Oncogene (1998) 16, 1701 ± 1712
 1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00
http://www.stockton-press.co.uk/onc
Alternatively spliced forms of cyclin D1 modulate entry into the cell cycle in
an inverse manner
Hiroki Sawa1, Teruyo Arato Ohshima2, Hiroyuki Ukita1,3, Hiromi Murakami3, Yukie Chiba3,
Hajime Kamada3, Mitsuhiro Hara1 and Isamu Saito1
Department of 1Neurosurgery and the 2Second Department of Biochemistry, Kyorin University School of Medicine, Mitaka, Tokyo
and 3Hokuto Hospital, Obihiro, Hokkaido, Japan
Alternative splicing of cyclin D1 gene mRNA has
recently been demonstrated. The novel transcript shows
no splicing at the downstream exon 4 boundary and
encodes a protein with an altered carboxyl-terminal
domain that is a cyclin D1 variant; exon 5 is not
included in the coding sequence which terminates downstream of exon 4. We here produced cells that
exogenously express each form of cyclin D1 and
analysed their cell cycle regulation. We found that (1)
alternative splicing forms of cyclin D1 modulated entry
into the cell cycle in an inverse manner; (2) both splicing
forms suppressed cell growth; and (3) cells overexpressing form [a] were inhibited from entry into and
completion of the S phase, although form [b]-expressing
cells showed no reduction of G1- to S transition. We also
found that overexpression of either cyclin D1 form upregulated Rb gene products, suggesting that this upregulation may be one of the causes of growth
suppression in cyclin D1 overexpressing cells.
Keywords: cyclin D1; cell cycle regulation; transfection;
FACS; Ki-67 antigen; alternative splicing
Introduction
There is increasing evidence that several types of
human tumors display abnormalities in the cyclin D1
gene (Arnold et al., 1989; Rosenberg et al., 1991a,b;
Jiang et al., 1992; Buckley et al., 1994; Motokura and
Arnold, 1993; Bartkova et al., 1994; Gillet et al., 1994;
Hunter and Pines, 1994; Michalides et al., 1995). This
gene, also termed prad 1 or bcl-1, is located at
chromosome 11q13. Chromosomal re-arrangements at
this locus in parathyroid tumors (Arnold et al., 1989;
Rosenberg et al., 1991a) or certain B cell lymphomas
(Rosenberg et al., 1991b) lead to increased expression
of this gene. Cyclin D1 gene product has been shown
to be overexpressed in a signi®cant fraction of primary
human breast carcinomas (Buckley et al., 1993; Gillet
et al., 1994), esophageal carcinomas (Jiang et al., 1992),
and several other types of tumors, both in the cell lines
and in the primary tumors (Motokura and Arnold,
1993; Hunter and Pines, 1994; Michalides et al., 1995;
Bartkova et al., 1995). It is well known that cyclin D1
regulates the G1- to S phase transition of the cell cycle
by activating cyclin-dependent kinases (CDKs) 4 and 6,
which phosphorylate the retinoblastoma protein (pRb)
Correspondence: H Sawa
Received 20 August 1997; revised 30 October 1997; accepted 31
October 1997
during G1 (Ewen et al., 1993), leading to the activation
of the transcription factor E2F (Cellappan et al., 1991;
Nevins, 1992; Lam and Watson, 1993; Weinberg,
1995). A recent study indicated that stable overexpression of cyclin D1 in rodent ®broblasts (Rat 6
embryo ®broblast) decreases the duration of the
G1 phase and enhances the ®broblasts' tumorigenicity
in nude mice (Jiang et al., 1993).
In addition, overexpression of cyclin D1 in mouse
NIH3T3 cells (Asano et al., 1995) and rat liver
epithelial cells (Zhou et al., 1996) creates the
conditions under which ampli®cation of CAD gene
message that can be resistant against N-phosphonoacetyl-L-aspartate (PALA), occurs more readily. Furthermore, overexpression of cyclin D1 in transgenic mice
has been shown to result in the formation of
mammary tumors after a latency of longer than 1
year (Wang et al., 1994). By contrast, serveral
overexpression studies have indicated that stable
overexpression of cyclin D1 prolongs the S phase
and inhibits growth in a human mammary epithelial
cell line (Han et al., 1995) and that it also inhibits
S phase entry in normal diploid ®broblasts (Atadja et
al., 1995). These con¯icting results on the celltransforming ability of the cyclin D1 gene might
result from the di€erent cell types and expression
systems used in each study, and/or from di€erences in
the extent and timing of the overexpression of cyclin
D1 protein.
Recent studies have demonstrated that cyclin D1
may be involved not only in cell-cycle regulation, as
described above, but also with the mechanisms
governing programmed cell death (apoptosis) (Freeman et al., 1994; Kranenburg et al., 1996) and
senescence (Dulic et al., 1993; Lucibello et al., 1993).
It is clear from these investigations that cyclin D1 is
involved in regulating cell proliferation, apoptosis, and
senescence by more than one mechanism.
Alternative splicing of cyclin D1 gene mRNA has
recently been demonstrated (Betticher et al., 1995). The
novel transcript shows no splicing at the downstream
exon 4 boundary, and encodes a protein with an
altered carboxyl-terminal domain that is a cyclin D1
variant without exon 5 and terminating downstream of
exon 4. Betticher et al. (1995) also found that there is a
polymorphism at the exon-4-intron-4 border (A or G)
and that splicing between exon 4 and exon 5 is
modulated by the frequency of the A/G polymorphism
in the splice donor region. This ®nding raised the
question of the functional di€erences between these
alternatively spliced forms. In this study, we attempted
to answer this question by exogenously expressing both
cyclin D1 forms in cultured cells and analysing their
Cell cycle regulation of cyclin D1 splice forms
H Sawa et al
e€ects on cell-cycle regulation. We found that the
alternatively spliced forms of cyclin D1 modulated
entry into the cell cycle in an inverse manner, and that
both forms suppressed cell growth. In addition, cells
overexpressing form [a] were inhibited from entering
and from completing the S phase. Finally we found
that overexpression of cyclin D1 up-regulates pRb,
which may be a cause of the growth suppression found
in cells overexpressing this gene.
Results
Expression of cyclin D1 alternatively spliced mRNAs in
human glioma cells
We studied the expression of cyclin D1 alternatively
spliced mRNAs in human glioma cells using RT ± PCR
analysis. As shown in Figure 1, each of the three
glioma cell lines studied (U251MG, T98G, U87MG)
expressed both mRNAs of cyclin D1, although the
ratio of form [a] to [b] mRNA di€ered among cell
types.
We also studied the DNA polymorphism of the
exon-4-intron-4 border of the cyclin D1 gene in each of
the three glioma cell types and in 94 Japanese
individuals using PCR ± RFLP. Betticher et al. (1995)
suggested that the allele encoding A (adenine) at
nucleotide position 870 is the major source of form
[b] mRNA, whereas mRNA from the G (guanine)
coding allele is either spliced to form [a] and [b] mRNA
to an apparently equal extent or favors form [a]
mRNA. Our ®nding that expression of form [b]
mRNA in T98G cells (genotype AA) was higher than
that of the other two cell types (genotypes AG) may
support the above hypothesis. In addition, we assessed
the population frequencies of A- (0.47) and G- (0.53)
coding alleles in 94 Japanese individuals using PCR ±
RFLP, and found that the ratio between these
frequencies was almost identical to that in European
individuals (Betticher et al., 1995).
