNA revealed many genes differentially regulated by this demethylating agent in normal versus cancer cells using
the microarray technique (20). Several criteria were used to select
probe-sets of interest: (a) results derived from drug-treated cells had
to be at least 2-fold higher than those derived from control or remethylated cells; (b) such differences between control, drug-treated,
and remethylated signals had to be statistically significant (P ⱕ 0.05);
and (c) for positive regulation with the drug, gene expression had to
be considered present in the probe set in all experiments (specialized
software was used to determine whether a given gene set differed
significantly from the background hybridization signal).
Fig. 2 is an example of the genes that are up-regulated by drug
treatment in relation to control and remethylated-derived RNA. The X
and Y axes correspond to the fraction between the signal from the
treated probe sets and that of the control and remethylated probe sets,
respectively. High values in the abscissa indicate that demethylation
resulted in an increase in gene expression. Probe sets corresponding to
high Y-axis values indicate that the removal of zebularine and the
recovery of the cells in IL-6-containing medium down-regulates gene
expression. The black diamonds (䉬) correspond to the total probe sets
and, overlaying them in white squares (䡺, are the probe sets that
correspond to transcripts considered significantly different between
both demethylated and control, and demethylated and remethylated
data (P ⱕ 0.05). Although statistical programs consider the differences significant, from a biological perspective differences of at least
2-fold are most likely relevant. Therefore, only the probe sets corresponding to white boxes in the upper right quadrant of the plot were
used for further analyses, because they indicate genes up-regulated by
zebularine and down-regulated in the absence of the drug and in the
presence of IL-6.
The overall results obtained from the microarray studies are summarized in Table 3. The number of probe sets up-regulated by demethylation represents a small fraction of the total probe sets: 0.19%
and 0.58% of genes up-regulated by drug in experiments 1 plus 2 and
3, respectively (U133A chip). These values corresponded to 0.11 and
0.39% for chip U133B. The results were always higher for experiment
3 because experiments 1 and 2 were analyzed together, which led to
the exclusion of duplicates that were not considered compatible with
each other. The overlap of up-regulated probe sets by zebularine
between all of the experiments was of 9 (0.04%) and 2 (0.0088%) for
chips U133A and U133B, respectively (Table 4A and Table 4B).
To evaluate the function of the genes up-regulated by demethylation, several databases were searched, particularly those from the
Gene Ontology (GO) Consortium. A main function was assigned to
each gene shown in Table 5. The regulated probe sets were most
associated with metabolism, intracellular signaling, transcriptional
regulation, and proliferation (with the exception of one chip) genes.
Genes associated with RNA binding/processing, chromatin/DNA,
replication/DNA repair, and cell structure, exhibited relatively few
changes on demethylation.
Some of the genes regulated by methylation, as assessed by the
microarray analysis, matched the genes identified by RLGS . These
include RLGS fragment 5E2, which codes for a microtubule-associated protein. The fragment was present exclusively in the drugtreated-derived samples, indicating that the promoter for this gene was
demethylated facilitating transcription. There are two probe sets corresponding to this gene, 208785_s_at and 208786_s_at, both from
chip U133A. The fractions obtained between demethylated and control or remethylated-derived samples were as follows: 1.32/1.19 (experiments 1 ⫹ 2), 1.52/1.46 (experiment 3) for 208765_s_at; and
1.18/0.98 (experiments 1 ⫹ 2), 1.4/1.68 (experiment 3) for set
208786_s_at. The restriction landmark genomic sequencing fragment
3F53, corresponding to SETBP1 overlaps the promoter corresponding
Table 3 Genes up-regulated by zebularine and identified by microarray analysis
Microarray chip
Total number of probe setsa
Up-regulated genes (average expb 1⫹2)
Up-regulated genes (exp 3)
Common up-regulated genes (exp 1–3)
Up-regulated genes with CPG islands (exp 1⫹2)
Up-regulated genes with CPG islands (exp 3)
U133A
22,283
42
129
9
30 (71% of up-regulated)
88 (68%)
U133B
22,645
23
89
2
15 (65% of up-regulated)
56 (63%)
a
There may be more than one probe set per gene, and the same gene may be present on different chips. The total of genes present in the microarrays U133A plus U133B is 33,000.
The presence of CpG islands was analyzed within the 1,500-bp upstream from the transcription initiation signal of the gene associated with each probe set, by using the “CpGProD
(CpG Island Promoter Detection)” program (52), which is available for public use at http://pbil.univ-lyon1.fr/software/cpgprod㛭query.html.
b
Exp, experiment.
3468
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 2004 American Association for Cancer
Research.
EPIGENETIC TRANSCRIPTION SILENCING IN MYELOMA CELLS
Table 4 Gene expression in probe sets up-regulated by zebularine
The microarray expression data are shown for the probe sets up-regulated by zebularine according to three criteria: presence of signal in zebularine-derived samples, increase in
signal of at least 2-fold, and significant difference between the zebularine signal and that of control and remethylated samples. Furthermore, only genes that fitted all of the criteria
in all of the experiments are shown. The columns show, respectively, the probe set identification number (id); the average value obtained, respectively, with control, zebularine-treated
(Zebul), and remethylated (Remeth) samples from experiments 1 and 2 [(1⫹2)]. The fraction between the value obtained with zebularine and that of the control or remethylated samples
are in the two following columns, respectively. Control, zebularine-treated, and remethylated values, as well as the fraction values obtained in experiment 3 [(3)] follow in columns
7–11, respectively. Column 12 contains a brief description of the gene associated with the probe set, and the last column indicates whether there is a CpG island in the promoter of
the gene (see Table 3). The value 1 is assigned when at least one significant CpG island is found in the promoter region of the gene associated with the probe set; otherwise, the value
0 is assigned.
