Cyclin D3 compensates for loss of cyclin D2 in mouse B-lymphocytes activated via the antigen receptor and CD40.

Cyclin D2 is the only D-type cyclin expressed in mature mouse B-lymphocytes, and its expression is associated with retinoblastoma protein (pRB) and pRB-related protein phosphorylation and induction of E2F activity, as B-cells enter the cell cycle following stimulation via surface IgM and/or CD40. Cyclin D-dependent kinase activity is required for cell proliferation, yet cyclin D2(-/-) mice have normal levels of mature B-lymphocytes. Here we show that B-lymphocytes from cyclin D2(-/-) mice can proliferate in response to anti-IgM and anti-CD40, but the time taken to enter S-phase is longer than for the corresponding cyclin D2(+/+) cells. This is due to the compensatory induction of cyclin D3, but not cyclin D1, which causes pRb phosphorylation on CDK4-specific sites. This is the first demonstration that loss of a D-type cyclin causes specific expression and functional compensation by another member of the family in vivo and provides a rationale for the presence of mature B-lymphocytes in cyclin D2(-/-) mice.

Cell proliferation requires successful transition through cell cycle checkpoints. One of the most important checkpoints, called the restriction point (R), 1 is located prior to the G 1 /S boundary, after which cells become insensitive to extracellular growth inhibitory signals (1). In mammalian cells, G 1 cyclins are rate-limiting for transition through R and are made up of two families of cell cycle regulatory proteins as follows: D-type cyclins (cyclin D1, D2, and D3) and cyclin E (cyclin E1 and E2) (2,3). The two types of G 1 cyclins have specificity for different cyclin-dependent kinase (CDK) subunits. The D-type cyclins bind and activate CDK4 and CDK6 preferentially, whereas cyclin E interacts predominantly with CDK2 (3,4). The expres-sion of D-type cyclins is regulated predominantly by extracellular mitogenic signals. Thus, cyclin D plays a central role in integrating the extracellular proliferative signals with the cell cycle machinery. The expression of cyclin E is regulated by an autonomous mechanism and induced in a cyclical fashion, peaking periodically at the G 1 /S boundary of the cell cycle (3,4). Cyclin E has been demonstrated to function downstream of cyclin D, and its expression can rescue the loss of cyclin D1 in cyclin D1 null mice (5). The analyses of mice with homozygous deletions of D-type cyclin genes showed only limited phenotypic defects. Mice lacking cyclin D1 develop normally, apart from certain cells in the retina and breast epithelium, where cyclin D1 is the predominant D-type cyclin (6,7). Similarly, disruption of the cyclin D2 gene loci only affects particular cells in the ovaries and testes (8). This limited spectrum of abnormalities suggests that there is a high level of functional redundancy among D-type cyclins, but the molecular mechanism for this functional compensation is unclear. It is possible that the usual functions of an individual D-type cyclin are compensated by another member of the family, by another cyclin (e.g. cyclin E), or by a downstream effector.
The principal physiological substrates of the G 1 CDKs are members of the retinoblastoma protein (pRB) family of pocket proteins (pRB, p107, and p130) (9 -11). It has been demonstrated that overexpression of any one of pRB, p107, or p130 can trigger cell cycle arrest which can be overcome by G 1 cyclin-dependent kinase activity. Although both the D-type cyclins and cyclin E contribute to pRB phosphorylation in vivo and are equally effective in overcoming a pRB-induced cell cycle block, only the cyclin D⅐CDK4 complex, but not the cyclin E⅐CDK2 complex, can phosphorylate p107 and p130 in vivo, dissociate p107 and p130 containing E2F complexes, and effectively alleviate growth suppression by p107 and p130 (12)(13)(14)(15)(16). In their hypophosphorylated forms, these pocket proteins bind to members of the E2F family of transcription factors, thereby negatively regulating transcription of E2F-regulated genes. E2F activity is important for the transcription of genes required for entry into the S-phase of the cell cycle (9,(17)(18)(19). Overexpression of E2F activity can induce cells to progress from G 0 through G 1 into S-phase. Thus, G 1 cyclins mediate hyperphosphorylation of the pocket proteins, leading to derepression of E2F-dependent gene transcription and progression into S-phase.