Exogenous expression of both cyclin D1 splice forms in
transfectants
We transfected all three cell lines (U87MG, T98G,
and U251MG cells) with cDNAs of both of the cyclin
D1 forms and successfully established pooled and
subclonal cell lines overexpressing each form. Our
preliminary cell cycle analysis of these cell lines gave
similar results, so we decided to study the U87MG
lineage cells in more detail. We examined the U87MG
parent cells and mock transfected cells, as well as cells
transfected with either form of the cyclin D1 gene, by
method of immunoblotting using a monoclonal
antibody against cyclin D1 (DCS 6) and a form [a]speci®c polyclonal antibody. We con®rmed that
pooled cells (U87-a pool and U87-b pool) of
transfectant expressed exogenous cyclin D1 protein,
as shown in Figure 2a and b. We then subcloned cells
from the pooled transfected cells and established six
subclonal cell lines (a-1 to a-3 for cyclin D1 form [a]
a
b
kDa
kDa
205 —
140 —
83 —
205 –
140 –
83 –
45 —
32.6 —
– form [a]
– form [b]
c
RT–PCR analysis of human glioma cells
(T98G, U251, U87 cells)
DNA polymorphism AG
AA
AG
2
3
1
U87-a pool
U87
bp
– form [a]
– form [b]
45–
32.6 –
18 –
7.5 –
18 —
1
2 3
U87-b pool
Rb
Rb
p53
p53
Cyclin D1
—
—
—
—
—
–
–
–
–
a
b
–
1353
1078
872
603
310
–
1702
G3PDH
U251
T98G
U87
Figure 1 RT ± PCR analysis in human glioma cell lines. All three
human glioma cells (U251MG, T98G, U87MG) express both
alternatively spliced mRNAs forms of the cyclin D1 gene. DNA
polymorphism of the exon 4-intron 4 border (nucleotide
position 870) of the cyclin D1 gene is evident at the top of
Figure. The expression of form [b] mRNA in T98G cells
(genotype AA) is higher than that of either U251MG or
U87MG (genotypes AG). The results of RT ± PCR for human
G3PDH (1.000 bp) are shown at the bottom of Figure. Note that
the intensities of the bands do not di€er among the lanes
Figure 2 Exogenous expression of both cyclin D1 splice forms
and up-regulation of pRb in the transfectants. (a) Fifty mg of total
protein were run in 15% SDS ± PAGE. Immunoblotting was done
with anti-cyclin D1 mouse monoclonal antibody (DCS 6). Lane 1,
U87MG parent cell; Lane 2, U87-a pool, (the pooled transfected
cells with cDNA for form [a]); Lane 3, U87-b pool, (the pooled
transfected cells with cDNA for form [b]. (a) shows exogenous
expression of form [a] in the U87-a pool (Lane 2) and of form [b]
in the U87-b pool (Lane 3). (b) Immunoblotting was done with
anti-cyclin D1 polyclonal antibody that was speci®c for form [a].
Lane 1, U87MG parent cell; Lane 2, U87-a pool; Lane 3, U87-b
pool. The immunoblotting shows an exogenous expression of
form [a] in the U87-a pool. (c) Expression of Rb gene products
(pRb), p53 in U87MG parent cell, and U87-a and -b pools.
Fifty mg of total protein were run in 15% SDS ± PAGE. The
results show up-regulation of pRb in the pooled transfected cells
(U87-a and -b pools). p53 protein of U87MG cells was easily
degraded, but that of either transfectant became stable
Cell cycle regulation of cyclin D1 splice forms
H Sawa et al
transfectants and b-1 to b-3 for form [b] transfectants). The results of the immunoblotting (Figure 3a)
demonstrated that a-1 cells strongly expressed cyclin
D1 form [a] protein, but strong expression of cyclin
D1 was not detected in a-2, a-3, b-1, b-2 or b-3
subclonal cell lines. We thought that cells strongly
a
—
—
—
—
—
—
Cyclin D1
(DCS 6)
—
U87
Mock a-1 a-2 a-3 b-1 b-2 b-3
— form [a]
— form [b]
Cyclin D1
(polyclonal)
form [a]
—
— form [b]
Rb
p53
p21
p27
Immunoblotting for related oncogene and suppressor
gene products in cells expressing cyclin D1 [a] and [b]
forms
CDK4
PCNA
b
U87
Mock a-1 b-2
U87
Mock a-1 b-2
— form [a]
— form [b]
— form [a]
— form [b]
Cyclin D1
(DCS 6)
Rb
Rb
p53
p53
p21
p21
Starvation
expressing cyclin D1 as detected by immunoblotting
could not be cloned, except for the a-1 cell lines. Next
we assessed the subclonal cell lines immunohistochemically using the mouse monoclonal antibody
against cyclin D1 (DCS 6), because this antibody
reacts with both forms of cyclin D1.
Immunocytological analysis (Figure 4A ± G) showed
that the intensities of immunostaining of the transfectants were stronger than that of the mock transfected
cell. The intensities of staining were a-14a-24a3=mock transfected cells for form [a] transfectants
and b-24b-1=b-34mock transfected cells for form [b]
transfectants. As immunoblotting results using the
cyclin D1 form [a]-speci®c polyclonal antibody were
the same for the mock transfected cells and form [b]
transfectants, we concluded that the immunostaining of
b-1, b-2 and b-3 cells may have been due to exogenous
expression of form [b]. In addition, the immunocytological staining of these cells was localized in the
nucleus and nucleolus, suggesting that both cyclin D1
forms were sorted correctly. Using these cells, we
studied the functional roles of both forms of cyclin D1
in cell cycle regulation by cell kinetics analysis,
including Ki-67 antigen expression, BrdU labeling,
and Flow cytometry.
Serum stimulation
(10% serum)
Figure 3 Immunoblotting for U87MG mock transfected cells
and subcloned cells from the pooled transfected cells (U87-a and b pool). (a) Immunoblotting for cyclin D1, pRb, p53, p21/WAF1,
p27/Kip1, CDK4, and PCNA in U87MG mock transfected cells
and subcloned cells from U87-a and -b pools. Fifty mg of total
protein were run in 15% and 10% (for pRb and p53) SDS ±
PAGE and immunoblotted with the indicated antibodies. The
results with anti-cyclin D1 antibodies (DCS 6 and form [a]-speci®c
polyclonal antibody) show that a-1 cells strongly express cyclin
D1 form [a], but that a-2, a-3, b-1, b-2 and b-3 subcloned cells do
not strongly express this form. Up-regulation of pRb was seen in
a-1 and b-1 cells. In particular, a-1 cells that showed the strongest
expression of cyclin D1 form [a] also showed the strongest
expression of pRb. Increased levels of p21 were also observed in
a-1 cells. The expression levels of p53, p27, CDK4, and PCNA
did not di€er in U87MG mock transfected cells and subcloned
cells from the pooled transfected cells (U87-a and -b pools). (b)
pRb, p53 and p21/WAF1 expression of U87MG mock transfected
cells and a-1 and b-2 cells under starvation and 10% serum
conditions. The hyperphosphorylated pRb (arrow) can be
observed in a-1 cells, even under starvation conditions. Twentyfour hours after serum stimulation, increased levels of p21 were
seen in mock transfected cells and a-1 cells, but not in b-2 cells.
The expression of cyclin D1 and p53 did not di€er under the two
conditions
Immunoblotting showed that overexpression of either
form of cyclin D1 in pooled transfectant cells was
associated with up-regulation of pRb (Figure 2c). p53
protein of U87MG cell was easily degraded, but that of
either transfectant was stable (Figure 2c). The
expression of p21/WAF1, p27/Kip1 and CDK4 did
not di€er between U87MG cells and transfectants
(data not shown). To examine whether up-regulation of
pRb and stabilization of p53 protein occurred by
transfection of a PCI neo vector or by overexpression
of either cyclin D1 form, mock transfected and
subclonal cells were studied by immumoblotting.
Stabilization of p53 protein was observed even in
mock transfected cells, suggesting that this was due to
transfection of the PCI neo vector. Up-regulation of
pRb was observed in the subclonal cells a-1 and b-1,
but not in mock transfected, a-2, a-3, b-2 or b-3 cells
(Figure 3a). In particular, a-1 cells that showed the
strongest expression of cyclin D1 form [a], also showed
the strongest expression of pRb, suggesting that
overexpression of cyclin D1 may induce up-regulation
of pRb.
Immunoblotting for p53, p27, CDK4 and proliferating cell nuclear antigen (PCNA) in mock transfected
cells and subclonal cells transfected with form [a] or
form [b] did not show distinct di€erences (Figure 3a).
Up-regulation of p21 was detected in a-1 cells that
overexpressed form [a] (Figure 3a). To investigate the
e€ects of starvation and serum stimulation, we
compared the expression of cyclin D1, pRb, p53, and
p21 under these two conditions (Figure 3b). The
expression of cyclin D1 and p53 did not di€er among
mock transfected a-1 and b-2 cells, but hyperphosphorylated pRb was observed in a-1 cells under
starvation conditions. After serum stimulation, all cell
types expressed hyperphosphorylated pRb (Figure 3b,
1703
Cell cycle regulation of cyclin D1 splice forms
H Sawa et al
1704
arrow). Under starvation conditions, p21 protein
expression was not observed in mock transfected, a-1
or b-2 cells. Twenty-four hours after serum stimulation, increased levels of p21 were seen in mock
transfected cells and a-1 cells, but not in b-2 cells.
Growth suppression of cells expressing either form of
cyclin D1
To investigate how the expression of each form of
cyclin D1 in¯uences cell proliferation in di€erent
serum concentrations, U87MG cells and pooled cells
overexpressing form [a] or form [b] were cultured in
the presence of 10%, 3%, 0.5% or 0.1% fetal bovine
serum (FBS). Figure 5a shows that the growth of
both U87-a and -b pooled cells was suppressed in all
serum concentrations, when compared with that of
U87MG parent cells. The growth suppression of form
[a] transfectant subclonal cells was clearly related to
the level of exogenous cyclin D1 form [a] expression
as shown in Figure 5b. Subclonal cells transfected
with form [b] (b-1, b-2 and b-3 cells) also showed
growth suppression (Figure 5b). These results indicate
that the increased levels of exogenous cyclin D1 in
the transfectants reduced the cells' capacity for
growth.
