An important role for the retinoblastoma protein in staurosporine-induced G1 arrest in murine embryonic fibroblasts.

In this study, we investigated the molecular basis of the ability of staurosporine to induce G1 arrest in murine embryonic fibroblasts (MEFs). We used MEFs from transgenic mice lacking several negative regulators of the G1/S phase transition including cells from mice lacking p53, p21, retinoblastoma (Rb), or p16 genes. We found that p53 function was not essential for staurosporine-induced G1 arrest. In contrast, MEFs from mice lacking Rb genes showed approximately a 70% reduced capacity to arrest in the G1 phase following staurosporine treatment. In support of a role for Rb in staurosporine-induced G1 arrest, rat embryonic fibroblasts stably overexpressing cyclin D1/Cdk4(R24C) exhibited approximately a 50% reduced G1 arrest response to staurosporine. The role of Rb in determining the degree of staurosporine-induced G1 arrest did not depend on the function of the cyclin-dependent kinase inhibitors p16 or p21 because MEFs lacking either of these genes were still capable of undergoing G1 arrest following staurosporine exposure. Our studies provide evidence of an important role for the Rb protein in determining the degree of staurosporine-induced G1 arrest in the first cell cycle.

In this study, we investigated the molecular basis of the ability of staurosporine to induce G 1 arrest in murine embryonic fibroblasts (MEFs). We used MEFs from transgenic mice lacking several negative regulators of the G 1 /S phase transition including cells from mice lacking p53, p21, retinoblastoma (Rb), or p16 genes. We found that p53 function was not essential for staurosporine-induced G 1 arrest. In contrast, MEFs from mice lacking Rb genes showed approximately a 70% reduced capacity to arrest in the G 1 phase following staurosporine treatment. In support of a role for Rb in staurosporine-induced G 1 arrest, rat embryonic fibroblasts stably overexpressing cyclin D1/Cdk4 R24C exhibited approximately a 50% reduced G 1 arrest response to staurosporine. The role of Rb in determining the degree of staurosporine-induced G 1 arrest did not depend on the function of the cyclin-dependent kinase inhibitors p16 or p21 because MEFs lacking either of these genes were still capable of undergoing G 1 arrest following staurosporine exposure. Our studies provide evidence of an important role for the Rb protein in determining the degree of staurosporine-induced G 1 arrest in the first cell cycle.
Staurosporine, a broad spectrum protein kinase inhibitor, has been shown to induce a G 1 and G 2 phase arrest in normal cells, and importantly the G 1 arrest response has been shown to be selectively lost in a variety of tumor cell lines (1,2). The mechanistic basis underlying the loss of the G 1 arrest response in tumor cells is presently unresolved. In this study we focused our attention on the importance of the p53 and/or retinoblastoma (Rb) 1 tumor suppressor proteins in the mechanism of staurosporine-induced G 1 arrest. These studies were prompted by the findings that the p53 and/or Rb pathways are commonly altered in human tumor cells (3)(4)(5)(6), raising the possibility that such disruptions might account for the lack of staurosporineinduced G 1 arrest in tumor cell lines.
Interestingly, the effects of staurosporine on the cell cycle are reminiscent of DNA damage in that both induce a G 1 and G 2 arrest in normal cells, but the G 1 arrest is lost in many tumor cell lines (1,2). The p53 tumor suppressor is essential for G 1 arrest induced by ionizing radiation and other DNA damaging agents, as well as G 1 arrest following suppression of ribonucleoside triphosphate pools (3,7). In fact, loss of p53 activity by gene mutation, gene deletion, or overexpression of dominantnegative acting factors such as mutant-p53, Mdm2, or the human papillomavirus E6 protein blocks G 1 arrest following DNA damage (8 -10). Although the lack of DNA damage-induced G 1 arrest in tumor cells has been linked to p53 dysfunction, a role for p53 in staurosporine-induced G 1 arrest still remains to be established. An important mediator of p53-induced G 1 arrest is the p21 Waf1/Cip1 gene product (p21) (11,12). The p21 protein binds to and inhibits a variety of G 1 /S phase cyclin-dependent kinases (Cdk), which in turn are prevented from phosphorylating and inactivating the Rb protein (13). Indeed, studies conducted in the bladder carcinoma cell line 5637 indicated that p21 may participate in staurosporine-induced G 1 arrest (14). Definitive evidence of this possibility was, however, not obtainable in these earlier studies.
