Cell, Vol. 119, 591–602, November 24, 2004, Copyright 2004 by Cell Press
A Single Nucleotide Polymorphism in the MDM2
Promoter Attenuates the p53 Tumor Suppressor
Pathway and Accelerates Tumor Formation in Humans
The tumor suppressor protein, p53, is activated upon
cellular stresses such as DNA damage and oncogene
activation and initiates a transcriptional program which
apoptosis (Jin and Levine, 2001). The p53 stress re-
sponse pathway has been shown to be crucial for the
prevention of tumor formation. For example, both mice
and humans harboring a germline inactivating mutation
in one allele of the p53 gene develop tumors very early
in life and at dramatically high frequencies (Donehower
et al., 1992; Garber et al., 1992; Li et al., 1990). Somatic
inactivating mutations of the p53 gene are also found
in over 50% of all human tumors (Lain and Lane, 2003).
Together, these observations and many others support
the importance of the p53 pathway in tumor suppres-
sion. It is therefore reasonable to assume that naturally
occurring polymorphic genetic variants in critical nodes
of the p53 pathway might underlie the variation seen
between individuals in their susceptibility to cancer and
the progression of their disease.
The search for genetic variation in the p53 pathway
an important negative regulator of p53. MDM2 directly
binds to and inhibits p53 by regulating its location, sta-
bility, and activity as a transcriptional activator (Michael
and Oren, 2003). MDM2 is an essential gene in murine
tation in the uterus. This lethal phenotype is rescued by
knocking out the p53 gene, clearly demonstrating an
important genetic interaction between these two genes
in murine development (Jones et al., 1995; Montes de
Oca Luna et al., 1995). Mendrysa et al. (2003) demon-
strated the importance of this interaction in the adult
mouse by genetically altering mice to express reduced
levels of Mdm2. These mice are small, lymphopenic,
and radiosensitive, with increased apoptosis in both
lymphocytes and epithelial cells. These phenotypes
were all shown to be p53 dependent, thereby further
demonstrating that Mdm2 is a key negative regulator of
p53 in both the developing and mature mouse. In hu-
mans, a subset of tumors overexpress MDM2 mRNA
and protein; this overexpression is associated with ac-
celerated cancer progression and lack of response to
therapy (Freedman and Levine, 1999). In a subset of
these tumors, overexpression of MDM2 was mutually
exclusive to p53 mutation, which could suggest that
overexpression of MDM2 can substitute for inactivating
p53 by mutation (Leach et al., 1993; Oliner et al., 1992).
As MDM2 expression levels seem to be vital to a well-
regulated p53 response, naturally occurring sequence
variations in the MDM2 promoter may result in altered
expression of the MDM2 protein, thereby impacting p53
tumor suppression and potentially cancer in humans. In
this report, data are presented which support this hy-
Gareth L. Bond,2,8Wenwei Hu,2,8
Elisabeth E. Bond,2Harlan Robins,1
Stuart G. Lutzker,2Nicoleta C. Arva,7
Jill Bargonetti,7Frank Bartel,4Helge Taubert,4
Peter Wuerl,5Kenan Onel,6Linwah Yip,3
Shih-Jen Hwang,3Louise C. Strong,3
Guillermina Lozano,3and Arnold J. Levine1,2,*
1School of Natural Sciences
Institute for Advanced Study
Princeton, New Jersey 08540
2Cancer Institute of New Jersey
University of Medicine and Dentistry of New Jersey
New Brunswick, New Jersey 08903
3Section of Cancer Genetics
M.D. Anderson Cancer Center
University of Texas
Houston, Texas 77030
4Institute of Pathology
Faculty of Medicine
University of Ulm
6Department of Medicine
Weill College of Medicine
New York, New York 10021
7Department of Biological Sciences
Hunter College and Graduate School
The City University of New York
695 Park Avenue
New York, New York 10021
The tumor suppressor p53 gene is mutated in mini-
mally half of all cancers. It is therefore reasonable to
assume that naturally occurring polymorphic genetic
variants in the p53 stress response pathway might
determine an individual’s susceptibility to cancer. A
central node in the p53 pathway is the MDM2 protein,
a direct negative regulator of p53. In this report, a
single nucleotide polymorphism (SNP309) is found in
ity of the transcriptional activator Sp1, resulting in
quent attenuation of the p53 pathway. In humans,
SNP309 is shown to associate with accelerated tumor
formation in both hereditary and sporadic cancers. A
model is proposed whereby SNP309 serves as a rate-
limiting event in carcinogenesis.
8These authors contributed equally to this work.
Figure 1. The MDM2 Promoter Contains a Single Nucleotide Polymorphism Which Alters the Affinity of the Transcriptional Activator Sp1
(A) A schematic diagram depicts the intronic promoter of the MDM2 gene. The position of SNP309 is indicated in relation to the exon/intron
boundaries and the transcription factor binding sites for p53 and Ets/AP-1. The region analyzed for sequence variation is marked by the bar
drawn below the diagram.
(B) The analysis of transcription factor binding sites in the region containing SNP309 is depicted. Potential Sp1 sites are underlined.
(C)An autoradiographof anelectrophoretic mobilityshift assayis depictedwhere varyingconcentrations ofpurifiedSp1 proteinwere incubated
with biotin-labeled oligonucleotides containing either SNP309 (G/G; lanes 6–10), wild-type sequence (T/T; lanes 1–5), or an Sp1 optimal binding
site (lane 11, oligo alone, and lane 12, oligo and Sp1). The arrow indicates the specific Sp1-DNA complex.
(D) Chromatin immunoprecipitations (ChIP) were preformed using antibodies against either Sp1 or lamin A. The presence of the MDM2
promoter was assayed for using PCR. The PCR products are depicted here after electrophoreses on an 8% nondenaturing polyacrylamide
gel and subsequent staining with EtBr.
(E) The relative measured luciferase levels are depicted in a bar graph, whereby pGL2 luciferase reporter plasmids containing either SNP309
sequence or wild-type sequence were cotransfected with pPac (control) or pPac-Sp1 expression plasmids. Each value is the average of at
least three independent experiments and the error bars represent the standard deviations.
Analysis of this region of the MDM2 promoter using a
computer algorithm (AliBaba) revealed several putative
binding sites for the transcription factor Sp1 (Figure 1B).
