[Cancer Research 60, 2598-2601, May 15, 2000]
© 2000 American Association for Cancer Research
Monoallelic Amplification of Estrogen Receptor-
Expression in Breast Cancer1
Eric R. Schuur and
Ronald J. Weigel2
Department of Surgery, Stanford University School of Medicine, Stanford, California 94305-5494
 |
ABSTRACT
|
|---|
Gene amplification and loss of heterozygosity are alterations to
chromosomal structure whereby tumor cells alter patterns of gene
expression. We have identified a novel mechanism of gene regulation in
which cancer cells predominantly express one of the two alleles of a
gene. Estrogen receptor (ER)-
is overexpressed in hormone-responsive
breast cancer compared with normal breast epithelial cells. Using a
polymorphism of codon 10, we examined allele-specific expression of the
four different ER promoters in MCF-7 breast cancer cells and primary
tumors. Monoallelic amplification of expression (MAX) for all four ER
promoters was identified, resulting in an allelic preference of
>100-fold. MAX was the result of an amplification of allele copy
number and a preference to transcribe the amplified allele. The effect
of MAX was most significant for the promoters clustered near the 1'
exon, whereas the expression from the distant H promoter mirrored
template copy number. MAX of the ER gene was not found to occur in
normal endometrial or breast tissue. As a novel mechanism in cancer
genetics, MAX can result in functional homozygosity at a gene locus.
|
Introduction
|
|---|
Gene amplification and overexpression are important factors in the
pathophysiology of cancer. Many breast cancers overexpress the
ER3
gene compared with normal breast epithelial cells
(1)
. In particular, patients with breast tumors
that overexpress ER have an improved prognosis relative to patients
with ER-negative breast tumors (2)
. Its prognostic
importance suggests a causal role for ER in the progression of breast
cancer. However, the molecular mechanisms controlling the expression of
ER in normal and disease states are as yet unclear.
ER is expressed in a wide variety of cell types in addition to
estrogen-responsive tissues such as mammary epithelium and endometrium.
The regulation of ER transcription is complex, involving multiple
independent promoters and multiple alternative 5' exons (3
, 4)
. The majority of ER mRNA transcripts appear to originate from
the originally defined promoter, termed P1, at the 5' end of exon 1.
The primary regulatory element influencing transcription initiation
from this promoter is an AP2
binding site found near position
+200 in the 5' untranslated leader (5)
. In addition, the
entire region encompassing P1 and exon 1 comprise a well-defined CpG
island (6)
, again suggesting an important regulatory
function in ER expression.
However, a significant proportion of ER transcripts originate at cap
sites associated with alternative promoters located farther upstream.
Most of these transcripts use a splice acceptor site located at
position +163 to generate ER transcripts with alternative 5' exons
(7)
. These alternative promoters account for at least some
cases of tissue-specific patterns of ER expression (8)
.
Detailed analysis of the interplay of these regulatory elements may
provide insights into the control of ER expression and its role in
breast malignancies. This study sought to determine whether unequal
transcription of the two alleles of the ER gene could contribute to
overexpression of ER in cancer.
|
Materials and Methods
|
|---|
Cell Lines and Tissue Samples.
MCF-7 cells were obtained from the American Type Culture Collection and
were maintained in MEM containing 10% fetal bovine serum, 6
ng/ml insulin, 100 units/ml penicillin, and 100 µg/ml streptomycin.
Normal endometrial and breast tissue samples were obtained from the
Cooperative Human Tissue Network, Western Division.
DNA and RNA Samples.
Genomic DNA was prepared from MCF-7 cells or fresh-frozen tissue
samples as described (9)
. Total RNA was prepared from
these same sources using a guanidinium-phenol method (Trizol reagent;
Life Technologies, Gaithersburg, MD). For each cell line or tissue
sample, 1 µg of total RNA was reverse transcribed using oligo(dT) and
Moloney murine leukemia virus reverse transcriptase (Superscript II;
Life Technologies) to generate cDNA for amplification. cDNA samples
from breast tumors were kindly provided by Devon A. Thompson, Stanford
University, Stanford, CA. Plasmids representing the two alleles of ER
in MCF-7 cells were constructed by ligating PCR products generated by
amplifying MCF-7 genomic DNA with ERPRO18 (5'-CGGGAGACCAGTACTTAAAG-3')
and ERPRO22 (5'-CCCTTGGATCTGATGCAGTA-3') into pCR2.1 (Invitrogen, La
Jolla, CA). The two selected clones were sequenced by dideoxy
termination using Sequenase T7 DNA polymerase (Amersham, Northbrook,
IL) to confirm the polymorphisms at position +262.