Previous observations have indicated that staurosporine promotes the hypophosphorylation of Rb in cell types susceptible to G 1 arrest (2,14,15). Hypophosphorylated Rb would retain cells in G 1 phase by virtue of its ability to repress genes that are regulated through E2F-dependent consensus elements in their promoters (16,17). When Rb function is inactivated through Rb gene mutation, gene deletion, expression of certain viral, or cellular proteins, the cells lose the capacity to G 1 arrest in response to some negative regulators of the G 1 /S phase transition, including the Cdk inhibitor protein p16 (5, 9, 18 -20). Previous studies in bladder carcinoma cells, which lack Rb and fail to undergo staurosporine-induced G 1 arrest (14), showed that ectopic expression of wild-type Rb delayed progression in the G 1 phase, and such cells were also sensitized to staurosporine-induced G 1 arrest. This study suggested a potential role for Rb in staurosporine-induced G 1 arrest, but definitive evidence of this possibility was not obtainable in these earlier studies.
In this study we investigated the mechanistic basis of staurosporine-induced G 1 arrest by using murine embryonic fibroblasts (MEFs) as a model system. Our choice of MEFs in this respect was fueled by the availability of cells from transgenic mice lacking important negative regulators of the G 1 /S phase transition. Such a model system enabled us to determine the importance of p53, p21, Rb, and p16 in the staurosporineinduced G 1 arrest mechanism in this cell type. Here we report that p53 function was not essential for staurosporine-induced G 1 arrest in MEFs. In contrast, the Rb protein played an important role in determining the degree of G 1 arrest observed in the first cell cycle following exposure to staurosporine. Supporting evidence of a role for Rb in staurosporine-induced G 1 arrest was provided by the observation that rat embryonic fibroblasts stably overexpressing cyclin D1/Cdk4 R24C exhibited a reduced G 1 arrest response to staurosporine. Staurosporine did not induce a Rb-dependent G 1 arrest in the first cell cycle by solely acting through the Cdk inhibitor proteins p16 or p21, because MEFs lacking either of these genes were still capable of undergoing G 1 arrest following staurosporine treatment. Our studies provide evidence of an important role for the Rb protein in determining the degree of staurosporine-induced G 1 arrest in the first cell cycle.

MATERIALS AND METHODS
Chemical Treatments-Staurosporine was obtained from Sigma and prepared as a 500 M stock solution in Me 2 SO, and aliquots were stored at Ϫ20°C until needed. Nocodazole (NOC) was purchased from Sigma.
Cell Culture-Early passage MEFs and rat embryonic fibroblasts were grown at 37°C in 95% air/5% CO 2 in Dulbecco's modified Eagle medium (Life Technologies, Inc.) supplemented with 10% heat-inactivated fetal bovine serum (Sigma), 50 units/ml penicillin G, 50 g/ml streptomycin (Life Technologies, Inc.). The wild type/wild type and p16 Ϫ/Ϫ MEF cells were graciously supplied by Dr. Flow Cytometry-Samples were prepared for flow cytometry essentially as described previously (21). Briefly, cells were washed with 1 ϫ phosphate-buffered saline, pH 7.4, and then fixed with ice-cold 70% ethanol. Samples were then washed with 1 ϫ phosphate-buffered saline and stained with propidium iodide 60 g/ml (Sigma) containing RNase 100 g/ml (Sigma) for 30 min at 37°C. Cell cycle analysis was performed using a Becton Dickinson flourescence-activated cell analyzer and Cell Quest version 1.2 software (Becton Dickinson Immunocytometry Systems, Mansfield, MA). For each sample at least 10,000 cells were analyzed, and quantitation of the cell cycle distribution was performed using the ModFit LT version 1.01 software (Verity Software House Inc., Topsham, ME).

Staurosporine-induced G 1 Arrest Does Not
Operate through a p53/p21-dependent Pathway-We initially investigated if staurosporine-induced G 1 arrest required the function of the p53 tumor suppressor. For this purpose we investigated the effects of p53 gene disruption on the ability of MEFs to arrest in G 1 and G 2 phases of the cell cycle following staurosporine treatment. Dose-response curves (5-50 nM STP) were performed to establish the dose of staurosporine that completely inhibited cell division within 24 h of addition to the different MEFs we studied. Choosing this dose enabled us to focus on events associated with the actions of staurosporine on the first cell cycle. The maximal growth inhibitory dose (50 nM) was then chosen to investigate cell cycle responses to staurosporine treatment (Fig. 1B). In the case of MEFs from normal littermates we found that staurosporine induced both a G 1 and a G 2 arrest, and this was evident within 24 h of staurosporine addition (Fig. 1, A and C). Comparable cell cycle responses were observed in MEFs from mice lacking p53 genes (Fig. 1, A and  C), suggesting that p53 function was not essential for staurosporine-induced G 1 or G 2 arrest.
