| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
Department of Molecular Biology, Cell Biology, and Biochemistry, Brown University, Providence, Rhode Island 02912
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In most cells, loss of anchorage results in the suppression of cyclin D expression (8) . In anchorage-deprived human fibroblasts and mouse NIH-3T3 cells, cyclin D1 mRNA levels were reduced severalfold, and cyclin D1 protein synthesis was almost completely inhibited, indicating that both transcriptional and translational mechanisms are involved in the down-regulation (9 , 10) . Constitutive expression of cyclin D1 overcame the G1 arrest observed in anchorage-deprived cells. Cyclin E expression was not significantly affected; however, cyclin E-Cdk6 2 complexes were not activated in response to serum stimulation (11 , 12) . This was attributable partly to an increase in the expression levels of the cyclin-dependent kinase inhibitor proteins p21Cip1 and p27Kip1 and partly to the absence of cyclin D1-Cdk4/6 complexes, which caused a redistribution of the p21 and p27 pools to cyclin E-Cdk2 (8) . Because in anchorage-deprived cells both Cdk4/6 and Cdk2 activation is suppressed, the Rb family proteins remain unphosphorylated, and the E2F-mediated induction of several delayed-response genes does not take place.
Cyclin A expression is also strongly adhesion-dependent (13 , 14) . The cyclin A promoter contains a motif that binds inhibitory E2F-4/p107 complexes. In attachment-deprived cells, p107 remains unphosphorylated and bound to E2F-4, thus suppressing cyclin A expression. Adhesion may also regulate cyclin A expression via an E2F-independent mechanism; anchorage-deprived NRK fibroblasts expressed cyclin D1 and exhibited Cdk2 activity, but cyclin A expression remained anchorage-dependent, and ectopic expression of cyclin A caused S-phase entry in the absence of attachment (15) .
It is well established that extracellular matrix proteins such as fibronectin can trigger a rapid, transient activation of the mitogen-activated protein kinase pathway (16, 17, 18) . In NIH-3T3 cells, activation of Ras in response to platelet-derived growth factor or epidermal growth factor stimulation was unaffected by anchorage deprivation, but activation of Erk was significantly reduced (19 , 20) . Lin et al. (19) implicated Raf-1 activation as the key step affected by matrix attachment, whereas Renshaw et al. (20) reported that Raf-1 activation was normal, but mitogen-activated protein/Erk kinase (MEK) activation was impaired. It is interesting to note that even in v-Ras- or v-Raf-transformed cells, Erk activity was decreased significantly by anchorage deprivation; however, because of the increased signaling elicited by the oncoproteins, Erk activity was maintained at levels similar to that seen in nontransformed attached cells (20) .
Expression of oncogenic Ras in anchorage-deprived NRK, NIH-3T3, or Rat-6 cells resulted in the induction of cyclin D1 and activation of cyclin E-Cdk2, leading to the phosphorylation of Rb family proteins, release of E2F, expression of cyclin A, and entry into S phase (21 , 22) . Experiments using Ras "effector loop" mutants (23) that selectively signal through only one of the major Ras targets, Raf, phosphatidyl inositol-3 kinase, or Ral guanine nucleotide dissociation stimulator, showed that none of these pathways was sufficient by itself to elicit anchorage-independent growth. A pairwise activation of signaling pathways synergized in producing anchorage-independent growth, but the robustness of this response was highly cell line-dependent. These results suggest that individual Ras effector pathways may not uniquely mediate anchorage dependence but may rather cooperate to provide a general signal to overcome a threshold requirement, which may be different in different cell lines.
Deregulation of c-myc and mutation of ras are commonly found in many human tumors, and several lines of evidence indicate that Myc and oncogenic Ras cooperate in causing malignant transformation (24) . Neither c-Myc nor oncogenic Ras alone was capable of transforming primary rodent fibroblasts, whereas together they did so effectively (25 , 26) . In later studies, transfection of oncogenic Ras alone elicited a conversion to a tumorigenic but not metastatic phenotype, whereas no tumors were observed in animals that received injections with c-myc-transfected cells (27) . Similar phenomena have been observed with hematopoetic and epithelial cells. Although immortalized murine lymphoid cell lines can be obtained by infection of bone marrow with v-ras alone, coinfection with v-myc resulted in cell lines that grew to higher densities, had reduced dependence on feeder cells, and grew more robustly in soft agar (28) . Along the same lines, transfection with c-myc and v-raf, but not with either alone, elicited efficient immortalization of p53 -/- hematopoetic cells (29) . In rat liver epithelial cells, transformation with both c-myc and oncogenic ras resulted in cultures that grew to much higher densities and formed 15–20 times more colonies in soft agar than cells transformed with ras alone (30) .
