Interdependence of cdk2 Activation and Interleukin-2R a Accumulation in T Cells*

We have shown previously that serum promotes T cell proliferation by acting with T cell receptor (TCR) agonists to efficiently down-regulate p27 Kip1 and activate cdk2-containing complexes. In the studies described here, the effect of serum on the expression of the a subunit of the interleukin-2 receptor (IL-2R a ) was examined. We found that serum was required for maximal and sustained IL-2R a protein expression and consequent IL-2 signaling in TCR-activated splenocytes. Serum had no effect on IL-2R a mRNA levels and thus modulates IL-2R a expression post-transcriptionally. Unlike wild-type splenocytes, splenocytes exhibiting se-rum-independent cdk2 activation due to loss of p27 Kip1 efficiently expressed IL-2R a in serum-deficient medium. Conversely, serum did not promote IL-2R a accumulation in conditions in which cdk2 activity was blocked. These findings demonstrate that cdk2 activation is necessary and sufficient for IL-2R a accumulation in TCR-stimulated splenocytes. On the other hand, IL-2 signaling was required (at least in part) for cdk2 activation in these cells. Thus, cdk2 activation, IL-2R a expression, and IL-2 signaling are interdependent events, and we suggest that this feed-forward regulatory loop plays a key role in T cell mitogenesis.

tion in conditions in which cdk2 activity was blocked. These findings demonstrate that cdk2 activation is necessary and sufficient for IL-2R␣ accumulation in TCR-stimulated splenocytes. On the other hand, IL-2 signaling was required (at least in part) for cdk2 activation in these cells. Thus, cdk2 activation, IL-2R␣ expression, and IL-2 signaling are interdependent events, and we suggest that this feed-forward regulatory loop plays a key role in T cell mitogenesis.
Engagement of T cell receptors (TCRs) 1 initiates a complex program of events that leads to the proliferation and differentiation of mature resting T cells (1). A key player in this program is interleukin-2 (IL-2), a lymphokine first identified in supernatants of antigen-primed T cells (2)(3)(4). The receptor for IL-2 (IL-2R) is noncatalytic and consists of three subunits termed IL-2R␣, IL-2R␤, and IL-R2␥ (also known as ␥c) (5). IL-2R␤ and IL-2R␥ are also components of other lymphokine receptors, whereas IL-2R␣ is unique to the IL-2R and thus is responsible for substrate specificity. IL-2R␣ is also required for high affinity (and presumably biologically relevant) IL-2 binding. In vitro, IL-2Rs lacking IL-2R␣ are functional in human T cells (albeit at a lower affinity than trimeric receptors) but not in mouse T cells (6 -10). On the other hand, IL-2R␣ has an extremely short cytoplasmic domain and does not participate in intracellular signaling per se. Instead, signaling is mediated by IL-2R␤ and IL-2R␥, which interact with a variety of cytoplasmic effector proteins (11). These include the Janus kinases (JAKs), Jak1 and Jak3, which associate constitutively with IL-2R␤ and IL-2R␥, respectively. When activated by ligandinduced receptor oligomerization, JAKs phosphorylate IL-2R␤ and IL-2R␥ on specific tyrosine residues to create docking sites for other effectors. Additional JAK substrates include the JAKs themselves and members of the STAT (signal transducer and activator of transcription) family of transcriptional regulators.
Of the three IL-2R subunits, IL-2R␣ exhibits the most variable expression. IL-2R␣ is not present in resting T cells and is transcriptionally up-regulated by TCR agonists (12). Such agonists include cognate antigen, anti-CD3, and concanavalin A (ConA). IL-2R␤ and IL-2R␥, on the other hand, are expressed constitutively, and TCR-induced changes in their expression are less dramatic (5). TCR activation increases IL-2 production (13), and IL-2 signaling further up-regulates IL-2R␣ expression in a STAT-dependent manner (14 -16). As a further prelude to IL-2 actions, TCR stimulation induces a variety of metabolic responses that allow quiescent T cells to exit G 0 (11). IL-2 then promotes continued G 0 /G 1 traverse and the initiation of DNA synthesis. T cell proliferation is also influenced by co-stimulatory signals delivered by CD28 (17) and by mitogens contained in serum, which is indispensable for T cell propagation in vitro (18). The capacity of co-stimulatory signals to up-regulate IL-2R␣ transcription has been reported (19).