The role of the cyclin-dependent kinase inhibitor p21, an important mediator of p53-induced G 1 arrest (11,(22)(23)(24) that can also be activated through p53-independent mechanisms (25), was also investigated. When MEFs from mice lacking p21 genes were exposed to staurosporine (50 nM) for 24 h, they underwent both a G 1 and a G 2 arrest (Fig. 1, A and C). These results suggested that p21 was similar to p53 in the respect that it also was not essential for staurosporine-induced G 1 or G 2 arrest. Thus, although it has previously been suggested that p21 participates in the mechanism by which staurosporine induces G 1 arrest (14), our studies in MEFs show that p21 is not essential for G 1 arrest in the first cell cycle of this cell type.
The Rb Protein Is an Important Mediator of Staurosporineinduced G 1 Arrest in the First Cell Cycle-Having established that the p53 and/or p21 gene products were not essential for staurosporine-induced G 1 arrest, we next turned our attention to the role of the Rb pathway in the staurosporine-induced G 1 arrest mechanism. For this purpose we conducted studies in MEFs with intact versus disrupted Rb genes. Again, doseresponse curves (5-50 nM STP) were performed to establish the dose of staurosporine that completely inhibited cell division within 24 h of addition. The maximal growth inhibitory dose (50 nM) was then chosen to investigate the cell cycle responses following addition of staurosporine (Fig. 2). In the case of MEFs from normal littermates, we found that staurosporine induced both a sustained G 1 and a G 2 arrest, and this arrest was evident within 24 h of addition (Fig. 1, A and C). In contrast, MEFs lacking Rb almost completely lost the ability to G 1 arrest in the first cell cycle following exposure to staurosporine (Fig. 2,  A and C). Cell cycle analysis of Rb Ϫ/Ϫ MEFs 24 h after a chronic exposure to 50 nM staurosporine showed that the G 1 FIG. 1. Staurosporine-induced G 1 and G 2 arrest in MEFs does not require the function of the p53 or p21 gene products. Exponentially growing MEFs with intact (wt/wt) or disrupted p53 or p21 genes were left untreated or exposed to vehicle (0.005% v/v Me 2 SO) or staurosporine for 24 or 48 h. A, cell cycle distribution of control (top) or STP-treated (50 nM, 24 h, bottom) samples. B, cell counts were determined for each sample using a Coulter counter over a 48-h time period. Samples shown were from a representative experiment that was repeated two or three times with similar results. C, the percentage of G 1 was determined by flow cytometry for each sample, and each point in C represents either the mean and S.D. from three independent experiments or data from two individual experiments. population was 14 Ϯ 3% (n ϭ 3) compared with 41 Ϯ 5% (n ϭ 3) for the vehicle-treated control cells. This amounted to approximately a 70% reduction in the degree of G 1 arrest seen in the presence of staurosporine (p Ͻ 0.01, Student's t test).
To delineate whether staurosporine-induced G 1 arrest was a transient or sustained event, we chronically exposed either wild-type or Rb Ϫ/Ϫ MEFs to staurosporine for up to 72 h and performed flow cytometric analysis at 24-h intervals over this time course. As indicated in Fig. 3, staurosporine induced a sustained G 1 arrest in wild-type MEFs, whereas a substantial proportion of Rb Ϫ/Ϫ cells were able to escape the G 1 arrest in the first cell cycle. These cells went on to arrest at the G 2 /M junction, a response that did not require Rb function. Interestingly, although Rb disruption enabled approximately 70% of the original G 1 population to escape the G 1 arrest normally induced by staurosporine, approximately 30% of the G 1 fraction of Rb Ϫ/Ϫ MEFs still remained in G 1 phase over this time course. This result suggested that a subpopulation of G 1 cells, presumably in early G 1 phase, might still be capable of G 1 arrest despite the absence of Rb genes (see "Discussion").
We also investigated the effects of staurosporine in REF52 cells infected with viruses expressing the Rb-inactivating kinase complex, cyclin D1-Cdk4. A Cdk4 mutant (Cdk4 R24C ) that is resistant to the inhibitory influence of p16 was used in these studies (26,27). Our view here was that if we could inactivate Rb through cyclin D1-Cdk4-mediated hyperphosphorylation, these cells would show a reduced capacity to G 1 arrest in response to staurosporine. In support of a role for Rb in the staurosporine-induced G 1 arrest mechanism, we found that REF52 cells expressing cyclin D1-Cdk4 R24C were less susceptible to staurosporine-induced G 1 arrest as compared with the empty vector control cells (Fig. 4). Cell cycle analysis of REF52 cells 24 h after a chronic exposure to staurosporine showed that approximately 49 Ϯ 11% (n ϭ 3) of the original G 1 population of cyclin D1-Cdk4 R24C overexpressing cells remained in G 1 phase compared with 96 Ϯ 5% (n ϭ 3) for the empty vector control cells. This amounted to approximately a 50% reduction in the degree of staurosporine-induced G 1 arrest in cyclin D1-Cdk4 R24C overexpressing cells as compared with empty vector control cells (p Ͻ 0.01, Student's t test).