Probe-set id
Control Zebul Remeth Zebul/Control Zebul/Remeth Control
(1⫹2) (1⫹2) (1⫹2)
(1⫹2)
(1⫹2)
(3)
Zebul
(3)
Remeth Zebul/Control Zebul/Remeth
(3)
(3)
(3)
Descriptions
CpG
NM㛭022568.1 aldehyde
dehydrogenase 12 (ALDH12)
ESTs, moderately similar to cytokine
receptor-like factor 2; cytokine
receptor CRL2 precursor
NM㛭000516 GNAS complex locus
(GNAS)
NM㛭006026 H1 histone family,
member X (H1FX)
NM㛭006191 proliferation-associated
2G4, 38 kD (PA2G4)
NM㛭023080 hypothetical protein
FLJ20989
NM㛭002896 RNA binding motif
protein 4 (RBM4)
NM㛭017858 hypothetical protein
FLJ20516 (FLJ20516)
NM㛭002394 Homo sapiens solute
carrier family 3 (activators of
dibasic and neutral amino acid
transport), member 2 (SLC3A2)
0
A. Gene expression in U133A-chip probe sets
220148㛭at
25.1
138.5
30.3
5.5
4.6
9.3
113.3
40.7
12.2
2.8
221988㛭at
116.4
332.6
150.9
2.9
2.2
163.3
354.9
176.8
2.2
2.0
214157㛭at
735.9
1973.8
703.7
2.7
2.8
615.6
1340.8
607.0
2.2
2.2
1770.0
4620.6
2095.4
2.6
2.2
1366.3
4270.0
1691.1
3.1
2.5
552.4
1308.3
452.9
2.4
2.9
605.6
1446.7
722.7
2.4
2.0
218187㛭s㛭at
2521.0
5591.0
2560.4
2.2
2.2
2478.9
6299.1
2503.3
2.5
2.5
200997㛭at
1148.5
2403.0
1114.3
2.1
2.2
1195.0
2684.4
973.4
2.2
2.8
219258㛭at
413.6
862.7
383.1
2.1
2.3
229.1
659.3
290.3
2.9
2.3
200924㛭s㛭at
868.4
1767.4
872.7
2.0
2.0
945.8
2279.4
825.0
2.4
2.8
232165㛭at
227474㛭at
228.5
454.5
1266.9
1306.9
384.8
508.4
5.5
2.9
3.3
2.6
204805㛭s㛭at
214794㛭at
1
1
1
1
1
1
1
1
B. Gene expression in U133B-chip probe sets
232.7
653.6
2044.0
1518.2
to probe set 205933_at. The expression fractions were: 1.12/1.16
(experiments 1 ⫹ 2) and 1.43/1.56 (experiment 3).
RT-PCR. To confirm data obtained in the microarrays, semiquantitative reverse-transcription PCRs (RT-PCRs) were carried out for
nine probe sets (Fig. 3 and Table 6). We found up-regulation of all
chosen genes in cells treated with zebularine, and the relative change
correlated well with the microarray results; therefore, microarray
analysis alone is reliable, at least for initial studies, to globally assess
the affects of zebularine on gene expression. The most pronounced
regulation by zebularine was found for aldehyde dehydrogenase 12
(ALDH12), a gene that metabolizes retinoic acid to its active retinoicacid receptor-binding form. All of the other results from RT-PCR and
microarrays indicate an up-regulation of ⬃2–3-fold on zebularine
treatment.
55.4
641.8
8.8
2.3
36.9
2.4
NM㛭031308 Epiplakin 1 (EPPK1)
Clone CDABP0036
0
0
RNase Protection Assay. We noted some effects of demethylation
on the expression levels of several apoptotic related genes in the
microarrays and used the RNase protection assay, with probes associated with apoptotic markers, to confirm our microarray data. As
shown in Fig. 4, up-regulation of BCL-W, BAD, BAK, and BAX by
demethylation was detected, whereas there was no change in BCL-X
and MCL-1 expression. This result was consistent with our microarray
data and reinforced the validity of the results. BCL-W expression
values induced by drug versus control, and drug versus remethylated
fractions were 1.15/1.36 (experiments 1 ⫹ 2) and 1.29/1.28 (experiment 3), respectively. The BAD gene corresponds to three probe sets,
one of which remained unchanged in both experiments, another that
was not detected in any of the experiments, and a third that, although
not detected in experiments 1 ⫹ 2, showed up-regulation in experi-
Table 5 Gene function of up-regulated genes
Each probe set found to be up-regulated by zebularine was assigned a main function based on Gene Ontology (GO) Consortium annotations (53), GeneBank and UniGene
annotations, publications, and bioinformatics prediction tools, such as SMART. “(1⫹2)” indicates the combined results from experiments 1 and 2 (using the same RNAs) and “(3)”
corresponds to experiment 3 (using RNAs obtained in an independent experiment). The second line, GO annotations, indicates the number and percentage of probe sets with any GO
annotation, according to the Affymetrix download of July 2003 (updated every 3 months).
