(Received for publication, January 25, 1996, and in revised form, October 4, 1996)
From the Institute for Experimental Cancer Research, Tumor Biology Center, Freiburg 79106, Germany
Stimuli that are mitogenic for mature T-cells induce cell cycle arrest in some T-cell tumors and T-cell hybridomas. The molecular mechanism of this growth inhibition is poorly understood. In this report, we show that in EL4, a murine T-lymphoma cell line, stimulation with concanavalin A or treatment with phorbol 13-myristate 12-acetate (PMA) inhibit growth, due to cell cycle arrest at both the G1 and the G2/M phases. The block at the G1 phase is accompanied by the appearance of a hypophosphorylated form of the retinoblastoma protein (pRb), due to the inhibition of G1 cyclin-Cdk complexes. However, the molecular mechanisms leading to this G1 cell cycle arrest differ between concanavalin A and PMA: concanavalin A inhibits both cyclin E-Cdk2 and cyclin D-Cdk4 complexes, while PMA inhibits only cyclin E-Cdk2. We demonstrate that concanavalin A inhibits cyclin D-Cdk4 activity by decreasing the amount of cyclin D. The inhibition of cyclin E-Cdk2 by both concanavalin A and PMA is due to increased binding of the Cdk inhibitor p21 to this complex. However, while stimulation of the cells with concanavalin A did not result in an evident increase of the total level of p21, treatment of the cells with PMA increased p21 levels significantly. Our results indicate, furthermore, that the G2/M block results from the inhibition of cyclin A- and cyclin B1-associated kinase activities. As for cyclin E-Cdk2, the inhibition of the cyclin A-Cdk2 complex is due to increased binding of the p21 inhibitor.
Activation of mature T lymphocytes through their T-cell receptor by either specific antigens or a wide range of agonists, including monoclonal antibodies and lectins, leads to the induction of several intracellular events, such as tyrosine phosphorylation of proteins, increase in phosphatidylinositol hydrolysis, activation of protein kinase C, and elevated cytosolic Ca2+ level (1-3). This, in turn, activates transcription and secretion of lymphokines such as interleukin 2 (4), and ultimately results in cell proliferation (4, 5). A similar effect is obtained when a combination of phorbol esters and calcium ionophores is used to directly activate protein kinase C and increase intracellular Ca2+ levels. Conversely, immature thymocytes, spontaneously dividing T-cell tumors and hybridomas exhibit a dramatic inhibition of growth following activation (6-9). Such growth inhibition was shown to be characterized by a G1/S block in the cell cycle followed by programed cell death (apoptosis). These two events are induced by different signaling pathways: activation-induced cell death requires extracellular Ca2+ and is prevented by cyclosporin A, whereas cell cycle block does not require extracellular Ca2+ and is not inhibited by cyclosporin A (9-11).
Passage through the cell cycle is a highly regulated process involving
sequential activation of a series of cell cycle control proteins, the
cyclins and their catalytic subunits, the cyclin-dependent kinases (Cdks)1 (12). Activation of a Cdk
requires its association with a cyclin. Different cyclin-Cdk complexes
are required at different phases in the cell cycle. Cyclin D- and
cyclin E-associated kinases are necessary for the G1/S
transition (12), whereas cyclin A- or cyclin B-activated kinases are
necessary for progression through S phase and entry into mitosis
(13-15). The cyclin-Cdk complexes can be further regulated by proteins
which bind to the complexes and inhibit their activity (16). A number
of growth inhibitory signals have been shown to promote cell cycle
arrest by activating such inhibitors. This is the case for the
transforming growth factor 1, which causes a G1/S block
in mink lung epithelial cells (17) by inducing the p27Kip1
inhibitor and in turn inhibiting the cyclin E-Cdk2 complex (18). DNA-damaging agents lead to cell cycle arrest due to the
transcriptional induction of another inhibitor protein, termed
variously p21, Waf1, and Cip1 (19-21).
