Download Human Tumor Cell Lines

[CANCER RESEARCH58. 95-101, January 1, 19981
Inhibition of DNA Methylation by 5-Aza-2'-deoxycytidine
Suppresses the Growth of
Human Tumor Cell Lines'
Christina M. Bender, Martha M. Pao, and Peter A. Jones2
Urological Cancer Research Laboratory, USC/Norris Comprehensive Cancer Center, Unis'ersity of Southern California School of Medicine, ii)s Angeles, California 90033
ABSTRACT
have been shown to coexist during oncogenic transformation (9, 10),
and de novo methylation
Alterations in DNAmethylationpatterns accompanythe establishment
of immortal cell lines. De novo methylation of CpG islands within the
control regions ofgrowth-regulatory
genes may Inactivate their transcrip
tion, giving cells selective growth advantages in culture. We exposed seven
human
tumor
cell
lines
and
two
human
fibroblast
cell
strains
to the
demethylating agent, 5-aza-2'-deoxycytidine (5-Aza-CdR), to determine
whether the silencing of growth-regulatory genes by de novo methylation
in immortalized cell lines could be reversed, possibly restoring growth
control. After recovery from the Immediate cytotoxic effects of 5-Aza
CdR, this agent suppressed
cellular growth in all seven tumor lines but not
in either fibroblast strain. Because alterations In the p16 (CDKN2/MTSJ)
cell cycle regulatory
gene are associated
with numerous
cancers,
we
analyzed expression of this gene before and after S-Aza-CdR treatment.
The gene was reactivated by 5-Aza-CdR treatment in three of four tumor
cell lines not expressing p16, whereas the fourth tumor line contained a
p16 homozygous deletion. pitS was shown to be hypermethylated only in
the cell lines and its up-regulation
by 5.Aza-CdR
was associated
with
demethylation of thepl6 promoter. The remaining tumor lines expressed
p16 at constant levels before and after S-Aza-CdR treatment and showed
minimal pi6 promoter methylation, suggesting that other growth-regula
tory genes may have been silenced by de novo methylation in these cells.
p16 expression, cell growth inhibition, and G1 cell cycle arrest by S-Aza
CdR in the T24 bladder tumor cell line were also heritable after prolonged
passage In culture.
Furthermore,
a dormant
pitS gene was reactivated
in
T24 cells growing in nWnu rats, and 5-Aza-CdR treatment of T24 cells
before inoculation into nWnu mice decreased the rate of tumor growth.
These results suggest that S-Aza-CdR
may slow the growth of tumor cells
by reactivating growth-regulatory genes silenced by de nova methylation.
INTRODUCTION
Tumorigenesis and increased cellular proliferation arise via the
accumulation of specific genetic and epigenetic changes that provide
tumor cells with selective growth advantages over neighboring normal
cells. Identifying molecular alterations (including mutation, deletion,
and abnormal methylation) associated with a particular cancer will
facilitate the early detection of tumors and the development of ther
apies that target these specific alterations. Aberrant DNA methylation
patterns accompany the neoplastic process in vivo (1, 2), and de novo
methylation of CpG islands is observed during the establishment of
immortal cell lines in vitro (3, 4).
DNA methylation normally occurs at cytosine residues within CpG
dinucleotides which are represented at a lower than expected fre
quency throughout the eukaryotic genome (5, 6). In contrast, regions
of the genome defined as CpG islands contain the predicted frequency
of CpG dinucleotides that are usually unmethylated, except for those
on the inactive X chromosome and parentally imprinted genes (7, 8).
Global hypomethylation and regional CpG island hypermethylation
@
of DNA has been demonstrated
in cancer
derived cell lines ( 11) and uncultured tumors ( 12, 13). This epigenetic
modification may increase cell proliferation via hypermethylation of
growth-regulatory genes, especially because de novo methylation
within the regulatory sequences of tumor suppressor genes has been
proposed to silence their transcription in specific cancers. For exam
ple, hypermethylation of the RBJ gene has been observed in retino
blastomas (14, 15), and transcriptional silencing has been associated
with hypermethylation of the VHL gene in sporadic renal cell card
nomas (16), the estrogen receptor gene in breast cancers ( 17), the p15
(INK4B/MTS2) gene in leukemias (18), the p16 gene in human tumor
cell lines (19, 20), and the H19 gene in Wilms' tumor (21). Genes
silenced by hypermethylation can be reactivated by 5-Aza-CR3 or
5-Aza-CdR (22), which inhibit DNA methylation and are also effec
tive antileukemic agents (23, 24).
In this report, we analyzed p16 activation because this gene is
altered in numerous cancers. The p16 gene on chromosome 9p2 1
encodes a Mr 16,000 cell cycle regulatory protein that belongs to the
cyclin-dependent kinase inhibitory protein family and regulates the
G1/S-phase cell cycle transition, preventing phosphorylation of the
retinoblastoma gene product (25). The p16 gene is especially inter
esting because it is frequently lost via homozygous deletion or loss of
heterozygosity in many human malignancies (26, 27), and its pro
moter and first two exons have the characteristics of CpG islands.
Evidence for pitS hypermethylation has also been documented in
several human cancers including: head and neck squamous cell car
cinoma (28), non-small cell lung cancer (28), bladder transitional cell
carcinoma (20), gliomas (29), and nasopharyngeal carcinoma (30).
p16 reactivation by 5-Aza-CdR has also been demonstrated in vitro
(19, 28, 31). Thus, methylation ofthe 5' regulatory region ofp/6 may
represent an alternative means for its inactivation besides intragenic
mutation or homozygous deletion.