As in all cell types, T cell proliferation is governed by the ordered activation of cyclin-dependent kinases (CDKs) (20). CDK activity is controlled by cyclins, which are positive regulators, and CDK inhibitors, which repress activity. Cyclins, which are expressed periodically, combine with CDKs to form active complexes at distinct points in the cell cycle. During G 0 /G 1 , for example, complexes containing the D cyclins and cdk4 or cdk6, cyclin E and cdk2, and cyclin A and cdk2 are sequentially assembled and activated. Activation of these complexes is required for G 0 /G 1 traverse and S phase entry, and key substrates include the Rb family of transcriptional repressors. The CDK inhibitor family includes the Cip/Kip proteins, which inactivate complexes containing cdk2 and, according to some reports, cdk4 and cdk6 (21)(22)(23). Of these inhibitors, p27 Kip1 is thought to play a particularly important role in T cell proliferation. p27 Kip1 is present at high levels in resting T cells and is down-regulated in response to mitogenic stimulation (24,25). Failure to reduce p27 Kip1 levels below a critical threshold precludes cdk2 activation and arrests T cells in G 0 /G 1 . Moreover, T cells lacking p27 Kip1 exhibit dysregulated cdk2 activation and proliferate in conditions that do not support the * This work was supported by the Cortner-Couch Endowed Chair for Cancer Research and National Institutes of Health Grants CA72694 and CA67360. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Given the pivotal role of IL-2 in T cell mitogenesis, a full understanding of the mechanisms regulating the expression of IL-2 and IL-2R is imperative. IL-2R density is a critical determinant of T cell proliferation (27), and previous studies have shown that IL-2R density is dictated by serum concentration (18). As monitored by IL-2 binding assays, serum (at 10%) significantly increased the expression of surface-localized IL-2Rs in antigen-treated human T cells (18). Building on this observation, data presented here demonstrate that serum selectively facilitates the post-transcriptional expression of IL-2R␣ in primary splenic T cells exposed to TCR agonists. Serummediated IL-2R␣ accumulation was accompanied by the induction of IL-2 signaling pathways and was dependent on cdk2 activation. Because cdk2 activity both contributed to and resulted from IL-2 signaling, we suggest that these events are interdependent. We propose a model of T cell proliferation in which serum-dependent p27 Kip1 down-regulation initiates a feed-forward loop consisting of cdk2 activation, IL-2R␣ accumulation, and IL-2R signaling.

Isolation of Splenocytes and Cell
Culture-A single cell suspension of splenocytes was prepared by passage through nylon mesh, and red cells were depleted using a whole blood erythrocyte lysing kit (R&D Systems). For purification, spleen cell suspensions were loaded onto T cell-enrichment columns (R&D Systems), and T cells were isolated by high affinity negative selection as specified by the manufacturer. Splenic and purified T cells were plated at 10 7 cells/ml and 5 ϫ 10 6 cells/ml, respectively, in RPMI 1640 supplemented with 50 units/ml penicillin, 2 mM L-glutamine, and 10% fetal calf serum. p27-47 fibroblasts were prepared and maintained as described previously (28). The percentage of S phase cells in a population was determined by FACS analysis of propidium iodide-stained cells as detailed previously (43).