Functional e€ects of alternatively spliced cyclin D1
forms on the G0 to G1 transition
To study the e€ect of the constitutive expression of
each cyclin D1 form on the G0 to G1 transition, parent
U87MG cells and transfected cells were stained with
MIB-1 antibody (Key et al., 1993) against Ki-67
antigen as described in the Materials and methods.
a
b
Figure 4 Immunocytological staining of U87MG mock transfected cells and subcloned cells from either transfected pooled
cells, using anti-cyclin D1 mouse monoclonal antibody (DCS 6).
(A) U87MG mock transfected cell; (B) a-1; (C) a-2; (D) a-3; (E) b1; (F) b-2; (G) b-3. The relative intensities of immunostaining of
U87MG mock transfected cells and subcloned cells were as
follows: a-14a-24a-3=mock transfected cells for form [a]
transfectants, and b-24b-1=b-34mock transfected cells for
form [b] transfectants. Since the immunoblotting results using
form [a]-speci®c polyclonal antibody were the same for the mock
transfected cells and subcloned cells (b-1, b-2 and b-3) from the
pooled transfected cells with form [b] cDNA, the immunostaining
of b-1, b-2 and b-3 may have been due to exogenous expression of
form [b]. Note that the positive staining localizes in the nuclei and
nucleoli of transfected cells, suggesting that exogenously expressed
cyclin D1 proteins were sorted correctly. Bar=25 mm
Figure 5 (a) Growth suppression of cells expressing either form
of cyclin D1 in di€erent serum concentrations. U87MG parent
cells and pooled cells overexpressing form [a] or [b] (U87-a pool
and U87-b pool) were cultured in DMEM containing 10%, 3%,
0.5% or 0.1% FBS. The growth of U87-a and -b pools was
suppressed under all serum conditions, when compared with that
of U87MG parent cells. Cells were grown in duplicate wells and
the cell numbers of each well were counted three times with a
Coulter cell counter. Each point represents the average of
duplicate wells. (b) Growth suppression in the subcloned cells
from pooled cells overexpressing form [a] or form [b] of cyclin
D1 under condition of 10% serum. All three subcloned cells (a1, a-2 and a-3) from U87-a pool (left) and three subcloned cells
(b-1, b-2 and b-3) from U 87-b pool (right) showed growth
suppression. The growth suppression of form [a] transfectant
subclonal cells is related to the level of exogenous cyclin D1
form [a] expression. Cells were grown in duplicate wells and the
cell numbers of each well were counted three times with a
Coulter cell counter. Each point represents the average of
duplicate wells
Cell cycle regulation of cyclin D1 splice forms
H Sawa et al
As shown in Figure 6a, the proportion of Ki-67
antigen-negative cells (G0 cells) was reduced in the
U87-a pool (P50.05), whereas the proportion of G0
cells increased in the U87-b pool (P50.05). Subcloned
cells (a-1 to a-3 and b-1 to b-3) and U87MG mock
transfected cells were also examined. The G0 cell ratio
increased in the subclonal form [b]-expressing cells
(b-1, b-2 and b-3) (P50.001), but there was no
statistically signi®cant di€erence between the G0 cell
ratios of U87MG mock transfected and form
[a]-expressing cells (a-1, a-2 and a-3) (Figure 6b). To
further investigate the functional roles of the cyclin D1
forms in the G0 to G1 transition, cells were serum
starved and stained with MIB-1 antibody to detect Ki67 antigen at the indicated time. Figure 6c shows that,
during starvation, synchronization to G0 was observed
time-dependently in U87MG parent cells, whereas the
U87-a pool kept a lower ratio of G0 cells and the U87-
a
b pool had a relatively high ratio of G0 cells. Pooled
cells overexpressing either form of cyclin D1 may be
dicult to synchronize to G0 by serum starvation. In
particular, a-1 cells, among the cyclin D1 form [a]expressing cells, continued to have a low number of G0
cells, whereas a-3 cells, which express low levels of
cyclin D1 form [a] and U87MG mock transfected cells,
showed time-dependent increases in the G0 ratio by
serum starvation (Figure 6d). This tendency was closely
related to the extent of form [a] expression. In cyclin
D1 form [b]-expressing subclonal cells, the G0 cell
ratios in all three subclonal cell types were higher than
that of U87MG mock transfected cells in medium
containing 10% FBS (P50.001). During starvation,
the G0 cell ratios of b-1 and b-2 cells increased time
dependently, but b-3 cells could not be synchronized to
the G0 phase (Figure 6e). These results suggest that
cyclin D1 form [a] may modulate the entrance of G0
b
f
c
d
e
Figure 6 (a) Ki-67 antigen-negative cell ratio (%) (G0 cells) in U87MG (U87) and cells expressing either form of cyclin D1 (U87-a
and -b pools). The proportion of Ki-67 antigen-negative cells (G0 cells) was reduced in the U87-a pool, whereas the proportion of
G0 cell increased in the U87-b pool, when compared with that of U87MG parent cells. Results represent the mean+s.d. of three
di€erent wells. Values marked by asterisk(s) di€er signi®cantly from values in U87MG parent cells: *P50.05, **P50.01,
***P50.001. (b) Ki-67 antigen-negative cell ratio (%) (G0 cells) in U87MG mock transfected cells (U87M) and subcloned cells
from U87-a and -b pools. The G0 cell ratio increased signi®cantly in subclonal form [b]-expressing cells (b-1, b-2 and b-3), but there
was no signi®cant di€erence between the G0 cell ratios of U87MG mock transfected cells and subcloned cells (a-1, a-2 and a-3)
from the U87-a pool. Results represent the mean+s.d. of three di€erent wells. Values marked by asterisk(s) di€er signi®cantly from
values in U87MG mock transfected cells: *P50.05, **P50.01, ***P50.001. (c, d, e) Ki-67 antigen-negative cell ratio (%) under
starvation conditions. (c) Synchronization to G0 can be time-dependently observed in U87MG (U87) cells, whereas the U87-a pool
maintains a lower ratio of G0 clels during starvation. The U87-b pool has a relatively high ratio of G0 cells. (d) Synchronization to
G0 can be time-dependently observed in U87MG mock transfected cells (U87 mock) and a-3 cells that express low levels of cyclin
D1 form [a], but a-1 and a-2 cells that express high levels of cyclin D1 form [a] continue to have a low number of G0 cells. (e)
Seventy-two hours after starvation, U87MG mock transfected cells (U87 mock) and b-1 and b-2 subcloned cells are synchronized to
G0, but b-3 cells are not. The results represent the mean+s.d. of three di€erent wells. Values marked by asterisk(s) di€er
signi®cantly from values in U87MG or mock transfected cells: *P50.05, **P50.01, ***P50.001. (f) Inhibitory e€ect of form [b]
against form [a] in the G0-G1 transition. The a-1 cells that expressed high levels of form [a] were double-transfected with pZeoSV2
(+), or with pZeoSV2 (+) that was inserted form [b] cDNA (pBF). These pooled transfectants were cultured with 250 mg/ml Zeocin
for 2 weeks and then starved for 72 h. The cells were stained with MIB-1 antibody. The transfection with pBF vector constructs
signi®cantly increased the G0 cell ratio induced by serum starvation (P50.001). a-1pZeo, a-1 cells transfected with pZeoSV2 (+). a1pBF, a-1 cells transfected with pBF vector construct. Results represent the mean+s.d. of three di€erent wells
1705
Cell cycle regulation of cyclin D1 splice forms
H Sawa et al
1706
cells into G1 and form [b] may play an important role
in keeping culture cells in the G0 phase. To further
analyse the roles of the two forms of cyclin D1 in the
G1-G0 transition, a-1 cells were double-transfected
with pZeoSV 2 (+) containing cDNAs of form [b] as
described in the Materials and methods. The doubletransfected cells were cultured with 250 mg/ml Zeocin
for 2 weeks and then were starved for 72 h and stained
with MIB-1 antibody. Figure 6f shows that transfection with vectors of form [b] signi®cantly increased the
G0 cell ratio induced by serum starvation, suggesting
that form [b] competes in function with form [a] in the
G0- to G1 transition (P50.001).
Further FACS analysis of subclonal cells of cyclin
D1 form [a] and form [b] transfectants was performed.
The S phase cell ratio of the a-1 cells which strongly
expressed cyclin D1 form [a], was higher than that of
U87MG mock transfected cells until 10 h after serum
stimulation, but was lower than those of the other cell
types (U87MG mock transfected cells and a-3, b-1 and
b-2 cells) 24 h after stimulation. The S phase cell ratio
of the a-1 cells at 32 and 48 h was still higher than that
a
b
BrdU incorporation in cells expressing form [a] or form
[b] cyclin D1 in 10% serum and starvation conditions
BrdU incorporation was assessed as described in the
Materials and methods to determine whether overexpression of either form of cyclin D1 modulates the
transition from the G1 to S phase. Surprisingly, in
10% serum, U87-a pool and a-1 cells had lower 24 h
BrdU labeling indices (P50.01 and P50.001, respectively) than U87MG cells or U87-b pool (Figure 7a,c).