We investigated cellular growth suppression in various human
tumor cell lines following treatment with 5-Aza-CdR after complete
recovery from the immediate cytotoxic effects of this agent. 5-Aza
CdR suppressed cellular growth in all seven tumor cell lines analyzed
but not in human fibroblasts treated under the same conditions. p16
expression levels were measured before and after treatment to deter
mine whether up-regulation of this gene was associated with cell
growth inhibition in the tumor cell lines. 5-Aza-CdR induced pitS
expression in three tumor cell lines not expressing this gene, but the
three remaining tumor cell lines expressed p16 at steady levels before
and after
5-Aza-CdR
treatment,
implicating
that growth-regulatory
genes outside the pitS pathway might have been silenced by de novo
methylation in these cells. The pitS promoter was also shown to be
hypermethylated only in the three cell lines not expressing the gene,
and significant demethylation of this region was associated with gene
activation by 5-Aza-CdR. In the T24 bladder carcinoma-derived cell
line, pitS expression, cellular growth suppression, and 0 arrest were
heritable for up to 10 passages following 5-Aza-CdR treatment. A
Received 9/16/97; accepted 10/29/97.
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
18U.S.C.Section1734solelyto indicatethis fact.
I Supported
by USPHS
2 To
requests
whom
Grant
for
reprints
R37
CA49758
should
be
from
the National
addressed,
Urological
Cancer
Institute.
Cancer
3 The
Research
abbreviations
used
are:
5-Aza-CR,
5-aza-deoxycytidine;
5-Aza-CdR,
5-Aza-2'-
deoxycytidine; Ara-C, l-@-D-arabinofuranosylcytosine; RT-PCR, reverse transcription
Laboratory, USC/Norris Comprehensive Cancer Center, University of Southern California
School of Medicine, 1441 Eastlake Avenue, Los Angeles, California 90033. Phone:
(213) 764-0816; Fax: (213) 764-0102.
PCR; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; Ms-SNuPE. methylation
sensitive single nucleotide extension assay.
95
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1998 American Association for Cancer Research.
INHIBITION OF DNA METhYLATION RESTORES GROWTH CONTROL
Determination
dormant p16 gene was also reactivated in T24 cells growing in nu/nu
rats, and T24 cells treated with 5-Aza-CdR before introduction into
nu/nu
mice formed
tumors
that were significantly
smaller
of Cell Doubling Time
Cells (10―/60-mm
dish) were plated approximately 9 days after 5-Aza-CdR
treatment, and the cell number/dish was counted with a Coulter counter each
than those
induced by untreated T24 cells. These results suggest cancer therapies
in which growth-regulatory genes (like p16) silenced by de novo
methylation are reactivated by 5-Aza-CdR, effectively suppressing the
growth of human tumor cells.
day for 5 consecutive
days. Untreated
cells were analyzed
under similar
conditions as a control. The average cell number from two plates was deter
mined, and the mean cell numbers were plotted to define the cell population
doubling times.
RT-PCR
MATERIALS
AND
METHODS
Total RNA was isolated from 2 x 106 cells lysed in 2 ml of buffer
Cell Lines
containing guanidine isothiocyanate
T24, J82, and 5637 (bladder transitional carcinoma-derived
cell lines) were
tific, Fair Lawn, NJ) and 2-mercaptoethanol
obtained from the American Type Culture Collection (Rockville, MD). HCT15
and HCT1 16 (colon carcinoma-derived
@
cell lines) were also obtained
Inc., Palo Alto,
(0.1 M; Sigma). RNA was pre
cipitated (I h at —20°C)
in 50% isopropanol/50% lysis buffer following
from the
standard phenol-chloroform extraction of the cell lysate. After centrifugation
(10 mm at 10,000 x g), the supernatant was decanted, and the RNA pellet was
American Type Culture Collection. SK-Mel-173 and SK-Mel-28 (melanoma
cell lines) were obtained from Memorial Sloan Kettering (New York, NY).
LD98 (human fibroblast cell strain) was established in our laboratory, and 1-1
(human fibroblast cell strain) was obtained from Children's Hospital (Los
Angeles, CA). T24, SK-Mel-l73, SK-Mel-28, and LD98 were cultured in
DMEM supplemented with 10% FCS and 5% penicillin/streptomycin. J82,
5637, HCT15, HCT1 16, and T-l were cultured in MEM supplemented
(4 M; Life Technologies,
CA), N-laurylsarcosine(0.5%;Sigma), sodiumcitrate(25 mM;FisherScien
washed twice in 70% ethanol prepared with diethylpyrocarbonate-treated
double-distilled
water. The RNA pellet was dissolved in 100% diethylpyro
carbonate water. Two
of total RNA was reverse-transcribed using random
hexamers, deoxynucleotide
triphosphates
(Boehringer
Mannheim, Germany),
and Superscript II reverse transcnptase (Life Technologies, Inc., Palo Alto,
CA) in a 25-@tlreactionas described(20). cDNA was amplifiedwith primers
with
10% FCS, 5% penicillin/streptomycin, nonessential amino acids, and sodium
specific for either p16 or GAPDH. pitS and GAPDH PCR primer sequences,
pyruvate.
probe sequences, and PCR conditions were used as described (20). PCRs were
performed with cDNA template concentration
equivalent to 100 ng of RNA in
5-Aza-CdR and Am-C Treatments
10% DMSO. All reactions were analyzed in the linear range of amplification.
PCR products were resolved on 2% agarose gels and subsequently transferred
Cell Lines and Cell Strains. Cells were plated (3 X l0@cells/lOO-mm to a nylon membrane (Zetaprobe; Bio-Rad, Richmond, CA) under alkaline
dish) and treated 24 h later with 5 X l0@ M, l0@ M, 5 X l0@ M, or lO_6 M
conditions. All blots were hybridized with digoxigenin-labeled internal oligo
5-Aza-CdR or l0@ M Ara-C (Sigma Chemical Co., St. Louis, MO). The
nucleotide probes (Genius Hybridization Method; Boehringer Mannheim).
medium was changed 24 h after drug treatment and every subsequent 3 days.
RNA and DNA were isolated 9 days after treatment as described (20).
Determination
Nude Rats. T24 cells (3 X 106)were injectedvia laparotomydirectlyinto
the bladder walls of4-week-old nw'nurats with a 27-gauge needle. 5-Aza-CdR
treatments
of Cell Cycle Profile
Cells (2 x 105/lOO-mm dish) were plated and treated with i0@ M or
5 X l0@ M 5-Aza-CdR or l0@ M Ara-C. Cells were fixed after 7 days with
were initiated 4—6 weeks later, allowing time for macroscopic
bladder tumors to form. Each animal received one daily i.p. injection of 150,
75% ethanol, treated with RNasin (20
300, 600, or 900
iodide (50 p.g/ml; Sigma). DNA content at each cell cycle stage was deter
@gof 5-Aza-CdR
for 5 days. Animals
were sacrificed
24 h
after the last treatment, and total bladder RNA was isolated to analyze pitS
expression.