Interestingly, the presence of SNP309, a T to G change,
extended the length of one of the putative Sp1 DNA
binding sites, suggesting that the presence of SNP309
could increase the affinity of Sp1 to this region of the
identified by computer analysis and to investigate the
functional consequences of SNP309, electro-mobility
shift assays (EMSAs) were performed. EMSAs were car-
ried out with purified recombinant human Sp1 protein
and labeled double-stranded oligonucleotides contain-
ing either the wild-type sequence (T/T) or the SNP309
sequence (G/G). Interestingly, as predicted by the com-
puter analysis, the binding affinity of oligonucleotides
containing SNP309 to a range of concentrations (150
ng–450ng) ofpurifiedSp1 ismuchhigher (2–4-fold)than
The search for genetic variation in the MDM2 promoter
was focused to a well-characterized region in the first
intron of the intronic promoter, which is utilized by both
the p53 and ras pathways to activate MDM2 transcrip-
tion (Ries et al., 2000; Zauberman et al., 1995). To look
for sequence variation in this region, 300 base pairs
from genomic DNAs isolated from 50 healthy volunteers
were amplified by PCR and sequenced (Figure 1A). Two
single nucleotide polymorphisms (SNPs) were found in
this region. One of these, SNP309 (a T to G change at
the 309thnucleotide in the first intron), was found at
relatively high frequency both in the heterozygous state
(T/G, 40%) and in the homozygous state (G/G, 12%).
The other SNP, SNP344 (a T to A change at the 344th
nucleotide in the first intron), is rare and is not further
studied (found only in the heterozygous state in 8%
A SNP in Mdm2, the p53 Pathway, and Tumor Formation
that of the wild-type sequence (Figure 1C, lanes 2–5
versus lanes 7–10). The same assays were carried out
using HeLa cell nuclear extracts as a source of protein
(data not shown). These data suggest that Sp1 can bind
to its putative consensus site in the MDM2 promoter
and that SNP309 can greatly enhance its binding affinity
of this site.
To verify the presence of Sp1 on the MDM2 promoter
formed. Lysates were prepared from growing Manca
cells, homozygous for SNP309 (G/G), and immunopre-
cipitations were carried out using antibodies against
either Sp1 or lamin A. After extensive washing and elu-
tion, the DNA was purified and the presence of the
MDM2 promoter was assayed for using PCR. As seen
in Figure 1D, the MDM2 promoter was detected in the
ChIP using the Sp1 antibody (lane 3) but not using the
lamin A antibody (lane 2). These data suggest that Sp1
binds this region of the MDM2 promoter in vivo.
Drosophila SL2 cells, which are deficient in Sp-related
proteins, were utilized. SL2 cells were transiently trans-
fected with an Sp1 expression vector (pPac-Sp1) and a
luciferase reporter plasmid driven by the MDM2 pro-
moter either wild-type (T/T) or homozygous (G/G) for
SNP309. As shown in Figure 1E, cotransfection of Sp1
strongly stimulatedluciferase expressionof thereporter
plasmid driven by the MDM2 promoter, as measured by
luciferase activity, suggesting that Sp1 can bind to this
Interestingly, the presence of SNP309 in the reporter
plasmid consistently showed higher Sp1-induced lucif-
erase expression (?50%) over the presence of wild-
type sequence in the reporter plasmid, as the increased
otides would have predicted. Similar results were ob-
tained when both reporter plasmids were transfected
into the mammalian HeLa cell line which has an abun-
dance of Sp1. The reporter plasmid containing SNP309
(G) yielded significantly higher luciferase levels (60%)
than the plasmid containing the wild-type sequence (T)
(data not shown).
Together, these data suggest that Sp1 can bind to
the MDM2 promoter and activate transcription and that
Sp1 by enhancing its DNA binding affinity to the MDM2
promoter. If true, individuals homozygous for SNP309
(G/G) should show heightened levels of MDM2 when
compared to individuals wild-type for SNP309 (T/T). To
tumor-derived cell lines were genotyped for SNP309.
SNP309 was present at afrequency similar to that found
in the normal volunteers. MDM2 RNA and protein levels
were compared in four cell lines homozygous for
SNP309 (G/G) to the levels found in four cell lines wild-
type for SNP309 (T/T). To compare MDM2 RNA levels,
total RNA was isolated from growing cells. MDM2 RNA
levels were measured by real-time PCR (TaqMan). The
presence of SNP309 correlated with high expression of
to the levels seen in cells wild-type for SNP309 (T/T)
(Figure 2A). MDM2 protein levels were also found to
be significantly higher (on average 4-fold) in cell lines
homozygous for SNP309 (G/G), as seen in the Western
blot analysis of total cell lysates (Figure 2B). Interest-
(A875, CCF-STTG1, and T47D) have been previously
reported to overexpress MDM2 when compared to cell
lines derived from similar tumor types (Landers et al.,
1997; Lu et al., 2002; Phelps et al., 2003). MDM2 protein
levels were found to be intermediate in four heterozy-
gous (T/G) cell lines (MDA-231, MCF-7, WM-9, and
CACL-7336): on average 1.9-fold higher than T/T cells
(data not shown). Thus, together these data support an
association of SNP309 with the increased levels of
gous for SNP309, endogenous Sp1 was inhibited, and
the resulting effect on endogenous MDM2 levels was
analyzed. Sp1 was inhibited using two different ap-
siRNAs specific to Sp1 RNA. Second, Sp1 activity was
inhibited by treating cells with mithramycin A, an aureo-
lic antibiotic that has been shown to selectively inhibit
Sp transcription factor-mediated transcriptional activa-
tion (Blume et al., 1991).
Sp1 siRNA reduced the protein levels of Sp1 over
2-fold in all three cell lines tested, as shown in Figures
3A and 3B, when compared to Sp1 levels in cells trans-
fected with either a nonspecific siRNA (NS, Figure 3A,
lanes 3, 6, and 9) or siRNAs targeted against lamin A/B
transcripts, although lamin A/B protein levels were dra-
matically reduced (Figure 3A, lanes 1, 4, and 7). Sp1
siRNA had no effect on another Sp family member, Sp3
(Figure 3A). As expected, reduction of Sp1 levels lead
to the reduction of the protein levels of one of its known
target genes, cyclin D1, in all three cell lines tested
(Figure 3A; Grinstein et al., 2002). As predicted, reduc-
tion of Sp1 levels dramatically decreased MDM2 levels,
up to 3-fold in cells homozygous for SNP309 (A875 and
tion of Sp1 levels had no significant effect on MDM2
levels (Figures 3A and 3B). Inhibition of Sp1 activity by
treating cells with mithramycin A showed similar effects
on MDM2 levels. A cell line homozygous for SNP309
(T47D) and a cell line wild-type for SNP309 (HL60) were
treated with various concentrations of mithramycin A
for 24 hr, and MDM2 protein levels were analyzed as
shown in Figure 3C. In cells homozygous for SNP309
MDM2 levels. Specifically, 200 nM mithramycin A re-
duced MDM2 levels more than 5-fold in T47D, while no
significant effect was seen in cells wild-type for SNP309
(HL 60). Similar effects were seen in other cell lines as
shown in Figure 3D. In two more cell lines homozygous
for SNP309 (Manca and A875), mithramycin A treatment
significantly decreased MDM2 levels. However, mithra-
mycin A only slightly reduced the MDM2 levels in two
more cell lines wild-type for SNP309 (HeLa and ML-1).