Dephosphorylation of a tumor suppressor gene product, the retinoblastoma protein (pRb), seems to be one effect of the extracellular signals that induce cell cycle arrest and differentiation (22, 23). In G0/G1, pRb is underphosphorylated and complexed to the E2F transcription factor. This prevents the activation of some E2F-regulated genes required for DNA replication (24-27). Phosphorylation of pRb, during mid to late G1 by cyclin D- and cyclin E-associated kinases (28, 29), is accompanied by release of E2F and activation of transcription of E2F-regulated genes, resulting in entry into S phase.
In this report, we show that stimulation of EL4 cells with concanavalin A or PMA induces a similar inhibition of growth, due to cell cycle arrest at both the G1 and the G2/M phases. However, concanavalin A and PMA activate different molecular mechanisms to inhibit growth.
The human pRb specific monoclonal antibody was purchased from Pharmingen. The rabbit polyclonal antibodies anti-Cdk2 and anti-cyclin E were obtained from UBI. The antibodies against p21, cdc2, cyclin B1, and Cdk4 were purchased from Santa Cruz Biotechnology. The antibodies specific for cyclin D1 and cyclin A were kindly provided by G. Draetta (Mitotix, Boston, MA). The GST-pRb fusion protein was a generous gift from L. Meijer (CNRS, Roscoff, France). The antiserum against ERK1 and ERK2 were a generous gift from D. Fabbro (Ciba, Basel, Switzerland). The MEK inhibitor, PD 98059, was obtained from the Warner-Lamber Co. (Ann Arbor, MI).
Cell CultureEL4 is a C57BL/6-derived murine T-cell lymphoma. EL4 cells were maintained in logarithmic growth in suspension culture of Dulbecco's modified Eagle's medium (BioWhittakers) supplemented with 10% heat inactivated fetal bovine serum (BioWhittakers), 2 mM glutamine, 100 units/ml penicillin, 100 µg/ml streptomycin, and 50 µM 2-mercaptoethanol (complete medium).
Proliferation AssayCell proliferation was measured as the incorporation of [3H]thymidine into DNA. EL4 cells were cultured in triplicate in 96-well plates in a volume of 200 µl of complete medium at a density of 2 × 104 cells/well. Cells were stimulated with either 20 µg/ml concanavalin A (Sigma) or 10 ng/ml phorbol 13-myristate 12-acetate (PMA; Sigma) and pulsed with 1 µCi of thymidine/well (methyl-3H, 6.7 Ci/mmol; DuPont NEN) for the final 2-4 h. Cells were harvested onto glass filter strips using a semiautomated cell harvester (PHD Cell Harvester, Cambridge Technology), and incorporation of [3H]thymidine was determined by liquid scintillation counting. Coefficient of variation of the triplicates were less than 10%.
Flow Cytometry2.5 × 106 cells were cultured in 5 ml of complete medium in the absence or presence of either 20 µg/ml concanavalin A or 10 ng/ml PMA. At the indicated times, the cells were washed with Dulbecco's modified Eagle's medium and resuspended in 1 ml of Dulbecco's modified Eagle's medium + 50% fetal bovine serum. The cells were then fixed by slowly adding chilled ethanol to a final concentration of 50% and incubated on ice for 30 min. After fixing, the cells were pelleted, washed with phosphate-buffered saline and incubated for 30 min on ice, in phosphate-buffered saline containing 250 µg/ml RNase A (Boehringer Mannheim) and 25 µg/ml propidium iodine (Sigma). DNA content analysis of stained nuclei was performed with a Becton Dickinson FACSort using 488 excitation at 15 milliwatts. Doublets and higher aggregates were excluded from analysis by selecting a singlet gate on width versus area dot plot. Cell cycle analysis was done by using the Lysis II software.