Nude Mice. T24 cells (2 X 106) were injecteds.c. into the rightand left
flanks of 4-week-old
@g/ml),and stained with propidium
mined via flow cytometry.
Determination
of Cytotoxicity
nu/nu mice 11 days after treatment with 5-Aza-CdR
(5 X l0@ M) in vitro. The cells were passed once in culture to facilitate
recovery from the cytotoxic effects of the drug before harvesting for injection.
T24 cells
Untreated T24 cells were injected s.c. into individual mice under the same
conditions as a control. Tumor sizes were measured 4 weeks after injection
using a caliper, and total tumor RNA was extracted
(200/60-mm
dish)
were
plated
in triplicate
sets for a colony
formation assay. Cells were treated with l0@ M, 5 X l0@ M, or lO_6 M
5-Aza-CdR or l0@ MAra-C 24 h later. Once cell colonies were visible (after
7 days), cells were fixed in 100% methanol and stained with 10% Giemsa
stain. The percentage cell survival was assessed by dividing the mean colony
number on the treated plates divided by the mean colony number on the
to analyze p16 expression
via RT-PCR.
untreated plates X 100.
Quantitation
ofpl6
Promoter
Methylation
RESULTS
p16 promoter methylation was measured using a Ms-SNuPE assay as
described (32). All cell lines and cell strains were treated with 5 X i0@ M
5-Aza-CdR, and DNA was harvested 9 days after treatment (by standard
proteinase K digestion, phenol/chloroform extraction, and ethanol precipitation
Effects
procedures). Genomic DNA was treated with sodium bisulfite as described by
Frommer's group (33, 34) to convert unmethylated cytosines to uracil, leaving
5-methylcytosine unchanged, and a region of the pitS 5' promoter was ampli
fled using PCR primers specific for bisulfite-converted DNA as described by
Gonzalgo and Jones (32). Single nucleotide primer extension was performed
using
internal
terminating
primers
immediately
specific
for the region
amplified,
with each primer
5' of the CpG site to be assayed. The methylation
status of three CpG sites were analyzed, residing in a region approximately
200-bp upstream from exon 1 in the p16 promoter. The internal primer
sequences, specific CpG locations, Ms-SNuPE extension conditions, and meth
ods for detecting 32P-labeleddCTP or dTTP incorporation (to quantitate the
percentage
methylation
per CpG)
are described
by Gonzalgo
and Jones
of 5-Aza-CdR
on Cell Growth
Inhibition,
p16 Expres
sion, andpl6 Promoter Methylation in Human Tumor Cell Lines.
During the establishment of non tumorigenic or tumorigenic cell lines,
hypermethylation within the 5' CpG islands of growth-regulatory
genes may cause cells to proliferate more rapidly, giving these cells a
selective advantage in culture (3, 4). Various tumor cell lines were,
therefore, treated with 5-Aza-CdR to determine whether this inhibitor
of DNA methylation could restore growth control to tumor cell lines
or alter the growth characteristics of diploid fibroblasts (Fig. 1). After
recovery from immediate drug cytotoxicity, significant growth mlii
bition was observed with all seven tumor lines treated with 5 X iO@
M 5-Aza-CdR
but not with human
fibroblasts.
Thus tumor
cells, but
not normal cells, were responsive to the growth-suppressive effects of
this DNA methylation inhibitor.
The results obtained for the expression of pitS before and after
(32).
The methylation percentage values determined for the three CpG sites were
averaged and recorded as the percentage of pi6 promoter methylation.
96
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1998 American Association for Cancer Research.
1N@B@ON OF DNA METHYLATION RESTORES GROWTH CONTROL
Fig. 1. Effects of 5-Aza-CdR on growth suppres
sion, p16 expression, and p16 methylation in tumor
Human
Time(Hrs)
Expression
Growth
Tumor
5AzaCdR
Suppressionp16 5AzaCdR
Cell LineDoubling
(+)i
(-)
(+)Fold
(-) (+)%
cell lines.Afterrecoveryfromimmediatedrugcy
T24
totoxicity (approximately 9—11days), the effects of
5 x i0@ M 5-Aza-CdR on cell growth inhibition
wereanalyzedin bladdercancer-derivedcell lines
(F24, J82, and 5637), colon cancer-derived cell lines
(HCTI5 and HCT1 16), melanoma cell lines (SK
Mel-173 and SK-Mel-28), and flbroblast cell strains
(LD98 and T-1) by comparing the doubling times of
these cells before and after dnig treatment (as de
scribed in “Materialsand Methods―).p16 expression
@
was analyzed in these cells by RT-PCR analysis
following RNA isolation approximately 9 days after
drug treatment. p16 methylation levels were also
measured in the cells before and after treatment with
5-Aza-CdR. DNA was isolated 9 days after drug
treatment, treated with sodium bisulfite, and a region
of the p16 promoter (approximately 200-bp up
p16 Promoter
Methylation
5AzaCdR
(-)
56
35.
J82
41
65
1.6
5
563721
29
501.7
1.7@J93
8
I:
5
.@
.@
7.@
0
I-.