Protein Analysis-Western blots were performed as described previously (43). For ELISA assays, flat bottom 96-well microtiter plates were coated overnight at 4°C with 6 g/ml IL-2R␣ monoclonal antibody (PC61, PharMingen) and blocked with 3% bovine serum albumin for 1 h at 37°C. Coated plates were treated consecutively as follows: serial dilutions of cell lysates, 1 h at 37°C; 1 g/ml biotin-conjugated IL-2R␣ monoclonal antibody (7D4, PharMingen), 1 h at 4°C; streptavidinbiotinylated peroxidase complex (PharMingen), 1 h at 37°C. Reactions were developed for 30 min with 100 l/well 3⅐5Ј-5⅐5Ј tetramethylbenzidine substrate (Dako) and stopped with 0.5 N H 2 SO 4 . Absorbance was measured at 450 nm on a Titertek ELISA reader. This protocol is a modified version of Osawa et al. (29). For analysis of cell surface markers, T cells were incubated in phosphate-buffered saline containing 2% mouse IgG (Dako) and fluorescein isothiocyanate-or phycoerythrinconjugated antibodies (PharMingen) for 30 min in the dark at 4°C. Corresponding isotype-specific conjugated antibody was used for detection of nonspecific binding. Analysis was performed on a FACScan flow cytometer with Cell Quest software (Becton Dickinson).
mRNA Analysis-Total mRNA was isolated using TRIzol, and Northern blots were performed as described previously (30). For RNase protection assays, mRNA (20 g) was hybridized overnight at 56°C with 32 P-labeled probes (10 5 cpm) corresponding to the mCR-1 probe set (PharMingen). Samples were then digested with RNase T1 and RNase A for 45 min at 30°C and proteinase K for 15 min at 37°C. Samples were extracted with phenol/chloroform, collected by sodium acetate/ ethanol precipitation, denatured at 90°C for 3 min, and electrophoresed on a 5% polyacrylamide gel. Gels were dried and exposed to x-ray film.
In Vitro Kinase Assay-Cell extracts were incubated with antibody to cyclin E or cyclin A for 4 -12 h at 4°C and subsequently with protein A-agarose beads. Immune complexes were washed twice with lysis buffer (43) and once with histone kinase buffer (50 mM Tris (pH 7.4), 10 mM MgCl 2 , 1 mM dithiothreitol). Washed complexes were incubated in 15 l of kinase buffer containing 20 M ATP, 0.1 Ci/ml [␥-32 P]ATP and 100 g/ml histone H1 (Roche Molecular Biochemicals) for 10 min at 37°C. Reactions were stopped by boiling in Laemmli buffer, and proteins were separated on SDS gels. Radiolabeled proteins were visualized by autoradiography.
Materials-ConA was purchased from Sigma Chemical Co., and anti-CD3, recombinant IL-2, and IL-2R␣ blocking antibody were obtained from PharMingen. Antibodies to IL-2R␣, IL-2R␤, IL-2R␥, Jak1, and phosphotyrosine were purchased from Santa Cruz Biotechnologies. p27 Kip1 and cdk2 antibodies were from Transduction Laboratories, and Jak3 antibody and cyclin A antibody were from Upstate Biotechnology and Neomarker, respectively. p27 Kip1 -deficient mice were provided by A. Koff.

RESULTS
Serum Modulates IL-2R␣ Expression at a Post-transcriptional Level-Although the serum dependence of IL-2R expression has been described previously (18), the receptor component(s) targeted by serum has yet to be identified. To address this issue, purified splenic T cells derived from Balb/c mice were stimulated for 30 h with mitogenic concentrations of anti-CD3 and either 10% or 0.1% serum, and the cell surface expression of IL-2R␣ and IL-2R␥ was determined by FACS analysis. These serum concentrations were chosen because they allow maximal (10% serum) and minimal (0.1% serum) amounts of DNA synthesis in anti-CD3-treated T cell cultures (Fig. 1A, top panel, and accompanying article (43)). As shown in Fig. 1B, resting T cells contained little if any surface-localized IL-2R␣, and a higher percentage of cells expressed this protein when stimulated with anti-CD3 and 10% as compared with 0.1% serum (76% versus 33%, respectively). Moreover, there were significantly more receptors per cell in the population receiving 10% serum as compared with 0.1% serum (Fig. 1C). On the other hand, one-third of the cells in the quiescent population were IL-2R␥-positive, and anti-CD3 increased this percentage ϳ2-fold irrespective of serum concentration (Fig. 1B). These data show that serum regulates the overall expression and/or the cell surface localization of IL-2R␣, but not of IL-2R␥, in TCR-stimulated T cells. Levels of IL-2R␤ detected by this assay were too low to be accurately quantitated.