Cell ExtractsCells were harvested, washed twice in
phosphate-buffered saline at 4 °C and the cell pellet resuspended in
extraction buffer (25 mM Tris-HCl, pH 7.5, 60 mM -glycerophosphate, 15 mM
MgCl2, 15 mM EGTA, 0.1 mM sodium
fluoride, 1 mM dithiothreitol, 15 mM 4-nitrophenyl phosphate, 0.1 mM phenylmethylsulfonyl
fluoride, 0.1% Nonidet P-40) and after vortexing, centrifuged at
10,000 × g for 10 min at 4 °C. The supernatant was
collected and protein concentration determined with the BCA protein
assay kit (Pierce). Samples were stored at
80 °C or immediately
used for immunoblotting and immunoprecipitations.
Proteins (3-20 µg) were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto Immobilon-P membranes (Millipore). Membranes were then incubated with specific antibodies and proteins were visualized using the ECL detection kit (Amersham).
Immunoprecipitations and Kinase AssaysCell extracts were
incubated with immunoprecipitating antibody in extraction buffer for
2-4 h at 4 °C. The immune complexes were precipitated with 10 µl
of protein A-Sepharose for 30 min, and washed three times with
extraction buffer and once with assay buffer (50 mM
Tris-HCl, pH 7.5, 10 mM MgCl2, 1 mM
dithiothreitol). Protein A-Sepharose beads were dried and then
resuspended in 20 µl of assay buffer. Kinase activities were assayed
for 15 min at 30 °C, in a final volume of 25 µl containing 5 µM [-32P]ATP (specific activity 50 µCi/nmol), and 40 µM peptide substrate. The synthetic
peptide employed was
p70s6k408-427,
corresponding to amino acids 408 to 427 of rat S6 kinase
(p70s6k) (30). Reactions were terminated by the addition of a
5-µl solution containing 0.1 M EDTA and 6 mM
adenosine. Incorporation of [32P]phosphate into the
peptide substrate was determined by spotting 20-µl aliquots on p81
paper (Whatman) as described previously (31). Cdk4 kinase activities
were measured similarly, except that 10 µg of a GST-pRb fusion
protein was used as substrate. The kinase reactions were terminated by
addition of Laemmli sample buffer, the phosphorylated GST-pRb
fusion protein were resolved on 7.5% SDS-PAGE and visualized by
autoradiography.
100 µg of total proteins was
immunoprecipitated with anti-ERK1 or anti-ERK2 antibodies for 2 h
in a lysis buffer containing 50 mM Tris-HCl, pH 8.0, 120 mM NaCl, 20 mM sodium fluoride, 1 mM benzamidine, 1 mM EDTA, 6 mM
EGTA, 15 mM sodium pyrophosphate, 30 mM
4-nitrophenyl phosphate, 0.1 mM phenylmethylsulfonyl
fluoride, and 1% Nonidet P-40. Immunoprecipitates were mixed with
protein A-Sepharose beads for 30 min, and beads were washed three times with 1 ml of the same lysis buffer. Kinase reactions were performed for
15 min at 37 °C in a 20-µl reaction mixture (20 mM
Hepes pH 7.0, 10 mM -glycerophosphate, 1 mM
dithiothreitol, 10 mM MgCl2, 20 µM [
-32P]ATP (specific activity 5 µCi/nmol), 0.1 mg/ml myelin basic protein (Sigma)).
The reactions were stopped by the addition of Laemmli SDS sample
buffer, and phosphorylated myelin basic protein was detected by
SDS-PAGE followed by autoradiography.
As opposed to normal mature T lymphocytes, which are
induced to proliferate by stimulation via the antigen specific receptor (T-cell receptor) or by using a combination of phorbol ester and calcium ionophores, T-cell tumors exhibit a dramatic growth inhibition upon stimulation (6, 8). The response of EL4 cells to stimulation with
a combination of PMA and the calcium ionophore ionomycin was measured
as incorporation of [3H]thymidine (Fig.
1A). As expected, we observed a dramatic
decrease in proliferation when cells were treated with PMA and
ionomycin, showing only about 10% of growth after 10 h of
stimulation compared to nontreated cells. Interestingly, a similar
growth inhibition was observed when EL4 cells were stimulated with PMA
in the absence of ionomycin, indicating that growth inhibition can be
mediated by a PMA-dependent pathway. The effect of PMA on
the proliferation of EL4 cells was rapid and pronounced (Fig.