Microarray
U133A (1⫹2)
U133A (3)
U133B (1⫹2)
U133B (3)
GO annotations
Adhesion molecule
Apoptosis
Chromatin (DNA repair; replication; methylation; acetylation)
Intracellular signaling
Metabolism
Proliferation
Chaperone
Proteolysis (proteolysis inhibition; ubiquitin pathway)
RNA-associated (splicing; ribosomal proteins; snRNAs)
Structural
Transcription factor (transcription factor binding protein;
transcription factor inhibitor)
Transporter
Other
27 (64.3%)
0
1 (2.9%)
1 (2.9%)
6 (17.6%)
2 (5.9%)
4 (11.8%)
2 (5.9%)
1 (2.9%)
6 (17.6%)
0
4 (11.8%)
86 (66.7%)
4 (4%)
2 (2%)
12 (12%)
13 (13%)
14 (14%)
9 (9%)
4 (4%)
4 (4%)
8 (8%)
9 (9%)
14 (14%)
7 (30.4%)
1 (6.3%)
1 (6.3%)
0
2 (12.6%)
2 (12.6%)
0
0
2 (12.6%)
1 (6.3%)
4 (25%)
1 (6.3%)
61 (68.5%)
0
2 (5%)
4 (10%)
6 (15%)
4 (10%)
4 (10%)
1 (2.5%)
1 (2.5%)
1 (2.5%)
3 (7.5%)
8 (20%)
0
2 (12.6%)
0
6 (15%)
2 (5.9%)
5 (14.7%)
3 (3%)
4 (4%)
3469
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 2004 American Association for Cancer
Research.
EPIGENETIC TRANSCRIPTION SILENCING IN MYELOMA CELLS
gene. As for MCL-1, corresponding to six probe-sets, there was no
regulation in five sets and no detection in the last.
DISCUSSION
Fig. 3. Semiquantitative reverse transcription-PCR for microarray genes up-regulated
by zebularine. RNAs from control, zebularine-treated, and remethylated cells were extracted and used as templates for the synthesis of cDNAs (the RNA was the same as that
used in microarray experiments 1 and 2). After normalization of cDNA concentrations,
PCR reactions were carried out with primers designed to hybridize with different exons of
each gene. The number of cycles and PCR conditions were optimized to obtain semiquantitative data (see “Materials and Methods”). The PCR reaction products for glyceraldehyde 3-phosphate (GAPDH) are shown as control for equal total cDNA template
added to each PCR reaction.
ment 3; the fractions found were 1.70/1.87. These results suggest BAD
might be alternatively spliced or merely that the probes used were not
optimum. BAK was also found to be up-regulated in experiment 3,
fractions were 1.59/1.40. Isoform ␦ of BAX was found to be upregulated in experiments 1 ⫹ 2, with fractions of 2.95/2.85.
BCL-X showed no significant change in microarray expression data
in experiments 1 ⫹ 2 and 3 for all three probe-sets representing this
Epigenetic silencing of important cell differentiation, tumor suppressor, and apoptotic genes represents a significant impediment to
cancer treatment. Here we show that a number of genes the silencing
of which could potentially lead to the development of a neoplastic
phenotype were reactivated by the use of a single dose of the methylation inhibitor zebularine. Reactivation of these genes using methylase inhibitor drugs may play an important role in cancer treatment
(14, 15). Zebularine, a compound found recently to inhibit DNA
methylation, effectively reversed the methylation status of epigenetically silenced genes in our cell model, suggesting that many tumors
probably select for methylation-induced silenced genes as they develop. To evaluate the methylation status of cellular genes, we assessed overall changes in DNA methylation by RLGS, and we systematically searched for specifically targeted genes using the
expression microarray technique. Genes of interest were then further
analyzed by RT-PCR and RNase protection assay.
The RLGS assay indicated that zebularine caused genome-wide
changes in DNA methylation. The analysis of 1360 DNA fragments
revealed that 1.7% were demethylated. The patterns of methylation in
control cells and in cells subjected to remethylation were unchanged,
indicating that the changes induced by zebularine were transient. After
drug treatment and rescue (removal of drug), KAS-6/1 cells survived
and proliferated only in the presence of IL-6 (16), suggesting that the
cytokine plays an important role in the DNA remethylation process.
Because cells rescued from drug treatment in the presence of IL-6
exhibited DNA methylation patterns similar to the untreated control
cells, IL-6 apparently plays some role in the epigenetic gene silencing
process. This role, particularly as noted by the IL-6-associated remethylation pattern, is consistent with our previous observations that
IL-6 up-regulates the transcription of human DNMT-1, an activity
essential for DNA methylation (10).