To ensure that T cells remained viable when stimulated in serum-deficient medium, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assays were performed. When examined 40 h after stimulation, T cells treated with anti-CD3 and 0.1% serum were only slightly less viable than those receiving anti-CD3 and 10% serum (Fig. 1A, bottom panel). This change in viability is not sufficient to account for the low levels of IL-2R␣ expression and S phase entry in T cell populations exposed to anti-CD3 and 0.1% serum. We also found that the serum requirement for DNA synthesis and IL-2R␣ expression could not be overridden by high levels of IL-2 or by antibody to CD28 (Fig. 1A, top panel, and data not shown). These results suggest that these signals are not rate-limiting in cells in serum-deficient medium. Although cells receiving anti-CD3 and 0.1% serum did not initiate DNA synthesis, they did exhibit increased expression of the early T cell activation markers, CD69 and CD28 (Fig. 1B). This finding is consistent with our earlier report showing that TCR-stimulated T cells partially traverse G 0 /G 1 in serum-deficient medium (43).
To determine if serum concentration affected total (as well as surface-localized) levels of IL-2R␣, ELISAs and Western blots were performed on whole cell extracts. To expedite these studies, unfractionated splenocyte populations were used. In our previous study, which measured the effects of serum concentration on several cell cycle-related parameters, splenocytes and purified T cells exhibited identical responses (43). As shown in Fig. 2, A and B (left panel), IL-2R␣ was barely detectable in resting splenocytes and was weakly induced by ConA at 10 h regardless of serum concentration. After this time, IL-2R␣ continued to accumulate in cells co-stimulated with ConA and 10% serum, with peak expression occurring at 30 -40 h. In contrast, IL-2R␣ essentially disappeared from cells receiving ConA and 0.1% serum. The capacity of serum to enhance IL-2R␣ expression was also evident in splenocytes stimulated with anti-CD3 (Fig. 2B, right panel). These findings indicate that serum increases the cell surface expression of IL-2R␣, at least in part, by increasing total cellular levels of IL-2R␣. On the other hand, serum did not affect the expression of IL-2R␤ or IL-2R␥ (Fig. 2, B and C).
To assess the functional consequences of serum-dependent IL-2R␣ up-regulation, we assayed the activity of the IL-2 signaling intermediate, Jak3, in splenocytes treated with ConA and either 10% or 0.1% serum. To detect active Jak3, Jak3 immunoprecipitates were immunoblotted with antibody to phosphotyrosine. As presented in Fig. 3A, increases in IL-2R␣ expression were paralleled by increases in Jak3 phosphorylation. Quiescent cells contained little if any IL-2R␣ or phosphorylated Jak3. Cells receiving ConA and 0.1% serum exhibited minor increases in IL-2R␣ expression and Jak3 phosphorylation, whereas both responses were substantially elevated in cells exposed to ConA and 10% serum. Serum (at 10%) also increased IL-2R␣ levels and activated Jak3 when added to cells 20 h after addition of ConA and 0.1% serum. This result shows that cells incubated with ConA in serum-deficient medium remain viable and retain the capacity to initiate IL-2-dependent events when subsequently exposed to 10% serum.