1B): only 30% of the cells remained proliferating after
3 h of PMA addition, this percentage decreasing to less than 5%
after 12 h. A similar effect was observed when the cells were
stimulated with concanavalin A (Fig. 1B): after a
progressive decrease of proliferation, less than 5% of growth was
observed after 12 h of treatment.
The observed inhibition could be explained by activation-induced cell death. To address this possibility, EL4 cells were stimulated with concanavalin A or PMA and cell death was measured using a 51Cr release assay (8). We could not observe any 51Cr release following stimulation of the cells with either concanavalin A or PMA at the early time points when our experiments were performed. Significant release of 51Cr could only be detected 24 h after stimulation (data not shown). These results indicate that the growth inhibition observed in response to both stimuli used is due to a growth arrest of the cells rather than to activation-induced cell death.
PMA and Concanavalin A Cause a Double Block in the Cell CycleIn order to characterize the cell cycle block, EL4 cells
were treated with either PMA or concanavalin A, incubated with
propidium iodine, and analyzed by fluorescence-activated cell sorter
for DNA content. As shown in Fig. 2, we observed a
decrease of the number of cells entering S phase in cells stimulated
with either PMA or concanavalin A for 6 h. This indicates the
existence of a block in the cell cycle at the G1 phase. In
addition, the percentage of cells in the G1 phase remained
constant, although the amount of cells in the G2 and M
phases doubled. These results show the existence of a second block at
the G2/M phase. Treatment of EL4 cells for 12 h with
both stimuli resulted in an almost complete transition of cells from S
to the G2/M phase of the cell cycle, rendering the blocks
at the G1 and G2/M phases even more
evident.
Hypophosphorylation of pRb following Stimulation with Concanavalin A and PMA
Functional inactivation of pRb by phosphorylation is
required for the G1 to S transition. Since PMA and
concanavalin A arrested EL4 cells before they reached S phase (Fig. 2),
we determined whether concanavalin A and PMA affect the phosphorylation
level of pRb. EL4 cells were stimulated with either PMA or concanavalin A and pRb was examined in whole cell lysates by Western blotting (Fig.
3). Differentially phosphorylated forms of the protein
can be distinguished by their relative mobilities in a gel. In
proliferating EL4 cells, pRb was found mainly in the
hyperphosphorylated form. When these cells were treated with PMA or
concanavalin A, pRb became hypophosphorylated as judged by the
appearance of more rapidly migrating forms. The appearance of
hypophosphorylated pRb after stimulation of the cells with concanavalin
A and PMA explains the G1 block observed in Fig. 2.
Activation-induced Growth Inhibition Is Mediated by a MAP Kinase-independent Pathway
The signaling pathway leading to
growth inhibition in T-cell tumors and T-cell hybridomas is obscure. It
has been previously shown that activation-induced growth arrest occurs
independently of extracellular Ca2+ and is insensitive to
cyclosporin A (10, 11). Since the MAP kinase pathway has been
implicated in both cell proliferation (32-37) and growth arrest (38,
39), we wanted to investigate whether the growth inhibitory signals
delivered by PMA and concanavalin A were mediated by the MAP kinase
pathway. To address this possibility, EL4 cells were stimulated with
concanavalin A or PMA in the presence or absence of PD 98059, a
compound which prevents MAP kinase activation by specifically blocking
MEK activity (37, 40, 41). As shown in Fig. 4,
A and B, the inhibitor had no effect on the
concanavalin A- or PMA-induced growth inhibition nor on the
hypophosphorylation of pRb, although the drug completely blocked the
concanavalin A-induced activation of the two MAP kinases, ERK1 and ERK2
(Fig. 4C). These results demonstrate that the MEK/MAP kinase
pathway is not involved in mediating the signals leading to
hypophosphorylation of pRb and growth inhibition in cells stimulated
with concanavalin A. Interestingly, this inhibitor of MEK had only a
little effect on PMA-induced MAP kinase activation (Fig.