Despite a large body of literature, very little is known about the mechanism of c-Myc and Ras interaction. Is Myc required for Ras-induced transformation? What molecular events lead to the observed synergy of Ras and c-Myc? To begin to address these questions, we developed a cell culture model system in which one or both copies of the c-myc gene were inactivated by gene targeting in an immortalized rat fibroblast cell line (31, 32, 33) . In this study, we demonstrate that although c-Myc is not absolutely essential for the malignant transformation of immortalized rat fibroblasts by oncogenic Ras or Raf, a relatively modest 50% reduction in c-Myc expression results in a greater than 10-fold reduction in susceptibility to transformation. Furthermore, we show that this effect is correlated with the synergistic action of c-Myc and Ras in down-regulating high levels of p27 induced by anchorage deprivation.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2 (35)
was obtained from Richard Mulligan (Harvard Medical School, Boston,
MA). Cells were cultured in DMEM containing glutamine, pyruvate,
high glucose, and 3.7 g/liter sodium bicarbonate, supplemented with
10% calf serum and penicillin/streptomycin. Cells were maintained in a
5% CO2 atmosphere at 37°C. Electroporation
conditions were as described (34)
. All of the drug
selections were performed by seeding cells at 30–40% confluence
immediately before selection. v-Raf retrovirus was obtained from the
American Type Culture Collection. Ha-Ras(G12V) cloned in the
hygromycin-resistant retrovirus vector LXSH (36)
was
obtained from Dan DiMaio (Yale University, New Haven, CT). c-Myc
and c-Myc (
106–143) retrovirus vectors were obtained from Charles
Sawyers (Jonsson Comprehensive Cancer Center, University of California
at Los Angeles, Los Angeles, CA) (37)
. Construction
of LXSH vectors that express cyclin D1, cyclin E, or cyclin A has been
described (38)
.
Soft Agar Growth Assay.
Noble agar (2% w/v; DIFCO Laboratories) was autoclaved and cooled down
to 45°C. The sterile agar solution was mixed with 2 x DMEM, calf serum, and water to a final concentration of 1 x DMEM, 10% calf serum, and 0.8% or 0.4% agar for the lower
and upper layers, respectively, and held at 45°C. The lower layer
solution (5 ml) was aliquoted into 6-cm tissue culture dishes and
allowed to solidify at room temperature. Cells were trypsinized, spun
through medium supplemented with 10% calf serum to neutralize the
trypsin, resuspended in 0.5 ml of fresh medium, and counted in a
Coulter counter. Indicated amounts of cells were rapidly mixed with 1.5
ml of the upper solution and immediately plated onto the solidified
lower layer. After the upper layer had solidified, the dishes were
transferred to 37°C. Plates were fed weekly by overlaying with 1.5 ml
of fresh upper layer solution. Colonies were counted at the indicated
times either microscopically or visually.
Methyl Cellulose Assay.
The assay was performed as described by Assoian et al.
(39)
and modified by Kume et al.
(40)
. Methyl cellulose (4000 centipoises; Sigma
Chemical Co.) was made up to 7% (w/v) in sterile water heated to
80°C while stirring gently. When the particles were thoroughly wetted
and evenly dispersed, ice cold water was added to bring the methyl
cellulose concentration to 2.6% (w/v). The suspension was allowed to
cool to 4°C for at least 1 h with constant gentle stirring. An
equal volume of 2 x DMEM, 20% calf serum was added,
and the suspension was stirred gently overnight at 4°C. This methyl
cellulose/medium suspension could be stored at 4°C for several weeks.
Cells were trypsinized and spun through medium supplemented with 10%
calf serum to neutralize the trypsin. Cells (5 x 105 to 1 x 106
) were resuspended in 10 ml of the methyl
cellulose/medium suspension prewarmed to 37°C, transferred into
polypropylene tubes, and incubated at 37°C for 72 h (the caps of
the tubes were loosened to allow equilibration of the medium with the
CO2 atmosphere). At the end of the incubation
period, 40 ml of PBS were added and mixed well; cells were spun down at
1500 rpm for 5 min at 4°C and gently washed once with 10 ml of PBS.
Harvested cells were resuspended in a small volume of PBS, counted, and
analyzed by immunoblotting as indicated.
Immunoblotting Analysis.