The CpG island-associated DNA fragments isolated from the
RLGS assay showed that promoter methylation was inversely correlated with gene expression, i.e., the higher the methylation, the lower
the gene expression. We decided to search for other genes regulated
by methylation using the expression microarray technique. Although
the RLGS is appropriate to show the relationship between the action
of zebularine and DNA methylation, as well as to assess DNA
methylation nonspecifically, i.e., irrespective of the genes involved or
whether the DNA is associated with coding or noncoding regions, the
microarray assay allows a quicker scanning of thousands of tran-
Table 6 Microarray and reverse transcription-PCR (RT-PCR) comparison
The bands obtained by RT-PCR were quantified by densitometry and corrected for the glyceraldehyde 3-phosphate (GAPDH) levels. The fraction between values obtained with
zebularine and control or zebularine and remethylated samples were then calculated. These fractions were then compared with those obtained in the microarray experiments [averages
of experiments (exp) 1 and 2]. Genes depicted in bold correspond to those considered significantly altered in microarray experiments 1, 2, and 3.
RT-PCR densitometry
a
Microarrays, average exp 1 and 2
Genes
Zebularine/Control
Zebularine/Remethylated
Zebularine/Control
Zebularine/Remethylated
ALDH12
AML1
BIK
ICAM1
JTV1
PA2G4
RASSF1
RNF7
SLC3A1 (upper)
SLC3A1 (lower)
10.8
2.5
3.8
2.2
2.6
1.9
4.7
2.2
2.0
1.9
14.2
2.3
3.3
2.6
2.3
1.7
2.2
2.5
5.5
Nda
12.2
3.9
3.2
2.1
2.4
2.4
1.7
2.5
2.4
2.8
3.2
2.9
2.4
2.4
2.9
1.5
2.4
2.8
Nd, not detected.
3470
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 2004 American Association for Cancer
Research.
EPIGENETIC TRANSCRIPTION SILENCING IN MYELOMA CELLS
Fig. 4. RNase protection assay for apoptosis-related genes. Probes for different
apoptosis-related genes consisting of radioactively labeled cRNAs were hybridized with
total RNA from control and from zebularine-treated and remethylated cells. On addition
of RNases specific for single-stranded RNAs, only double-stranded RNAs remained
integral. These were then run on a 6% polyacrylamide gel. After gel drying, the probes
were detected on X-ray films. Nonhybridized probes were run in parallel (Lane 4) to
localize the bands of interest. These nonhybridized probes are larger than the probes
mixed with the RNAs and submitted to RNase digestion because cRNAs don’t hybridize
completely to probes, leaving single-stranded probe regions that are shortened by RNases.
Probes for L32 and glyceraldehyde 3-phosphate (GAPDH) were added to indicate even
loading of samples on the gel.
scripts, albeit regulation cannot be directly attributed to methylation,
because it may result from an indirect effect of the drug on, for
instance, a transcription factor. Nevertheless, some sequences identified by RLGS overlapped promoters of regulated transcripts found by
using the microarray approach, thus showing that these techniques are
correlated and complementary.
Methylation usually occurs in CpG islands, defined as regions
greater than 200 bp with a CG content greater than 50% and an
observed to predicted ratio of CG greater than or equal to 0.6 (21).
Gene up-regulation by methylation inhibitors generally occurs by
demethylation of the promoters, allowing gene transcription, or by
some other unidentified, indirect effect on the gene. Indeed, although
the expected ratio of CpG islands in the promoter regions and first
exons is around 50% (22), in our microarray results, we found that
63–71% of the genes regulated by zebularine contained a CpG island
within the first 1500 bases upstream from the transcription initiation
site, indicating that zebularine preferentially targeted CpG islands.
The regulation of many genes devoid of CpG islands could be attributed to indirect regulation by transcription factors that do contain CpG
islands and that are regulated by methylation. Additionally, there are
reports that methylation may occur on cytosines adjacent to nucleotides other than Gly, particularly Ala (23), but it is not clear whether
they must be in CpG islands. Furthermore, CpG islands may also be
found in exons and introns, interfering with gene transcription, and in
mRNA splicing. Because we did not search for exon and intron CpG
islands, the 29 – 47% of zebularine-regulated genes, i.e., those without
CpG islands in their promoters, may have been directly affected by
CpG islands downstream of their promoter regions. Careful examination of our microarray data reveals many unknown open reading
frames, representing potential genes, that contain CpG islands in their
promoters. These unknown open reading frames have not yet been
assigned functional roles and are under investigation.
Although microarray analysis has the advantage of allowing the
screening of a large number of genes, the results must be verified by
another method to ensure accuracy. For this reason, several genes
were chosen for semiquantitative RT-PCR or RNA protection assay to
assess the reliability of the microarray data. The relative values
obtained by RT-PCR were very similar to those found using microarrays. Among the criteria to select the genes to be analyzed were the
following: consistency between the different microarray assay results;
the selection of genes associated with tumorigenesis, such as genes
that regulate apoptosis, cell proliferation, DNA repair, and cell differentiation; and the selection of genes associated with IL-6 or inflammation. Several genes found to be silenced after drug removal
and IL-6 addition are associated with metabolism, intracellular signaling, transcriptional regulation, and proliferation. The last three
functions are closely interconnected and associated with cell transformation and cancer. Genes controlling metabolism are usually not
prone to transcriptional regulation because most of them are “housekeeping” genes. However, some cancer cells are known to alter their
metabolic pathways to overcome difficulties such as lack of nutrients
or hypoxia, and to escape the effects of chemotherapy (24, 25).