The activities of CDKs are in turn negatively regulated by two families of CDK inhibitors (CKIs): the WAF/CIP family of proteins (p21 Cip1 , p27 Kip1 , and p57 Kip2 ), which control a broad range of CDKs by binding to the cyclin⅐CDK complexes, and the INK4 family (p15 INK4b , p16 INK4a , p18 INK4c , and p19 INK4d ), which specifically inhibit cyclin D⅐CDK complexes through di-rect association with CDK4 and CDK6 (3,4). The p27 KIP1 is the CKI most directly associated with restriction point regulation (3,4). The expression level of p27 is high in quiescent (G 0 ) cells but is down-regulated during G 1 as the cells enter the cell cycle (20,21).
In B-lymphocytes, two receptors play a central role in transducing signals that drive their activation and clonal expansion in response to T-dependent antigens: the B-cell antigen receptor (BCR) and CD40. The latter is engaged during cognate interactions with preactivated CD4 helper cells, which express the counterreceptor for CD40, CD154 (the CD40 ligand). CD40-CD154 interactions are essential for generating switched Ig isotypes, germinal centers, and B memory cells (22). B-cell activation in vivo can be mimicked by stimulating cultured mouse B-lymphocytes with antibodies to the antigen receptors (e.g. IgM) and co-receptors (e.g. CD40). We and others (23)(24)(25)(26) have shown that signals via the BCR synergize with those transduced via CD40 to drive B-cell proliferation. By using this B-cell activation system, we have demonstrated that the pRB/ E2F pathway integrates proliferative signals emanating from BCR and CD40 receptors (27). In the present study, we identify cyclin D2 as the key regulator of cell cycle progression and pocket protein phosphorylation in response to anti-IgM and/or anti-CD40 treatment in B-cells. We have also investigated the consequences of the loss of cyclin D2 on B-lymphocyte proliferation, and we show that cyclin D3 is up-regulated to compensate for the lack of cyclin D2 in cyclin D2 null B-lymphocytes to mediate cell proliferation.
Cell Proliferation Analysis-Cell proliferation was monitored by [ 3 H]thymidine incorporation assays. The cells were cultured at 10 5 cells/well in 200-l cultures in supplemented RPMI 1640 medium, and 5% FCS and [ 3 H]thymidine was added for the final 4 h of a 72-h culture period. Cells were collected using a PHD cell harvester (Cambridge Technology, Cambridge, MA), and [H 3 ]thymidine incorporation into DNA was quantified by scintillation counting.
Western Blot Analysis and Antibodies-Western blot extracts were prepared by lysing cells with 4 times packed cell volume of lysis buffer (20 mM HEPES, pH 7.9, 150 mM NaCl, 1 mM MgCl 2 , 5 mM EDTA, pH 8.0, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM NaF, 5 mM sodium orthovanadate) on ice for 20 min. Protein yield was quantified by Bio-Rad Dc protein assay kit (Bio-Rad). Fifty g of lysate was separated by SDS-polyacrylamide gel electrophoresis, transferred to nitrocellulose membranes, and recognized by specific antibodies.  (17), and p27 (C-19) were purchased from Santa Cruz Biotechnology. Anti-p130 monoclonal antibody (anti-pRB2) was acquired from Transduction Laboratories. The anti-phospho-pRB(Ser 807/811 ) antibody was purchased from New England Biolabs, and the anti-phospho-pRB (Thr 821 ) antibody was a kind gift from Dr. Yoichi Taya, Tokyo, Japan. Both the anti-DP1 and DP2 mouse monoclonal antibodies have been described previously (31) and were raised against His 6 -tagged (Qiagen) peptides corresponding to unique amino-terminal regions of the corresponding human proteins. The antibodies were detected using horseradish per-oxidase-linked goat anti-mouse or anti-rabbit IgG (Dako) or mouse absorbed goat anti-rat IgG (Southern Biotechnology Associates, Inc) and visualized by the enhanced chemiluminescent detection system (Amersham Pharmacia Biotech).