Cells were grown to 80–90% confluence at the indicated temperatures
and were harvested by rapid lysis in Laemmli sample buffer
(32)
. Laemmli sample buffer was supplemented with protease
inhibitors (10 µg/ml aprotinin, 1 µg/ml leupeptin, 1 mM
phenylmethylsulfonyl fluoride). Equal loading of lanes was established
as described (32)
, except that Coomassie Blue-stained gels
were quantitated using the Gel Doc digital gel documentation system and
Molecular Analyst software (Bio-Rad). Gels were transferred to
Immobilon-P membranes (Millipore), and the immunoblots were probed
using standard procedures (32
, 41)
. Signals were
visualized with the enhanced chemiluminescence system (Amersham) or the
Super Signal R chemiluminescence system (Pierce Chemical Co.).
Antibodies used were: cyclin A and cyclin D1 obtained from Upstate
Biotechnology, Inc. (06-138, 06-137); cyclin E, p27, and p16 obtained
from Santa Cruz Biotechnology (sc-481, sc-528, sc-1661); Ras, LA069
hybridoma cell line obtained from Quality Biotech (Camden, NJ); and
Raf, PBB1 hybridoma cell line (42)
obtained from
Ulf Rapp (University of Würzburg, Würzburg,
Germany).
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
106–143), and clonal cell lines
overexpressing the c-Myc proteins were obtained by G418 selection and
confirmed by immunoblotting (data not shown).
Interference with intracellular c-Myc activity either by reducing the
expression of the endogenous protein by 50% or by expressing
dominant-defective c-Myc protein caused in both cases an approximately
10-fold reduction in the number of soft agar colonies (Fig. 1)
. Overexpressing wild-type c-Myc resulted in 1.6-fold more colonies
relative to the TGR-1 control. Infection with Ha-Ras(G12V) produced the
same results, although the colonies were larger and grew faster than
those produced by v-Raf (data not shown).
|
106–143) cells proved to be unstable, with expression of
the dominant-defective Myc protein decreasing upon subsequent passage,
and were thus not used in additional experiments. In contrast, HET15
cells were very stable, and the striking haplo-insufficient
c-myc transformation phenotype was highly reproducible. In
repeat experiments, reductions in transformation efficiency of up to
40–50-fold were recorded (e.g., see Fig. 2
).
|
. Clonal derivatives of
TGR-1 and HET15 infected with empty pLXSH retrovirus vector were
established as controls. Cyclin A expression did not affect the growth
rate or elicit anchorage-independent growth in any of the cell lines
(data not shown).
The transformation efficiency of the various cell lines was determined
in soft agar assays after infection with retrovirus vectors expressing
either v-Raf or Ha-Ras(G12V). In the experiment shown in Fig. 2
, the
transformation potential of HET15/LXSH cells was approximately 40-fold
below that of TGR/LXSH cells in response to both v-Raf and
Ha-Ras(G12V). TGR/CycA cells gave somewhat fewer colonies (2.8-fold at
most) than TGR/LXSH control cells. This result may be attributable to
the known toxicity of cyclin A overexpression (15)
. In
contrast, colony formation in HET15 cells was strongly rescued by the
ectopic expression of cyclin A (10–30-fold relative to HET/LXSH
controls). Consistent results were obtained in two independent trials
of this experiment.
Cyclin A Lowers the v-Raf Expression Threshold Required for
Transformation of c-myc ± Cells.
To investigate whether the observed difference in the transformability
of HET15 cells with or without ectopic cyclin A corresponds to a
differential requirement for oncogene expression, well developed
colonies were picked from soft agar, expanded into clonal cell lines,
and the expression level of v-Raf protein was examined by
immunoblotting. Four clones derived from HET/LXSH cells and six clones
derived from each of HET/CycA-1, HET/CycA-2, and HET/CycA-3 cells were
examined (Fig. 3)
. Although there was variability in v-Raf expression levels between
individual cell lines, clones derived from HET15 cells with ectopic
cyclin A clearly expressed significantly less v-Raf than the LXSH empty
vector controls.
|
Ectopic Expression of Cyclins D1 and E in
c-myc ± Cells Has a Minimal Effect on
Restoring Transformation Potential.
Activity of cyclin D1-Cdk4/6 and cyclin E-Cdk2 complexes has been shown
to be required for anchorage-independent growth (10
, 14) .
To determine whether ectopic expression of either cyclin D1 or cyclin E
can increase the transformation potential of c-myc ± cells, stable clonal cell lines were derived after infection of
HET15 cells with LXSH retrovirus vectors expressing either cyclin D1 or
E. As described above, for cyclin A cell lines, hygromycin B-resistant
colonies were expanded into cell lines, and cyclin expression was
determined by immunoblotting.