These results suggest that serum stimulation may
reduce the G1 to S transition in form [a]-overexpressing cells, but not in form [b]-expressing or U87MG
cells. On the other hand, under serum starvation
conditions, the proportion of the cells entering S phase
in either 24 or 44 h increased in cells overexpressing
either form of cyclin D1 (U87-a and U87-b pools),
when compared with U87MG cells (Figure 7b). In cells
subcloned from the pooled cells, the same results were
obtained (Figure 7c,d). We also studied the ratio of
pulse labeled (30 min) BrdU incorporation after serum
stimulation at the indicated time. The proportion of
BrdU-positive U87MG cells and U87-b pooled cells
increased 12 h after serum stimulation and then
quickly fell 48 h after stimulation. By contrast, U87-a
pooled cells showed a slow increase in the proportion
of BrdU labeled cells, which accumulated as shown in
Figure 8. These BrdU incorporation patterns suggest
that entrance to the S phase and completion of DNA
synthesis may be delayed in the U87-a pooled cells.
These results were con®rmed by Flow cytometry.
Fluorescence activated cell sorting (FACS) analysis
after serum stimulation
U87MG cells and transfectants were starved for 72 h
and then growth was stimulated by adding serum to
the medium as described in the Materials and methods.
The cell-cycle distribution was assessed from the DNA
contents by propidium iodide (PI) staining. FACS
analyses of U87MG cells and U87-a and -b pooled
cells con®rmed the results of BrdU pulse labeling after
serum stimulation (Figures 8 and 9a). Seventy-two
hours after serum starvation, the ratio of S phase cells
in the U87-a pool was higher than those of U87MG
and U87-b pooled cells, but 12 h after serum
stimulation, the G1 to S phase transition was
inhibited. Forty-eight hours after serum stimulation,
the ratio of S phase cells did not fall in the U87-a pool
as was observed for the other two cell types, suggesting
that the S phase was elongated in the U87-a pool
(Figure 9).
c
d
Figure 7 BrdU incorporation in cells expressing form [a] or form
[b] under 10% serum and starvation conditions. (a) Forty-eight
hours after plating U87MG cells (U87) and pooled cells
overexpressing either form of cyclin D1 (U87-a and -b pools),
the cells were labeled with 10 mM BrdU for 24 h. The BrdU
labeled cells were stained as described in the Materials and
methods. The results represent the mean+s.d. of three di€erent
wells. Note that the percentage of BrdU positive cells of U87-a
pool is signi®cantly lower than that of U87MG parent cells
(**P50.01). (b) U87MG cells (U87) and pooled cells overexpressing either form of cyclin D1 (U87-a and -b pool) were
cultured under starvation conditions for 24 h and then labeled
with 10 mM BrdU for 24 or 44 h. BrdU positive cell ratios (%) of
U87-a and -b pools were signi®cantly higher than that of U87MG
parent cells. Results represent the mean+s.d. of three di€erent
wells. Values marked by asterisk(s) di€er signi®cantly from values
in U87MG parent cells: **P50.01, ***P50.001. (c) U87MG
mock transfected cells (U87M) and subcloned cells from pooled
transfected cells were cultured under starvation conditions for
48 h and then culture media were replaced to medium containing
10% FBS and then the cells were labeled with 10 mM BrdU for
24 h. Results represent the mean+s.d. of three di€erent wells.
Note that a-1 cells which express high levels of form [a] have
lower 24 h BrdU labeling indices than U87 mock transfected cells:
***P50.001. (d) U87MG mock transfected cells (U87M) and
subcloned cells from pooled transfected cells with either form
were cultured under starvation conditions for 24 h and then
labeled with 10 mM BrdU for 24 h. BrdU positive cell ratios (%)
of subcloned cells were signi®cantly higher than that of U87MG
mock transfected cells. Results represent the mean+s.d. of three
di€erent wells. Values marked by asterisk(s) di€er signi®cantly
from values in U87MG mock transfected cells: *P50.05,
**P50.01, ***P50.001
Cell cycle regulation of cyclin D1 splice forms
H Sawa et al
1707
a
Figure 8 BrdU labeling index after serum stimulation. After
serum starvation of 48 h, U87MG and cells overexpressing either
form of cyclin D1 (U87-a and -b pools) were cultured in medium
containing 10% FBS. The cells were labeled with 10 mM BrdU for
30 min at the indicated time. The proportion of BrdU-positive
U87MG cells and U87-b pool increased 12 h after serum
stimulation and quickly fell 48 h after stimulation. By contrast,
the U87-a pool showed a slow increase in the proportion of BrdU
labeled cells. Results represent the mean+s.d. of three di€erent
wells. Values marked by asterisk(s) di€er signi®cantly from values
in U87MG parent cells: *P50.05, **P50.01
of U87MG mock transfected cells (Figure 9b). The
S phase cell ratios of b-1 and b-2 cells at 24 h were
much higher than that of U87MG mock transfected
cells and decreased at a slower rate than that of
U87MG mock transfected cells (Figure 9b). These
results suggest that entrance to the S phase and the
completion of DNA synthesis were inhibited in a-1
cells and that, in b-1 and b-2 cells, completion of the
S phase was inhibited, whereas entrance to the S phase
was increased. a-3 cells expressing low levels of cyclin
D1 form [a] showed a similar pattern to that of
U87MG mock transfected cells. The changes in the
S phase cell ratio were inversely related to changes in
the G1/G0 cell ratio in these cells (a-1 and b-1 and b-2
cells), suggesting that G2/M phase may not be
in¯uenced by expression of exogenous cyclin D1
forms [a] and [b].
Discussion
In this study we observed the following. First, that cell
proliferation was blocked by overexpression of either
splice form of cyclin D1. Second, that overexpression
of form [a] increased the proportion of cells of
proliferation pool (G1, S, G2/M), but overexpression
of form [b] reduced the proportion of these cells. Third,
that a high level of expression of cyclin D1 form [a]
inhibited cells from entering and completing the
S phase in the presence of serum, but in serumdepleted conditions, expression of either form of
cyclin D1 accelerated cell entry into the S phase. The
most important ®nding of this study is that these
alternatively spliced forms of cyclin D1 may modulate
the G0 to G1 phase transition in an inverse manner.
b
Figure 9 FACS analysis after serum stimulation. After the cells
were starved for 72 h, growth was stimulated by adding 10%
serum to the medium. The cells were harvested with a TrypsinEDTA solution at the indicated time and were stained with 50 mg/
ml of PI. The cells were analysed on an EPICS XL SYSTEM II
¯ow cytometer for DNA content. The percentage of cells in
di€erent phases of the cell cycle was determined using a
MULTICYCLE computer program. (a) FACS analysis of
U87MG and pooled cells overexpressing either form of cyclin
D1 (U87-a and -b pools) after serum stimulation. The ratio of
S phase cells in the U87-a pool was higher than that of S phase
cells in the U87MG (U87) or U87-b pool until 12 h after serum
stimulation, whereas, at 24 h, this ratio was lower in U87-a pool
than in U87MG and U87-b pool. At 48 h, this ratio was higher in
U87-a pool than in the U87MG and U87-b pool. (b) FACS
analysis of U87MG mock transfected cells (U87 mock) and
subcloned cells (a-1, a-3, b-1 and b-2) from U87-a and -b pools
after serum stimulation. The ratio of S phase cells in a-1 cells that
express high levels of cyclin D1 form [a] was higher than that of
U87MG mock transfected cells until 10 h after serum stimulation,
but the ratio of S phase cells in a-1 cells at 24 h was lower and at
32 and 48 h was higher than that of U87MG mock transfected
cells. By contrast, 24 h after stimulation, the ratio of S phase cells
in b-1 and b-2 was much higher than that of U87MG mock
transfected cells, and these two cell types show a slow reduction in
the ratio of S phase cells. The a-3 cells that express low levels of
cyclin D1 form [a] show a pattern similar to that of U87MG
mock transfected cells. The changes in the S phase cell ratio were
inversely related to changes in the G1/G0 cell ratio
Ki-67 antigen is associated with cell proliferation and
occurs throughout the cell cycle (G1, S, G2/M phases),
except in resting (G0) cells (Key et al., 1993). Cells
expressing exogenous form [a] (U87-a pool, and a-1
and a-2 subcloned cells) maintained a high proportion
of Ki-67 antigen-positive cells during serum starvation.