4B). Activation of the MAP kinases in response to PMA could
not be blocked even when higher concentrations of the MEK inhibitor
were used (data not shown), making it difficult to determine the role
of the MAP kinase pathway in the PMA-induced growth inhibition, using
this MEK-specific inhibitor.
Differential Inhibition of G1 Cyclin-Cdk Complexes by PMA and Concanavalin A
The appearance of a hypophosphorylated
form of pRb could result from either the activation of a phosphatase or
the inactivation of a kinase. Two protein phosphatases, types 1 and 2A
(PP1 and PP2A), have been implicated in the control of cell cycle
events in yeast and mammalian cells (42, 43). Indeed, at the end of
mitosis, pRb has been shown to be dephosphorylated in an okadaic acid-sensitive manner (44). The authors suggested a role for PP1 in
late mitotic dephosphorylation of pRb. To investigate whether pRb
hypophosphorylation was due to the activation of these phosphatases, EL4 cells were preincubated with okadaic acid and then stimulated for
2 h with PMA. Western blotting of pRb showed that okadaic acid at
concentrations up to 10 µM, which is well over the
IC50 for both PP1 and PP2A, did not have any effect on the
phosphorylation state of pRb (Fig. 5). These results
suggest that, in EL4 cells, pRb hypophosphorylation is not due to
activation of PP1 or PP2A.
Alternatively, the appearance of the dephosphorylated pRb could be due
to inhibition of a kinase. In mammalian cells, cyclin D-Cdk4 and cyclin
E-Cdk2 complexes have been shown to phosphorylate pRb, regulating
G1 progression and entry into S phase (28, 29, 45, 46).
Therefore, we wanted to analyze whether pRb hypophosphorylation was due
to a decreased activity of either one of these complexes. Cdk4 kinase
activity was measured from cells stimulated in the presence or the
absence of either PMA or concanavalin A. As shown in Fig.
6A, Cdk4 antibody precipitated a significant
amount of kinase activity in unstimulated cells. PMA had no effect on
the basal activity of this kinase. On the contrary, stimulation of the
cells with concanavalin A resulted in a significant reduction in the
kinase activity of Cdk4. We next measured the cyclin E-associated kinase activity from cells stimulated with concanavalin A or PMA (Fig.
6B). Cyclin E-associated kinase activity was high in
unstimulated cells. Addition of either concanavalin A or PMA resulted
in a dramatic decrease in this kinase activity. These results show that
concanavalin A inhibit both the cyclin E-associated kinase and Cdk4,
while interestingly, stimulation with PMA resulted in a specific
inhibition of the cyclin E-associated kinase.
Molecular Mechanisms of the G1 Block Induced by Concanavalin A and PMA Are Different
Activation-induced
G1/S block is a common property of many transformed T-cells
(6, 8). In order to further analyze this cell cycle block we set out to
investigate the mechanism of the inhibition of Cdk4. Cdk4 is known to
form a complex with cyclin D, its regulatory subunit. To determine
whether Cdk4 retained the ability to form a complex with cyclin D in
the activated cells, Cdk4 immunoprecipitates were probed with
anti-cyclin D antibodies. As shown in Fig.
7A, Cdk4 co-precipitated with cyclin D in
unstimulated EL4 cells. Stimulation of the cells with PMA did not
change this association. However, stimulation of the cells with
concanavalin A reduced the association of the two proteins. This could
be due to a decrease in the total level of either protein. As shown in Fig. 7A, the total level of cyclin D was reduced in cells
treated with concanavalin A but not in cells treated with PMA, while
the amount of Cdk4 remained constant after either stimulation. Taken together, these results show that the concanavalin A-mediated inhibition of cyclin D/Cdk4 kinase activity is due to a reduction in
the total level of cyclin D.