PCR-RFLP Assay.
One µg of genomic DNA or cDNA derived from 80 ng of total RNA was
used per reaction. In addition, each reaction contained 300
nM exon-specific upstream primer, 300 nM
unlabeled ERPRO22, 0.5 nM ERPRO22 kinased with
32P, 200 mM nucleotides, 2% DMSO,
and 1.8 units of Expand High Fidelity Polymerase mixture (Roche
Molecular Systems) in a 100-µl volume. After an initial denaturation
step at 95°C for 2 min, samples were cycled 25 times as follows:
94°C for 30 s, 57°C for 1 min, and 72°C for 1 min. The final
extension was for 5 min at 72°C. The exon-specific upstream primers
used were ERPRO100 (5'-GAAAGGTCCATGCTCCTTTC-3'), ERPRO89
(5'-GCCCATGGAACATTTCTGGA-3'), ERPRO92 (5'-AGCCTCTATCCAGCAGCGAC-3'), and
ERPRO23 (5'-GGGAGTCGACGAGCTGGCGGAGGGCGTTCGT-3').
Following amplification, excess nucleotides and primers were removed
using Qiaquick PCR Purification columns (Qiagen, Valencia, CA). Ten
µl of each reaction were cleaved with 10 units each of
NciI and SacII in 20 µl for 1 h at 37°C.
Loading dye was added to each sample, and the samples were
electrophoresed through a 10% acrylamide/0.5x Tris-borate-EDTA gel.
Gels were dried and then exposed to film to detect bands corresponding
to each allele. To quantitate the amount in each band, the gels were
exposed to phosphorimager screen (Molecular Dynamics, Sunnyvale, CA)
for 6–24 h.
|
Results and Discussion
|
|---|
A PCR-based assay was devised (see "Materials and Methods") to
measure mRNA transcript abundance from the ER promoters
(4)
associated with exons Ha, 1', E, and 1 (Fig. 1
). By taking advantage of a single nucleotide polymorphism
(10)
in exon 1 of the human ER located at nucleotide +262
[numbering according to Greene et al. (11)
]
this assay enabled measurement in informative samples of the relative
contribution of each allele to the mRNA pool. This polymorphism occurs
at the third base of codon 10 of the ER open reading frame, changing
the sequence of codon 10 from TCT to TCC and creating a NciI
site. The TCT allele has a population frequency (10)
of
55%, whereas the TCC allele has a population frequency of 45%.
PCR amplification with an exon-specific upstream oligonucleotide and a
common downstream oligonucleotide in exon 1 yields a DNA fragment that
can be cleaved with SacII to yield a 70-bp restriction
fragment spanning the codon 10 polymorphism. In cells heterozygous for
the codon 10 polymorphism, cleavage with NciI can be used to
distinguish between the PCR products derived from each allele.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 1. PCR-RFLP assay for ER alleles. a, genomic
organization of the 5' region of the human ER locus. The genomic
organization of the human ER gene 5' region was determined (details to
be published elsewhere). The locations of the ER exons
Ha, Hb, 1',
E, and 1 are shown to scale. The
horizontal line represents genomic DNA, and the exons
are represented by filled vertical rectangles.