HCTI5
48SK@MEL173
g
.3
HCT1I6
53
48
701.5
1.5IsSI90
C.)35
stream from exon 1) was amplified with PCR prim
ers specific for bisulfite-converted DNA. Methyla
tionat eachofthreespecificCpUsin thisregionwas
quantitated using Ms-SNuPE as described (32), and
the values were averaged and recorded as the per
centage of p16 promoter methylation. 5-Aza-CdR
treatment suppressed growth in all seven cell lines,
@
@
up-regulatedpi6expressionintheT24,HCT15,and
HCT116cell lines,anddecreasedp16 promoterhy
permethylation in these same three cellilnes. Growth
suppressionwas not detectedin the LD98andT-l
fibroblasts
71
C
I 6@@1
SK-MEL-2844
49
.11
2.2I
Strain
I LD98
@i
1101!6
1
Io.I8
T-1
@j__32
27
@--
5-Azu-C4Rtreatmentare also shownin Fig 1.This gene1whichwas
32
27
-@::
I..1
—@3
5
7
and RT-PCR was performed to determine whether p16 was induced
by drug treatment, The T24 bladder carcinoma-derived cell line con
tains a transcriptionally silent, hypermethylated p16 gene with a
wild-type sequence (data not shown),4 Fig 2 shows that l0@@M
5-Aza-CdR induced p16 expression in a dose-dependent manner, The
124 population doubling time increased after treatment with iO@@M
5-Aza-CdR with a dose-dependent effect at higher concentrations
(5 X l0@ M and lO_6 M), directly correlating with p16 induction,
Cells were also treated with the chemotherapeutic agent Ara-C, which
incorporates into replicating DNA but does not inhibit DNA methy
lation. Am-C (l0@ M) did not induce growth suppression in T24 cells
at a dose equitoxic to 5 X l0@ M S-Aza-CdR (data not shown),
showing that p)6 reactivation did not result from the general cyto
toxicity of 5-Aza-CdR.
selected for study because it is known to frequently undergo de novo
methylation during tumorigenesis. was up-regulated in T24, HCT1S,
and HCT1 16 cells but not in J82, 5637, and SK-Mel-28 cells or
normal fibroblasts. SK.Mel-173 cells contain a p16 homozygous
deletion, Because 5-Aza-C4R is known to be highly effective at
inducing the expression of genes inappropriately silenced by de novo
methylation (35), these results suggested that the growth-suppressive
effects of 5-Aza-CdR in cancer cells were due to the up-regulation of
p16 in the T24, HCT15, and HCT1 16 cells and possibly other growth
regulators in the four other cancer cell types@In contrast, tibroblasts
that also showed Initial sensitivity to drug toxicity (data not shown)
did not respond to treatment by showing a heritable alteration in cell
doubling time, possibly because abnormal methylation patterns may
not exist in these normal somatic cells,
Fig 1 also shows the quantitative measurements of pitS promoter
methylation immediately downstream from a region previously shown
Cell Cycle Arrest after Treatment with 3-Aza.CdR, T24, 5637,
and LD98 cells treated with 5-Aza-CdR (5 X l0@ M) were analyzed
by flow cytometry to identify any changes in their cell cycle profiles
Cells were maintained in drug-free medium and harvested two pas
sages after treatment to ensure complete recovery from the immediate
toxic effects of 5-Aza-CdR, Interestingly. exposure of T24 cells to
5-Aza-CdR resulted in a 2-fold increase in the proportion of cells in
G1 (from 34 to 64%; Fig. 3)@In contrast, 5637 cells, which express
p16, showed a more subtle increase in the proportion of cells in G1,
from 34 to 44%, after treatment with 5-Aza-CdR. LD98 fibroblasts,
which also express p16, showed a negligible change in cell cycle
profile (Fig 3), Treatment of T24 cells with an equitoxic dose of
Ara-C (l0@ M)did not increase the proportion of cells in G1 (data not
to contain putative transcription initiation sites (36)@Hypermethyla
tion of p)6 was demonstrated in the T24, HCT1@,and HCT1 16 cell
lines, and demethylation by 5-Aza-CdR was associated with the
up-regulation of p16 exclusively in these three lines, The 5637, J82,
and SK-Mel-28 cell lines, which express p16, showed low levels of
p16 methylation as cxpected@ These results suggest that the growth
suppressive effects of 5-Aza-CdR observed in these cell lines may
arise from the demethylation of other growth-regulatory genes besides
p16, which are presumably not hypermethylated in the normal fibro
blasts,
Effects of 5.Aza.CdR on p16 ExpressIon. Previous studies have
demonstrated a strong correlation between hypermethylation of the
p16 5' CpG island and transcriptional silencing of this gene in
4 M. L Gonzalgo,T. Haysahida,M. M. Pan,C. M. Bender,Y. C. Tsui,F. A. Gonzales,
uncultured tumors and cancer-derived cell lines (19, 20, 28, 29, 31).
H.0. Nguyen,T.T. Nguyen,andP. A.Jones.Theroleof DNAmethylationinexpression
of the pJQ/pIólocus in human bladder cancer cell lines, submitted for publication.
T24 cells were treated with increasing concentrations of 5-Aza-CdR.
97
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1998 American Association for Cancer Research.
INHIBITIONOF DNA METHYLATIONRESTORESGROWTHCONTROL
after 5-Aza-CdR treatment to determine whether growth suppression
and pitS expression following drug treatment were heritable in vitro.
Growth inhibition after treatment with l0@ M 5-Aza-CdR was heri
table for nine passages, but cells exposed to 5 X l0@ M 5-Aza-CdR
showed growth suppression even after the tenth passage (Fig. 4a). In
addition, RT-PCR-based expression analyses showed significant pitS
induction by both l0@ M and 5 X l0@ M 5-Aza-CdR (Fig. 4b).
Expression levels gradually diminished during each subsequent pas
sage up to the tenth passage, possibly arising from de novo methyl
ation of p16 or the selection of cells unaffected by the drug. These
observations demonstrate that p16 induction and growth suppression
by 5-Aza-CdR treatment are heritable and dose dependent, supporting
findings by Costello et a!. (29) in which 5-Aza-CdR-mediated pitS
expression was heritable in vitro. The induction of pitS by inhibition
of DNA methylation may explain why T24 cells treated with 5-Aza
CdR proliferated more slowly; however, reactivation of other growth
regulatory genes silenced by de novo methylation may have also
contributed to this change.
5AzaCdR-Mediatedp16 InductionIn Vitro
@3 4
@z
5AzaCdR(M)
oUTT@
C.)
p16mRNA
GAPDHmRNA
—0@
—0
Doubling time (h) —0
21 21 26 35 72
Fig. 2. Effects of 5-Aza-CdR on p16 expression and cell doubling times in T24 cells.