RESULTS
Expression of Cyclins, CDKs, and CKIs following Anti-IgM and/or Anti-CD40 Treatment-Our previous results suggested that the pRB/E2F pathway integrates proliferative signals emanating from the sIgM and CD40 receptors (27). By having defined the molecular events that regulate E2F activity in response to BCR and CD40 signaling, we then proceeded to investigate the upstream signaling molecules that mediated these changes in the phosphorylation state of the pRB family and E2F activity. Since it is known that cyclins and CDKs as well as the CKIs integrate positive and negative proliferative signals and are responsible for controlling pocket protein phosphorylation (3, 4), we therefore studied the expression patterns of these in B-lymphocytes stimulated with anti-IgM and/or anti-CD40 ( Fig. 1). Western blot analysis results showed that, as expected, no G 1 cyclin was detected in unstimulated Blymphocytes. Upon anti-IgM and/or anti-CD40 stimulation, cyclin D2 was first detected at 24 h, concomitant with the initial hyperphosphorylation of pRB and the related pocket proteins p107 and p130 (27). CDK4 and CDK6, the specific kinase partners for cyclin D2, were present in unstimulated cells, and their expression levels were elevated gradually as the cells progressed toward S-phase. Neither cyclin D1 (data not shown) nor D3 expression was detectable in activated B-cells; hence, these cyclins are not involved in mediating the phosphorylation of the pRB-related pocket proteins and cell cycle progression in this cell type. Both cyclin A and E and their kinase subunits, CDC2 and CDK2, were not expressed in resting B-lymphocytes, but all four proteins were up-regulated at 48 h after treatment, approximately 24 h after the induction of cyclin D2.
Cyclin-CDK activity is negatively regulated by various CKIs (3,4). Western blot analysis detected significant levels of p27 in resting B-lymphocytes. The expression of the p27 was downregulated following stimulation with anti-CD40 and/or anti-IgM. Interestingly, the down-regulation of p27 Kip1 after receptor stimulation closely trailed the up-regulation of cyclin D2 expression. This suggests that the cyclin D2 could potentially influence the expression of p27 Kip1 in response to anti-CD40 and/or anti-IgM stimulation. However, it is also possible that p27 Kip1 is down-regulated as the cells enter the cell cycle by some other mechanisms independent of cyclin D2. Extracts prepared from small dense B-cells after anti-CD40, anti-IgM, or anti-IgM ϩ anti-CD40 stimulation for 0, 6, 12, 24, 48, and 72 h were separated on SDS-polyacrylamide gels. Following transfer onto nitrocellulose membrane, proteins were detected by antibodies against p27 KIP1 , CDC2, CDK2, -4, and -6 and cyclin D2, D3, E, and A. As a positive control for cyclin D3 expression, an extract from the T-lymphoma cell line EL4 was included in the Western blot.
Accumulation of Cyclin D2 Coincides with Induction of CDK4/6 Activity, Pocket Protein Phosphorylation, and Precedes Cell Cycle Entry-pRB is the key physiological substrate for cyclin/CDKs, and specific sites are phosphorylated in vivo by distinct G 1 /S cyclin/CDK inhibitors, CKIs (3,4). Therefore, we used antibodies that specifically recognize CDK4/6 or CDK2-phosphorylated residues on pRB to monitor the in vivo D-and E-type cyclin-dependent kinase activity in resting primary B-lymphocytes stimulated with anti-IgM and/or anti-CD40. This method has considerable advantages over in vitro kinase assays, particularly for studying CDK4/6 activity. First, the results obtained with in vitro kinase assays are frequently inconsistent, mainly due to phosphorylation by contaminating kinases and changes of CDK4/6 activities under the conditions used. Moreover, the immunoprecipitation procedures may also alter the kinase activities of the complexes. In contrast, using the CDK-specific anti-pRB antibodies permits the monitoring of the actual in vivo kinase activity of the cyclin and dependent kinase complexes in its physiological environment. We used a pRB phospho-specific antibody, anti-pRB(Ser 807/811 ), which recognizes the CDK4/6 kinase-phosphorylated residues Ser 807 and Ser 811 (32). Fig. 2 shows that pRB is phosphorylated at Ser 807/811 at 24 h, concomitant with the first detectable sign of pocket protein phosphorylation in mid-G 1 (27). In contrast, another phospho-pRB antibody (anti-pRB(Thr 812 )), recognizing pRB only when phosphorylated on Thr 812 by CDK2 (32), detected hyperphosphorylated forms of pRB which only appeared later in the time course (48 h post-stimulation), corresponding to late G 1 and S. The accumulation of this CDK2-phosphorylated forms of pRB coincided with the expression of cyclin E and A and their preferred kinase subunits, CDK2 and CDC2. This further confirms that both cyclin E-and A-dependent kinase activity are activated at late G 1 /S, presumably as a result of up-regulation of the expression of their respective cyclin and kinase subunits. Since cyclin D2 is the first G 1 cyclin and the only D-type cyclin detected in activated normal B-cells, our finding that its expression level parallels the accumulation of the in vivo CDK4/6 activity and pocket protein hyperphosphorylation strongly suggested that cyclin D2-CDK4/6 mediates the initial and continuous hyperphosphorylation of pRB, and perhaps also p107 and p130 during G 1 . In addition, since CDK4 and 6 are constitutively expressed in B-lymphocytes, these results also indicate that cyclin D2 protein expression could be rate-limiting for cyclin D-CDK4/6 activity and, therefore, the initiation of pocket protein phosphorylation.