This result agrees in part with the study of Zwijsen et
al. (1995), who reported that overexpression of cyclin
D1 form [a] in MCF7 breast tumor cells results in
continued proliferation under low serum conditions;
these authors suggested that cyclin D1 form [a]overexpressing cells show reduced exiting from G1 to
G0. We here demonstrated that exogenous expression
of form [b] (U87-b pool, and b-1, b-2 and b-3
subcloned cells) reduced the proportion of Ki-67
antigen-positive cells even in the presence of serum.
These results support the idea that form [a] plays an
important role for cell-cycle entry, and form [b] for
cell-cycle exit.
Among the molecules that are speci®cally expressed
by G0 cells, the growth-arrest-speci®c (gas) genes
(Schneider et al., 1988; Ciccarelli et al., 1990;
Brancolini et al., 1992) and statin (Pellicciari et al.,
Cell cycle regulation of cyclin D1 splice forms
H Sawa et al
1708
1995) have been characterized. The protein products of
gas genes have been reported to have an inhibitory
e€ect on S phase entry (Schneider et al., 1988;
Ciccarelli et al., 1990; Brancolini et al., 1992). Statin,
a 57 kDa nuclear protein, has been recognized as a
unique marker of quiescent G0 cells (Pellicciari et al.,
1995), although generally the functional roles of these
proteins in the G0 phase have not been clari®ed. When
resting, quiescent G0 cells are stimulated by serum
and/or growth factors such as platelet derived growth
factor (PDGF) and epidermal growth factor (EGF);
products of immediate early gene families such as c-fos
and c-jun become activated within a short period of
time after the addition of the growth stimuli without a
requirement for protein synthesis (Bravo, 1990). These
proteins interact with genomic sites containing the
activating protein-1 (AP-1) sequence and modulate
gene transcription (Ryder et al., 1989). Moreover, it
has been indicated that AP-1 binding activity during
the G0-G1 transition is ®nely regulated and complex,
involving changes both in protein expression and
posttranslational modi®cation (Lallemand et al.,
1997). In addition, evidence for cross-talk between
G0-G1 and G1-S transitions has also been reported
(Eblen et al., 1995). It would appear that more than
one mechanism may be involved in regulating the G1
and G0 phase transitions. Our data may provide an
important clue to the molecular basis of these
mechanisms, and the reduction of cell proliferation
may prove to be an important novel target in cancer
therapy. Accordingly, further biochemical and molecular biological studies of the splice forms of cyclin D1
gene may be clinically relevant.
We also performed a cell kinetics study, using BrdU
incorporation and FACS analysis to determine why
cells expressing exogenous cyclin D1 forms [a] or [b]
showed suppressed growth. We found that incorporation of BrdU di€ered in the presence and absence of
serum for form [a]-overexpressing cells (a-1 and U87-a
pool cells). Although, under starvation conditions,
form [a]-overexpressing cells incorporated BrdU at a
higher rate than they incorporated U87MG cells and
U87MG mock transfected cells, after serum stimulation the situation was reversed, with U87MG parent
and mock transfected cells incorporating more BrdU
than form [a]-overexpressing cells. Our FACS and
BrdU labeling experiments suggest that form [a]overexpressing cells were inhibited from entering and
completing S phase. Our observations are partly
consistent with previous reports (Han et al., 1995;
Atadja et al., 1995). Atadja et al. (1995) also showed
that normal human diploid ®broblasts that endogenously or exogenously express high levels of cyclin D1
form [a] are unable to enter S phase in response to
serum stimulation. Han et al. (1995) have reported that
stable overexpression of cyclin D1 in HBL-100 cells
increases their doubling time and lengthens the
duration of the S phase, although in serum depleted
conditions, our 24 and 44 h BrdU labeling study
showed that S phase entry of form [a]-overexpressing
cells was not inhibited but rather increased when
compared with that of U87MG cells and mock
transfected cells (Figure 7). The di€erences in the
culture conditions may mean that an unknown factor
present in the serum is responsible for this discrepancy.
Cyclin D1 form [a] associates in a ternary complex
with PCNA, a 21 kDa protein (WAF-1), and cyclin
dependent kinase (Xiong et al., 1992; El-Deiry et al.,
1993, 1994; Waga et al., 1994). One possible
mechanism by which high levels of cyclin D1 form [a]
might negatively regulate G1-S transition is through
stabilization of p21 and/or by the targeting of p21 to
other cyclin/CDK protein complexes. Indeed, overexpression of p21 is known to inhibit entry to the
S phase and is growth suppressive (El-Deiry, 1994). In
our immunoblotting study, form [a]-overexpressing
cells (a-1) expressed higher levels of p21 protein.
Therefore, it is possible that the inhibition of S phase
entry of the a-1 cells may have been partly due to high
expression of p21 protein. The mechanisms serving to
regulate p21 expression in these p53-independent
circumstances are not known, but recent accumulated
evidence (Liu et al., 1996) suggests a role for the
mitogen-activated protein kinase (MAPK) signaling
pathway in mediating transcriptional activation of p21
in response to growth factor stimulation. Since cyclin
D1 promoter activity is also stimulated by MAPK
through AP-1 binding proteins (Albanese et al., 1995),
cross-regulation and activation between p21 and cyclin
D1 syntheses may be suggested. For example, it has
been shown that p21 induces cyclin D1 synthesis (Chen
et al., 1995) and that increased cyclin D1 protein
appears necessary, in concert with p21, for the
induction of a proper G1 arrest as mediated by wild
type p53 (Del Sal et al., 1996). The other possible
mechanism of growth suppression in our transfectants
is that of the CDKs that are capable of associating
with cyclin D1; a CDK4/Cyclin D1 complex is known
to phosphorylate Rb protein (Matsuoka et al., 1994;
Kato et al., 1994), but CDK2 is also known to
associate with cyclin D1 in an apparently catalytically
inactive form (Xiong et al., 1992). CDK2 is also known
to associate with cyclins A and E (Walker and Maller,
1991). Cyclin A/Cdk2 complex is probably involved in
the completion of DNA synthesis (Pagano et al., 1992)
and mitosis, and a role in entry to the S phase has been
reported (Girard et al., 1991). The cyclin E/CDK2
complex may play a role in G1-S phase transition
(Dulic et al., 1992; Ohtsubo et al., 1995). Thus cyclins
A and E might not bind to CDK2 when a high level of
cyclin D1 form [a] is associated with CDK2 in a
catalytically inactive form. As a result, entry to and
completion of the S phase might be impaired. The
other factor may be the PCNA protein; PCNA plays
an essential role in DNA replication and repair (Bravo
and MacDonald-Bravo, 1987). Thus, overexpression of
both cyclin D1 forms may be involved in inhibiting the
function of PCNA.
Our experiments showed up-regulation of pRb in
U87-a and -b pooled cells and in a-1 and b-1 subcloned
cells. The characterized target for regulation of pRb is
the transcription factor E2F. During the early
G1 phase, pRb binds to E2F and suppresses its
activity, but as the cells progress from G1 into
S phase, pRb becomes inactivated and dissociates
from E2F, allowing E2F to regulate transcription
(Cellappan et al., 1991; Nevins, 1992; Ewen et al.,
1993; Lam and Watson, 1993; Weinberg, 1995).
Inactivation of pRb at G1-S transition is achieved
through hyper-phosphorylation by the cyclin D- and Edependent kinases. More recent data (Cavanaugh et
al., 1995; White et al., 1996) demonstrated that pRb
Cell cycle regulation of cyclin D1 splice forms
H Sawa et al
can also inhibit transcription of RNA polymerase I
(pol I) and III (pol III). Pol I synthesizes large
ribosomal RNAs (rRNA) whereas pol III synthesizes
a variety of small stable RNAs, including 5S rRNA
and transfer RNA. Together pols I and III are
responsible for 70 ± 80% of all nuclear transcription.
Up-regulation of pRb followed by the exogenous
expression of cyclin D1 may be responsible in part
for the growth suppression of our transfected cells.
Cyclin D1 expression is high in early and mid G1
and decreases by about 6 ± 7-fold before onset of
replication in normal human and rat ®broblasts
(Pusch et al., 1996). Since stably transfected cells may
lack such cyclin D1 ¯uctuation throughout the cell
cycle (Pusch et al., 1996), these inhibitory e€ects of the
cyclin D1 forms on the completion of S phase may be
non-physiological.
The CDK-cyclin complexes not only have positive
roles in determining when S phase or mitosis begins,
but are important in preventing the initiation of an
untimely cell-cycle event. For example, in ®ssion yeast,
the cdc2-cdc3 complex in the G2 phase prevents more
than one round of DNA replication per cell cycle
(Hayles et al., 1994). Experiments designed to examine
phase speci®c expression of cyclin D1 are a promising
method for studying cell cycle regulation by these
proteins in the future.