b, exons analyzed by the PCR-RFLP assay are diagrammed
along with the oligonucleotides used and the locations of the relevant
restriction sites relative to the previously published cap site for ER
mRNA. atg, ER initiation codon located at position +233;
SA, splice acceptor site for Hb, 1', and E at position
+163; N, NciI site located at position
+262; S, SacII site located at position
+221; * adjacent to ERPRO22 indicates the location of
the radioactive label on this oligonucleotide. The sizes and locations
of the restriction fragments detected in the assay are shown by
solid bars below exon 1. KB, kilobases
|
|
Five ER-positive breast carcinoma cell lines were screened for
heterozygosity at the codon 10 polymorphism (data not shown). Only one,
MCF-7, was found to be heterozygous. In addition to being heterozygous,
it appeared that the copy numbers for each allele were unequal (Fig. 2
). To validate that the difference in copy number seen in MCF-7 cells
was accurately measured by the PCR-RFLP assay, cloned PCR products
representing the sequences between the oligonucleotides ERPRO18 and
ERPRO22 from each allele in MCF-7 (275N for the TCT allele
and 275n for the TCC allele) were subjected to the PCR-RFLP
assay either individually or in defined mixtures of the two clones
(Fig. 2
). Amplification of the two clones individually yielded only the
expected band [31 bp for n (Fig. 2
, Lane 1), 70
bp for N (Fig. 2
, Lane 2)] with no trace of the
alternative. Assay of the defined mixtures yielded ratios of these two
fragments in proportion to the starting abundance of the clones (Fig. 2
, Lanes 3–5). Quantitation of the amount of each fragment
present in each sample by phosphorimaging allowed the calculation of a
ratio of the two bands to estimate the relative abundance of the two
alleles. The N/n ratio for the 1:9 mixture was 0.16 with a
SD of 0.063; for the 1:1 mixture, the average was 1.28 ± 0.156, and for the 9:1 mixture, the average was 25.66 ± 8.68. This assay consistently produces N/n values of
1 for normal heterozygous individuals (see Fig. 3
, Lanes 5, 10, 15, and
20). These data show that, within experimental error,
the amount of N or n fragment detected in the
PCR-RFLP assay is directly proportional to the input copy number over a
100-fold range.
View larger version (57K):
[in this window]
[in a new window]
|
Fig. 2. ER monoallelic expression in MCF-7 cells. Cloned PCR
products representing the two ER codon 10 polymorphic alleles
(275n and 275N) were PCR amplified using
primers ERPRO18 and ERPRO22 (Fig. 1
) either individually (Lanes
1 and 2) or mixed at the molar ratios indicated
(N/n, Lanes 3–5).
Individual cDNA samples from MCF-7 cells were amplified with
exon-specific upstream primers plus a common radiolabeled downstream
primer, ERPRO22 (Fig. 1
). Exon-specific upstream primers used were
ERPRO100 for exon Ha, (Lane 6), ERPRO89 for exon 1'
(Lane 7), ERPRO92 for exon E (Lane 8),
ERPRO23 for exon 1 (Lane 9), and ERPRO18 for genomic DNA
(Lane 10). Each sample was analyzed in triplicate; one
representative lane is shown in the autoradiograph in the top
panel. The average N/n ratio is
displayed along with the SD in the column chart in the bottom
panel. The N/n ratio for genomic
DNA was statistically different from the
N/n ratio for the 1:1 mixture
(P < 0.01). The
N/n ratios for exon-specific PCR products
were significantly different from that of genomic DNA when similarly
analyzed (Ha versus genomic, P < 0.01; 1' versus genomic,
P = 0.015; E versus
genomic, P < 0.01; 1
versus genomic, P < 0.01). * below Lanes 1 and 2 indicate
that no meaningful N/n ratio could be
calculated.
|
|
View larger version (51K):
[in this window]
[in a new window]
|
Fig. 3. ER is expressed biallelically in normal endometrium.
Genomic (G) DNA and total RNA were isolated from
fresh-frozen tissue specimens, and then analyzed by the PCR-RFLP assay
as described in Fig. 2
. The top panel shows the
autoradiograph for four representative samples. All samples were
analyzed in the same experiment. The locations of the bands
corresponding to the N allele and the n
allele are indicated. The N/n ratio was
calculated for each sample and charted in the bottom
panel. Bars indicate the 20% intrasample
variation seen with this assay.
|
|
Because the copy numbers for the two ER alleles are unequal, the
question arose whether the two ER alleles are equally active in
transcription. To determine the relative usage of the two alleles, cDNA
from MCF-7 was PCR amplified using exon-specific upstream
oligonucleotides paired with ERPRO22. Upstream oligonucleotides
specific for the alternative 5' ER exons Ha (4)
, 1'
(12
, 13)
, E (4)
and 1 (11
, 14)
were used (Fig. 1
). Genomic DNA from MCF-7 was amplified with ERPRO18
and ERPRO22 in parallel. Analysis of MCF-7 genomic DNA revealed that
the N allele was 2.6-fold more abundant than the
n allele, which was significantly different from the 1:1
mixture of the 275N and 275n clones
(P < 0.01). cDNA analyzed for exon Ha
transcripts revealed an N/n ratio of 4.0, a difference from
genomic DNA that was highly significant (P = 0.001). This implies that 1.5 mRNA transcripts were derived from the
N allele for each transcript from the n allele.