Delayed Entry into S-phase in Cyclin D2 Null B-lymphocytes following Anti-IgM and/or Anti-CD40 Stimulation-We next evaluated the role of cyclin D2 in B-lymphocyte proliferation induced by anti-IgM and/or anti-CD40, or LPS. The results showed that cyclin D2 Ϫ/Ϫ B-lymphocytes respond as well as wild-type cells to LPS, indicating that cyclin D2 is not essential for mediating proliferation to this mitogen (Fig. 3). In striking contrast, cyclin D2 Ϫ/Ϫ B-cells gave significantly lower proliferative responses to anti-IgM in the presence or absence of anti-CD40. The [ 3 H]thymidine uptake of costimulated cyclin D2 Ϫ/Ϫ cells was less than 30% of the normal controls. The responses of wild-type cells reached a plateau after 48 h, whereas the proliferative response of cyclin D2 Ϫ/Ϫ B-cells continued to increase between 48 and 72 h. Hence, the lack of this cyclin causes a delay in entry into S-phase in response to stimuli derived from CD40 and the BCR, indicating that cyclin D2 plays a key role in mediating normal proliferative responses following ligation of these receptors. Cells from either cyclin D2 Ϫ/Ϫ or normal mice stimulated with anti-CD40 alone gave only low levels of proliferation, as expected.
Cyclin D2 Is Responsible for the Initial Phosphorylation of Pocket Protein and the Induction of E2F Activity following Anti-IgM and/or Anti-CD40 Stimulation-In order to examine the role of cyclin D2 in mediating pocket protein hyperphosphorylation and up-regulation of E2F activity, we compared the expression levels and phosphorylation status of pocket proteins and E2F-regulated gene products in both wild-type and cyclin D2 null B-lymphocytes. Western blot analysis showed that in both unstimulated normal and cyclin D2 Ϫ/Ϫ cells, all three pocket proteins, pRB, p107, and p130, were present in their respective hypophosphorylated forms (Fig. 4). Although the hypophosphorylated pRB and p107 were detected at comparable levels in resting normal and cyclin D2 Ϫ/Ϫ B-lymphocytes, the expression level of p130 in cycling cyclin D2 Ϫ/Ϫ cells was noticeably higher than in their wild-type counterparts. Moreover, whereas IgM and CD40 stimulation induced pRB, p107, and p130 hyperphosphorylation in wild-type B-lymphocytes at 48 h, the majority of the pocket proteins were in their respective hypophosphorylated forms in the cyclin D2 Ϫ/Ϫ cells. This suggests that cyclin D2 induction is indeed essential for pocket protein phosphorylation. Furthermore, after 24 h of anti-IgM and/or CD40 stimulation, the levels of pocket proteins, most evidently p130, were down-regulated in wild-type B-lymphocytes, but this was not observed in cells from cyclin D2 Ϫ/Ϫ mice. The reduction in p130 level observed in normal B-lymphocytes has previously been shown to be associated with the hyperphosphorylation of the protein (33,34). It is notable that incubation of B-lymphocytes for 48 h with anti-CD40 alone resulted in substantial levels of cell death, particularly in cyclin D2 Ϫ/Ϫ cells (data not shown). This could account for the low levels of proteins detected in these samples.