To determine if serum regulated IL-2R␣ mRNA expression, RNase protection assays were done on splenocytes stimulated with ConA and 10% versus 0.1% serum. As shown in Fig. 3B, IL-2R␣ mRNA levels rose within 5 h of addition of ConA to cells and remained elevated for up to 25 h regardless of serum concentration. This finding demonstrates that maximal increases in IL-2R␣ mRNA levels are not sufficient for maximal expression of IL-2R␣ protein and, consequently, that serum Results are expressed as absorbance at 550 nm Ϯ the standard deviation. B and C, resting T cells were stimulated with 5 g/ml anti-CD3 and either 10% or 0.1% serum for 30 h. The cell surface expression of CD69, CD28, IL-2R␣ (CD25), and IL-2R␥ (CD132) was determined by FACS analysis using the appropriate antibodies. B, data are expressed as the percentage of positive cells. C, the histogram for IL-2R␣ is shown. Background fluorescence (left peak) was determined using an isotype control antibody, and specific fluorescence (right peak) was determined using an IL-2R␣ antibody.
cdk2 Activation and Post-transcriptional IL-2R␣ Expression controls IL-2R␣ expression at a post-transcriptional level. Similar to protein levels, mRNA levels of IL-2R␤ and IL-2R␥ were unaffected by serum concentration.
Serum-dependent IL-2R␣ Expression Requires cdk2 Activity-The data presented above show that T cells do not appreciably express IL-2R␣, activate JAKs, or enter S phase when stimulated with medium containing ConA or anti-CD3 and 0.1% serum. On the other hand, we have found that splenocytes derived from p27 Kip1 -null C57b1/6 mice are capable of initiating DNA synthesis when exposed to a TCR agonist and either 10% or 0.1% serum ( Fig. 4 and accompanying article (43)). This finding implies that cells lacking p27 Kip1 either efficiently express IL-2R␣ in serum-deficient medium or no longer require IL-2 signaling for proliferation. To distinguish between these alternatives, we examined IL-2R␣ expression and Jak1 and Jak3 activity in p27 Ϫ/Ϫ splenocytes exposed to ConA and either 10% or 0.1% serum. IL-2R␣ was not detectable in quiescent p27 Ϫ/Ϫ splenocytes but was present at high levels in ConAtreated p27 Ϫ/Ϫ splenocytes regardless of serum concentration (Fig. 4). Moreover, both Jak1 and Jak3 were active in p27 Ϫ/Ϫ splenocytes receiving ConA and either 10% or 0.1% serum. Similar to Balb/c splenocytes, wild-type C57b1/6 splenocytes required 10% serum for both IL-2R␣ accumulation and JAK activation. These data indicate that the capacity of p27 Ϫ/Ϫ splenocytes to proliferate in serum-deficient medium results (at least in part) from the capacity of these cells to optimally express IL-2R␣ in a serum-independent manner.
In wild-type splenocytes, cdk2 activation requires a pronounced and persistent decrease in p27 Kip1 levels, and both ConA and 10% serum are needed to achieve this effect (43). In contrast, in p27 Kip1 -null splenocytes, cyclin E-cdk2 activation is constitutive and cyclin A-cdk2 activation is serum-(although not ConA-) independent (26,31,43). Thus, conditions that promote cdk2 activation are the same as those that promote IL-2R␣ expression, and it is possible, therefore, that cdk2 activity contributes to IL-2R␣ expression. In support of this hypothesis, we found that IL-2R␣ did not accumulate in Balb/c splenocytes treated with ConA and 10% serum in the presence of roscovitine, a potent and selective inhibitor of cdk2 activity (Fig. 5A and Ref. 32). Roscovitine did not inhibit the expression of cyclin E and thus does not nonspecifically block protein synthesis. Similar to naive splenocytes, roscovitine also precluded IL-2R␣ expression, as well as cdk2 activity, in quiescent T lymphoblasts restimulated with ConA and 10% serum (Fig. 5B).