Since we observed a decrease in cyclin E-associated kinase activity in cells treated with either concanavalin A or PMA, we determined whether this was due to a decrease in the total level of one of the proteins forming the cyclin E-Cdk complex, as it is the case for the inhibition of cyclin D-Cdk4. Cdk2 co-precipitated with cyclin E in unstimulated cells (Fig. 7B). Treatment of the cells with concanavalin A or PMA did not affect the formation of the cyclin E-Cdk2 complex. In agreement with these results, the total cellular levels of cyclin E and Cdk2 proteins remained constant after stimulation of the cells with concanavalin A or PMA (data not shown). Another possible explanation for inhibition of the cyclin E/Cdk2 activity could be the induction of an inhibitor of cyclin-Cdk complexes. To address this question, whole cell lysates of unstimulated cells or cells stimulated with concanavalin A or PMA were blotted with an antibody directed against the cyclin/Cdk inhibitor, p21 (Fig. 7C). Proliferating EL4 cells expressed a low amount of p21. Treatment of the cells with concanavalin A had no detectable effect on the basal level of p21. On the contrary, in cells stimulated with PMA, expression of the p21 inhibitor was significantly induced. To determine whether the induction of p21 is responsible for the inhibition of the cyclin E/Cdk2 kinase activity, cyclin E was immunoprecipitated from unstimulated and cells stimulated with concanavalin A or PMA. p21 did not co-precipitate with cyclin E from unstimulated cells. However, cyclin E complexes contained the p21 inhibitor in cells stimulated with PMA, strongly indicating that inhibition of the cyclin E-associated kinase activity in PMA-treated cells is mediated by p21. Surprisingly, although concanavalin A had no detectable effect on the total cellular level of p21, this inhibitor was also co-precipitated with cyclin E complexes after stimulation of the cells with concanavalin A. This indicates that p21 is also involved in the concanavalin A-mediated inhibition of the cyclin E/Cdk2 activity.
Activation-induced Inhibition of Cyclin A- and Cyclin B-associated KinasesTwo cyclins are important for progression through the
G2 and M phases of the cell cycle, cyclin A and cyclin B
(14, 15, 47). To investigate whether the G2/M block
observed could be correlated with the inhibition of the Cdk complexes
containing either one of these cyclins, we immunoprecipitated cyclin A
and cyclin B1 from unstimulated cells or from cells stimulated with either concanavalin A or PMA and determined in vitro kinase
activity of the immunoprecipitates. Both concanavalin A and PMA
stimulation resulted in a significant decrease of cyclin A- and cyclin
B1-associated kinase activity (Fig. 8A).
These data strongly suggest that inhibition of the cyclin A and cyclin
B1-associated kinase activities is involved in the G2/M
block induced by both PMA and concanavalin A.
In order to better understand the mechanism of inhibition cyclin A- and cyclin B1-associated kinase activities, we analyzed the composition of cyclin A and cyclin B1 complexes. Cyclin A and cyclin B1 immunoprecipitates were blotted with antibodies directed against Cdk2 or cdc2. As shown in Fig. 8B, Cdk2 is the preferential partner of cyclin A, while only a small proportion of cdc2 is associated with this cyclin (data not shown). In contrast, cyclin B1 is bound to cdc2 (Fig. 8B) and only a negligible amount of cyclin B1-Cdk2 complexes could be detected (data not shown). We observed that binding of Cdk2 and cdc2 to cyclin A and cyclin B1, respectively, was not affected by either concanavalin A or PMA stimulation of the EL4 cells (Fig. 8B). This indicates that inhibition of complex formation is not the mechanism by which these kinases are inhibited. Next, we addressed the question of whether p21 could also inhibit cyclin A- and cyclin B1-associated kinase activities. Cyclin A and cyclin B1 were immunoprecipitated from lysates of unstimulated cells or cells stimulated with concanavalin A or PMA and the presence of p21 in the immunoprecipitates was detected by Western blotting. As shown in Fig. 8B, p21 did not co-precipitate with cyclin A from unstimulated cells. However, after stimulation with either concanavalin A or PMA, p21 was found associated with cyclin A. On the contrary, p21 was not associated with cyclin B1 whether the cells were stimulated or not, suggesting that inhibition of cyclin B1/cdc2 kinase activity occurs independently of p21.