In normal B-lymphocytes, E2F1, pRB, and p107 expression was up-regulated at 48 h following anti-IgM and/or anti-CD40 treatment, which corresponded with cells entering late G 1 , and coincided with hyperphosphorylation of p130 and other pocket proteins (Fig. 5). This observation is in agreement with previous findings that hyperphosphorylation of p130 in mid-G 1 plays a major role in relieving E2F-mediated repression of G 1 /S genes, including E2F1, pRB, and p107 (18,(35)(36)(37). In contrast, the expression levels of E2F1, p107, and pRB were low at 48 h post-stimulation in the cyclin D2 Ϫ/Ϫ cells. This further reinforces the concept that cyclin D2 plays an important role in Fig.  1 were subjected to Western blotting with anti-phospho-pRB antibodies specific for CDK4/6 or CDK2 phosphorylation. The anti-phospho-pRB (Ser 807/811 ) antibody recognizes pRB phosphorylated by CDK4/6 kinase on Ser 807 or Ser 811 , whereas the anti-phospho-pRB (Thr 821 ) antibody detects pRB phosphorylated by CDK2 kinase on Thr 821 . initiating pocket protein hyperphosphorylation and the subsequent induction of E2F activity in response to BCR and CD40 signaling. However, the detection of low levels of pocket protein phosphorylation and E2F-regulated gene expression in cyclin D2 Ϫ/Ϫ B-lymphocytes after anti-IgM and anti-CD40 stimulation also implied that other G 1 /S cyclins could partially compensate for the absence of cyclin D2 to induce pocket protein phosphorylation and, thereby, S-phase entry. To investigate this, we again monitored the in vivo activity of cyclin-dependent kinase activity in cyclin D2 Ϫ/Ϫ and normal B-lymphocytes using CDK4/6 and CDK2 phospho-specific pRB antibodies. The timing of phosphorylation at Ser 807/811 in pRB in wild-type cells coincided with cyclin D2 expression and the initial appearance of the hyperphosphorylated forms of pRB. Since cyclin D2 is the only G 1 cyclin present in these cells and its up-regulation coincides with the phosphorylation at Ser 807/811 in pRB, we reasoned that cyclin D2 expression is rate-limiting for CDK4/ 6-dependent kinase activity. Surprisingly, we also detected CDK4/6-dependent kinase activity in cyclin D2 Ϫ/Ϫ B-lymphocytes after 48 h of stimulation with anti-IgM or anti-IgM ϩ anti-CD40. This indicates that the expression of alternative D-type cyclin(s) at 48 h must (partially) substitute for the loss of cyclin D2. Low levels of CDK2-dependent kinase activity were also observed at 48 h. These findings led us to examine the cyclin D2 Ϫ/Ϫ B-lymphocytes for other D-type cyclins.

FIG. 2. Expression of phospho-specific pRb following anti-IgM and/or anti-CD40 stimulation. Extracts prepared as detailed in
Deregulation of Cyclin D3 and p27 Expression in Cyclin D2 Null B-lymphocytes-Consistent with earlier results, in normal B-lymphocytes cyclin D2 expression was first observed after 24 h of anti-IgM and/or anti-CD40 treatment (i.e. in mid-G 1 ), whereas cyclin D3 was not detectable. However, in cyclin D2 null B-lymphocytes we detected up-regulation of cyclin D3 protein after 48 h of anti-IgM or anti-IgM ϩ anti-CD40 treatment (Fig. 6). This could therefore explain the low levels of pocket protein phosphorylation and E2F-dependent gene product expression, and the induction of CDK4/6-dependent kinase activity detected in cyclin D2 Ϫ/Ϫ lymphocytes after 48 h of stimulation. The delayed expression of cyclin D3 and thus CDK4/6-dependent kinase activity presumably accounts for the fact that these cells take longer to enter S-phase. The Western blot analysis results also showed that p27 Kip1 was only down-regulated in wild-type but not in cyclin D2 Ϫ/Ϫ lymphocytes following anti-IgM and/or CD40 treatment, suggesting that cyclin D2 has a role in regulating p27 Kip1 expression. To investigate the expression patterns of D-type cyclins after prolonged anti-IgM and/or anti-CD40 stimulation, we extended our experiment to 72 h. Surprisingly, the extended time course (Fig. 6, lower  panel) showed cyclin D3 was detectable in wild-type lymphocytes at 72 h after stimulation with anti-IgM and anti-IgM ϩ anti-CD40 treatment, although its level was lower than that in the parallel cyclin D2 Ϫ/Ϫ cells with the same treatment. Again, we failed to detect the expression of either cyclin D2 or D3 in the anti-CD40-stimulated cells. However, it is also notable that the anti-cyclin D3 antibody detected a protein with slightly lower molecular weight than full-length cyclin D3 following 72 h of anti-CD40. In these samples we observed a high level of apoptosis (data not shown), which could be the cause of cyclin D3 degradation detected on the blot. This could again help to explain our failure to detect either cyclin D2 or D3 after treatment of anti-CD40 alone.