During starvation conditions, BrdU incorporation in
form [b]-expressing cells was higher than that in
U87MG parent cells, but was lower than that in form
[a]-expressing cells. This suggests that form [b] can also
accelerate S phase entry under starvation conditions,
although to a lesser extent than form [a]. By contrast,
after serum stimulation, entry into the S phase was not
inhibited in form [b]-expressing cells, but was inhibited
in form [a]-expressing cells. The reason for this
di€erence is not clear at present. One possible
explanation is that form [b] might not be associated
with, or a€ected by, the factors that serum induces.
Indeed, we observed lower p21 expression in b-2 cells
after serum stimulation than in U87MG mock
transfected and a-1 cells.
The molecular di€erence between forms [a] and [b] is
33 amino acids of the C-terminal sequence (Betticher et
al., 1995). Form [a] possesses a PEST rich region
(destruction box) that is responsible for rapid turnover
(Rogers et al., 1986), but form [b] does not contain the
PEST rich sequence. This suggests that the half-life of
form [b] might be longer. It is important to analyse the
di€erent interactions between the two forms of cyclin
D1 and their associating proteins. These interactions
are currently under investigtion.
The U87MG cells that were used in this study do
not express p16 protein (INK 4/CDKN2) (Tenan et al.,
1995), which is a tumor suppressor protein that
regulates the G1-S check point by inhibiting CDK4
activity (Lukas et al., 1995; Koh et al., 1995).
Therefore, we could not study the role of p16 protein
in G0-G1 and G1-S transitions in cells with a high level
of expression of cyclin D1. Further transfection studies
of the p16 gene in a-1 cells are in progress.
In this report, we demonstrated that alternatively
spliced forms of cyclin D1 might modulate G0-G1
transition in an inverse manner. As interactions with
other gene products remain to be examined, our series
on cyclin D1-overexpressing cells should be useful for
studying the regulatory mechanisms of G0-G1 transition and the molecular mechanisms of cell growth and
proliferation.
Materials and methods
Cell culture
The human glioma cell lines U87MG, T98G, and U251MG
were cultured in Dulbecco's modi®ed essential medium
(DMEM) (Nihonseiyaku, Japan) containing 10% heat
inactivated fetal bovine serum (FBS) and were maintained
in a 378C incubator with 5% CO2, 95% air.
PCR ± RFLP analysis of a polymorphism within the cyclin D1
coding sequence in glioma cells and in a cohort of Japanese
subjects
To study a single base pair A/G polymorphism of the
exon-4-intron-4 border of cyclin D1 coding sequence in
human glioma cells and in a separate Japanese cohort,
polymerase chain reaction (PCR)-restriction fragment
length polymorphism (RFLP) analysis was performed as
reported previously (Betticher et al., 1995). Genomic DNA
was isolated from cultured glioma cells and the blood of 94
Japanese individuals according to Sambrook et al. (1989).
For PCR ampli®cation, the following primers were used:
forward primer, 5'-GTGAAGTTCATTTCCAATCCGC-3';
reverse primer, 5'-GGGACATCACCCTCACTTAC-3'.
One mg of genomic DNA was added to a PCR mixture
containing dNTP at 250 mM, 0.5 mg of each primer,
16PCR bu€er (Takara, Japan) and 2.5 U.Taq polymerase
(Takara, Japan). Thirty PCR cycles of 1 min denaturation
at 948C, 1 min annealing at 558C and 1 min elongation at
728C, were performed in an automated thermocycler. The
PCR products were digested with the restriction enzyme
ScrFI (Wako, Japan) and run on agarose gels consisting
3% Nusieve agarose (Takara, Japan) and 1% HS agarose
(Wako, Japan) containing 0.01% ethidium bromide.
RT ± PCR analysis of alternatively spliced forms of cyclin D1
gene in human glioma cell lines U87MG, T98G and U251MG
To study whether mRNAs of cyclin D1 forms are expressed
in human glioma cells, reverse transcriptase (RT) ± PCR
analysis was performed using speci®c primer sets for both
forms of cyclin D1: for form [a], reverse primer, 5'CAAGGAGAATGAAGCTTTCCCTT-3' (3' UTR beyond
exon 5); and for form [b], reverse primer, 5'-GGGACATCACCCTCACTTAC-3' (intron 4). The foward primer, 5'CCTACTTCAAATGTGTGCAGAAG-3' (exon 1), was
common to both forms. mRNA isolation from culture
cells and RT ± PCR conditions are described below. The
resultant PCR products were run on 1.5% agarose gel
containing 0.01% ethidium bromide. A RT ± PCR control
ampli®er set for human glyceraldehyde 3-phosphate
dehydrogenase (G3PDH) (Clontech, CA) was used as a
positive control for cDNA synthesis and PCR ampli®cation.
Expression vector construction and transfection
The 1.3 kb human cyclin D1 cDNA (form [a]) containing
the entire coding sequence was kindly provided by Dr Y
Xiong (Xiong et al., 1992) and was inserted at the EcoRI
site of pCI neo Mammalian Expression Vector (Promega,
WI). The cDNA of the alternatively spliced form [b] of
cyclin D1 (Betticher et al., 1995) was cloned by RT ± PCR
for the mRNA of T98G cells because these cells expressed
the highest levels of form [b] mRNA among the human
glioma cell lines examined (Figure 1). Brie¯y, total RNA
was prepared from T98G cells, using Isogen solution
1709
Cell cycle regulation of cyclin D1 splice forms
H Sawa et al
1710
according to the manufacturer's protocol (Wako, Japan).
T98G cells were resolved in Isogen solution and an equal
volume of chloroform was added. After centrifugation for
5 min at 18 000 g, the aqueous phase was precipitated with
isopropanol. Poly(A)+ RNA was puri®ed, using oligo-dTlatex (Oligo-dT 30 super, Takara, Japan) beads. Using 1 mg
of poly(A)+ RNA as template, cDNA was then produced
with reverse transcriptase (Promega, WI) and random
priming. The resultant cDNA was ampli®ed by 40 PCR
cycles of 45 s at 958C, 45 s at 558C, and 45 s at 728C, using
a GeneAmp RNA system (Perkin Elmer, CT) according to
manufacturer's protocols. The primer sets used for
ampli®cation of cyclin D1 consisted of forward primer 5'GCGAATTCATGAACACCAGCTCCTGTG-3', and reverse primer 5'-GCGAATTCCAACCTTGTCACCCTTGGGG-3'. The ampli®ed DNA fragment was subcloned by
TA cloning vector (Invitrogen, WI) and sequenced by the
chain termination method (Sanger et al., 1977). The
0.9 kbp cyclin D1 form [b] cDNA was insertd at the
EcoRI site of pCI neo vector.
cDNA of cyclin D1 forms ([a] and [b]) were also inserted
at the EcoRI site of mammalian expression vector pZeoSV2
(+) (Invitrogen, WI) containing a gene that confers
resistance to the antibiotic Zeocin (Invitrogen, WI).
Transfection of cyclin D1 expression vector was performed
using lipofectoamine (Gibco BRL, MD) as follows. Two
micrograms of cyclin D1 expression vector, 20 ml of
lipofectoamine and 200 ml of opti-MEM (Gibco BRL, MD)
were mixed and left for 30 min at room temperature. One ml
of opti-MEM was added to the mixture. The liquid was
overlaid to human glioma cells that were washed three times
with opti-MEM, and the cells were incubated at 378C for 5 h.
The transfected cells were then cultured for 2 weeks in the
DMEM containing 10% FBS containing 1 mg/ml (active) of
geneticine (Wako, Japan). The pooled transfected cells were
subcloned as follows. One hundred of the geneticine-resistant
pooled cells were plated in the 10 cm culture dish (Iwaki,
Japan) and cultured for 7 days. The colonies were marked
with a color pen under microscopy and were detached using a
small piece (363 mm) of soft paper soaked in TrypsinEDTA solution (Gibco BRL, MD). The cloned cells were
cultured in a 24-well culture dish (Nunc, IL) and expanded
for further analysis. U87MG cells were also transfected with
the pCIneo vector alone, as described above and the pooled
transfectants (U87 mock) was used for the control study.
To examine whether form [b] competes in function with
form [a] in the G0- to G1 transition, a-1 cells that expressed
high levels of form [a] were double-transfected with pZeoSV2
(+), or with pZeoSV2 (+) that was inserted form [b] cDNA
(pBF). These pooled transfectants were cultured in DMEM
containing 10% FBS and 250 mg/ml Zeocin for 2 weeks. Each
Zeocin resistant pooled transfected cell was then starved for
72 h as described below in `Starvation experiments'. The cells
were stained with MIB-1 antibody as described in
`Immunocytochemistry'.