Five cell lines were recovered from the cyclin D1 infection, three of
which expressed moderate levels of cyclin D1. The cell line that in the
initial screen expressed the highest level of cyclin D1 protein grew
slowly and proved unstable upon further passage (data not shown);
therefore, it could not be used in the subsequent analysis. The two
remaining cell lines as well as two TGR-1-derived cell lines expressing
similar levels of cyclin D1 protein (Fig. 4C)
were infected with either Ha-Ras(G12V) or v-Raf retrovirus
vectors, and their transformation efficiency was measured in soft agar
assays. One of the two HET15-derived cell lines (HET/CycD-2) exhibited
an increase in transformation efficiency, whereas the other
(HET/CycD-1) was not significantly different from the HET15/LXSH
control cells (Fig. 4, A and B)
. Contamination of
the HET/CycD-2 cell line with TGR-1 cells was excluded by Southern blot
analysis of clones recovered from soft agar (data not shown). Both
TGR-1-derived cell lines exhibited a modest but reproducible increase
in soft agar colony formation. Although cyclin D1 levels in the two
HET15-derived cell lines were similar, the HET/CycD-2 cell line
reproducibly displayed a slightly higher level of expression. These
results indicate that the window of cyclin D1 expression that can
rescue transformation of c-myc ± cells may
be very narrow; slightly lower than optimal expression does not rescue,
whereas even a very modest increase above the optimum level is toxic to
cell growth.
|
. When this cell line was infected
with either Ha-Ras(G12V)-expressing or v-Raf-expressing retrovirus
vectors and analyzed in soft agar assays, its transformation potential
was indistinguishable from that of HET15/LXSH control cells (data not
shown).
To address the influence of clonal variation on transformation
potential, the cyclin D1 and E rescue experiments were repeated with
large pools of infected cells. After infection of HET15 cells with the
cyclin D1-expressing or cyclin E-expressing retrovirus vectors, at
least 100 hygromycin B-resistant colonies were pooled and reinfected
with either Ha-Ras(G12V)-expressing or v-Raf-expressing retrovirus
vectors. The time between the harvest of the pool and reinfection was
kept as short as possible to minimize the overgrowth of rapidly
dividing clones. An aliquot of cells was analyzed for cyclin D1 and
cyclin E expression by immunoblotting (Fig. 5B)
; both were found to be clearly overexpressed.
However, neither cyclin D1 nor cyclin E changed significantly the
transformation efficiency of HET15 cells in this experiment (Fig. 5A)
. In summary, these results indicate that, in contrast to
cyclin A, ectopic expression of cyclins D1 and E cannot efficiently
compensate for a reduction in c-Myc expression. Cyclin D1 appears to be
capable of rescue in a very narrow range of expression, but the levels
required for rescue are very close to those that cause inhibition of
growth.
|
. Because Ha-Ras(G12V) is known to antagonize this response
(44)
and loss of c-Myc expression has been observed to
result in elevation of p27 levels (38)
, it was of interest
to investigate the interplay of Ha-Ras(G12V) and c-Myc in regulating
p27 expression during anchorage-independent growth. TGR-1, HET15, and
HO15.19 cells were infected with a Ha-Ras(G12V)-expressing
retrovirus vector, selected with hygromycin B, and at least 100
colonies of each cell line were pooled together. The pools were
analyzed by immunoblotting and found to express similar levels of
Ha-Ras protein (Fig. 6B)
. The Ha-Ras(G12V)-expressing pools,
along with parental TGR-1, HET15, and HO15.19 controls, were then
deprived of anchorage for 72 h by culture in a methyl cellulose
suspension medium (see "Materials and Methods"). After 72 h,
the cells were harvested from the methyl cellulose, and p27 expression
was analyzed by immunoblotting (Fig. 6C)
. Introduction of
Ha-Ras(G12V) strongly decreased p27 expression in c-myc +/+
cells; however, this down-regulation was almost completely absent in
c-myc ± and -/- cells. It is noteworthy
that a mere 2-fold reduction in c-Myc expression (c-myc +/+
versus c-myc ± cells) resulted in
a significantly amplified defect in p27 down-regulation (5–10-fold).
These results indicate that c-Myc and Ha-Ras(G12V) may cooperate in
reducing p27 levels in anchorage-deprived cells.
|
). HLRaf-1,
HA2Raf-5, and parental HET15 cells were anchorage deprived for 72 h by growth in methyl cellulose suspension medium, and cell lysates
were analyzed by immunoblotting (Fig. 6D)
. Although p27 was
up-regulated by anchorage deprivation in all of the cases, cells
expressing high levels of v-Raf (HLRaf-1) displayed the lowest levels
of p27. In cells expressing low levels of v-Raf in combination with
ectopic cyclin A (HA2Raf-5), p27 levels were similar to those in HET15
control cells. These results suggest that the extent of p27
down-regulation during anchorage-independent growth depends on the
level of v-Raf expression and that cyclin A acts downstream to allow
anchorage-independent growth despite high p27 levels.