Activation of T-cell hybridomas and T-cell lymphomas through their T-cell receptor or with a combination of calcium ionophores and phorbol esters has been shown to induce a cell cycle arrest (6-9). The present study extends these previous observations, showing that activation of EL4 cells with either concanavalin A or PMA results in a double block in the cell cycle, at both the G1 and G2/M phases. This growth arrest is accompanied by the appearance of a hypophosphorylated form of pRb. Although they cause a similar growth arrest, concanavalin A and PMA utilize different mechanisms to inhibit the activity of G1 and G2/M cyclin-Cdk complexes. Furthermore, the concanavalin A-mediated growth inhibitory signal does not involve activation of the MEK/MAP kinase pathway.
Following stimulation of asynchronously growing EL4 cells, we observed two cell cycle blocks. Other reports describe a unique block either at the G1 phase or at the G2/M phase of the cell cycle (6, 8, 48, 49). This discrepancy could be explained by cell line-specific variations. Moreover, most of the studies so far reported were conducted on cells undergoing apoptosis after only a few hours of stimulation, making it more difficult to observe the two blocks. This is not the case for EL4 cells. Finally, the drugs used in order to synchronize the cells in some studies may have influenced the experiments.
Transition of cells from G1 to S phase in the cell cycle is controlled by pRb activity (for a review, see Refs. 50 and 51). In G1, pRb is hypophosphorylated and exerts its antiproliferative function. Hyperphosphorylation of pRb by cyclin D-Cdk4 and cyclin E-Cdk2 complexes inactivates it at the G1/S transition, allowing the cells to proceed into S phase (28, 29, 46, 52, 53). Several lines of evidence indicate that cell cycle arrest at the G1 to S transition is linked to cyclin D-Cdk4 or cyclin E-Cdk2 complex inhibition (17, 18, 54-56) and pRb hypophosphorylation (22). Our results with EL4 cells confirm this model, showing that G1 cell cycle arrest in response to either concanavalin A or PMA is coupled to pRb hypophosphorylation. During anaphase, pRb has been shown to be dephosphorylated in an okadaic acid-sensitive manner (44), suggesting a role for PP1 or PP2A in late mitotic dephosphorylation of pRb. We reasoned that this could have also been the case in our study. However, we found that addition of okadaic acid to EL4 cells did not prevent PMA-induced pRb hypophosphorylation. We conclude that the accumulation of hypophosphorylated pRb is rather due to inhibition of the kinase activity of cyclin-Cdk complexes than to activation of PP1 or PP2A. In addition, these results suggest that the cells very likely accumulate before anaphase when blocked in G2/M by PMA, a time at which PP1 and PP2A are still not involved in dephosphorylating pRb.
Our results show that concanavalin A and PMA are equally potent inhibitors of EL4 cell proliferation. However, the mechanisms by which they inhibit growth are not similar. Concanavalin A inhibits the activity of the two G1 complexes, cyclin D-Cdk4 and cyclin E-Cdk2, while PMA specifically inhibits cyclin E-Cdk2. We find that the inhibition of the cyclin E/Cdk2 kinase activity in response to both concanavalin A and PMA results from binding of the p21 Cdk inhibitor to the complex. Moreover, in the case of concanavalin A, the observed decrease in the amount of cyclin D-Cdk4 complexes may also contribute to the inactivation of cyclin E/Cdk2. It has been proposed that cyclin D-Cdk4 complexes sequester a Cdk2 inhibitor (18, 56). In that case, disruption of cyclin D-Cdk4 complexes by concanavalin A, may release one (or several) inhibitor which is then free to bind to and inhibit the activity of the Cdk2-containing complexes. In agreement with this hypothesis, we find that both concanavalin A and PMA induced binding of p21 to the cyclin E-Cdk2 and the cyclin A-Cdk2 complexes. However, only PMA led to an induction of the total cellular level of p21, suggesting that concanavalin A acts by retargeting pre-existing p21 protein to cyclin E-Cdk2 and cyclin A-Cdk2 complexes. An alternative explanation is that induction of the p21 protein by concanavalin A is low and thus, below the limits of sensitivity of our detection system. This low level of induction of p21 might, however, be sufficient to allow its binding to proteins with high affinity like the cyclin-Cdk2 complexes.