To obtain further evidence that cyclin D3 compensates for the loss of cyclin D2 in cyclin D2 Ϫ/Ϫ B-lymphocytes and to investigate the reason underlying the observation that cyclin D2 Ϫ/Ϫ B-cells responded normally to LPS, we treated both normal and cyclin D2 Ϫ/Ϫ lymphocytes with LPS and monitored the expression of cyclin D2 and cyclin D3 in these cells. The results demonstrate that cyclin D3 was induced in cyclin D2 null B-lymphocytes with kinetics similar to that of cyclin D2 and D3 in normal B-cells after LPS stimulation, suggesting that LPS mediates a normal proliferative response in cyclin D2 Ϫ/Ϫ lymphocytes through up-regulation of cyclin D3. These results also suggest that while LPS induces proliferation through both cyclin D2 and D3, the BCR and/or CD40 signals specially target cyclin D2. DISCUSSION In this study we have continued our investigations into the mechanisms whereby signals generated by ligation of the BCR and/or CD40 on B-cells drive B-cell proliferation. Here we have focused on the role of the D-type cyclins⅐CDK complexes in FIG. 4. Altered expression of pRB, p107, p130, and E2F1 in cyclin D2 ؊/؊ B-cells following anti-IgM and/or anti-CD40 stimulation. B-cells were cultured with anti-IgM, anti-IgM ϩ anti-CD40, or anti-CD40 for 0, 24, and 48 h, respectively. Extracts prepared as described were used for Western blotting with antibodies against pRB, p107, p130, and E2F1. wt, wild type. regulating pocket protein phosphorylation in mouse B-cells stimulated via the BCR and/or CD40. We show that normal B-cells activated in this way only express cyclin D2 and not cyclin D1 or cyclin D3 (Fig. 1). This is consistent with previous findings that cyclin D1 is not expressed in lymphoid lineages (38,39). At odds with some previous reports (40), we were unable to detect the expression of cyclin D3 in wild-type Blymphocytes, although its expression in proliferating cyclin D2 Ϫ/Ϫ B-cells and the EL4 T-lymphoma cells (30) is evident. The basis for this discrepancy is unclear. Nevertheless, the expression pattern of cyclin D3 in cyclin D2 Ϫ/Ϫ B-cells is consistent with the CDK4/6-dependent phosphorylation of pRB and the kinetics of S-phase entry in these cells. Our Western blot analyses showed that the kinetics of expression of cyclin D2 coincided with pocket protein phosphorylation and the induction of E2F activity (Figs. 1 and 2). This therefore suggests that cyclin D2 expression is responsible for the initial and the continuous phosphorylation of pRB, p107, and p130 after anti-IgM and/or anti-CD40 treatment. Since the catalytic partners of cyclin D2, CDK4 and CDK6, are expressed constitutively in B-lymphocytes, it is likely that cyclin D2 expression is ratelimiting for cyclin D-dependent kinase activity in these cells. The late up-regulation of cyclin E and A and their catalytic subunits, CDC2 and CDK2, also indicates that they are unlikely to be involved in the initial phosphorylation of pocket proteins. These results are consistent with the concept that the initial cyclin D/CDK4-or cyclin D/CDK6-induced pRB phosphorylation induces E2F-dependent expression of cyclin E which, in turn, mediates further pRB phosphorylation, culminating in entry into S-phase (18).