The necessity of cdk2 activity for IL-2R␣ accumulation was further demonstrable in experiments in which IL-2R␣ was transiently expressed in a fibroblast cell line (termed p27-47) that inducibly expresses p27 Kip1 in response to isopropyl-␤-Dthiogalactopyranoside (IPTG) (Fig. 5C and Ref. 28). As we reported previously, induction of p27 Kip1 in sparse p27-47 cells represses cdk4 and cdk2 activity but does not result in growth inhibition (28). As shown in Fig. 5C, IL-2R␣ mRNA and protein were apparent in p27-47 cells transfected with a plasmid containing IL-2R␣ cDNA but not with vector alone. Although IL-2R␣ mRNA levels were approximately equal in cells treated with or without IPTG, IL-2R␣ protein levels were substantially lower in IPTG-treated as compared with untreated cultures. On the other hand, levels of ectopically expressed ␤-galactosidase were similar in both IPTG-treated and untreated cultures, thus indicating that p27 Kip1 overexpression does not globally inhibit protein expression. Collectively, the data in Fig. 5 show that cdk2 activity is required for IL-2R␣ expression. Our studies, therefore, establish a series of events in which serum facilitates cdk2 activation, which in turn modulates IL-2R␣ expression at a post-transcriptional level.
cdk2 Activity and IL-2 Signaling Comprise a Regulatory Loop-Previous studies have shown that IL-2 stimulates cyclin E-cdk2 and cyclin A-cdk2 activity in activated T cells (33)(34)(35). Because these investigations place cdk2 activation downstream of IL-2R activation, the need for cdk2 activity for IL-2R␣ accumulation seems paradoxical. It is possible, however, that these processes are interdependent; i.e. cdk2 activity enhances IL-2R␣ expression and consequent IL-2 signaling promotes cdk2 activation. To assess the dependence of cdk2 activation on IL-2 signaling in our system, we assayed cdk2 activity in splenocytes treated with ConA and 10% serum in the presence or absence of an IL-2R␣ blocking antibody. As shown in Fig. 6A, cdk2 activity was substantially lower in antibodytreated cultures, as was IL-2R␣ expression. AG490, a selective inhibitor of Jak activity (36), also blocked cdk2 activation and IL-2R␣ expression when presented to cells 12 h after stimulation with ConA and 10% serum (Fig. 6B). AG490 also repressed the expression of cyclin A and cdk2; we have shown previously that the expression of these proteins in T cells requires cdk2 activity (43). On the other hand, AG490 had no effect on cyclin FIG. 3. Effect of serum on IL2R␣, IL-2R␤, and IL-2R␥ mRNA levels and JAK activity in splenocytes. A, splenocytes were treated continuously with 2.5 g/ml ConA and either 10% or 0.1% serum or were exposed to ConA and 0.1% serum for 20 h before receiving 10% serum for an additional 20 h (indicated by ϩ). Cell extracts were immunoblotted with antibody to IL-2R␣ or immunoprecipitated with antibody to Jak3 and immunoblotted with antibody to Jak3 or phosphotyrosine (Jak3-P). B, quiescent splenocytes received 2.5 g/ml ConA and either 0.1% or 10% serum for the indicated times. mRNA levels of the IL-2R components and of L3T4 and GAPDH (loading controls) were determined by RNase protection assay.
FIG. 4. Serum-independent IL-2R␣ expression and JAK activation in p27 Kip1 -deficient splenocytes. Quiescent p27 ϩ/ϩ and p27 Ϫ/Ϫ splenocytes were treated with 2.5 g/ml ConA and either 0.1% or 10% serum for 36 h. Cell extracts were immunoblotted with antibody to IL-2R␣ or immunoprecipitated with antibody to Jak1 or Jak3 and immunoblotted with antibody to phosphotyrosine (Jak1-P, Jak3-P). To ensure that all samples contained equal amounts of Jak1 or Jak3, immunoprecipitates were also immunoblotted with antibodies to these proteins. The percentage of S phase cells was determined by FACS analysis of propidium iodide-stained cells.
cdk2 Activation and Post-transcriptional IL-2R␣ Expression E levels and thus does not inhibit protein expression in general. Together, the above findings show that cdk2 activation both results from and contributes to IL-2R␣ expression and that cdk2 activity, IL-2R␣ expression, and IL-2 signaling comprise a regulatory loop.