The gene showing the greatest change in expression levels after
drug treatment was ALDH12, important in the metabolism of retinoic
acid, a known ligand for the RAR nuclear receptor, which is involved
in cell differentiation (26). There exists an inverse correlation between
hematopoietic cell differentiation and carcinogenesis, with highly
differentiated, end-stage cells rarely observed as undergoing unregulated proliferation. In contrast, the gene expression profiles of many
hematopoietic tumors are more closely compared with those of developing precursor cells, suggesting that many tumors of this type
probably arose by the clonal expansion of cells “frozen” in a partially
differentiated precursor stage (27). Given the great changes in the
levels of ALDH12 expression, it is probable that the genes regulated
by retinoic acid were involved in promoting the differentiation of the
KAS-6/1 cells. Furthermore, retinoic acid has been found to induce
apoptosis in many cell types, and suppression of its metabolism in this
myeloma cell line may promote tumor cell survival (28). Curiously,
ALDH12 was the only gene of the nine chosen for RT-PCR not to
contain CpG islands in its promoter (first 1500 bp upstream from
transcription initiation site). It may be that methylation occurs despite
this, that there are CpG islands upstream or downstream from the
region studied, or, most probably, that ALDH12 regulation is indirect,
3471
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 2004 American Association for Cancer
Research.
EPIGENETIC TRANSCRIPTION SILENCING IN MYELOMA CELLS
via the regulation of some transcription factor regulated by methylation that binds to the ALDH12 promoter.
Acute myeloid leukemia 1 (AML1), also known as runt-related
transcription factor 1 (RUNX1) is one of the transcription factors
regulated by zebularine, and is known to be an important factor in
developmental hematopoiesis. The AML1/RUNX1 gene has a tumor
suppressor activity and is mutated in several types of cancer (29).
AML1 inhibits the transcription of macrophage inflammatory protein
1 (MIP1), which is overexpressed in 70% of the patients with multiple
myeloma (30). Because KAS-6/1 cells are derived from a multiple
myeloma, it is possible that they repress AML1 expression to allow
overexpression of MIP1, which may protect cells from death and
promote their proliferation (31).
The protein encoded for by the JTV1 (p38) gene (32), which
contains a glutathione S-transferase (GST) domain, was shown to bind
damaged DNA, and act as a chaperone. It is possible that JTV1
silencing occurs because it exerts some DNA protection effect. Therefore, silencing of DNA-protective or repair proteins, such as JTV1,
may promote cancer. Furthermore, lowered expression of JTV1, with
its antioxidant glutathione S-transferase domain, may lead to an accumulation of reactive oxygen species, promoting cell proliferation or
mutagenesis.
The ring finger protein 7 (RNF7), also known as Sensitive-toApoptosis-Gene (SAG), was also confirmed by RT-PCR assays.
RNF7 also has antioxidant activity, being capable of inactivating
peroxynitrite, hydrogen, and fatty acid peroxides (33, 34). The presence of antioxidants in cancer cells may hinder their proliferation and
their mutation rate, depending on the presence of reactive oxygen
species (35), which may explain the fact that RNF7 is down-regulated
in control KAS-6/1 cells.
The intercellular adhesion molecule 1 (ICAM1), also known as
CD54, the proliferation-associated 2G4 (PA2G4), and the solute carrier family 3 member 2 (SLC3A2) genes are also greatly up-regulated
by demethylation. ICAM1 is an adhesion protein found in many cells
of the immune system and binds to integrins, which then activate
intracellular signaling pathways. This protein is highly expressed in
colorectal cancer and breast cancer and is an indicator of inflammation in the latter (36, 37). The expression of ICAM1, a p53-responsive
gene, may have been induced directly by demethylation or indirectly
by p53 transactivation, which was also reactivated in drug-treated
KAS-6/1 cells (Ref. 38 and manuscript submitted).4 Furthermore,
whereas the expression of ICAM1 mRNA in zebularine-treated cells
may seem paradoxical, it could be explained, in part, by the fact that
its background expression level in control cells was relatively high,
possibly indicating only partial methylation but still sufficient for
tumor promotion.
The multifunctional protein, PA2G4, also known as Ebp-1, was first
identified as having strong homology with a mitogen-inducible, murine cell cycle gene of the same name (39). Later work revealed that
the coding sequence of PA2G4 was identical to a protein, designated
as Ebp-1, that bound to Erb-B3, an inactive tyrosine kinase. PA2G4/
Ebp-1 has been shown to participate in the differentiation of human
ErbB receptor-positive breast and prostate cancer cells, and is capable
of inducing antiproliferative effects by virtue of its interactions with
the retinoblastoma (Rb) tumor suppressor protein. The ability of
PA2G4/Ebp-1 to complex with Rb enables the resulting heterodimer
to recruit histone deacetylases (HDACs), inhibiting the transcription
from the cyclin E promoter, an important cell cycle initiator protein
(40). Epigenetic silencing of PA2G4/Ebp-1 could affect cell cycle
regulation by impeding this activity.
Lastly, SLC3A2 is involved in cell activation, transformation, inflammation and carcinogenesis. The SLC3A2 protein, also known as
CD98, is similar to ICAM1, in that it signals through integrins (41).
CD98 functions also as a L-phenylalanine transporter, and in some
myeloma cells, its reduced expression has been correlated with resistance to the drug melphalan (42). Thus, SLC3A2/CD98 might share
similar properties with the major drug resistance protein (MDR-1),
also known as P-glycoprotein, which removes chemotherapeutic
drugs by active transport.