In order to verify that cyclin D2 is responsible for pocket protein phosphorylation, we used CDK4/6-and CDK2-specific anti-pRB antibodies to gauge the in vivo G 1 /S cyclin-dependent kinase activity; this approach enables the in vivo phosphorylation status of pRB to be investigated and indirectly identified the CDK and thus cyclin involved (Fig. 2). Since cyclin D2 is the only D-type cyclin expressed in wild-type B-lymphocytes after anti-IgM and/or anti-CD40 treatment, the finding that D-type cyclin activity detected by the CDK4/6-specific anti-pRB antibodies correlated with the expression pattern of cyclin D2 confirmed that cyclin D2 expression is responsible for the initial phosphorylation of pocket proteins. These results led us to examine the role of cyclin D2 in regulation of pocket protein phosphorylation, E2F activity, and cell cycle entry following BCR stimulation.
We have shown elsewhere that B-cells from mice lacking cyclin D2 proliferate poorly when stimulated via the BCR with monoclonal antibodies, while responding normally to LPS. 2 Here we have extended the previous study, and we demonstrated that the kinetics of cell cycle entry in cyclin D2 Ϫ/Ϫ B-lymphocytes activated by anti-IgM and anti-IgM plus anti-CD40 is significantly slower (Fig. 3). In contrast, LPS-stimulated cyclin D2 Ϫ/Ϫ cells entered S-phase at a normal tempo, indicating that cyclin D2 is not indispensable for B-cell proliferation. Rather, cyclin D2 is apparently a specific mediator of proliferative signals emanating from the BCR and/or CD40 in normal B-lymphocytes. In addition, the kinetics of pocket protein phosphorylation, and the expression patterns of E2F-regulated genes, including E2F1, p107, pRB, are all deregulated in cyclin D2 Ϫ/Ϫ B-lymphocytes following anti-IgM and/or anti-CD40 stimulation (Figs. 4 and 5). This reinforces the concept that cyclin D2 plays an important role in regulating pocket protein phosphorylation and consequently the expression of these E2F target genes in response to BCR stimulation (27,29).
Although our results clearly demonstrate that cyclin D2 is important for mediating proliferative signals from the BCR, anti-IgM ϩanti-CD40 did stimulate some cyclin D2 Ϫ/Ϫ cells to enter S-phase. It emerged that this was due to the induction of cyclin D3 in the cyclin D2 null cells, which occurred with slower kinetics than the induction of cyclin D2 in wild-type cells. The possibility that cyclin D3 functionally substituted in part for cyclin D2 was confirmed by the findings that the expression of cyclin D3 was associated with induction of D-type cyclin/CDK activity, as revealed by the CDK4/6-specific anti-pRB antibodies (Fig. 5). In addition, the delayed induction of cyclin D3-dependent kinase activity also resulted in a corresponding later onset of pocket protein hyperphosphorylation and E2F gene expression (Fig. 4).
Taken together, these data support the idea that cyclin D3 partially compensates for the loss of cyclin D2 in the mutant B-cells stimulated via the BCR and/or CD40, and its delayed expression culminates in a correspondingly tardier S-phase entry, compared with wild-type B-lymphocytes. The ability of cyclin D3 to compensate for the loss of cyclin D2 presumably explains why cyclin D2 Ϫ/Ϫ mice do not manifest any overt abnormalities in the development of conventional (so-called B2) B-lymphocytes. However, these animals cannot generate normal numbers of peritoneal (B1) B-cells, which indicates that this cyclin plays a more crucial role in the development and/or self-renewal of this B-cell subset. 2 Indeed, a recent report has shown that the rapid cell cycle progression of B1 B-lymphocytes is related to the unusually early expression of cyclin D2 in these cells following mitogenic stimulation (41).
Interestingly, the expression of the CKI p27 Kip1 appeared to be influenced only by the expression of cyclin D2 but not cyclin D3, since the level of p27 Kip1 is not down-regulated in cyclin D2 Ϫ/Ϫ B-cells, even in the presence of cyclin D3 (Fig. 6). Moreover, the down-regulation of p27 Kip1 is unlikely to be an event associated merely with cell cycle entry and not influenced by cyclin D2 expression, as the level of p27 Kip1 remained high in the mutant B-cells, even after entry into S-phase. The significance of this is unclear. This CKI is known to play an important role in inhibiting the CDK activity in G 0 and G 1 phases of these cells (3,4). Alternatively, it could be a vital activator of cyclin D-dependent kinases through acting as an assembly factor, a novel role being revealed in recent studies (42,43). Notably, the expression levels of CDK4 and CDK6 did not alter drastically after anti-IgM and/or anti-CD40 treatment. This further emphasizes the fact that D-type cyclin expression is rate-limiting for cyclin D-dependent kinase activity and thus S-phase entry and is essential for mediating the proliferative signals emanating from IgM and CD40.