In summary, both methods of Sp1 inhibition can prefer-
entially reduce the heightened levels of MDM2 in cells
homozygous for SNP309, thereby supporting the hy-
pothesis that the Sp1 transcription factor is indeed re-
Figure 2. SNP309 Associates with the Overexpression of MDM2
(A) The levels of MDM2 RNA from tumor-derived cell lines are shown as measured by real-time PCR (TaqMan). All values have been normalized
to the level ofGAPDH and are the averages of three independentreadings. The standard deviation is given in theparenthesis next to each value.
(B) A Western blot analysis of MDM2 levels found in total cell lysates is shown. The upper panel is a Western blot using a monoclonal MDM2
antibody (SMP-14). Full-length MDM2 protein runs as a doublet and is indicated by arrows. The middle panel is a Western blot of the same
filter using a monoclonal antibody to p53 (Do-1) and the lower panel is a Western blot of the same filter using a monoclonal antibody to actin.
sponsible for the heightened levels of MDM2 in cells
homozygous for SNP309.
Overexpression of MDM2 can lead to the inactivation
of the p53 pathway (Lain and Lane, 2003). We therefore
reasoned that the SNP309-homozygous cell lines, with
higher levels of MDM2, should have an attenuated p53
response. To test the p53 response, these cells were
treated with the chemotherapeutic drug etoposide
(VP16) to induce DNA damage. This DNA damage acti-
vates the p53 pathway, leading to DNA repair, cell cycle
arrest, and apoptosis. Significant death (20%–35% of
the total cell population) was observed in cells that are
wild-type at p53 and wild-type for SNP309 (T/T) (H460,
ML-1, Tera-2: Figure 4A, lanes 1–3). Interestingly, cells
with wild-type p53 that are homozygous for SNP309
(G/G) (Manca, CCF-STTG1, and A875: Figure 4A, lanes
6–8) showed much lower death rates (on average
2%–3% of the total cell population). In fact, the percent-
age of cells dying after treatment was similar to that
seen in cells either mutant (T47D) or null (Saos2) for the
p53 gene itself (compare lanes 5 and 4 with 1–3). Four
cell lines heterozygous for SNP309 (T/G) and wild-type
for p53 were also assayed for their response to etopo-
side and found to be intermediate in their response: on
average 5%–7% of the total cell population (data not
shown). Together, these data demonstrate that the DNA
damage response pathway is attenuated in the SNP309
homozygous cell lines.
These data support the model that the heightened
If true, reducing the MDM2 levels should allow for the
cells lines. To test this, MDM2 levels were reduced by
blocking Sp1-mediated transcription by mithramycin A
treatment, and the activity of the p53 pathway was as-
sessed by measuring the apoptotic response after DNA
damage (etoposide treatment). Specifically, Manca (ho-
mozygous for SNP309) andML-1 (wild-type for SNP309)
A and etoposide for 48 hr, after which the percentage
of cells undergoing apoptosis was measured. Interest-
ingly, as shown in Figure 4B, the percentage of etopo-
in Manca (SNP309 homozygous cell line) with mithra-
mycin A treatment. However, in ML-1 (SNP 309 wild-
type cell line), mithramycin A treatment had no obvious
effect on etoposide-induced cell death. Similar results
were obtained using various incubation times (data not
shown). These data demonstrate that reduction of
MDM2 levels by inhibiting Sp1 activity can reverse the
attenuated p53 response in the SNP309 homozygous
cell lines, thereby supporting the model that the height-
ened MDM2 levels due to increased Sp1 activity on the
of the p53 pathway.
p53 responds to DNA damage by activating a tran-
scriptional program (Jin and Levine, 2001). If the low
zygous SNP309 cell lines, is due to inhibition of p53,
the p53 transcriptional program should be weakened in
these cells. To test this, cells with wild-type p53 were
treated with etoposide, and then the induction of 27
known p53-responsivegenes was compared inthe cells
either homozygousfor SNP309or wild-typefor SNP309.
RNA was isolated from cells before and after etoposide
treatment, and gene expression was analyzed using the
A SNP in Mdm2, the p53 Pathway, and Tumor Formation
Figure 3. Inhibition of Endogenous Sp1 Can Reduce the Heightened Levels of Endogenous MDM2 in Cells Homozygous for SNP309
(A) Cell lines homozygous for SNP309 (G/G; A875 and T47D) and wild-type for SNP309 (T/T; HeLa) were transfected with siRNA specific for
Sp1, for lamin A/B, or a nonspecific siRNA (NS). Depicted here are autoradiographs of Western blots using antibodies specific for Sp1, Sp3,
MDM2, cyclin D1, lamin A/B, and Gapdh. Arrows indicate the positions of the proteins.
(B) The bar graphs represent the quantification of Western blot analysis in (A). The relative expression levels in cells transfected with siRNA
for Sp1 was determined as the percentage of either Sp1 or MDM2 in the cells transfected with Sp1-siRNA to those transfected with nonspecific
siRNA. Each value is the average of at least three independent experiments and the error bars represent the standard deviations.
(C and D) Cell lines either homozygous for SNP309 (G/G; T47D, A875, and Manca) or wild-type for SNP309 (T/T; HL60, H460, and Ml-1) were
treated with various concentrations of mithramycin A for 24 hr. Autoradiograms of Western blots using either MDM2 or Gapdh antibodies are
shown. Arrows indicate the positions of the proteins.
AffyMetrix GeneChip array (Human Genome U95A Array).
responsive genes above 5-fold (Figure 4C). Specifically,
nine genes in the lung carcinoma cell line, H460 (Figure
4C, lane 1), six genes in the myeloid leukemic cell line,
ML-1 (Figure 4C, lane 2), and four genes in the testicular
teratoma cell line, Tera-2 (Figure 4C, lane 3), were in-
duced. Incontrast,cells homozygous forSNP309showed
a much-weakened p53-dependent transcriptional re-
did not induce any known p53-responsive gene above
5-fold and both the astrocytoma cell line, CCF-STTG1
(Figure 4C, lane 6), and the Burkitt’s lymphoma cell line,
Manca (Figure 4C, lane 5), only induced one gene above
5-fold. These data demonstrate that the ability of p53
to act as a transcriptional activator for known target
genes is attenuated in cells homozygous for SNP309.