Very High Levels of Ha-Ras(G12V) Can Allow Anchorage-independent
Growth in Complete Absence of c-Myc.
Despite the fact that Ha-Ras(G12V) failed to transform c-myc -/-
(HO15.19) cells in a standard soft agar assay, the possibility remained
that very high Ha-Ras(G12V) levels could effectively down-regulate p27
despite a complete absence of c-Myc. Because previous data indicated
that clones expressing high enough Ha-Ras(G12V) levels would be rare,
HO15.19 cells were infected with the LXSH retrovirus vector expressing
Ha-Ras(G12V), and a large number of hygromycin-resistant colonies were
screened microscopically for signs of morphological transformation.
Several candidate clones were obtained, and the expression of Ha-Ras
was analyzed by immunoblotting (Fig. 7A)
. Three cell lines that expressed very high (HO/Ras-1),
high (HO/Ras-8), and moderate (HO/Ras-7) levels of Ha-Ras were selected
for further analysis.
|
. Furthermore, the ability to keep p27 expression low
when deprived of attachment strongly correlated with transformation
potential (Fig. 7C)
. Moderate expression of Ha-Ras(G12V;
HO/Ras-7) resulted in very low plating efficiency in either focus
formation (0.07%) or soft agar (0.48%) assays. Intermediate
expression of Ha-Ras(G12V; Ho/Ras-8) elicited good focus formation
(18.2%) but only very modest growth in soft agar (3.2%). Finally,
high Ha-Ras(G12V, Ho/Ras-1) expression allowed both focus formation
(11.1%) and soft agar growth (31.1%). For comparison, the plating
efficiency of TRG/Ras cells in either assay was in the 20–30% range
(data not shown). These results indicate that high enough levels of
Ha-Ras(G12V) can sufficiently down-regulate p27 to allow
anchorage-independent growth even in the complete absence of c-Myc. It
is important to note that although Ha-Ras(G12V) was capable of
transforming c-myc -/- cells to anchorage-independent
growth, it did not affect their slow growth phenotype; the doubling
time of HO/Ras-1 cells was in the range of 50–55 h, which is
indistinguishable from that of the parental HO15.19 cells
(33)
.
c-myc Haplo-insufficiency Affects Predominantly the
Promotion of Transformation.
A large body of work has identified three principal biological
activities of c-Myc: promotion of proliferation, transformation, and
apoptosis. Given that a 2-fold reduction in c-Myc levels leads to a
very significant reduction of transformation potential, it was
therefore of interest to investigate the cell cycle and apoptotic
phenotypes of c-myc ± cells. The cell cycle
effects have been investigated in detail in the past
(31, 32, 33)
; these studies showed that
c-myc ± cells display minimal cell cycle
defects, with the exponential phase growth rate (doubling time)
increasing at most by 20%.
Overexpression of c-Myc induces apoptosis, especially when combined
with a reduction of serum supplementation (45
, 46)
. When
TGR-1 cells were incubated under subconfluent conditions in medium
supplemented with 0.1% serum, 40–50% of the cells rounded up and
detached within 24–36 h (Fig. 8A)
. Examination of the floating cells for DNA laddering
indicated that they were apoptotic (Fig. 8C)
. Additional
assays of apoptosis, such as terminal deoxynucleotidyl
transferase-mediated dUTP-biotin nick end labeling, annexin staining,
or poly(ADP-ribose) polymerase cleavage, were also positive
(data not shown). Likewise, apoptosis could be induced in TGR-1 cells
by incubation in the presence of etoposide (Fig. 8B)
.
c-myc ± cells displayed a reduced apoptotic
response to both serum deprivation and etoposide, but the defect was
modest (less than 2-fold) under both conditions (Fig. 8, A and B)
. Interestingly, c-myc -/- cells were
profoundly defective in some (but not all) of the apoptotic responses;
a detailed mechanistic examination of these phenomena will be the topic
of future studies. In summary, the haplo-insufficient c-Myc phenotype
is restricted almost exclusively to transformation potential, namely, a
greater than 10-fold defect in susceptibility to transformation by
oncogenic Ras or Raf, whereas the cell cycle and apoptotic phenotypes
are relatively minor (20% decrease in growth rate and less than 2-fold
reduction in apoptosis).