RNA was isolated from T24 cells 9 days after 5-Aza-CdR (5 X l0@ M)treatment (as
described in @Materialsand Methods―).Untreated T24 cells were analyzed under similar
conditions as a control. p16 expression was measured via RT-PCR. The cell doubling
times were also analyzed 9 days after drug treatment to allow recovery from the cytotoxic
effects of the agent. p16 expression and growth suppression were observed after treatment
Reactivation
with at least l0@ M5-Aza-CdR. GAPDH expression was measured to control for cDNA
input and integrity.
shown); therefore, the general cytotoxic effects of 5-Aza-CdR did not
influence the change in cell cycle profile observed in these cells.
Heritability
of 5-Aza.CdR-mediated
Growth
Suppression
of p16 by 5-Aza-CdR
in Vivo. Bladder
tumors
and
pl6 Expression in Vitro. Cell population doubling times and pitS
mRNA levels were measured in T24 cells during successive passages
CellType5AzaCdR3%3%Cell
5AzaCdR+
T24@34%
Line
63%@@°“5637
Fig. 3. Effects of 5-Aza-CdR on cell cycle pro
files in T24, 5637, and LD98 cells. Cells were
exposed to 5-Aza-CdR (5 X l0@ M) for 24 h,
followed by drug removal. Cells were maintained
@34%
for 7 days and passaged once before harvesting. On
the tenth day after treatment, cells were fixed in
75% ethanol, stained with propidium
iodide, and
analyzed by flow cytometry to determine relative
DNA content at specific cell cycle stages. A pro
nounced increase ofcells in G1 was observed in T24
cells treated with 5-Aza-CdR, but a more subtle
Cell Line19%
47%LD982%
effect was observed in 5637 cells. A negligible
effect was observed in LD9S cells.
27%Cell
in
duced in nu/nu rats, via laparotomy and direct implantation of T24
cells into bladder walls, were analyzed to determine whether 5-Aza
CdR could activate pi6 expression in vivo. Approximately 4—6weeks
later, when macroscopic bladder tumors were evident, 5-Aza-CdR
was administered i.p. every 24 h for 5 days. Animals were sacrificed
on the sixth day for the harvesting of total bladder RNA. Reactivation
of pitS was detected via RT-PCR in five of the six bladder tumors in
animals treated with 5-Aza-CdR (Fig. 5), demonstrating that a dor
mant pitS gene could be effectively reactivated by the drug in vivo.
30%
@-‘‘4%
Strain
@
=G2/M
=G1
I=s
98
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1998 American Association for Cancer Research.
INHIBITION OF DNA METHYLATION RESTORES GROWTH CONTROL
A
1.65
C
0
I1O-7M 5AzaCdR
U 5x1 O-7M
5AzaCdR
1.55
11.0.
0.
Cl)
1.45
p.-.
II
@0 1.35
1.25
L0
+
EE
1.15
.E.E
1.05
00
,@
11
n o@
1
2
3
Passage
B
@
@
4
5
8
9
10
with 5AzaCdR
g o
Passage
#after
Treatment
with1O-7M
5AzaCdR
+C
0 1 2 3 4 5 6 7 8 9 10
I
—@•.....—._
•1
@
.@
@
¶@
g
@
7
# after Treatment
p16 mRNA
GAPDH mRNA
6
‘ft
.@ pass.9
S after
Treatment
with
5x10-7M
5AzaCd@!@@
o ‘i
2 3 4 5 6 7 8 g io
p16mRNA
j
@
GAPDHmRNA
—0
I
Fig. 4. Heritability of 5-Aza-CdR-mediated p16 induction in T24 cells. In A. T24 cells were treated with either l0@ Mor 5 X l0@ M5-Aza-CdR for 24 h. After drug removal,
cells were maintained in culture and passaged every 3 days prior to approaching confluency. Cell doubling times were determined at each passage after drug treatment (as described
in “Materials
and Methods―).
In B, cells were harvested for total RNA isolation, and RT-PCR-based p16 expression analysis was performed during each passage after treatment (as
described in “Materialsand Methods―).The passage numbers after treatment are indicated (0—10).Growth suppression and p16 expression (induced by 5-Aza-CdR) were heritable for
at least nine passages, and these effects were dose dependent. GAPDH mRNA levels were measured to control for cDNA input and integrity.
Additional control experiments were performed in which total bladder
T24
RNA from rats treated with PBS showed no detectable p16 expression
culture to facilitate recovery from the immediate cytotoxic effects of
the drug, and then injected s.c. into nu/nu mice. Table 1 shows the
effects of 5-Aza-CdR on the tumorigenicity of T24 cells in vivo.
Tumors induced via the injection of untreated T24 cells were approx
imately 1.5 mm3 in size, whereas the tumors induced by T24 cells
(data not shown).
Effects of 5-Aza-CdR on Twnorigenicity
of T24 Cells in Vivo.
Tumors were induced in nu/nu mice via s.c. injection of either
untreated T24 cells or T24 cells treated with 5 X i0@ M 5-Aza-CdR.
cells
were
treated
1 1 days
prior
to injection,
passed
99
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1998 American Association for Cancer Research.
once
in
INHIBITION OF DNA METHYLATION RESTORES GROWTh CONTROL
5AzaCdR-Mediatedp16 InductionInVivo
0
@
.@
confirming previous studies by Merlo et a!. (28). We have also
demonstrated that both of the cell lines not expressing p16 (T24 and
HCT 15) were hypermethylated at this locus, suggesting that p16
methylation may be a common event during the establishment of cell
lines in vitro. p16 up-regulation in these three tumor cell lines may
also explain how 5-Aza-CdR suppressed cell growth; however, the
absence of p16 up-regulation in the remaining four tumor cell lines
suggests that other growth-regulatory genes in different pathways may
@øj
Tumor+5AzaCdR
+ C@1 2 3 4 56
p16
mRNA
be potential targets for de novo methylation. The induction of other
—0@
GAPDH mRNA -0@
genes, either alone or in combination with pitS, may also explain the
observed growth suppression in the T24, HCT15, and HCT116 cell
lines. Although de novo methylation of growth-regulatory genes may
accompany the development of primary tumors in vivo, the strong
selection pressure during the establishment of these tumors into im
mortal cell lines implicates that further de novo methylation of these
genes may also occur in vitro. This phenomenon may explain why few
primary tumors ever transform into immortal cell lines, and those
tumor cells which do survive selection to become established lines
often appear to divide more rapidly than they did in vivo.