DISCUSSION
Engagement of the TCR results in the transcriptional upregulation of IL-2R␣, and numerous studies have focused on the pathways and promoter elements involved in this response (5). Data presented here show that IL-2R␣ expression is also regulated post-transcriptionally and that serum and cdk2 play prominent roles in this process. In splenocytes exposed to ConA and 0.1% serum, IL-2R␣ was only weakly and transiently expressed. In contrast, in splenocytes receiving ConA and 10% serum, IL-2R␣ expression was robust and sustained. Serum also elevated the cell surface expression of IL-2R␣ in purified T cells, and IL-2R␣ accumulation was accompanied by activation of the IL-2 signaling intermediates, Jak1 and Jak3. On the other hand, serum had no effect on IL-2R␣ mRNA levels or on the expression of IL-2R␤ or IL-2R␥ at either the protein or message level. These findings clearly show that serum selec-tively modulates the expression of the ␣ subunit of the IL-2R at a post-transcriptional level and that serum-induced increases in IL-2R␣ levels are biologically relevant.
The requirement for cdk2 activity for IL-2R␣ accumulation was established using two different experimental approaches. First, endogenous IL-2R␣ levels were determined in T cells stimulated in the presence of the pharmacological cdk2 inhibitor, roscovitine. We found that roscovitine abolished the capacity of serum to increase IL-2R␣ levels in both ConA-treated splenocytes and ConA-treated T lymphoblasts. Second, levels of ectopically expressed IL-2R␣ were measured in p27-47 fibroblasts treated with or without IPTG. As reported previously, these cells are stably transfected with a plasmid that expresses p27 Kip1 under the control of the lac repressor (28). When exposed to IPTG, which alleviates lac repression, these cells produce high amounts of p27 Kip1 , and consequently, cdk2 activity is inhibited. As shown here, IPTG-induced cdk2 inactivation also markedly repressed the expression of IL-2R␣ protein. In contrast, IL-2R␣ mRNA levels were similar in IPTG-treated and untreated cells. These findings complement those obtained in the roscovitine experiments and show that cdk2 activity, like serum, modulates IL-2R␣ expression at a post-transcriptional level.
Because IL-2 signaling is thought to be a cause rather than an effect of cdk2 activation in T cells (24,25), we considered the possibility that these processes were interdependent. In support of this hypothesis, we found that abrogation of IL-2 signaling repressed both cdk2 expression and cdk2 activity in splenocytes stimulated with ConA and 10% serum. Two methods were used to inhibit IL-2 signaling: an IL-2R␣ blocking antibody and the JAK inhibitor, AG490. On the basis of these data and those discussed above, we propose the following regulatory loop (Fig. 7). We suggest that this loop begins with cyclin E-cdk2 activation, which results from persistent serum-dependent p27 Kip1 downregulation. As described by others (37)(38)(39), cyclin E-cdk2 complexes, together with cyclin D-containing complexes, phosphorylate Rb, and thus allow the E2F-mediated transcription of cyclin A and the consequent formation of cyclin A-cdk2 complexes. As shown here, activation of cyclin E-cdk2 also leads to the post-transcriptional accumulation of IL-2R␣ and FIG. 5. Inhibition of IL-2R␣ accumulation by roscovitine and p27 Kip1 overexpression. A, resting Balb/c splenocytes were treated for 12 h with 2.5 g/ml ConA and 10% serum prior to addition of either DMSO (vehicle control) or roscovitine (final concentration, 25 M). Cells were harvested at the indicated times. IL-2R␣ and cyclin E levels were determined by Western blotting. B, to prepare T lymphoblasts, splenocytes were treated with 5 g/ml ConA and 10% serum for 48 h and 10% serum alone for an additional 48 h. Quiescent lymphoblasts then received 2.5 g/ml Con A and 10% serum and either Me 2 SO (DMSO) or 25 M roscovitine for the indicated times. IL-2R␣ expression was determined by Western blotting, and cdk2 activity (A (H1)) was determined in cyclin A immune complexes by in vitro kinase assay. C, exponentially growing p27-47 fibroblasts treated with or without 1 mM IPTG for 20 h were transfected with a vector alone (pIRES2-EGFP) or vector containing IL-2R␣ cDNA by LipofectAMINE (Life Technologies). Both sets of cells were cotransfected with ␤-galactosidase (␤-gal) under control of the cytomegalovirus promoter. After transfection, cells were incubated with or without IPTG for an additional 20 h. Transfection efficiency was monitored by green fluorescence protein expression using flow cytometry. IL-2R␣ protein and mRNA levels were determined by Western blotting and Northern blotting, respectively. Equal RNA loading was ascertained by hybridization of the membrane with ␤-actin cDNA probe. cdk2 activity ((A) H1) was assessed as in B. Data showing ␤-galactosidase protein levels in mock-transfected cells and ␤-galactosidase-transfected cells are also presented.