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cyclin A expression in late G1 has been reported to be both adhesion- and c-Myc-dependent (14 , 15) , and ectopic expression of c-Myc restored cyclin A levels in adhesion-deprived Rat-1a fibroblasts (43) . These data, as well as the results reported here, indicate that cyclin A acts downstream of c-Myc in regulating anchorage-independent growth. Whereas in NRK and Rat-1a cells expression of cyclin A alone was sufficient to elicit anchorage-independent growth (15 , 43) , this was not the case with the TGR-1 cells used in this report. Likewise, NIH-3T3 cells cannot be induced to grow in soft agar by cyclin A alone.7 Although both Rat-1a and TGR-1 are subclones of the original Rat-1 cell line, Rat-1a was selected specifically for its ease of transformation and is the only rodent fibroblast cell line known to be transformed by c-myc alone (47) . The levels of cyclin A required to elicit anchorage-independent growth are likely to be somewhat variable in different cell lines, and deregulated cyclin A expression has been reported to be toxic by several groups (15 , 48 , 49) . Indeed, constitutive high level expression of cyclin A has been reported only in NRK and HeLa cells (15) , and all of the TGR-1-derived clones expressed only moderate levels of cyclin A.
The issue of toxicity likewise clouds the interpretation of cyclin E and D1 rescue of c-myc ± cells. Although in two different assays overexpression of cyclins E and D1 elicited clearly different results than overexpression of cyclin A, it could be argued that TGR-1 cells are sensitive to cyclin E and D1 overexpression and that toxicity set in before rescue became possible. This interpretation is supported by the isolation of one cyclin D1 cell line in which the c-myc haplo-insufficiency was clearly restored; however, the fact that despite repeated attempts only one cell line could be obtained and that in a large pool of clones cyclin D1 did not appreciably rescue indicates that such events must be rare and likely can only be elicited in a very narrow range of cyclin D1 expression. Overexpression of cyclin D1 but not of cyclin E using an inducible system was capable of supporting transient anchorage-independent growth in Rat-1 cells (50) . Inducible expression of cyclin E was reported to overcome p27-induced arrest in NIH-3T3 cells (51) , but retroviral overexpression of cyclin E did not alleviate a p27-induced arrest in Rat-1 cells (52) .
It is well established that anchorage deprivation triggers a strong accumulation of p27 protein (10 , 11) . This effect appears to be c-Myc independent, because c-myc +/+, ±, and -/- cell lines all exhibited a similar increase in p27 when grown in methyl cellulose. However, whereas introduction of oncogenic Raf or Ras significantly reduced p27 levels in attachment-deprived c-myc +/+ cells, this process occurred much less efficiently in c-myc ± and -/- cells. Several groups have recently demonstrated the importance of the Ras-Raf pathway in the down-regulation of p27 in growth factor-dependent cell cycle progression (53 , 54) , as well as in anchorage-deprived cells (44) . The regulation of p27 was reported to be posttranscriptional, and both inhibition of translation as well as induction of degradation were implicated. Overexpression of c-Myc has also been reported to reduce p27 levels via posttranscriptional mechanisms (55, 56, 57) , and c-Myc and Ras were observed to cooperate in reducing p27 levels in rat fibroblasts (58) .
The strong cooperation between c-Myc and Ha-Ras(G12V) in eliciting anchorage-independent growth was closely correlated with p27 expression; a mere 2-fold reduction in c-Myc expression (c-myc +/+ versus c-myc ± cells) resulted in a significantly amplified defect in p27 down-regulation (5–10-fold). The cooperation of c-Myc and Ras argues against these two genes acting in a linear pathway. Furthermore, the observation that high levels of Ras were capable of down-regulating p27 and causing anchorage-independent growth in the complete absence of c-Myc (c-myc -/- cells) indicates that c-Myc is not acting downstream of Ras. Although not explicitly addressed in the experiments reported here, there is no indication in the literature that c-Myc can act upstream of Ras. Taken together, the most plausible hypothesis is that oncogenic Ras and c-Myc act by independent pathways to reduce p27 levels, with the end result depending on the cumulative strength of the signal. What could be the mechanism of such cooperation?
Although the complete spectrum of Ras-signaling activities is complex and incompletely understood, one well documented effect is the transcriptional activation of cyclin D1 expression (21 , 22) . Ha-Ras(G12V) was also observed to significantly increase the levels of cyclin D1 in the studies reported here.8 c-Myc has been reported to regulate positively the expression of Cdk4 and Cdk-activating kinase and negatively the expression of p27 (38 , 59) . Thus, in c-myc -/- cells, the levels of Cdk4 and Cdk-activating kinase are down and the level of p27 is up, and although the changes in expression are modest (2–3-fold), they result in a strong (greater than 10-fold) reduction of Cdk4/6 activity.