Homozygous deletion or intragenic mutation of growth-regulatory
genes may also give cells a selective advantage in vivo or in vitro,
leading to increased cellular proliferation. Transcriptional silencing
via the de novo methylation of 5' promoter sequences of growth
@• ===::::=:@:I
Fig. 5. Reactivation of p16 by 5-Aza-CdR in vivo. p16 expression after 5-Aza-CdR
treatment was detected via RT-PCR (with primers specific for human p16 cDNA) in T24
cells growing in nu/nu rats. Daily 5-Aza-CdR treatmenLswere performed (4-6 weeks
after T24 cell inoculation) via i.p. injection into six individual nu/nu rats for 5 days
(columns1—6;
asdescribedin “Materials
andMethods―).
PBS wasinjectedinto additional
nu/nu rats as a negative control. Total bladder RNA was harvested for p16 RT-PCR
analysis 24 h after the last drug treatment (as described in “Materialsand Methods―),and
GAPDH mRNA levels were measured to control for cDNA input and integrity.
treated with 5-Aza-CdR averaged 0. 14 mm@in size (Table 1). The p16
gene was also shown to be expressed by RT-PCR analysis in six of six
tumors induced by 5-Aza-CdR-treated T24 cells and in zero of six
tumors induced by untreated T24 cells (data not shown). The activa
tion ofpl6 and/or other growth-regulatory genes may explain why the
tumors grew more slowly. These results demonstrate that the herita
ble, growth-inhibitory effects of 5-Aza-CdR treatment in vitro are still
observed in T24 cells subsequent to their implantation in vivo.
regulatory genes represents an alternative mechanism for their mac
tivation, and studies have shown that increased methylation of CpG
islands accompanies the establishment of immortal tumor cell lines (3,
4). Thus, hemizygous deletion of one allele coupled to hypermethy
lation of the remaining allele represents an alternative means for
complete inactivation of tumor suppressor genes in either primary
tumors or cancer cell lines. This is specifically demonstrated in the
T24 cell line, which contains only one copy of chromosome 9 (which
harbors the pitS locus), and contains a hypermethylated p16 gene on
the remaining allele.
Numerous investigations have described reactivation of genes by
5-Aza-CdR, which include loci on the inactive X chromosome (37),
the VHL gene (16), the E-cadherin gene (38), and the estrogen
receptor gene (17). Our studies ofpl6 reactivation in the T24 bladder
tumor cell line represents a good system for investigating the general
effects of 5-Aza-CdR during the reactivation of genes silenced by de
novo methylation. Other cell lines containing various hypermethylated
genes that are inducible by 5-Aza-CdR in vitro also represent good
systems for study. The demonstration of pitS expression and growth
suppression by 5-Aza-CdR after 41 cell population doublings also
suggests the potential of this drug as an effective chemotherapeutic
agent. Whether the gradual decrease in pitS expression during subse
quent cell passages arose from the de novo methylation of this gene
or the selection of cells unaffected by 5-Aza-CdR needs to be
determined.
01 arrest was also observed in T24 cells after drug treatment, and
this growth inhibition may have been achieved by reactivating a
dormant plo gene rather than introducing an exogenous gene as
described in previous studies (39, 40). It is possible, however, that the
DISCUSSION
Our data show that brief treatment of established human tumor cell
lines, but not diploid fibroblasts, with 5-Aza-CdR leads to a substan
tial increase in the doubling times of surviving cells. Because all seven
cell lines analyzed had the same response and 5-Aza-CdR is well
known for its ability to induce the expression of genes silenced by de
novo methylation, the most likely explanation for the result is that
reexpression of silenced growth-regulatory
genes inhibited cell
growth. Other genes besides growth-regulatory genes may also be
activated by 5-Aza-CdR, resulting in growth suppression. Alternative
explanations for these phenomena could be the cytotoxic activity of
the drug or the induction of other cellular physiological changes (i.e.,
changes in nucleotide pools), which normally slow cell growth. We
think that these explanations are unlikely because cells were allowed
to recover from the immediate cytotoxicity of 5-Aza-CdR before the
growth kinetics were measured, and in the case of the T24 cells, the
effects were apparent for up to 10 passages (41 population doublings)
after drug treatment. T24 cells treated with equitoxic doses of Ara-C
(which does not inhibit DNA methylation) showed no heritable
growth inhibition, confirming that the cytotoxic effects of 5-Aza-CdR
did not suppress cell growth. Furthermore, neither fibroblast strain
tested showed growth inhibition at equal drug doses, which caused
equivalent cytotoxicities and which should cause similar changes in
cellular physiology.
In three of the seven cases examined, we observed expression of the
pitS gene after treatment with 5-Aza-CdR. pitS negatively regulates
the G1/S transition of the cell cycle; therefore, loss of pitS function
should potentiate the transition of cells into S-phase and increase
cellular proliferation. Hypermethylation ofpl6 in various cancers has
also been reported in several studies (19, 20, 28—31),and our study
demonstrates that a dormant pitS gene can be reactivated in vitro,
Table 1 Effects of 5-Aza-CdR on tumorigenicity in vivo
T24 cells were treated for 24 h with 5-Aza-CdR (5 X lO@ M)in vitro. These cells
were injected s.c. into the right and left flanks of nu/nu mice after 11 days, allowing time
for recovery from the immediate cytotoxic effects of the drug. Untreated T24 cells were
used as a control. Individual tumors induced by either 5-Aza-CdR-treated or untreated T24
cells were measured 4 weeks after injection.