Taken together, these data support a model that cells
homozygous for SNP309 (G/G) express higher levels of
been shown to inhibit p53 minimally by three different
mechanisms. One well-studied mechanism is its ability
to serve as an E3 ubiquitin ligase, targeting p53 for
proteosamal degredation (Haupt et al., 1997; Honda et
al., 1997; Kubbutat et al., 1997). Upon certain cellular
stresses, such as DNA damage, p53 levels rise, as the
half-life of p53 dramatically increases. The increase of
the half-life of p53 has been attributed to the inability of
MDM2 to target p53 for degradation after DNA damage.
Specifically, immediately after DNA damage, MDM2 lev-
els decrease dramatically as does its affinity for p53
binding (Michael and Oren, 2003).
To address what effect the elevated MDM2 levels in
G/G cells have on p53 levels, the levels of wt p53 in
cells before and after stress (DNA damage) were moni-
ences in basal wild-type p53 levels between cells wild-
type or homozygous for SNP309 (compare p53 levels
in lanes 1, 2, and 3 to those in lanes 5, 7, and 8). To rule
out the possibility that MDM2 cannot target p53 for
degradation in G/G cells, MDM2 levels were artificially
reduced in nonstressed cells, and p53 levels were moni-
tored. As seen in Figure 4D, reduction of MDM2 levels,
using siRNAs, in the two G/G cell lines tested led to
the significant stabilization of p53. The stabilized p53 is
active, as p21 levels also rise, and p21 is a well-known
Figure 4. p53 Activity Is Attenuated in Cells Homozygous for SNP309
(A) In the bar graph, the percent of the population of cells which are dead after 24 hr of either 1 ?M (black) or 5 ?M (gray) etoposide treatment
is plotted for each cell line. Each value is the average of three independent experiments and the error bars represent the standard deviations.
(B) Inhibition of Sp1 activity by mithramycin A treatment increases DNA damage-induced cell death in cells homozygous for SNP309. Manca
(homozygous for SNP309) and ML-1 (wild-type for SNP309) were treated with either 1 ?M (black) or 5 ?M (gray) etoposide and various
concentrations of mithramycin A for 48 hr. The percentage of the dead cell induced by etoposide and mithramycin A treatment was measured
and subtracted with any death induced by mithramycin A treatment alone. Fold induction of cell death is depicted in the bar graph. Each
value is the average of at least three independent experiments and the error bars represent the standard deviation.
(C) In the bar graph, the number of known p53 target genes induced after DNA damage (etoposide treatment) above 5-fold from levels found
in untreated cells is plotted for each cell line.
(D) MDM2 can inhibit p53 in cell lines homozygous for SNP309 (G/G; A875 and CCF). Cells were transfected with siRNA specific for MDM2,
for lamin A/B or a nonspecific siRNA (NS). Depicted here are autoradiographs of Western blots using antibodies specific for MDM2, p53, p21,
p53 transcriptional target. These data suggest that
MDM2 can target p53 for degradation in G/G cells, but
that, in nonstressed cells, heightened levels of MDM2
in G/G cells do not further reduce the levels of wt p53.
As mentioned above, the heightened MDM2 levels
could also inhibit the proper stabilization of p53 in re-
sponse to stress, thereby attenuating the p53 pathway.
To address this possibility, the stabilization of p53 after
Cells were treated with etoposide and harvested after
by Western blotting and the fold induction of p53 levels
after etoposide treatment was calculated for each cell
line and each time point. The values are depicted in the
bar graph in Figure 5B. As expected, the p53 levels in
cells wild-type for SNP309 (T/T) increased 4-fold or
above after only 1 hr of etoposide treatment. Interest-
data support the hypothesis that the heightened levels
stabilize p53 in response to cellular stresses like DNA
damage but that the heightened levels of MDM2 do not
further reduce the levels of wt p53 in nonstressed cells.
The inability to properly stabilize p53 in response to
DNA damage provides one possible mechanism for the
observed attenuation of p53 pathway in cells homozy-
gous for SNP309 (G/G).
The data thus far support a model that cells homozy-
gous for SNP309 express higher levels of MDM2,
thereby weakeningthe p53 pathway. Humanswho carry
a germline p53 mutation in one allele (Li-Fraumeni Syn-
A SNP in Mdm2, the p53 Pathway, and Tumor Formation
Figure 5. The Stabilization of p53 after DNA Damage Is Impaired in Cells Homozygous for SNP309, and the p53 DNA Damage Response Is
Attenuated in Li-Fraumeni Fibroblasts Homozygous for SNP309 with High MDM2 Levels
(A) Cell lines either wild-type (T/T) or homozygous (G/G) for SNP309 were treated with etoposide and harvested after 1, 2, and 3 hr of treatment.
Depicted hereare autoradiographs of Westernblots using antibodies specificfor p53 and ran.The autoradiograms aredepicted, and exposures
were chosen to highlight the induction of p53 after etoposide treatment for each cell line. The p53 levels were measured and normalized to
ran levels from the autoradiograms in (A), and the fold induction of p53 after etoposide treatment from basal levels was calculated for each
cell line and each time point. The values are depicted in the bar graph in (B).
(C) A Western blot analysis of MDM2 levels found in total cell lysates is shown. The upper panel is a Western blot using a monoclonal MDM2
antibody. Full-length MDM2 protein is indicated an arrow. The lower panel is a Western blot of the same filter using a monoclonal antibody
(D) The fibroblastoid cell line derived from a germline p53 mutation carrier (Li-Fraumeni) and homozygous for SNP309 has an altered DNA
damage response when compared to Li-Fraumeni fibroblastoid cell lines wild-type for SNP309. Histograms are depicted of the DNA content
in cells after 24 hr of varying concentrations of etoposide treatment (0, 1, and 5 ?M, indicated above the panels).