Staurosporine-induced G 1 Arrest in the First Cell Cycle Does Not Require the Function of the p16 Tumor Suppressor Gene-
Because the Cdk inhibitor p16 has previously been shown to induce an Rb-dependent G 1 arrest in mammalian cells (18 -20), we investigated whether staurosporine might act through p16 to induce G 1 arrest in MEFs. Our investigations showed, however, that MEFs from mice lacking p16 genes underwent a combined G 1 and G 2 arrest comparable with that observed in wild-type MEFs (Fig. 2, A and C). These results suggested that p16, a well recognized regulator of Rb function, was not essential for staurosporine-induced G 1 arrest in MEFs. DISCUSSION In this study we investigated the mechanistic basis of staurosporine-induced G 1 arrest in murine cells derived from transgenic mice lacking important negative regulators of the G 1 /S phase transition. Such a model system enabled us to investigate the importance of the p53, p21, Rb, and p16 genes in the mechanism through which staurosporine induces G 1 arrest in otherwise normal cells. We found that in contrast to the p53 tumor suppressor, the Rb protein was an important mediator of  3. Time-dependent changes in the cell cycle distribution profiles of wild-type and Rb ؊/؊ MEFs chronically exposed to vehicle or staurosporine. Murine embryonic fibroblasts from either wild-type or Rb Ϫ/Ϫ mice were exposed to vehicle (0.005% v/v Me 2 SO) or STP (50 nM) for up to 72 h. Samples were isolated at the indicated times, and cell cycle distribution was analyzed by flow cytometry as described under "Material and Methods." Samples shown are from a representative experiment that was repeated twice with similar results. Black, % G 1 ; white, % S; shading, % G 2 /M. the degree of staurosporine-induced G 1 arrest observed in the first cell cycle. Further evidence implicating Rb in the mechanism of staurosporine-induced G 1 arrest was evidenced in rat embryonic fibroblasts stably overexpressing cyclin D1-Cdk4 R24C . Such cells exhibited a reduced G 1 arrest response to staurosporine compared with empty vector control cells. Staurosporine did not, however, induce an Rb-dependent G 1 arrest by acting through the Cdk inhibitor proteins p16 or p21, because MEFs lacking these genes were still capable of G 1 arrest following staurosporine treatment. Our studies provide evidence of an important role for Rb in determining the degree of staurosporine-induced G 1 arrest.
Our finding that MEFs lacking p53 still underwent a combined G 1 and G 2 arrest in response to staurosporine provided evidence that in this cell type, p53 was not essential for staurosporine-induced G 1 arrest. Because the Cdk inhibitor, p21, can be induced by p53-independent mechanisms (25) and because p21 has previously been implicated in the mechanism by which staurosporine induces G 1 arrest (14), we also examined the importance of p21 in staurosporine-induced G 1 arrest. However, we again found that MEFs lacking p21 still underwent G 1 arrest in response to staurosporine, indicating that in this cell type, p21 was not essential for the G 1 arrest response.
Studies indicating the common occurrence of Rb disruptions in tumor cells and studies indicating that a tumor cell line overexpressing a Rb transgene underwent G 1 arrest following staurosporine exposure (14) prompted us to also test the role of Rb in staurosporine-induced G 1 arrest. We found that Rb was indeed an important mediator of the degree of G 1 arrest observed following exposure to staurosporine: approximately 70% of the Rb Ϫ/Ϫ cells escaped G 1 arrest when exposed to staurosporine, whereas in contrast virtually all of the cells with intact Rb remained arrested in the G 1 phase (Ͼ95%). The function of the tumor suppressor, p16, was not essential for staurosporineinduced G 1 arrest because MEFs lacking the p16 gene still underwent G 1 arrest. Approximately 30% of Rb Ϫ/Ϫ MEFs remained in the G 1 phase in the presence of staurosporine. This residual subpopulation could conceivably represent an early G 1 fraction of cells that arrest through a Rb-independent mechanism; however, further studies outside the scope of this report will be needed to investigate any early G 1 effects of staurosporine. In conclusion we show that Rb function is important for the majority of G 1 arrest seen in the first cell cycle and that investigations aimed at determining the mechanism through which staurosporine maintains Rb in an active form could ultimately provide the target of staurosporine as it relates to Rb.