-5AzaCdR+5AzaCdRAverage
tumor size
Range1.5
mm3)Sample
size(n
mm@
mm@
(0—2.4
mm3)0.14
= 16)(n
(0-0.3
= 14)
100
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1998 American Association for Cancer Research.
INHIBI11ON OF DNA METHYLATION RESTORES GROWTH CONTROL
increased proportion
18. Herman, J. 0., Jen J., Merlo, A., and Baylin, S. B. Hypermethylation-associated
inactivation indicates a tumor suppressor role for pI5IINK4B. Cancer Res., 56:
722—727,
1996.
19. Herman, J. G., Merlo, A., Mao, L., Lapidus, R. G., Issa, J-P., Davidson, N. E.,
of cells in G@ resulted from the nonspecific
reactivation of other growth-regulatory genes by 5-Aza-CdR, which
@
may act alone or in combination with pitS. To address this issue, we
also investigated the cell cycle profile of 5637 cells, which express
pitS before and after 5-Aza-CdR treatment. These cells did not exhibit
a dramatic increase of cells in G@ and normal fibroblasts treated under
the same conditions showed a negligible change in cell cycle profile
(Fig. 3). Although these results are consistent with the hypothesis that
pitS induction by 5-Aza-CdR mediates G1 arrest in T24 cells, addi
Sidransky
Tornout, J. M., and Jones, P. A. Methylation of the 5' CpG island of the p16 tumor
suppressor gene in normal and transformed human tissues correlates with gene
silencing. Cancer Res., 55: 453 1—4537,1995.
21. Zhang, Y., Shields, T., Crenshaw, T., Hao, Y., Moulton, T., and Tycko, B. Imprinting
of human H19: allele-specific CpG methylation, loss of the active allele in Wilms'
tumor and potential for somatic allele switching. Am. 3. Hum. Genet., 53: 113—124,
1993.
22. Harris, M. Induction of thymidine kinase in enzyme deficient Chinese hamster cells.
Cell,29: 483—492,
1982.
23. Jones, P. A.. and Taylor. S. M. Cellular differentiation, cytidine analogs, and DNA
methylation.Cell,20: 85—93,
1995.
24. Pinto,A., andZagonel,V. 5-Aza-2'-deoxycytidine
(Decitabine)and 5-Aza-cytidine
in the treatment of acute myeloid leukemias and myelodysplastic syndromes: past,
present, and future trends. Leukemia (Baltimore), I: 51-60, 1993.
25. Seimno, M., Hannon, G. J., and Beach, D. A new regulatory motif in cell-cycle
control causing specific inhibition ofcyclin D/CDK4. Nature (Lond.), 386: 704—707,
REFERENCES
1993.
26. Nobori, T., Miura, K., Wu, D. J., Lois, A., Takabayashi, K., and Carson, D. A.
Deletions of the cyclin dependent kinase 4-inhibitor gene in multiple human cancers.
1. Jones, P. A. DNA methylation errors and cancer. Cancer Rca., 56: 2463-2467, 1996.
2. Baylin, S. B., Makos, M., Wu, J., Yen, R-W. C., deBustros, A., Vertino, P., and
Nelkin, B. D. Abnormal patterns of DNA methylation in human neoplasia: potential
consequencesfor tumorprogression.CancerCells.3: 383-390,1991.
3. Jones,
P.A.,Wolkowicz,
M.J.,Rideout,
W.M.,ifi, Gonzales,
F.G.,Marziasz,
C.M.,
Coetzee, G. A., and Tapscott, S. J. Dc novo methylation of the MyoDi CpG island
during the establishment of immortal cell lines. Proc. Nail. Aced. Sci. USA. 87:
6117-6121, 1990.
4. Antequera, F., Boyes, J., and Bird, A. High levels ofde novo methylation and altered
Nature (Lond.), 388: 753—756,1994.
27. van der Reit, P., Nawroz, H., Hruban, R. H., Cono, R., Tokino, K., Koch, W., and
Sidransky, D. Frequent loss of chromosome 9p2l—22early in head and neck cancer
progression. Cancer Res., 54: 1156—1
158, 1994.
28. Merlo, A., Herman, J. G., Mao, L, Lee, D. J., Gabrielson, E., Burger, P. C., Baylin.
S. B.,andSidransky,D. 5' CpGislandmethylationis associatedwithtranscriptional
silencing of the tumor suppressor pl6/MTS1 in human cancers. Nat. Med.. I:
686—692,1995.
29. Costello, J. F., Berger, M. S., Huang H-T.. and Cavenee, W. K. Silencing of
chromatinstructureat CpGislandsin cell lines.Cell,62: 503—514,
1990.
5, Bird, A. P. CpG-rich islands and the function of DNA methylation. Nature (Lond.),
209—213, 1986.
pJ6ICDKN2 expression in human gliomas by methylation and chromatin condensa
6. Gardiner-Garden, M., and Frommer, M. CpG islands in vertebrate genomes. J. Mol.
Biol., 196: 261-282, 1987.
tion. Cancer Res., 56: 2405—2410,1996.
30. Lo, K. W., Cheung, S-T., van Hasselt, A., Tsang Y-S.. Ma, K-F., Chung Y-F., Woo,
K. K. S., Lee, J. C. K., and Huang, D. P. Hypermethylation of the pitS gene in
nasopharyngeal carcinoma. Cancer Res., 56: 2721—2725,1996.
7, Li, E., Berad, C., and Jaenish, R. DNA methylation in genomic imprinting. Nature
(Lond.), 386: 382—385.1993.
8. Ferguson-Smith, A., Sasaki, C. H., Cattanach, B. M., and Surani, M. A. Parental
origin specific epigenetic modification of the mouse H19 gene. Nature (Lond.), 382:
31. Otterson,G. A.,Khlief,S. N.,Chen,W.,Coxon,A.B.,andKaye,F.J. CDKN2gene
silencing in lung cancer by DNA hypermethylation and kinetics of p16INK4 protein
induction by 5-aza-2'-deoxycytidine. Oncogene, Ii: 1211—1216,
1995.
751—755,1993.
9. Feinberg, A. P., and Vogelstein, B. Hypomethylation distinguishes genes of some
human cancers from their normal counterparts. Nature (Lond.), 301: 89—92,1983.