An attractive and simple hypothesis is that Ras and c-Myc signals converge to regulate the activity of cyclin D-Cdk4/6 complexes, the former by up-regulating the expression of cyclin D1 and the latter by affecting the activity of the complexes. The cooperation of Ras and c-Myc to activate Cdk4/6 activity would then set into motion mechanisms that result in the degradation of p27 and full activation of cyclin E-Cdk2 complexes, leading eventually to the activation of cyclin A-Cdk2 complexes, which appear to be required for anchorage-independent growth. The mechanisms by which Ha-Ras(G12V) and c-Myc up-regulate cyclin D1 expression and Cdk4 activity are incompletely understood, and this model does not rule out additional downstream functions for either Ras or c-Myc. The exquisite fine tuning of these regulatory mechanisms is underscored by the existence of haplo-insufficient phenotypes demonstrated in several previous studies: c-myc ± cells had a clear delay in the phosphorylation of Rb and activation of cyclins E and A (32) , and p27 ± mice were significantly more susceptible to tumors in multiple tissues (60) .
Myc and oncogenic Ras cooperate effectively to transform primary rodent fibroblasts but cannot transform normal human fibroblasts. Although immortalization of all cells requires a telomere maintenance mechanism, human cells activate telomerase very infrequently, whereas rodent cells do so quite readily (61) . In human cells, overexpression of c-Myc has been reported to activate the expression of the catalytic protein subunit of telomerase (62) ; to what extent this is relevant for immortalization during tumor development is not known. There is no evidence that c-Myc facilitates the expression of telomerase in rodent cells or that telomerase participates in transformation (63) , but it has been shown that deregulated expression of c-Myc establishes conditions favorable for the selection of immortalizing lesions in the p53 pathway (64) . Rat-1 cells used in this study do not express detectable telomerase activity9 and presumably maintain telomeres by a telomerase-independent mechanism; this has precluded the investigation of whether loss of c-Myc function affects telomerase expression in this system.
Expression of oncogenic Ras triggers a growth arrest associated with the up-regulation of p53 and p16INK4a in both rodent and human primary fibroblasts (65) . Whereas interference with either the p16-Rb or the p53 pathway is sufficient to abrogate arrest in rodent cells, both pathways need to be ablated in human cells. Rat-1 cells used in this study are not inhibited by oncogenic Ras, presumably because they have lost functional expression of the cyclin-dependent kinase inhibitor p21 (66) . Until recently, the lack of immortalized but nontransformed human cell lines equivalent to the NIH 3T3 rodent model (or the Rat-1 cell line used here) has seriously hampered the study of transformation-specific functions in human cells. Although the transformation of primary human cells by defined genetic steps has been recently achieved (67) , the specific mechanisms responsible for the bypass of senescence, the activation of telomerase, and the acquirement of transformed phenotypes such as anchorage-independent growth remain poorly delineated. However, it appears reasonable that once the growth inhibitory effects of oncogenic Ras have been bypassed by immortalization processes, Ras additionally needs to down-regulate p27 expression to facilitate anchorage-independent growth. Anchorage deprivation increases the levels of p27 in both rodent and human fibroblasts (11 , 12) . Activated Ras or Raf can down-regulate p27 in a number of rodent cell lines (68) . In one report (69) , activated Raf did not down-regulate p27 in human IMR-90 fibroblasts. Whether the Myc-Ras cooperation mechanism described in this study is operational in human cells remains to be investigated. It is important to note, however, that p27 expression has been found to be reduced but not eliminated in many human tumors (70 , 71) and that decreased levels of p27 correlate with poor prognosis and increased grade of many human cancers.
The observation that a 50% reduction of c-Myc expression results in a 20% decrease in proliferation but a greater than 10-fold reduction in susceptibility to oncogenic Ras transformation suggests the existence of two discrete thresholds for c-Myc activity, one involved in malignant transformation and the other in cell cycle progression. From a therapeutic standpoint, it is encouraging that, although transformation appears to be very sensitive to c-Myc expression levels, much larger reductions can be tolerated without causing significant cell cycle defects. Indeed, there have been several reports (72, 73, 74) of successful reversion of cancerous phenotypes using c-Myc antisense oligonucleotides in a variety of cell culture models, as well as in whole animal studies. These studies have mostly focused on tumor types in which c-Myc overexpression is believed to be causal to the development of malignancy. The data presented in this study suggest that targeting endogenous c-Myc expression in tumors where c-Myc is not overexpressed, but where the endogenous protein is likely cooperating with other activated oncogenes, may be an efficacious and hitherto overlooked strategy. For example, approximately 30% of human tumors contain activated Ras, but in most ras gene expression is not deregulated, and the mutant protein is found at a normal level of expression. This study indicates that in these cases transformation elicited by activated Ras may be highly sensitive to endogenous c-Myc expression levels. Furthermore, the normal low level of c-Myc expression and/or activity may be a better therapeutic target than those situations where c-Myc expression has been deregulated.