Although the levels of p27 Kip1 appear to be inversely correlated with those of cyclin D2, the molecular basis for this is unknown. It has been shown that cyclin D-CDK4/6 activity can activate cyclin E expression via the induction of pocket protein hyperphosphorylation and up-regulation of E2F activity and that phosphorylation of p27 Kip1 by cyclin E-CDK2 results in its degradation via the ubiquitin/proteasome pathway. It is therefore possible that expression of cyclin D2 induces down-regulation of p27 Kip1 through up-regulation of cyclin E. However, even though cyclin D3 can functionally compensate for cyclin D2, our data on p27 Kip1 show that they are not completely interchangeable in the way in which they act.
Although cyclin D3 is generally not expressed in wild-type B-cells, within the 72-h anti-IgM ϩ anti-CD40 time course (Fig.  1), we do occasionally detect low levels of cyclin D3 expression in wild-type B-cells after prolonged stimulation with anti-IgM and/or anti-CD40 (Fig. 6). The expression of cyclin D3 probably reflects how rapidly one particular batch of B-cells progress through the first cell cycle. It is noteworthy that the induction of cyclin D3 occurred about 48 h after the initial induction of cyclin D2. The significance of the expression of cyclin D3 in normal B-cells after anti-IgM and/or anti-CD40 stimulation is unclear, but its delayed expression (after the majority of activated cells have crossed G 1 /S) (27) indicates that cyclin D3 may not be necessary for the normal B-cells to traverse the restriction point R in response to signaling via BCR and/or CD40.
We also showed that the induction of cyclin D3 in response to LPS treatment in cyclin D2 Ϫ/Ϫ B-cells occurred with kinetics similar to the induction of cyclin D2 and cyclin D3 in normal B-cells following LPS stimulation (Fig. 6), confirming further that cyclin D3 functionally substitutes for cyclin D2 in cyclin D2 Ϫ/Ϫ B-cells and that cyclin D2 is specifically targeted by BCR and/or CD40 signaling. Moreover, this also helps to explain why cyclin D2 Ϫ/Ϫ B-cells responded normally to LPS stimulation (Fig. 3).
In conclusion, we have shown that in normal B-cells stimulated via the BCR and/or CD40, cyclin D2 plays a critical role in mediating pocket protein hyperphosphorylation, regulation of E2F activity, and S-phase entry. However, in cyclin D2 Ϫ/Ϫ cells, stimulation via these receptors induces the appearance of cyclin D3, which can partially compensate for the absence of cyclin D2 in inducing cell cycle entry. Our analysis of cyclin D2 Ϫ/Ϫ B-lymphocytes demonstrates not only compensatory expression of cyclin D3 but also defines a compensatory functional activity of cyclin D3. To our knowledge, this is the first demonstration that loss of function of a member of the D-type cyclin family leads to specific expression and functional compensation by another member in vivo. Nevertheless, cyclin D3 does not completely compensate for the absence of cyclin D2. This is highlighted by the delay in entry into S-phase, lower levels of DNA synthesis, and absence of p27 Kip1 down-regulation in cyclin D2 Ϫ/Ϫ B-lymphocytes following BCR stimulation. This incomplete compensation might contribute to the subtle but significant abnormalities in B-cell development seen in cyclin D2 Ϫ/Ϫ mice as follows: these include significant depletion of CD5-expressing B1 B-cells and the lower levels of IgA and IgG3 immunoglobulins. 2 Since D-type cyclins have overlapping functions and can compensate functionally for one another, it is therefore likely that uncontrolled ectopic expression of one member of the cyclin D family can lead to deregulated proliferation. Indeed, constitutive overexpression of cyclin D1 has been demonstrated in centrocytic lymphoma (44,45), whereas cyclin D2 is overexpressed in chronic lymphocytic leukemia (46). Future experiments will address the molecular basis for the up-regulation of cyclin D3 in cyclin D2 null B-cells. It would also be informative to study the effects of double cyclin D2 and D3 loss on pocket protein phosphorylation, regulation of E2F activity, and cell cycle proliferation.