The mechanism by which concanavalin A and PMA inhibit cyclin B1/cdc2 kinase activity is not yet clear. Our data indicate that complex formation between cyclin B1 and cdc2 is not impaired. Moreover, we could not detect significant amounts of p21 co-precipitating with the cyclin B1 complexes. This indicates that p21 is unlikely to be the cause of their inhibition. Two hypothesis accounting for the inhibition of these complexes can be formulated. The first would be binding of a novel inhibitory protein to the complexes. In alternative, a post-translational modification of either cyclin B or cdc2 might be the cause of the inhibition observed. Indeed, phosphorylation of the cdc2 protein has been shown to regulate its kinase activity (57-59). We are currently performing additional experiments to test these different possibilities and elucidate the mechanism of inhibition of the cyclin B-cdc2 complexes.
Limited information on the signaling pathway leading to activation-induced growth inhibition in T-cell lymphomas is available at the moment. It has been previously shown that growth inhibition is not dependent on extracellular Ca2+ and is insensitive to cyclosporin A (10, 11). Among other signal transduction pathways, we reasoned that the MAP kinase pathway could be involved in mediating the antiproliferative signals in EL4 cells. Indeed, although this cascade is usually engaged during proliferation (32-36), in yeast, activation of the pheromone response pathway, which is homologous to the mammalian MAP kinase pathway, results in cell cycle arrest (see for reviews, Refs. 38 and 39). We observed that a MEK-specific inhibitor was able to block the concanavalin A-induced activation of MAP kinase without, however, affecting the growth inhibition. This demonstrates that the MAP kinase pathway is not responsible for mediating the antiproliferative signals delivered by concanavalin A. Surprisingly, the MEK inhibitor did not prevent MAP kinase activation upon PMA stimulation. Two hypothesis can explain this phenomenon. First, MAP kinase activation could result from the existence of different isoforms of MEK, which are activated upon stimulation with PMA but not with concanavalin A and are insensitive to the inhibitor used. Alternatively, PMA could activate MAP kinase independently of MEK, revealing the existence of an alternative cascade leading to MAP kinase activation. Although this second hypothesis contradicts the established model in which MEK is the MAP kinase activator (60, 61), it is supported by recent findings indicating that activation of MEK does not necessarily lead to MAP kinase activation (62). This suggests that as yet undefined mechanisms may be involved in determining information flow through the MAP kinase pathway. Moreover, tissue- and cell line-specific variations in the composition of this signaling pathway may exist. Our preliminary results implicate the Ca2+-independent isoforms of protein kinase C in the PMA-induced growth inhibition.2 In addition, we considered the possible involvement of another signal transduction cascade, the stress-activated pathway, which is related to the MAP kinase pathway and has been shown to mediate antiproliferative signals upon cellular stress (63-65). In preliminary experiments, we could not see any effect of the specific inhibition of p38, a member of the stress kinases family, on the growth inhibition mediated by either concanavalin A or PMA. This indicates that the p38 kinase is not essential for the growth inhibition process. However, the involvement of other members of the stress kinases family in this process remains to be determined. A major challenge will now be to further elucidate the concanavalin A- and PMA-induced growth inhibitory pathways.
We are grateful to Drs. G. Draetta, D. Fabbro, and L. Meijer for the generous supply of antibodies and reagents and the Warner-Lamber Co. for the gift of the MEK inhibitor. We thank Drs. P. E. Shaw and Y. Nagamine for helpful discussion and Dr. H. Pahl for critical review of this manuscript.