Growth curves
One thousand cells were plated in each well of the 24 well
culture dish (Nunc, IL) in the presence of 0.1%, 1%, 3%,
or 10% FBS. Every 2 days, cells were detached with a
Trypsin-EDTA solution and counted using a Coulter cell
counter. The growth of U87MG mock transfected cells
(U87 mock) and subcloned cells (a-1, a-2, a-3, b-1, b-2 and
b-3) from the U87-a and -b pools was also studied in the
presence of 10% FBS.
Flow cytometery
One million cells were plated in 10 cm culture dish (Iwaki,
Japan) and cultured in DMEM containing 10% FBS for 3
days. The medium was then replaced with DMEM
containing 0.1% FBS and, after 48 h, again replaced with
medium containing 10% FBS. The cells were harvested at
the indicated time before and after the addition of 10%
serum. The cells were collected and washed twice with
phosphate bu€ered saline (PBS). Cell pellets were resuspended in 1 ml PBS, ®xed in 10 ml of 70% ethanol,
and stored at 48C. On the day of analysis, cells were
collected by centrifugation and the pellets were resuspended in cell densities of 16106/ml with PBS and
treated with 1 mg/ml of RNAse A (Sigma, MO) for 60 min
at 378C. The cell suspension was incubated with 50 mg/ml
of propidium iodide (PI) (Sigma, MO) in the dark at room
temperature for 30 min. The suspension was ®ltered
through a 60 mm mesh ®lter and analysed on an EPICS
XL SYSTEM II ¯ow cytometer (Coulter, FL) for DNA
content. The percentage of cells in di€erent phases of the
cell cycle was determined using a MULTICYCLE
computer program (Phoenix Flow Systems, CA).
Starvation experiments
Five thousand cells were cultured on triplicate Labtec
chamber slides (Nunc, IL) for 48 h in DMEM containing
10% FBS, after which the medium was replaced with
DMEM containing 0.1% FBS. At the indicated times, cells
were ®xed with 70% ethanol, and stained with an antibody
against Ki-67 (MIB-1) (Immunoteck, France).
The cells were also labeled with 10 mM 5'-bromodeoxyuridine (BrdU) for 24 h or 44 h under starvation conditions
and were stained with anti-BrdU mouse monoclonal antibody
(Br-3) (CALTAG, CA) as described in `Immunocytochemistry'. All experiments were performed in triplicate, and over
1000 cells were counted under light microscopy in each.
Growth stimulation experiments
Five thousand cells were plated on Labtec chamber slides
in DMEM medium containing 10% FBS and grown for
72 h. The medium was then replaced with DMEM
containing 0.1% FBS and the cells were maintained for
48 h under starvation conditions. The medium was again
replaced with DMEM containing 10% FBS. The cells were
then ®xed with 70% ethanol at the indicated time after
growth stimulation by adding serum. The cells were pulselabeled with 10 mM BrdU for 30 min at the indicated time
after growth stimulation. The cells were stained with antiBrdU antibody as described below in `Immunocytochemistry'. All experiments were performed in triplicate, and
over 1000 cells were counted under light microscopy in
each.
Immunocytochemistry
Cell cultures grown on Labtec chamber slides (Nunc, IL)
were washed with 50 mM Tris-HCl, pH 7.6, and 150 m M
NaCl (TBS), and ®xed with 70% ethanol for 10 min. The
intrinsic peroxidase activity was quenched by 0.3% H2O2 in
methanol for 20 min. Immunostaining was then performed
with the anti-human Ki-67 monoclonal antibody (MIB-1)
(diluted 1 : 500 with TBS) for 1 h at room temperature.
After washing three times with TBS, a 1 : 500 diluted
solution of biotinylated anti-mouse IgG (Dako, Denmark)
in TBS was reacted for 1 h at room temperature and
followed by avidin-biotin peroxidase complex (Dako,
Denmark) for 30 min. After washing three times as well,
the peroxidase reaction was performed with 3-3'-diaminobenzidine tetrachloride (Wako, Japan) in the presence of
0.03% H2O2. Nuclear staining was done with hematoxylin
solution. For BrdU staining, before anti-BrdU mouse
monoclonal antibody (Br-3) (diluted 1 : 3000 with TBS)
was applied to the glass slides, DNA denaturation with 4N
HCl was performed for 30 min, and 0.05% Tween 20
Cell cycle regulation of cyclin D1 splice forms
H Sawa et al
(Sigma, MO) was added in all solutions for washing and
antibody dilution. For cyclin D1 staining, anti-cyclin D1
mouse monoclonal antibody (DCS 6) (diluted 1 : 100 with
TBS) (Progen, Heidelberg, Germany) was used. IgG2a
(Gamma-2a-Chains), mouse [Negative Cotnrol] (Dako,
Denmark) was used as a control.
Immunoblotting
Protein extraction and immunoblotting analysis were
performed as follows. Exponentially growing cells were
collected with a cell scraper and washed three times with
ice-cold TBS. The cell pellets were re-suspended in lysis
bu€er (20 mM Tris-HCl, 2 mM EDTA, NP-40 and 1 mM
PMSF). This suspension was sonicated three times with a
Soni®er Cell Disruptor. Protein concentrations were
determined by the method of Lowry et al. (1951). For
immunoblotting, 50 mg of protein in the total cell lysates
were subjected to 10% or 15% SDS ± PAGE (Laemmli,
1970). The proteins were transferred to the nitrocellulose
membrane according to Towbin et al. (1979). Membranes
were blocked with 1% free fat milk in TBS for 60 min at
room temperature. The membranes were then incubated
with a 1 : 100 dilution of anti-cyclin D1 monoclonal
antibody (DCS-6) or anti-cyclin D1 form [a]-speci®c
polyclonal antibody (Upstate Biotechnology incorporated,
NY) overnight at room temperature. This polyclonal
antibody was produced by immunization with a synthetic
peptide corresponding to the 11 C-terminal amino acids
(residues 285 ± 295) of human cyclin D1 form [a]. Since the
C-terminal amino acid sequence of form [b] is di€erent
from that of form [a], it does not recognize form [b]. As the
monoclonal antibody DCS-6 was generated by polyethylene glycol-mediated fusion of NS-2 myeloma cells with
splenocytes of a Balb/c mouse immunized with the fulllength recombinant human cyclin D1 protein (Bartokova et
al., 1994), it may react both forms of cyclin D1. Anti-Rb
(Ab-5, mouse monoclonal, 1 : 50), p21/WAF-1 (Ab-1,
mouse monoclonal, 1 : 50), CDK4 (Ab-1, rabbit polyclonal, 1 : 200) and p27/Kip1 (Ab-1, rabbit polyclonal, 1 : 200)
were obtained from Calbiochem, MA, anti-p53 (DO-7,
mouse monoclonal, 1 : 100) from Dako, Denmark, and
PCNA (mouse monoclonal, 1 : 100) from Immunoteck,
France. These antibodies were also diluted with TBS and
incubated overnight. After washing with washing bu€er
(TBS containing 0.05% Tween 20), the membranes were
reacted with a 1 : 2000 dilution of alkaline phosphataselinked anti-mouse IgG or anti-rabbit IgG (Dako, Denmark) for 1 h. The membranes were then washed three
times with washing bu€er and immune detection was
performed using the mixture of nitroblue tetrazolium
(NBT) (Wako, Japan) and isopropyl-b-D-thiogalactopyranoside (BCIP) (Wako, Japan) in 100 m M Tris HCl,
pH 9.0, 150 mM NaCl, and 5 mM MgCl2.
Acknowledgements
This research was supported from a grant-in-aid for the
Science Research Promotion Fund of the Japan Private
School Promotion Foundation and by a grant-in-aid from
the Tokyo Institute of Medical Science. The authors thank
also Mr Isao Naito for photographical assistance, Ms
Hideko Sato Shimizu for technical assistance, and KN
International, Inc. and Mr DB Douglas for comments on
the manuscript and Mr Kohji Aoyama and Ms Kayoko
Sakuma of Coulter Co, Ltd for their help with the FACS
analysis.
References
Albanese C, Johnson J, Watanabe G, Eklund N, Vu D,
Arnold A and Pestell RG. (1995). J. Biol. Chem., 270,
23589 ± 23597.
Arnold A, Kim HG, Gaz RD, Eddy RL, Fukushima Y,
Byers MG, Shows TB and Kronenberg HM. (1989). J.
Clin. Invest., 83, 2034 ± 2040.
Asano K, Sakamoto H, Sasaki H, Ochiya T, Yoshoda T,
Oishi Y, Matida T, Kakizoe T, Sugimura T and Terada M.
(1995). Biochem. Biophys. Res. Commun., 217, 1169 ± 1176.
Atadja P, Wong H, Veillete C and Riabowol K. (1995). Expl.
Cell Res., 217, 205 ± 216.
Bartkova J, Lukas J, Strauss M and Bartek J. (1994). J.
Pathol., 172, 237 ± 245.
Bartkova J, Lukas J, Strauss M and Bartek J. (1995).