There are several reports on the repression of apoptosis in cancer
cells by DNA methylation (43– 46). Assuming that dependence on
IL-6 during neoplastic development occurs because of its role in
mediating gene silencing by methylation, survival in this case means,
among other things, evasion of cell death. Therefore, in cells in which
proapoptotic genes were epigenetically silenced by methylation, one
would expect an up-regulation of antiapoptotic genes by IL-6. We
analyzed by RNase protection assay the expression of some apoptosis-associated genes and correlated these results with those obtained
by microarray analysis. Genomic demethylation induced the increased
expression of BCL-W, an antiapoptotic gene functionally similar to
BCL-2. However, as assessed by microarray analysis and RNase
protection assay, zebularine induced BCL2-interacting killer (BIK), a
member of the BCL2 family that promotes cell death, and several
other death-promoting members of the BCL2 family, including the
proapoptotic proteins BAD, BAK, and BAX (47). This increase in
proapoptotic BCL-2 family members may reflect a combination of
drug-mediated reactivation of other apoptotic inducers such as p53,
and the cytotoxic effects of zebularine. Despite the increase in one
antiapoptotic protein, the increased expression of four proapoptotic
proteins indicates that this multiple myeloma cell line is vulnerable to
the removal of methylation-induced epigenetic silencing of apoptotic
genes. Because IL-6-mediated resistance to apoptosis is a significant
problem in multiple myeloma, the regulation of these apoptotic genes
represents an interesting correlation between apoptosis and epigenetic
regulation of gene expression.
Several other genes, previously identified as regulated by methylation in cancer are compiled on the website “Genes Affected by
Promoter CpG Island Methylation in Aging and/or Cancer.”5 This
collection contains 66 genes, 35 of which are present in the U133
series of Affymetrix microarrays, corresponding to 212 probe sets.
The comparison of the microarray data with the aforementioned
collection reveals no gene up-regulated by zebularine by a minimum
factor of 2. Only RASSF1 “Ras association (RalGDS/AF-6) domain
family 1” was found to be regulated by a minimum factor of 1.5,
considering all three experiments. RASSF1 is a gene similar to the
RAS genes in that it exerts tumor suppressor effects by interacting
with the repair protein XPA (48) and by inhibiting the accumulation
of cyclin D1 (49). The regulation of RASSF1 was confirmed by
RT-PCR. The RASSF1 gene has been reported to be epigenetically
regulated in many cancers, including multiple myeloma, thus demonstrating that its epigenetic regulation is an important factor in the
development and maintenance of neoplasia (50, 51).
We have presented an initial exploration into the effects of zebularine on gene expression, using a cellular model in which gene
resilencing can be achieved by a pro-inflammatory cytokine, IL-6.
The data shown here confirms the effects of epigenetic silencing of
important cell-regulatory genes that can be reversed by treatment with
DNA methylation inhibitors such as zebularine, as verified by RLGS,
microarray, RT-PCR, and RNase protection assay analysis. Also
revealed is the requirement for IL-6 in the post-drug recovery phase
and its ability to restore epigenetic gene silencing. Because remethylation/resilencing of the genes examined, and subsequent cell survival,
correlates with IL-6 treatment, we have demonstrated an important
5
Internet address: http://www3.mdanderson.org/leukemia/methylation/cgi.html.
3472
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 2004 American Association for Cancer
Research.
EPIGENETIC TRANSCRIPTION SILENCING IN MYELOMA CELLS
effect of IL-6 on the survival of multiple myeloma cells, and the
reestablishment of epigenetic gene silencing.
ACKNOWLEDGMENTS
We thank Dr. Joost Oppenheim for critical review of the manuscript; Dr.
Dianne Jelinek, of the Mayo Clinic, Rochester, MN, for kindly providing the
KAS-6/1; and Suneetha Betsy Thomas, for technical aid.
REFERENCES
1. Ryan BM, Russel MG, Langholz E, Stockbrugger RW. Aminosalicylates and colorectal cancer in IBD: a not-so bitter pill to swallow. Am J Gastroenterol 2003;98:
1682–7.
2. Schwartsburd PM. Chronic inflammation as inductor of pro-cancer microenvironment: pathogenesis of dysregulated feedback control. Cancer Metastasis Rev 2003;
22:95–102.
3. Genta RM. The gastritis connection: prevention and early detection of gastric neoplasms. J Clin Gastroenterol 2003;36(5 Suppl):S44 –9; discussion S61–2.
4. Heinrich PC, Behrmann I, Haan S, Hermanns HM, Muller-Newen G, Schaper F.
Principles of interleukin (IL)-6-type cytokine signalling and its regulation. Biochem
J 2003;374:1–20.
5. Kato H, Kinoshita T, Suzuki S, et al. Production and effects of interleukin-6 and other
cytokines in patients with non-Hodgkin’s lymphoma. Leuk Lymphoma 1998;29:
71–9.
6. Kovalchuk AL, Kim JS, Park SS, et al. IL-6 transgenic mouse model for extraosseous
plasmacytoma. Proc Natl Acad Sci USA 2002;99:1509 –14.
7. Maione D, Di Carlo E, Li W, et al. Coexpression of IL-6 and soluble IL-6R causes
nodular regenerative hyperplasia and adenomas of the liver. EMBO J 1998;17:
5588 –97.