drome) develop tumors at very high frequencies (Garber
et al., 1992; Li et al., 1990). These individuals develop
tumors on average at very young ages and can develop
multiple primary tumors throughout a lifetime. We hy-
pothesized that increased levels of MDM2 by SNP309
could further weaken the p53 pathway in Li-Fraumeni
individuals and further impact tumorigenesis. To test
this, the first approach was to analyze MDM2 levels
and the DNA damage response (etoposide treatment)
in fibroblasts derived from Li-Fraumeni individuals who
were either wild-type for SNP309 (T/T) or homozygous
for SNP309 (G/G). MDM2 levels were analyzed by West-
ern blotting for two fibroblast cell lines wild-type for
SNP309 (T/T) and one fibroblast cell line derived from
the Li-Fraumeni individual homozygous for SNP309
(G/G). As predicted by the analysis of tumor-derived cell
lines, the G/G fibroblasts showed significantly higher
MDM2 levels (above 3-fold) when compared to those
found in both T/T fibroblast cell lines (Figure 5C). Upon
etoposide treatment, the two fibroblast cell lines wild-
type for SNP309 (T/T) showed a characteristic p53-
mediated G2 arrest as noted by the accumulation of
the fibroblast cell line homozygous for SNP309 (G/G)
did not show such a response. In fact, no significant
difference in the DNA content of the SNP309 homozy-
gous (G/G) cells was seen. These data suggested that
both MDM2 levels and the p53 pathway in Li-Fraumeni
individuals, as measured by the DNA damage response
in fibroblasts, could be greatly affected by the presence
If SNP309 further debilitates the p53 pathway in Li-
Figure 6. SNP309 Associates with an Accelerated Age of Onset of Tumors and the Occurrence of Multiple Subsequent Tumors in Li-
(A) Of the 88 individuals in the Li-Fraumeni cohort, 66 were diagnosed with at least one cancer at a median age of 22 years old. The first
cancers the 66 individuals were diagnosed with are shown in the pie chart. Individuals either T/G or G/G are diagnosed on average 7 years
(circles) or wild-type for SNP309 (triangles) is plotted against the age of onset of STS (C) and breast cancer (D).
(E) The bar graph depicts the number of individuals who were first diagnosed with STS (black) and then with subsequent cancers (second,
third, fourth, and fifth) for individuals either wild-type (T/T), heterozygous (T/G), or homozygous (G/G) for SNP309.
Fraumeni individuals, tumor development might also be
affected. To address this possibility, 88 individuals who
are members of Li-Fraumeni families and have germline
mutations in one allele of p53 were studied, and the
frequency of SNP309 was found to be similar to the
frequencies found in the 50 normal volunteers (Hwang
et al., 2003). Sixty-six have been diagnosed with at least
uals), breast cancer (17), and osteosarcomas (13) were
the most prevalent first cancers (Figure 6A). Interest-
ingly, individuals who carried SNP309, in either the het-
erozygous or homozygous state, showed a significantly
earlier age of onset for all tumor types (Figure 6B). For
individuals with SNP309, the median age of tumor onset
was 18 years old, while in wild-type individuals the me-
dian age of tumor onset was 27 years old (p ? 0.031).
Specifically, the median age of onset of STS in wild-
typeindividuals wasfourteenyearsold, whileinSNP309
individuals the median age of onset was two years old
(p ? 0.019, Figure 6C). In those who developed breast
A SNP in Mdm2, the p53 Pathway, and Tumor Formation
cancer, the median age of onset in wild-type individuals
was 39 years old, and in SNP309 individuals the median
age of onset was 29 years old (p ? 0.01, Figure 6D).
Taken together, those carrying SNP309 developed all
tumor types 9 years earlier, STS on average 12 years
earlier and breast cancer on average 10 years earlier
than those who do not carry SNP309. These data sup-
port the hypothesis that the presence of SNP309 further
debilitates the p53 pathway in Li-Fraumeni individuals,
resulting in an even earlier tumor onset.
As mentioned above, another hallmark of the Li-
Fraumeni syndrome is the occurrence of multiple pri-
mary tumors in a lifetime. Interestingly, the presence of
SNP309 also correlated with the occurrence of multiple
0.0014, Fisher’s exact test, Figure 6E). Specifically, both
individuals homozygous for SNP309 (G/G) developed a
second and a third cancer, and one developed a fourth
and a fifth cancer. Of the nine people heterozygous for
SNP309 (T/G), seven developed a second cancer, four
a third, and one a fourth and a fifth cancer. In contrast,
of the nine people wild-type for SNP309 (T/T), only two
developed a second cancer and none a third, a fourth,
or a fifth cancer. These data indicate an association
with the presence of SNP309 and the occurrence of
independent subsequent cancers. Together, these data
support a model whereby the presence of SNP309 in
of the MDM2 protein, which then inhibits p53 and im-
namely the early age of onset of tumors and the occur-
rence of multiple primary tumors in a lifetime.
The next question raised was if SNP309 caused a
weakened p53 pathway, did this SNP act upon sporadic
cancers as well as genetically altered individuals with a
who develop sporadic adult STS and had no known
p53 mutation was studied. As seen in the pie chart in
Figure 7A, the sarcomas come from a variety of tissues.
Interestingly, as in the Li-Fraumeni individuals, SNP309
associates with an earlier age of onset of STS (Figure
7B). Specifically, those individuals homozygous for
SNP309 were diagnosed on average 12 years earlier
than those individuals without SNP309 (p ? 0.01, Figure
7C). As seen in Figure 7D, the frequency of the SNP309
G allele was greatly increased in those individuals who
developed STS at a young age. Those who developed
STS below the age of 41 years old had an allele fre-
quency for SNP309 of 50%, while the allele frequency
is only 33% for the whole group (p ? 0.0262, Figure 7D).
These data demonstrate that SNP309 does not require
the presence of an inactivating germline p53 mutation
to associate with earlier STS tumor formation.
in cell culture systems demonstrate that SNP309 in the
promoter of the MDM2 gene increases the binding affin-
ity of the transcriptional activator Sp1, which results in
high levels of MDM2 RNA and protein. The heightened
MDM2 levels were shown to lead to the attenuation of
observations that MDM2 is as a key negative regulator
of p53. The results of clinical epidemiological studies
demonstrated that SNP309 associates with minimally a
9-year earlier onset of tumors in both hereditary and
whereby SNP309 enhances the affinity of the transcrip-
tional activator Sp1 to the promoter of the MDM2 gene,
resulting in heightened transcription. Heightened levels
of MDM2 lead to the direct inhibition of p53, which re-
leases the cell from p53 tumor suppression. Inhibition
of endogenous Sp1 using either siRNA or mithramycin
A can significantly reduce SNP309-related overexpres-
sion of MDM2. The experiments presented here demon-
strate that reduction of SNP309-related overexpression
of MDM2 by mithramycin A treatment can reverse the
attenuated p53 response. Mithramycin A is an aureolic
antibiotic that has been used in humans to treat several
suggest that treating cancer patients homozygous for
to chemotherapeutic drugs by reducing MDM2 levels
to allow for a p53-dependent apoptotic response in-
duced by most chemotheraputic agents.
The model proposed here states that the heightened
levels of MDM2, due to SNP309, result in accelerated
tumor formation in humans. There is much evidence in
the literature to support the claim that high MDM2 levels
can positively impact tumorigenesis. For example,
Jones et al. (1998) created mice which overexpress
Mdm2 by using the entire MDM2 gene as a trans-gene.