Similar analysis of cDNA with oligonucleotides specific for the other
ER exons demonstrated N/n ratios of 57.3 for 1',
16.6 for exon E, and 19.0 for exon 1. These data imply even higher
transcriptional activity of the N allele relative to the
n allele. Therefore, in MCF-7 cells, not only are there more
copies of the N allele, but the N allele is also
significantly more active in transcription. As a result, expression of
ER is elevated from that in normal mammary epithelium, and the
expressed ER is derived predominantly from one allele. We coined the
term "monoallelic amplification of expression," or MAX, to describe
this phenomenon.
Monoallelic transcription due to genomic imprinting, allelic exclusion
or X-chromosome inactivation has been demonstrated for several genes,
including IGF2, immunoglobulin genes, and genes
on the X chromosome, respectively (15
, 16)
. However, MAX
has not been described before. Amplification of one allele of the
cyclin D1 gene has been reported in some breast cancer cell
lines, suggesting MAX of the cyclin D1 gene
(17)
. Therefore, MAX of the ER gene was examined in a
panel of primary breast cancers. Of 15 ER-positive tumors, 7 were
informative and 3 demonstrated amplification of one allele. The range
of amplification was from 5- to 65-fold; the N allele was
amplified in one sample and the n allele in two samples.
Each of the informative ER-positive tumor samples was analyzed for
allele-specific transcription. Of the seven samples, two samples
demonstrated evidence of monoallelic transcription (Fig. 4
). Tumor 15 displayed a pattern of amplification of DNA and preferential
transcription of the amplified allele similar to that of MCF-7 (Fig. 4
,
Lanes 1–5). The n allele was present at
8-fold higher copy number than the N allele. Transcripts
from the n allele were between 43- and 100-fold more
abundant than those from the N allele with respect to exons
1', E, and 1. Transcripts derived from the Ha promoter on the
n allele were 3-fold more abundant than similar
transcripts from the N allele.
View larger version (66K):
[in this window]
[in a new window]
|
Fig. 4. Monoallelic expression of ER occurs in breast cancer.
Genomic DNA and total RNA were isolated from fresh-frozen tumor
specimens, and then analyzed by the PCR-RFLP assay as described in Fig. 2
. The top panel shows the autoradiograph for two tumors
that evidenced allele-specific DNA amplification and/or allele-specific
transcription (T15 and T26), as well as
one tumor displaying equivalent transcription from both alleles
(T3). All samples were analyzed in the same
experiment. For each tumor analyzed, Ha, 1', E, and 1 were generated by
amplification of cDNA with primer pairs specific for each of these
exons. G, amplification of genomic DNA with primers
specific for genomic sequence. The locations of the bands corresponding
to the N allele and the n allele are
indicated. The N/n ratio was calculated
for each sample and charted in the bottom panel.
Bars indicate the 20% intrasample variation seen with
this assay.
|
|
In contrast, monoallelic transcription was restricted to exon 1' in
tumor 26 (Fig. 4
, Lane 7). The two alleles were present in
approximately the same copy numbers, and transcripts associated with
exons Ha, E, and 1 were symmetrically expressed from the two alleles
(Fig. 4
, Lanes 7–10). For comparison, a sample
demonstrating equal copy numbers for both the N and
n alleles and symmetrical expression from each of the
promoters is shown (T3; Fig. 4
, Lanes 11–15).
In a panel of 14 ER-negative tumors, 8 were heterozygous; of these, 5
demonstrated evidence of amplification of one allele. In two samples,
the N allele was amplified, whereas in three samples, the
n allele was amplified (data not shown), suggesting that ER
expression was extinguished after gene amplification had occurred. As
expected, when samples were analyzed for allele-specific
transcription, no conclusion could be drawn because only low levels of
cDNA and, therefore, mRNA were detected.