8. Geiman TM, Robertson KD. Chromatin remodeling, histone modifications, and DNA
methylation— how does it all fit together? J Cell Biochem 2002;87:117–25.
9. Hodge DR, Li D, Qi SM, Farrar WL. IL-6 induces expression of the Fli-1 protooncogene via STAT3. Biochem Biophys Res Commun 2002;292:287–91.
10. Hodge DR, Xiao W, Clausen PA, Heidecker G, Szyf M, Farrar WL. Interleukin-6
regulation of the human DNA methyltransferase (HDNMT) gene in human erythroleukemia cells. J Biol Chem 2001;276:39508 –11.
11. Nephew KP, Huang TH. Epigenetic gene silencing in cancer initiation and progression. Cancer Lett 2003;190:125–33.
12. Karpf AR, Jones DA. Reactivating the expression of methylation silenced genes in
human cancer. Oncogene 2002;21:5496 –503.
13. Goffin J, Eisenhauer E. DNA methyltransferase inhibitors—state of the art. Ann
Oncol 2002;13:1699 –716.
14. Driscoll JS, Marquez VE, Plowman J, Liu PS, Kelley JA, Barchi, JJ Jr. Antitumor
properties of 2(1H)-pyrimidinone riboside (zebularine) and its fluorinated analogues.
J Med Chem 1991;34:3280 – 4.
15. Zhou L, Cheng X, Connolly BA, Dickman MJ, Hurd PJ, Hornby DP. Zebularine: a
novel DNA methylation inhibitor that forms a covalent complex with DNA methyltransferases. J Mol Biol 2002;321:591–9.
16. Westendorf JJ, Ahmann GJ, Greipp PR, Witzig TE, Lust JA, Jelinek DF. Establishment and characterization of three myeloma cell lines that demonstrate variable
cytokine responses and abilities to produce autocrine interleukin-6. Leukemia (Baltimore)1996;10:866 –76.
17. Hatada I, Hayashizaki Y, Hirotsune S, Komatsubara H, Mukai T. A genomic scanning
method for higher organisms using restriction sites as landmarks. Proc Natl Acad Sci
USA 1991;88:9523–7.
18. Smiraglia DJ, Fruhwald MC, Costello JF, et al. A new tool for the rapid cloning of
amplified and hypermethylated human DNA sequences from restriction landmark
genome scanning gels. Genomics 1999;58:254 – 62.
19. Zacharski LR. Anticoagulants in cancer treatment: malignancy as a solid phase
coagulopathy. Cancer Lett 2002;186:1–9.
20. Liang G, Gonzales FA, Jones PA, Orntoft TF, Thykjaer T. Analysis of gene induction
in human fibroblasts and bladder cancer cells exposed to the methylation inhibitor
5-aza-2⬘-deoxycytidine. Cancer Res 2002;62:961– 6.
21. Gardiner-Garden M, Frommer M. CpG islands in vertebrate genomes. J Mol Biol
1987;196:261– 82.
22. Antequera F, Bird A. Number of CpG islands and genes in human and mouse. Proc
Natl Acad Sci USA 1993;90:11995–9.
23. Dodge JE, Ramsahoye BH, Wo ZG, Okano M, Li E. De novo methylation of MMLV
provirus in embryonic stem cells: CpG versus non-CpG methylation. Gene (Amst.)
2002;289:41– 8.
24. Mathupala SP, Rempel A, Pedersen PL. Glucose catabolism in cancer cells: identification and characterization of a marked activation response of the type II hexokinase
gene to hypoxic conditions. J Biol Chem 2001;276:43407–12.
25. Younes M, Lechago LV, Somoano JR, Mosharaf M, Lechago J. Wide expression of
the human erythrocyte glucose transporter Glut1 in human cancers. Cancer Res
1996;56:1164 –7.
26. Lin M, Napoli JL. cDNA cloning and expression of a human aldehyde dehydrogenase
(ALDH) active with 9-cis-retinal and identification of a rat ortholog, ALDH12. J Biol
Chem 2000;275:40106 –12.
27. Matsui WH, Huff CA, Wang Q, et al. Characterization of clonogenic multiple
myeloma cells. Blood 2004;103:2332– 6. Epub 2003 Nov 20.
28. Zheng A, Savolainen ER, Koistinen P. All-trans retinoic acid induces apoptosis in
acute myeloblastic leukaemia cells. Apoptosis 1997;2:319 –29.
29. Silva FP, Morolli B, Storlazzi CT, et al. Identification of RUNX1/AML1 as a classical
tumor suppressor gene. Oncogene 2003;22:538 – 47.
30. Choi SJ, Oba T, Callander NS, Jelinek DF, Roodman GD. AML-1A and AML-1B
regulation of MIP-1alpha expression in multiple myeloma. Blood 2003;101:
3778 – 83.
31. Lentzsch S, Gries M, Janz M, Bargou R, Dorken B, Mapara MY. Macrophage
inflammatory protein 1-alpha (MIP-1 alpha) triggers migration and signaling cascades
mediating survival and proliferation in multiple myeloma (MM) cells. Blood 2003;
101:3568 –73.