|
ACKNOWLEDGMENTS |
|---|
106–143) retrovirus, and
Drs. Alvaro J. Obaya and Kam C. Yeung for helpful discussions. |
FOOTNOTES |
|---|
1 Supported by Grant GM-R01-GM41690 to
J. M. S. 
2 Present address: Department of Molecular
Pharmacology, Stanford University, Stanford, California 94305.
3 Present address: Second Department of Medicine,
Tokyo Medical and Dental University, Tokyo 113, Japan.
4 Present address: Cor Therapeutics Incorporated,
South San Francisco, California 94080.
5 To whom requests for reprints should be
addressed, at Department of Molecular Biology, Brown University, Box
G-J223, J. W. Wilson Lab., Rm. 223, 69 Brown Street, Providence, RI
02912. Phone: (401) 863-7631; Fax: (401) 863-9653; E-mail: johnsedivy{at}brown.edu
6 The abbreviations used are: Cdk,
cyclin-dependent kinase; E2F, adenovirus early promoter 2 activating
factor; Erk, extracellular signal-regulated kinase; NRK, normal rat
kidney cells; Rb, retinoblastoma.
7 S. Li and J. M. Sedivy, unpublished
observations.
8 A. V. Bazarov and J. M. Sedivy, unpublished
observations.
9 S. Wei and J. M. Sedivy, unpublished
observations.
Received 7/19/00. Accepted 11/30/00.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Ota, Z.-Q. Zhou, J. M. Link, and P. J. Hurlin The role of senescence and prosurvival signaling in controlling the oncogenic activity of FGFR2 mutants associated with cancer and birth defects Hum. Mol. Genet., July 15, 2009; 18(14): 2609 - 2621. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. L. Iversen, V. Arora, A. Acker, D. H. Mason, and G. R. Devi Efficacy of Antisense Morpholino Oligomer Targeted to c-myc in Prostate Cancer Xenograft Murine Model and a Phase I Safety Study in Humans Clin. Cancer Res., July 1, 2003; 9(7): 2510 - 2519. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Wei, W. A. Jobling, W. Chen, W. C. Hahn, and J. M. Sedivy Abolition of Cyclin-Dependent Kinase Inhibitor p16Ink4a and p21Cip1/Waf1 Functions Permits Ras-Induced Anchorage-Independent Growth in Telomerase-Immortalized Human Fibroblasts Mol. Cell. Biol., April 15, 2003; 23(8): 2859 - 2870. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Petrenko, G. Fingerle-Rowson, T. Peng, R. A. Mitchell, and C. N. Metz Macrophage Migration Inhibitory Factor Deficiency Is Associated with Altered Cell Growth and Reduced Susceptibility to Ras-mediated Transformation J. Biol. Chem., March 21, 2003; 278(13): 11078 - 11085. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. M. Kenney, M. D. Cole, and D. H. Rowitch Nmyc upregulation by sonic hedgehog signaling promotes proliferation in developing cerebellar granule neuron precursors Development, January 1, 2003; 130(1): 15 - 28. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. A. Stewart, S. D. Thomas, C. A. Mayfield, and D. M. Miller Psoralen-modified clamp-forming antisense oligonucleotides reduce cellular c-Myc protein expression and B16-F0 proliferation Nucleic Acids Res., October 1, 2001; 29(19): 4052 - 4061. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Bowman, M. A. Broome, D. Sinibaldi, W. Wharton, W. J. Pledger, J. M. Sedivy, R. Irby, T. Yeatman, S. A. Courtneidge, and R. Jove Stat3-mediated Myc expression is required for Src transformation and PDGF-induced mitogenesis PNAS, June 7, 2001; (2001) 131568898. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Bowman, M. A. Broome, D. Sinibaldi, W. Wharton, W. J. Pledger, J. M. Sedivy, R. Irby, T. Yeatman, S. A. Courtneidge, and R. Jove Stat3-mediated Myc expression is required for Src transformation and PDGF-induced mitogenesis PNAS, June 19, 2001; 98(13): 7319 - 7324. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