These mice expressed an average of 4-fold more Mdm2
in various tissues relative to nontransgenic mice. Inter-
oped spontaneous tumors in a lifetime. Lundgren et
al. (1997) also showed that targeted overexpression of
Mdm2 in the murine mammary epithelium results in tu-
mors, albeit with a lower penitence (16%). These two
studies, together with numerous accounts of MDM2
overexpression or amplification in a variety of human
cancers, support the idea that heightened levels of
MDM2, by SNP309, could positively impact tumor for-
mation (Cordon-Cardo et al., 1994; Momand et al., 1998;
Oliner et al., 1992; Taubert et al., 2000).
MDM2 is a key negative regulator of p53. MDM2 can
targeting p53 for proteosamal degredation (Michael and
Oren, 2003). Observations in this report support the hy-
pothesis that the heightened levels of MDM2 in SNP309
cells (G/G) result in the inability to properly stabilize p53
in response to cellular stresses like DNA damage; this
hypothesisoffers onepossible mechanismof p53atten-
uation by MDM2 in SNP309 (G/G) cells. Interestingly,
the heightened levels of MDM2 do not further reduce
the levels of wild-type p53 in nonstressed cells. One
explanation for these observations is that the ability of
MDM2 to regulate p53 levels may not be limiting in
nonstressed cells, so heightened levels of MDM2 would
Experimental data have been presented in this report
which support the hypothesis that single nucleotide
polymorphisms in the tumor suppressor p53 pathway
Figure 7. SNP309 Associates with an Accelerated Age of Onset of Tumors in Sporadic Soft Tissue Sarcoma
(A) The subtypes of STS are depicted in the pie chart. Individuals homozygous for SNP309 showed a significantly earlier age of onset of STS,
on average 12 years earlier (B). The cumulative number of individuals either homozygous for SNP309 (circles) or wild-type for SNP309 (triangles)
is plotted against the age of onset of STS (C).
(D) SNP309 is at a higher frequency in the population with early onset STS. The cumulative number of individuals either homozygous for
SNP309 (circles), heterozygous for SNP309 (squares), or wild-type for SNP309 (triangles) is plotted against the age of onset of STS in those
who were diagnosed below the age of 41 years old (one standard deviation from the median age of onset).
(E) SNP309 could serve as a rate-limiting event in liposarcoma. The number of rate-limiting events required to create a cancer cell and
subsequent tumor was calculated using I(t) ? kt
genotype by plotting the cumulative number of individuals either homozygous for SNP309 (circles), heterozygous for SNP309 (squares), or
wild-type for SNP309 (triangles) against the age of onset of liposarcoma and performing a nonlinear regression as a power function to obtain
the exponent for each curve. The equations for the trend-line, R2value, and R value are displayed for each genotype.
r?1, whereby r is the number of rate-limiting events. The R value was calculated for each
have no further impact on p53 levels. Only when MDM2
activity is reduced after cellular stress (or artificially,
e.g., siRNA in nonstressed cells, Figure 4D) does MDM2
become limiting, where heightened MDM2 levels would
result in less induced p53 (Figures 5A and 5B).
The results presented in this report suggest that
SNP309 can impact tumorigenesis in humans who carry
a germline inactivating mutation in one p53 allele (Li-
Fraumeni Syndrome). The data suggest that SNP309
can further lower the age of onset of tumors in these
individuals on average 9 years and increase the occur-
ble scenario to explain these observations is that high
weakened p53 tumor suppressor pathway resulting in
a higher mutation rate, poorer DNA repair processes,
and reduced apoptosis leading to faster and more fre-
quent tumor formation.
Fifty years ago it was noted that age-specific inci-
dence of many tumors increases with the power of age,
be correlated to the number of rate-limiting steps in-
volved in the formation of a cancer cell and subsequent
tumor. Specifically, the function I(t) ? ktr?1was used to
describe the incidence of cancer (I) observed at a given
age (t), whereby r is the number of rate-limiting events
which have to occur in cells at the constant rate (k). This
equation or derivatives thereof have been successfully
used to describe many cancers (Knudson, 2001), most
notably in Alfred G. Knudson’s description of hereditary
and nonhereditary forms of retinoblastoma, which later
became known as the two hit theory. As shown in Fig-
ures 6 and 7, the age-incidence curves of those patients
with SNP309 (G/G) vary greatly from those individuals
with a genotype of T/T or even T/G. Interestingly, when
the number of rate-limiting events(R value) is calculated
for each genotype at the SNP309 locus in sporadic lipo-
sarcomas, T/T gives a value of 4.8, T/G 3.5, and G/G
2.5 (Figure 7E). The fact that the heterozygote reduces
A SNP in Mdm2, the p53 Pathway, and Tumor Formation
by tandemly inserting two copies of double-stranded oligonucleo-
tides containing the same sequence as used for the EMSAs into
pGL2 luciferase reporter plasmid (Promega). The clones were con-
firmed by DNA sequencing.
Drosophila Schneider’s SL2 cells were seeded at 1.5 ? 106cells
per well in a six-well plate 24 hr before transfection. Transient trans-
to the manufacturer’s instructions. The DNA transfection mixture
contained 250 ng reporter plasmid, varying amounts of pPac-Sp1
plasmid, and empty pPac vector to normalize DNA concentrations.
The expression plasmids pPac and pPac-Sp1 were kindly provided
by Dr. Richard D. Kolodoner (La Jolla, California). Cells were har-
vested 48 hr after transfection in reporter lysis buffer (Promega)
and assayed for luciferase activity. Each extract was analyzed in
duplicate, and at least three independent experiments were per-
formed. Luciferase activities were normalized to cellular protein,
measured by the Bio-Rad protein assay system.
that the G allele can serve as a rate-limiting event in the
formation of a sarcoma.
It is not difficult to propose a model to explain why
Over the past 50 years, it has become clear that age-
specific incidence of cancer is dependent minimally on
three factors: the number of rate-limiting mutations re-
quired for a given cancer, the mutation rate per mitosis,
and the net proliferation rate of the effected cells (cell
division rate minus cell death rate; Knudson, 2001). Inhi-
bition of the p53 pathway by SNP309 could potentially
affect all three of these factors (Jin and Levine, 2001;
Lain and Lane, 2003). The p53 gene itself is thought of
as a rate-limiting mutation in many cancer types, as it
is found mutated in over 50% of all human tumors, and
humans who carry a germline p53 mutation develop
cancer with increased incidence and on average early
in life. The wild-type p53 pathway is also thought to
reduce mutation rates per mitosis, as loss of p53 leads
mosomal abnormalities. Finally, the p53 pathway also
impacts the net proliferation rate of cells, as it functions
signals like DNA damage and oncogene activation.