The observed monoallelic expression could reflect a normal
physiological process that has deteriorated to varying degrees in the
tumors examined, as has been observed for the IGF2 gene in
colon carcinomas (18
, 19)
. Alternatively, the observed
monoallelic expression might reflect some aspect of the molecular
pathology of breast (and perhaps other) tumors. To begin
differentiating between these two possibilities, normal
estrogen-responsive tissues (endometrium and breast) were evaluated for
evidence of monoallelic transcription. Eight normal endometrium samples
that were heterozygous for the codon 10 polymorphism were analyzed by
the PCR-RFLP assay, four of which are shown in Fig. 3
. In each case,
the N/n ratio for the two alleles was not
statistically different from 1 (Fig. 3
, Lanes G). Likewise,
in each tissue sample, the N/n ratio of RNA from
the various 5' exons was not statistically different from 1. A similar
analysis of normal breast tissue samples likewise revealed identical
levels of transcription from both alleles and no evidence of
single-allele gene amplification (data not shown).
Methylation of the CpG island in the region of the promoter and exon 1
of the ER gene has been correlated with reduced or absent expression of
ER in breast (6)
and colon tumors (20)
.
Because methylation is also one mechanism of silencing a single allele
in genomic imprinting (16)
, the level of methylation of
each ER allele in MCF-7 cells was assessed. MCF-7 genomic DNA was
treated with two restriction enzymes, HpaII and
SacII, that do not cut DNA methylated at CpG dinucleotides.
HpaII cuts six times in the genomic region defined by
ERPRO18 and ERPRO22, whereas SacII cuts once. After exposure
to the restriction enzymes, the samples were subjected to the PCR-RFLP
assay. The resulting autoradiographs revealed that the two ER alleles
were equally sensitive to cleavage by the two methylation-sensitive
enzymes (data not shown). Therefore, it is unlikely that
allele-specific methylation underlies the transcriptional preference
for the N allele.
Because MAX was not observed in normal tissues, this process appears to
be specific to breast malignancies. Unequal template copy numbers can
lead to increased expression of one allele as is likely the case for
monoallelic expression of cyclin D1 observed in breast
cancer cell lines (17)
. In addition, there can be
preferential transcription of one allele, further contributing to the
degree of monoallelic expression—a phenomenon observed in one-third of
ER-positive tumor samples. The combination of these two factors
(allelic amplification and preferential transcription) can effectively
result in monoallelic expression. The resulting overexpression is an
additional element of MAX that distinguishes it from known mechanisms
of monoallelic transcription, such as imprinting. Loss of
heterozygosity and MAX as mechanisms in cancer genetics both
lead to homozygosity, but unlike the loss of function associated with
loss of heterozygosity, MAX results in retention of function. The
functional homozygosity resulting from MAX could have a profound effect
during oncogenesis by unmasking gene mutations with subtle alterations
in function.
|
Acknowledgments
|
|---|
We thank Devon A. Thompson for providing cDNA samples from
ER-positive and ER-negative breast cancers.
|
FOOTNOTES
|
|---|
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.
1 This work was supported by Grant 1RO1CA77350
from the NIH and a grant from the Clowes Foundation (to R. J. W.). 
2 To whom requests for reprints should be
addressed, at Department of Surgery, Stanford University School of
Medicine, 1201 Welch Road, MSLS P214, Stanford, CA 94305-5494. Phone:
(650) 723-9799; E-mail: ronald.weigel{at}stanford.edu
3 The abbreviations used are: ER, estrogen
receptor; RFLP, restriction fragment length polymorphism; MAX,
monoallelic amplification of expression.
Received 1/26/00.
Accepted 4/ 3/00.