32. Nicolaides NC, Kinzler KW, Vogelstein B. Analysis of the 5⬘ region of PMS2 reveals
heterogeneous transcripts and a novel overlapping gene. Genomics 1995;29:329 –34.
33. Kim SY, Lee JH, Yang ES, Kil IS, Park JW. Human sensitive to apoptosis gene
protein inhibits peroxynitrite-induced DNA damage. Biochem Biophys Res Commun
2003;301:671– 4.
34. Sun Y, Tan M, Duan H, Swaroop M. SAG/ROC/Rbx/Hrt, a zinc RING finger gene
family: molecular cloning, biochemical properties, and biological functions. Antioxid
Redox Signal 2001;3:635–50.
35. Halliwell BG, Gutteridge MC. Free radicals in biology and medicine. 3rd ed. Oxford:
Oxford University Press; 2001. p. 936
36. Maeda K, Kang SM, Sawada T, et al. Expression of intercellular adhesion molecule-1
and prognosis in colorectal cancer. Oncol Rep 2002;9:511– 4.
37. Blann AD, Byrne GJ, Baildam AD. Increased soluble intercellular adhesion molecule-1, breast cancer and the acute phase response. Blood Coagul Fibrinolysis
2002;13:165– 8.
38. Arnold JM, Cummings M, Purdie D, Chenevix-Trench G. Reduced expression of
intercellular adhesion molecule-1 in ovarian adenocarcinomas. Br J Cancer 2001;85:
1351– 8.
39. Radomski N, Jost E. Molecular cloning of a murine cDNA encoding a novel protein,
p38 –2G4, which varies with the cell cycle. Exp Cell Res 1995;220:434 – 45.
40. Zhang Y, Woodford N, Xia X, Hamburger AW. Repression of E2F1-mediated
transcription by the ErbB3 binding protein Ebp1 involves histone deacetylases.
Nucleic Acids Res 2003;31:2168 –77.
41. Rintoul RC, Buttery RC, Mackinnon AC, et al. Cross-linking CD98 promotes
integrin-like signaling and anchorage-independent growth. Mol Biol Cell 2002;13:
2841–52.
42. Harada N, Nagasaki A, Hata H, Matsuzaki H, Matsuno F, Mitsuya H. Downregulation of CD98 in melphalan-resistant myeloma cells with reduced drug uptake.
Acta Haematol 2000;103:144 –51.
43. Virmani A, Rathi A, Sugio K, et al. Aberrant methylation of TMS1 in small cell, non
small cell lung cancer and breast cancer. Int J Cancer 2003;106:198 –204.
44. Deocampo ND, Huang H, Tindall DJ. The role of PTEN in the progression and
survival of prostate cancer. Minerva Endocrinol 2003;28:145–53.
45. Nakatsuka S, Takakuwa T, Tomita Y, et al. Hypermethylation of death-associated
protein (DAP) kinase CpG island is frequent not only in B-cell but also in T- and
natural killer (NK)/T-cell malignancies. Cancer Sci 2003;94:87–91.
46. Teitz T, Lahti JM, Kidd VJ. Aggressive childhood neuroblastomas do not express
caspase-8: an important component of programmed cell death. J Mol Med 2001;79:
428 –36.
47. Zou Y, Peng H, Zhou B, et al. Systemic tumor suppression by the proapoptotic gene
bik. Cancer Res 2002;62:8 –12.
48. Dammann R, Li C, Yoon JH, Chin PL, Bates S, Pfeifer GP. Epigenetic inactivation
of a RAS association domain family protein from the lung tumour suppressor locus
3p21.3. Nat Genet 2000;25:315–9.
49. Shivakumar L, Minna J, Sakamaki T, Pestell R, White MA. The RASSF1A tumor
suppressor blocks cell cycle progression and inhibits cyclin D1 accumulation. Mol
Cell Biol 2002;22:4309 –18.
50. Ng MH, Lau KM, Wong WS, et al. Alterations of RAS signalling in Chinese multiple
myeloma patients: absent BRAF and rare RAS mutations, but frequent inactivation of
RASSF1A by transcriptional silencing or expression of a non-functional variant
transcript. Br J Haematol 2003;123:637– 45.
51. Horiguchi K, Tomizawa Y, Tosaka M, et al. Epigenetic inactivation of RASSF1A
candidate tumor suppressor gene at 3p21.3 in brain tumors. Oncogene 2003;22:
7862–5.
52. Ponger L, Mouchiroud D. CpGProD: identifying CpG islands associated with transcription start sites in large genomic mammalian sequences. Bioinformatics 2002;18:
631–3.
53. Creating the gene ontology resource: design and implementation. Genome Res
2001;11:1425–33.
3473
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 2004 American Association for Cancer
Research.
Microarray Analysis of Epigenetic Silencing of Gene
Expression in the KAS-6/1 Multiple Myeloma Cell Line
Celine Pompeia, David R. Hodge, Christoph Plass, et al.
Cancer Res 2004;64:3465-3473.
Updated version
Cited articles
Citing articles
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/64/10/3465
This article cites 52 articles, 19 of which you can access for free at:
http://cancerres.aacrjournals.org/content/64/10/3465.full#ref-list-1
This article has been cited by 15 HighWire-hosted articles. Access the articles at:
http://cancerres.aacrjournals.org/content/64/10/3465.full#related-urls
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 2004 American Association for Cancer
Research.