32. Gonzalgo, M. L., and Jones, P. A. Rapid quantitation of methylation at specific sites
using methylation-sensitive single nucleotide primer extension (Ms-SNuPE). Nucleic
Acids Res., 25: 2529—2531,1997.
10. Jones,P. A.,Rideout,W.M.,Shen,J. C.,Spruck,C. H.,andTsai,Y.C. Methylation,
mutation, and cancer. Bioessays, 14: 33—38,1992.
Gonzales, F. G., and Jones, P. A. Progressive increases in the methylation status and
heterochromatinization of the Myo Dl CpG island during oncogenic transformation.
Mol.Cell.Biol.,14: 6143—6152,
1994.
12. Baylin, S. B., Fearon, E. R., Vogelstein, B., de Bustros, A., Sarkis, S. J., Burke, P. J.,
Steal, S. P., and Nelkin, B. D. Hypermethylation of the 5' region of the calcitonin
gene is a property of human lymphoid and acute myeloid malignancies. Blood, 70:
412—417,
1987.
13. Vertino, P. M., Spillare, E. A., Harris, C. C., and Baylin, S. B. Altered chromosomal
methylationpatternsaccompanyoncogene-induced
transformationof humanbron
chial epitheial cells. Cancer Res., 53: 1684—1689,1993.
14. Greger, V., Passarge, E., Hopping, W., Messmer, E., and Horsethmke, B. Epigenetic
changes may contribute to the formation and spontaneous regression of retinoblas
toma. Hum. Genet., 83: 155—158,1989.
15. Sakai, T., Togushida, J., Ohtani, N., Yandell, D. W., Rapaport, J. M., and Dryja, T. P.
gene is
Cancer Res., 55: 4525—4530, 1995.
by 5-Aza-CdR in vivo and the decreased tumorigealcity of 5-Aza
CdR-treated T24 cells in vivo further suggest the chemotherapeutic
potential of 5-Aza-CdR. Reactivation of growth-regulatory genes
silenced by de novo methylation suggests clinical applications in
which genes inactivated by hypermethylation in human tumors can be
induced by 5-Aza-CdR during anticancer therapy.
11. Rideout, W. M., Eversole-Cire, P., Spruck, C. H., Hasted, C. M., Coetzee, G. A.,
of the CDKN2/p16/MTSI
20. Gonzalez-Zulueta, M., Bender, C. M., Yang, A. S., Nguyen, T., Beart, R. W., Van
tional studies must be performed to confirm this. The induction of p16
321:
D., and Baylin, S. B. Inactivation
frequently associated with aberrant DNA methylation in all common human cancers.
33, Clark, S. J., Harrison, J., Paul, C. L., and Frommer, M. High sensitivity mapping of
methylated cytosines. Nucleic Acids Res., 22: 2990—2997, 1994.
34, Frommer, M., McDonald, L E., Millar, D. S., Collis, C. M., Watt, F., Grigg, G. W.,
Molloy, P. L., and Paul, C. L. A genomic sequencing protocol that yields a positive
display of 5-methylcytosine residues in individual DNA strands. Proc. NatI. Aced.
Sci. USA, 89: 1827—1831,
1992.
35, Jones, P. A. Gene activation by 5-azacytidine. In: A. Razi, H. Cedar, and A. D. Riggs
(edt.), DNA Methylation:
Biochemistry
and Biological Significance,
pp. 165—187.
New York: Springer-Verlag, 1984.
36. Hara, G., Smith, R., Parry, D., Tahara, H., Stone, S., and Peters, G. Regulation of
p16/CDKN2 expression and its implications for cell immortalization
and senescence.
Mol. Cell. Biol., 16: 859—867,1996.
37. Jones, P. A., Taylor, S. M., Mohandas, T., and Shapiro, L. J, Cell cycle-specific
reactivation of an inactive X-chromosome locus by 5-azadeoxycytidine. Proc. Natl.
Aced.Sd. USA,79: 1215—1219,
1982.
Allele-specific hypermethylation of the retinoblastoma tumor suppressor gene. Am.
Hum. Genet., 48: 880—888,1991.
38. Yoshiura, K., Kanai, Y., OChiari,A., Shimoyama, Y., Sugimura, T.. and Hirohashi,
tumor-suppressor gene by DNA methylation in renal carcinoma. Proc. NatI. Aced.
Sci. USA. 91: 9700—9704, 1994.
17. Ottaviano, Y. L., Issa, J-P., Parl, F. F., Smith, H. S., Baylin, S. B., and Davidson, N. E.
inhibition of tumor cell proliferation by the pI6'@4 gene by an adenovirus vector.
Cancer Ret., 55: 3250—3253,1995.
S. Silencingof the E-cadherininvasion-suppressor
gene by CpG methylationin
human carcinomas. Proc. Natl. Aced. Sci. USA, 92: 7416—7419,1995.
16. Herman.J. G.,Latif,F.,Weng,Y.,Lerman,M.I.,Zbar,B.,Liu,S.,Samid,D.,Dusts,
D-S, Gnarra,3. R., Linehan,W. M., and Baylin,S. B. Silencingof the VHL 39, Jin, X., Nguyen, D., Zhang, W-W, Kiritis, A. P.. and Roth, J. A. Cell cycle arrest and
40. Amp, A., Nishikawa,
R., Furnari, F. B., Cavenee, W. K., and Su Huang, H-i.
Replacement of the pitS gene suppresses human glioma cell growth. Cancer Res., 55:
Methylation of the estrogen receptor gene CpG island marks loss of estrogen receptor
expression in human breast cancer cells. Cancer Rca., 54: 2552—2555,1994.
1371—1374,
1995.
101
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1998 American Association for Cancer Research.
Inhibition of DNA Methylation by 5-Aza-2′-deoxycytidine
Suppresses the Growth of Human Tumor Cell Lines
Christina M. Bender, Martha M. Pao and Peter A. Jones
Cancer Res 1998;58:95-101.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/58/1/95
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 16, 2017. © 1998 American Association for Cancer Research.