Therefore, inhibition of the p53 pathway by SNP309
could affect all three factors, which have been shown
to influence the age-specific incidence of cancer to ac-
celerate carcinogenesis in an individual.
To analyze MDM2 RNA levels, total RNA was isolated from cell
pellets using RNeasy (Qiagen). cDNAs were made using TaqMan
reverse transcription reagents from Applied Biosystems. Real-time
PCR was carried out using the ABI Prism 7000 sequence detection
system. Probe and primer sets for MDM2 and Gapdh were pur-
chased as predeveloped assays from Applied Biosystems.
To analyze protein levels, total cell extracts were made by using a
detergent lysis buffer (50 mM Tris [pH 7.5] 150 mM NaCl, 1% NP40,
0.1% SDS, 0.5% deoxycholic acid, 5mM EDTA, and a protease
protein was run on a 4%–20% tris-glycine gel (Invitrogen) and trans-
ferred to a PVDF membrane. MDM2 was detected using the mouse
monoclonal antibody SMP-14; Sp1 and Sp3 were detected using
the rabbit polyclonal antibody PEP2 and D-20, respectively. Lamin
clonal antibody 346, DCS-6, Do-1, and C-2, respectively. Gapdh
were purchased from Santa Cruz.
The MDM2 promoter was analyzed for sequence variation by PCR
amplification and subsequent sequencing, primer 1: CGGGAGTT
CAGGGTAAAGGT and primer 2: AGCAAGTCGGTGCTTACCTG.
Gene Silencing with siRNA
Sp1 siRNA targeted to AATGAGAACAGCAACAACTCC was used to
lower Sp1 expression. Two hundred picomoles siRNA duplex was
reagent (Invitrogen) according to the manufacturer’s instructions.
Control siRNA duplex has no known target in mammalian genomes
and was usedas follows: sense UUCUCCGAACGUGUCACGUdTdT,
antisense ACGUGACACGUUCGGAGAAdTdT. Lamin A/C siRNAs
were purchased from Qiagen and MDM2 siRNAs from Dharmacom.
Cells were lysed 48 hr after transfection, and protein levels of Sp1,
Sp3, and MDM2 were analyzed as described above.
Electrophoretic Mobility Shift Assays (EMSA)
EMSAs were performed with a LightShiftTM Chemiluminescent
EMSA Kit (PIERCE, Rockford, Illinois). The binding reactions were
performed for 20 min at room temperature in 10 mM Tris-HCl ([pH
7.5] at 25?C), 1 mM MgCl2, 50 mM NaCl, and 0.5 mM DTT, 4%
glycerol, 50 ?g/ml poly (dI-dC) (dI-dC), 10 fmol biotin 3?-end-labeled
double-stranded oligonucleotides, and purified recombinant Sp1
protein. After incubation, samples were separated on a native 4%
polyacrylamide gel and then transferred to a nylon membrane. The
positions of the biotin end-labeled oligonucleotides were detected
by a chemiluminescent reaction with streptavidin-horseradish per-
oxidase according to the manufacturer’s instruction and visualized
by autoradiography. The nucleotide sequence of the double-
stranded oligonucleotides with either wild-type sequence or the
SNP is as follows: 5?-CCGGGGGCTGCGGGGCCGCTT/GCGGCGC
Mithramycin A Treatment
Logarithmically growing cells were treated with various concentra-
tions of mithramycin A (Sigma) for 24 hr. After treatment, cells were
lysed and protein levels of MDM2 were analyzed.
Cell Viability Analysis
cells were treated with either 1 ?M or 5 ?M etoposide for 24 hr. To
analyze etoposide-induced cell death after inhibition of Sp1, grow-
ing cells were treated with various concentrations of mithramycin
A and etoposide for 48 hr. After treatment, cells were harvested and
viability was measured using the Guava ViaCount assay (Guava).
Proteins were crosslinked to DNA in 1% formaldehyde. After wash-
ing, cells were lysed in detergent lysis buffer. Lysates were washed
andsonicated.Two microgramsofantibodieswere addedandincu-
bated overnight. Protein A/G Plus beads (Santa Cruz) were used,
and after extensive washing, crosslinks were removed at 65?C over-
using the QIAquick PCR purification kit (Qiagen). Ten percent of
purified DNA was analyzed by PCR. The entire PCR reaction was
mide gel and subsequent ethidium bromide staining.
Cell Cycle Analysis
To analyze the DNA damage response in fibroblasts, all cells were
treated with two concentrations of etoposide (1 ?M and 5 ?M) for
24 hr. Cells were harvested and fixed with methanol and kept
at ?20?C for 40 min or overnight. After washing and equlibration in
PBS, the cell pellet was resuspended in 1 ml of staining solution
containing 50 ?g of RNase A and 0.5 ?g of propidium iodide per
ml in PBS and kept at room temperature for over 30 min. The cells
Luciferase Reporter Assays
The MDM2 promoter-luciferase reporter plasmids containing either
the wild-type sequence or the SNP309 sequence were constructed
Cell Download full-text
were then applied to the fluorescence-activated cell sorter (FACS-
Calibur; Becton Dickson). The FACSCalibur program was used to
sort and count the cells.
dent role for Mdm2 in tumorigenesis. Proc. Natl. Acad. Sci. USA
Knudson, A.G. (2001). Two genetic hits (more or less) to cancer.
Nat. Rev. Cancer 1, 157–162.
Kubbutat, M.H., Jones, S.N., and Vousden, K.H. (1997). Regulation
of p53 stability by Mdm2. Nature 387, 299–303.
Lain, S., and Lane, D. (2003). Improving cancer therapy by non-
genotoxic activation of p53. Eur. J. Cancer 39, 1053–1060.
Landers, J.E., Cassel, S.L., and George, D.L. (1997). Translational
enhancement of mdm2 oncogene expression in human tumor cells
containing a stabilized wild-type p53 protein. Cancer Res. 57, 3562–
Leach, F.S., Tokino, T., Meltzer, P., Burrell, M., Oliner, J.D., Smith,
S., Hill, D.E., Sidransky, D., Kinzler, K.W., and Vogelstein, B. (1993).
p53 Mutation and MDM2 amplification in human soft tissue sarco-
mas. Cancer Res. 53, 2231–2234.
Li, F.P., Strong, L.C., Fraumeni, J.F., Jr., Nelson, C.E., Kim, D.H.,
Kassel, J., Gryka, M.A., Bischoff, F.Z., and Tainsky, M.A. (1990).