FIG. 6. Inhibition of cdk2 activity by an IL-2R␣ blocking antibody and AG490. A, resting Balb/c splenocytes were treated with 2.5 g/ml ConA and 10% serum in the presence or absence of an IL-2R␣ blocking antibody (7.5 g/ml) for the indicated times. Cell extracts were immunoblotted with antibody to IL-2R␣ or immunoprecipitated with antibody to cyclin E for determination of cdk2 activity (H1 (E)) by in vitro kinase assay. B, quiescent splenocytes received 2.5 g/ml ConA and 10% serum for 12 h, followed by Me 2 SO (DMSO) (vehicle control) or 50 M AG490 for 24 h. Levels of the indicated proteins were determined by Western blotting. cdk2 activity (H1 (A)) was measured in cyclin A immunoprecipitates by in vitro kinase assay.
cdk2 Activation and Post-transcriptional IL-2R␣ Expression the subsequent induction of IL-2-mediated events, as exemplified by JAK activation. IL-2 signaling pathways, in turn, further optimize and sustain cyclin E-cdk2 and cyclin A-cdk2 activity. At this point in the cycle, it is likely that both of these activities promote the continued expression of IL-2R␣.
The capacity of serum to facilitate splenocyte proliferation by initiating the regulatory loop outlined in Fig. 7 is clearly demonstrated by our studies on p27 Kip1 -deficient splenocytes. In these cells, cdk2 activation, IL-2R␣ accumulation, and JAK activation were serum-independent, as was S phase entry ( Fig.  4 and accompanying article (43)). In p27 Ϫ/Ϫ cells, ConA was still required for transcriptional up-regulation of IL-2R␣ and for activation of D cyclin complexes and consequent expression of cyclin A (43). It is possible that the more proximal target of serum is IL-2R␣ expression (and consequent IL-2 signaling) rather than p27 Kip1 down-regulation (and consequent cdk2 activation). However, in support of our model, it is noted that submaximal increases in cdk2 activity have been observed in T cells treated with TCR agonists (and 10% serum) in conditions in which IL-2 is not produced (40,41). Thus, we propose that ConA activates cyclin E-cdk2 to a small extent in a serum-dependent but IL-2-independent manner and that this limited increase in activity is sufficient to set the cycle in motion.
Although translational regulation of IL-2R␣ has been reported previously (42), our studies at present do not distinguish between an effect of cdk2 activity on the synthesis versus the stability of IL-2R␣. Our transfection data indicate that the information required for maximal IL-2R␣ expression is contained within the IL-2R␣ coding region and a 123-bp upstream sequence. As described previously (42), this upstream sequence contains a translational start codon and an in-frame stop codon that are thought to reduce translational efficiency by causing leaky scanning. Whether this process is influenced by cdk2 activity remains to be determined. As a final point, we note that many of the effects described here have also been observed in splenocytes receiving ConA and (rather than serum) a serum substitute consisting of insulin, selenium, and transferrin. 2