|
REFERENCES
|
|---|
-
Ricketts D., Turnbull L., Ryall G., Bakhshi R., Rawson N. S., Gazet J. C., Nolan C., Coombes R. C. Estrogen and progesterone receptors in the normal female breast. Cancer Res., 51: 1817-1822, 1991.[Abstract/Free Full Text]
-
Knight W. A., Livingston R. B., Gregory E. J., McGuire W. L. Estrogen receptor as an independent prognostic factor for early recurrence in breast cancer. Cancer Res., 37: 4669-4671, 1977.[Abstract/Free Full Text]
-
Grandien K. Determination of transcription start sites in the human estrogen receptor gene and identification of a novel, tissue-specific, estrogen receptor-mRNA isoform. Mol. Cell. Endocrinol., 116: 207-212, 1996.[Medline]
-
Thompson D. A., McPherson L. A., Carmeci C., deConinck E. C., Weigel R. J. Identification of two estrogen receptor transcripts with novel 5' exons isolated from a MCF7 cDNA library. J. Steroid Biochem. Mol. Biol., 62: 143-153, 1997.[Medline]
-
deConinck E. C., McPherson L. A., Weigel R. J. Transcriptional regulation of estrogen receptor in breast carcinomas. Mol. Cell. Biol., 15: 2191-2196, 1995.[Abstract]
-
Ottaviano Y. L., Issa J. P., Parl F. F., Smith H. S., Baylin S. B., Davidson N. E. Methylation of the estrogen receptor gene CpG island marks loss of estrogen receptor expression in human breast cancer cells. Cancer Res., 54: 2552-2555, 1994.[Abstract/Free Full Text]
-
Weigel R. J., Crooks D. L., Iglehart J. D., deConinck E. C. Quantitative analysis of the transcriptional start sites of estrogen receptor in breast carcinoma. Cell Growth Differ., 6: 707-711, 1995.[Abstract]
-
Grandien K., Backdahl M., Ljunggren O., Gustafsson J. A., Berkenstam A. Estrogen target tissue determines alternative promoter utilization of the human estrogen receptor gene in osteoblasts and tumor cell lines. Endocrinology, 136: 2223-2229, 1995.[Abstract]
-
Strauss, W. M. Preparation of genomic DNA from mammalian tissue. In: F. M. Ausubel, R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidmen, J. A. Smith, and K. Struhl (eds.), Current Protocols in Molecular Biology, Vol. 1, pp. 2.2.1–2.2.3. New York: John Wiley and Sons, Inc., 1998.
-
Roodi N., Bailey L. R., Kao W. Y., Verrier C. S., Yee C. J., Dupont W. D., Parl F. F. Estrogen receptor gene analysis in estrogen receptor-positive and receptor-negative primary breast cancer. J. Natl. Cancer Inst., 87: 446-451, 1995.[Abstract/Free Full Text]
-
Green S., Walter P., Kumar V., Krust A., Bornert J. M., Argos P., Chambon P. Human oestrogen receptor cDNA: sequence, expression and homology to v-erb-A. Nature (Lond.), 320: 134-139, 1986.[Medline]
-
Keaveney M., Klug J., Dawson M. T., Nestor P. V., Neilan J. G., Forde R. C., Gannon F. Evidence for a previously unidentified upstream exon in the human oestrogen receptor gene. J. Mol. Endocrinol., 6: 111-115, 1991.[Abstract]
-
Piva R., Bianchi N., Aguiari G. L., Gambari R., del Senno L. Sequencing of an RNA transcript of the human estrogen receptor gene: evidence for a new transcriptional event. J. Steroid Biochem. Mol. Biol., 46: 531-538, 1993.[Medline]
-
Ponglikitmongkol M., Green S., Chambon P. Genomic organization of the human oestrogen receptor gene. EMBO J., 7: 3385-3388, 1988.[Medline]
-
Ohlsson R., Tycko B., Sapienza C. Monoallelic expression: ‘there can only be one. ’ Trends Genet., 14: 435-438, 1998.[Medline]
-
Tilghman S. M. The sins of the fathers and mothers: genomic imprinting in mammalian development. Cell, 96: 185-193, 1999.[Medline]
-
Hosokawa Y., Arnold A. Mechanism of cyclin D1 (CCND1, PRAD1) overexpression in human cancer cells: analysis of allele-specific expression. Genes Chromosomes Cancer, 22: 66-71, 1998.[Medline]
-
Cui H., Horon I. L., Ohlsson R., Hamilton S. R., Feinberg A. P. Loss of imprinting in normal tissue of colorectal cancer patients with microsatellite instability [see comments]. Nat. Med., 4: 1276-1280, 1998.[Medline]
-
Issa J. P., Vertino P. M., Boehm C. D., Newsham I. F., Baylin S. B. Switch from monoallelic to biallelic human IGF2 promoter methylation during aging and carcinogenesis. Proc. Natl. Acad. Sci. USA, 93: 11757-11762, 1996.[Abstract/Free Full Text]
-
Issa J. P., Ottaviano Y. L., Celano P., Hamilton S. R., Davidson N. E., Baylin S. B. Methylation of the oestrogen receptor CpG island links ageing and neoplasia in human colon. Nat. Genet., 7: 536-540, 1994.[Medline]
Top
ABSTRACT
Introduction