Germ line p53 mutations in a familial syndrome of breast cancer,
sarcomas, and other neoplasms. Science 250, 1233–1238.
tumor resistance to p53. Cancer Res. 62, 1305–1310.
Lundgren, K., Montes de Oca Luna, R., McNeill, Y.B., Emerick, E.P.,
Spencer, B., Barfield,C.R., Lozano, G., Rosenberg,M.P., and Finlay,
C.A. (1997). Targeted expression of MDM2 uncouples S phase from
mitosis and inhibits mammary gland development independent of
p53. Genes Dev. 11, 714–725.
Mendrysa, S.M., McElwee, M.K., Michalowski, J., O’Leary, K.A.,
Young, K.M., and Perry, M.E. (2003). mdm2 is critical for inhibition
Mol. Cell. Biol. 23, 462–472.
Michael, D., and Oren, M. (2003). The p53-Mdm2 module and the
ubiquitin system. Semin. Cancer Biol. 13, 49–58.
Momand, J., Jung, D., Wilczynski, S., and Niland, J. (1998). The
MDM2 gene amplification database. Nucleic Acids Res. 26, 3453–
Montes de Oca Luna, R., Wagner, D.S., and Lozano, G. (1995).
Rescue of early embryonic lethality in mdm2-deficient mice by dele-
tion of p53. Nature 378, 203–206.
Oliner, J.D., Kinzler, K.W., Meltzer, P.S., George, D.L., and Vo-
gelstein, B. (1992). Amplification of a gene encoding a p53-associ-
ated protein in human sarcomas. Nature 358, 80–83.
Phelps, M., Darley, M., Primrose,J.N., and Blaydes, J.P. (2003). p53-
independent activation of the hdm2–P2 promoter through multiple
transcription factor response elements results in elevated hdm2
expression in estrogen receptor alpha-positive breast cancer cells.
Cancer Res. 63, 2616–2623.
Ries, S., Biederer, C., Woods, D., Shifman, O., Shirasawa, S., Sasa-
zuki, T., McMahon, M., Oren, M., and McCormick, F. (2000). Oppos-
ing effects of Ras on p53: transcriptional activation of mdm2 and
induction of p19ARF. Cell 103, 321–330.
Taubert, H., Koehler, T., Meye, A., Bartel, F., Lautenschlager, C.,
Borchert, S., Bache, M., Schmidt, H., and Wurl, P. (2000). mdm2
mRNA level is a prognostic factor in soft tissue sarcoma. Mol. Med.
Zauberman, A., Flusberg, D., Haupt, Y., Barak, Y., and Oren, M.
(1995). A functional p53-responsive intronic promoter is contained
within the human mdm2 gene. Nucleic Acids Res. 23, 2584–2592.
A randomization test is employed to determine the statistical signifi-
cance of the age of onset of cancer between the groups with and
without SNP309. The two groups are compared pair-wise, and each
instance of an element from the second group greater than an ele-
ment of the first group adds one to the distance. This total distance
is the cutoff. The lists are then randomly permuted, holding fixed
the number of elements of each list. The calculated p value is the
percent of randomized groups that have a distance less than the
cutoff as determined by a large Monte Carlo simulation.
We would like to thank A. Teresky, H. Jaramillo, B. Wypior, and Z.
Feng for their expert technical assistance. We thank W. Liu and D.
Notterman in the CINJ Core Expression Array Facility. T. Gopen
kindly provided us with the status of the p53 gene in the Manca cell
line. We thank A.G. Knudson and S. Friend for helpful discussions.
We thank S. Harris for help with the preparation of this manuscript.
Received: April 14, 2004
Revised: September 23, 2004
Accepted: October 13, 2004
Published: November 23, 2004
Blume, S.W., Snyder, R.C., Ray, R., Thomas, S., Koller, C.A., and
Miller, D.M. (1991). Mithramycin inhibits SP1 binding and selectively
inhibits transcriptional activity of the dihydrofolate reductase gene
in vitro and in vivo. J. Clin. Invest. 88, 1613–1621.
Cordon-Cardo, C., Latres, E., Drobnjak, M., Oliva, M.R., Pollack, D.,
Woodruff, J.M., Marechal, V., Chen, J., Brennan, M.F., and Levine,
A.J.(1994). Molecularabnormalitiesof mdm2andp53genes inadult
soft tissue sarcomas. Cancer Res. 54, 794–799.
Donehower, L.A., Harvey, M., Slagle, B.L., McArthur, M.J., Mont-
gomery, C.A., Jr., Butel, J.S., and Bradley, A. (1992). Mice deficient
for p53 are developmentally normal but susceptible to spontaneous
tumours. Nature 356, 215–221.
Freedman, D.A., and Levine, A.J. (1999). Regulation of the p53 pro-
tein by the MDM2 oncoprotein–thirty-eighth G.H.A. Clowes Memo-
rial Award Lecture. Cancer Res. 59, 1–7.
Garber, J.E., Friend, S.H., Strong, L.C., Patenaude, A.F., Juengst,
E.T., Reilly, P.R., Correa, P., Fraumeni, J.F., Jr., and Frebourg, T.
(1992). Germ-line mutations of the p53 tumor suppressor gene in
patients with high risk for cancer inactivate the p53 protein. J. Natl.
Cancer Inst. 84, 1156–1160.
Sp1 as G1 cell cycle phase specific transcription factor in epithelial
cells. Oncogene 21, 1485–1492.
Haupt, Y., Maya, R., Kazaz, A., and Oren, M. (1997). Mdm2 promotes
the rapid degradation of p53. Nature 387, 296–299.
Honda, R., Tanaka, H., and Yasuda, H. (1997). Oncoprotein MDM2
is a ubiquitin ligase E3 for tumor suppressor p53. FEBS Lett. 420,
Hurtado, J., and Esbrit, P. (2002). Treatment of malignant hypercal-
caemia. Expert Opin. Pharmacother. 3, 521–527.
Hwang, S.J., Lozano, G., Amos, C.I., and Strong, L.C. (2003). Germ-
line p53 mutations in a cohort with childhood sarcoma: sex differ-
ences in cancer risk. Am. J. Hum. Genet. 72, 975–983.
Jin, S., and Levine, A.J. (2001). The p53 functional circuit. J. Cell
Sci. 114, 4139–4140.
Jones, S.N., Roe, A.E., Donehower, L.A., and Bradley, A. (1995).
Rescue of embryonic lethality in Mdm2-deficient mice by absence
of p53. Nature 378, 206–208.
A. (1998). Overexpression of Mdm2 in mice reveals a p53-indepen-