Oncogene, 10, 775 ± 778.
Betticher DC, Thatcher N, Altermatt HJ, Hoban P, Ryder
WDJ and Heighway J. (1995). Oncogene, 11, 1005 ± 1011.
Brancolini C, Ottega S and Schneider C. (1992). J. Cell. Biol.,
117, 1251 ± 1261.
Bravo R and MacDonald-Bravo H. (1987). J. Cell. Biol.,
105, 1549 ± 1554.
Bravo R. (1990). Growth Factors, Di€erentiation Factors, and
Cytokines. Habenicht A (ed). Springer-Verlag: Berlin
pp. 324 ± 343.
Buckley MF, Sweeney JE, Hamilton JA, Sini RL, Manning
DL, Nicholson RI, deFazio A, Watts CK, Musgrove EA
and Sutherland RL. (1993). Oncogene, 8, 2127 ± 2133.
Cavanaugh AH, Hempel WM, Taylor LJ, Rogalsky V,
Todorov G and Rothblum LI. (1995). Nature, 347, 177 ±
180.
Cellappan SP, Hiebert S, Mudryj M, Horowitz JM and
Nevins JR. (1991). Cell, 65, 1053 ± 1061.
Chen X, Bargonetti J and Prives C. (1995). Cancer Res., 55,
4257 ± 4263.
Ciccarelli C, Philipson L and Sorrentino V. (1990). Mol. Cell.
Biol., 4, 1525 ± 1529.
Del Sal G, Murphy M, Ruaro EM, Lazarevic D, Levine AJ
and Schneider C. (1996). Oncogene, 12, 177 ± 185.
Dulic V, Lees E and Reed SI. (1992). Science, 257, 1958 ±
1961.
Dulic V, Drullinger LF, Lees E, Reed SI and Stein GH.
(1993). Proc. Natl. Acad. Sci. USA, 90, 11034 ± 11038.
Eblen ST, Fausch MP, Anders RA and Leof EB. (1995). Cell
Growth Di€er., 6, 915 ± 925.
El-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons
R, Trent JM, Lin D, Mercer WE, Kinzler KW and
Vogelstein B. (1993). Cell, 75, 817 ± 825.
El-Deiry WS, Harper W, O'Connor PM, Velculescu VE,
Canman CE, Jackman J, Pietenpol JA, Burrell M, Hill
DE, Wang Y, Wiman KG, Mercer WE, Kastan MB, Kohn
KW, Elledge SJ, Kinzler KW and Vogelstein B. (1994).
Cancer Res., 54, 1169 ± 1174.
Ewen ME, Sluss HK, Scherr CJ, Matsushime K, Kato J and
Livingston D. (1993). Cell, 73, 487 ± 497.
Freeman RS, Estus S and Johnson Jr EM. (1994). Neuron,
12, 343 ± 355.
Gillet C, Fantl V, Smith R, Fisher C, Bartek J, Dickson C,
Barnes D and Peter G. (1994). Cancer Res., 54, 1812 ±
1817.
Girard F, Strausfeld U and Lamb J. (1991). Cell, 67, 1169 ±
1179.
Han EKH, Sgambato A, Jiang W, Zhang YJ, Santella RM,
Doki Y, Cacace AM, Schieren I and Weinstein IB. (1995).
Oncogene, 10, 953 ± 961.
Hayles J, Fisher D, Woollard A and Nurse P. (1994). Cell,
78, 813 ± 822.
Hunter T and Pines J. (1994). Cell, 79, 573 ± 582.
1711
Cell cycle regulation of cyclin D1 splice forms
H Sawa et al
1712
Jiang W, Kahn SM, Tomita N, Zhang YJ, Lu SH and
Weintstein IB. (1992). Cancer Res., 52, 2980 ± 2983.
Jiang W, Kahn SM, Zhou P, Zhang YJ, Cacace AM, Infante
AS, Doi S, Santella RM and Weintstein IB. (1993).
Oncogene, 8, 3447 ± 3457.
Kato JY, Matsuoka M, Strom DK and Sherr CJ. (1994).
Mol. Cell. Biol., 14, 2713 ± 2721.
Key G, Becker MH, Baron B, Duchrow M, Schluter C, Flad
HD and Gerdes J. (1993). Lab. Invest., 68, 629 ± 636.
Koh J, Enders GH, Dynlacht BD and Harlow E. (1995).
Nature, 375, 506 ± 510.
Kranenburg O, van der Eb AJ and Zantema A. (1996).
EMBO J., 15, 46 ± 54.
Laemmli UK. (1970). Nature, 227, 680 ± 685.
Lallemand D, Spyrou G, Yaniv M and Pfarr CM. (1997).
Oncogene, 14, 819 ± 830.
Lam EW and Watson RJ. (1993). EMBO J., 12, 2705 ± 2713.
Liu Y, Martindale JL, Gorospe M and Holbrook NJ. (1996).
Cancer Res., 56, 31 ± 35.
Lowry OH, Rosenbrough NJ, Farr A and Randall RJ.
(1951). J. Biol. Chem., 193, 265 ± 275.
Lucibello FC, Sewing A, BruÈsselbach S, BuÈrger C and MuÈller
R. (1993). J. Cell. Sci., 105, 123 ± 133.
Lukas J, Parry D, Aagaard L, Mann DJ, Bartkova J, Strauss
M, Peters G and Bartek J. (1995). Nature, 375, 503 ± 506.
Matsuoka M, Kato JY, Fisher RP, Morgan DO and Sherr
CJ. (1994). Mol. Cell. Biol., 14, 7265 ± 7275.
Michalides R, van Veelen N, Hart A, Loftus B, Wientjens E
and Balm A. (1995). Cancer Res., 55, 973 ± 978.
Motokura T and Arnold A. (1993). Biochem. Biophys. Acta,
1155, 63 ± 78.
Nevins JR. (1992). Science, 258, 424 ± 429.
Ohtsubo M, Theodoras AM, Schumacher J, Roberts JM and
Pagano M. (1995). Mol. Cell. Biol., 15, 2612 ± 2624.
Pagano M, Pepperkok R, Verde F, Ansorge W and Draetta
G. (1992). EMBO J., 11, 761 ± 771.
Pellicciari C, Mangiarotti R, Bottone MG, Danova M and
Wang E. (1995). Cytometry, 21, 329 ± 337.
Pusch O, Soucek T, Wawra E, HengstschlaÈger-Ottnad E,
Bernaschek G and HengstschlaÈger M. (1996). FEBS.
Letter, 385, 143 ± 148.
Rogers S, Wells R and Rechsteiner M. (1986). Science, 234,
364 ± 368.
Rosenberg CL, Kim HG, Shows TB, Kronenberg HM and
Arnold A. (1991a). Oncogene, 6, 449 ± 453.
Rosenberg CL, Wong E, Petty EM, Bale AE, Tujimoto Y,
Harris NL and Arnold A. (1991b). Proc. Natl. Acad. Sci.
USA, 88, 9639 ± 9642.
Ryder K, Lanahan A, Perez-Albuerne E and Nathaus D.
(1989). Proc. Natl. Acad. Sci. USA, 86, 1500 ± 1503.
Sambrook J, Fritsch EF and Maniatis T. (1989). Molecular
cloning: A laboratory manual, Ed. 2. Cold Spring Harbor
Laboratory Press, New York.
Sanger F, Nicklen S and Coulson AR. (1977). Proc. Natl.
Acad. Sci. USA, 74, 5463 ± 5467.
Schneider C, King RM and Philipson L. (1988). Cell, 54,
787 ± 793.
Tenan M, Benedetti S and Finocchiaro G. (1995). Biochem.
Biophys. Res. Commun., 217, 195 ± 202.
Towbin H, Staehelin T and Gordon T. (1979). Proc. Natl.
Acad. Sci. USA, 76, 4350 ± 4354.
Xiong Y, Zhang H and Beach D. (1992). Cell, 71, 505 ± 514.
Waga S, Hannon GJ, Beach D and Stillman B. (1994).
Nature, 369, 574 ± 578.
Walker DH and Maller JL. (1991). Nature, 354, 314 ± 317.
Wang TC, Cardi€ RD, Zukerberg L, Lee E, Arnold A and
Schmidt EV. (1994). Nature, 369, 669 ± 671.
Weinberg RA. (1995). Cell, 81, 323 ± 330.
White RJ, Trouche D, Martin K, Jackson SP and
Kouzarides T. (1996). Nature, 382, 88 ± 90.
Zhou P, Jiang W, Weghorst CM and Weinstein IB. (1996).
Cancer Res., 56, 36 ± 39.
Zwijsen RML, Klompmarker R, Wientjens EBHG, Kristel
PMP, Burg B and Michalides RJAM. (1995). Mol. Cell.
Biol., 15, 2554 ± 2560.