We speculate on some potential targets through which staurosporine might act to induce a Rb-dependent G 1 arrest in Fig.  5. We represent Rb in an equilibrium between its hyperphosphorylated (relatively inactive) state and its hypophosphorylated (active) state. This equilibrium is governed by cyclindependent kinases (28) and probably a type 1 phosphatase, respectively (29,30). Rb kinases may include cyclin D1-Cdk4, cyclin D1-Cdk6, cyclin E-Cdk2, or cyclin A-Cdk2, all of which are known to directly phosphorylate Rb (17,31,32). Cyclin E-Cdk2 and cyclin A-Cdk2 are probably not the G 1 phase targets of the action of staurosporine because microinjection of neutralizing antibodies to either cyclin E or cyclin A induces a Rb-independent G 1 arrest (33,34). Also, overexpression of a dominant-negative Cdk2 transgene induces a Rb-independent G 1 arrest, providing evidence against a role for Cdk2 in staurosporine-induced G 1 arrest (35). This leaves cyclin D1-Cdk4 open as a possible target, and it has been reported that microinjection of anti-cyclin D1 antibodies or cyclin D1 antisense cDNA into cells induces G 1 arrest in cells containing wild-type Rb but not defective Rb (4). Furthermore, staurosporine does inhibit cyclin D1-Cdk4 kinase activity in vitro in the dose range used in our G 1 arrest experiments (50% inhibitory concentra- FIG. 4. REF52 cells overexpressing cyclin D1-Cdk4 R24C exhibit a reduced capacity to G 1 arrest in response to staurosporine. Rat embryonic fibroblasts stably overexpressing cyclin D1-Cdk4 R24C were left untreated or exposed to NOC (0.5 g/ml) or a combination of NOC (0.5 g/ml) and STP (12.5 nM) for 24 h. Samples were isolated following exposure to the chemicals, and cell cycle distribution was analyzed in the empty vector control and cyclin D1-Cdk4 R24C -overexpressing cells by flow cytometry as described under "Material and Methods." Samples shown are from representative experiments that were repeated three times with similar results.
FIG. 5. Potential mechanisms by which staurosporine might induce a Rb-dependent G 1 arrest. The Rb gene product in its hypophosphorylated form inhibits cell cycle progression from G 1 to S phase (17). Rb is inactivated by phosphorylation that can be directed by G 1 /S phase cyclin-dependent kinases such as cyclin D1-Cdk4, cyclin D1-Cdk6, cyclin E-Cdk2, and cyclin A-Cdk2 (17,31,32). The state of Rb phosphorylation is probably also regulated at the level of Rb dephosphorylation by an as yet poorly defined phosphatase (29,30). Although DNA damage-induced G 1 arrest has been linked to a pathway involving the p53 tumor suppressor and the p21 Cdk inhibitor, staurosporine does not elicit G 1 arrest through a pathway involving p53 or p21 function. Another Cdk inhibitor, p16, which induces a G 1 arrest in a Rb-dependent fashion (18 -20) was also found not to be required for staurosporine-induced G 1 arrest. Staurosporine might induce a Rb-dependent G 1 arrest through direct inhibition of G 1 /S phase cyclin-dependent kinase activity. Important here would be cyclin D1-Cdk4 or cyclin D1-Cdk6 complexes whose inactivation causes a Rb-dependent G 1 arrest (4), and in these studies we found that overexpression of cyclin D1-Cdk4 R24C could reduce the ability of staurosporine to induce G 1 arrest. Alternatively, staurosporine could conceivably inhibit a kinase that normally negatively regulates a phosphatase that maintains Rb in an active, hypophosphorylated state. A potential Rb-independent arrest point in early G 1 phase is represented by X. Whether this arrest point is actually present or not will require further studies outside the scope of this study.
tion of approximately 20 nM, data not shown), and we found that overexpression of cyclin D1-Cdk4 R24C in REF52 cells reduced the ability of staurosporine to induce G 1 arrest compared with the empty vector control cells. Such results, although suggestive of cyclin D1-Cdk4 as the direct target of the actions of staurosporine, are not by themselves conclusive. Indeed, staurosporine in G 1 phase might induce a Rb-dependent G 1 arrest through the activation of a phosphatase that dephosphorylates and activates Rb (Fig. 5). This could occur if staurosporine inhibited a kinase that normally negatively regulated this Rb phosphatase (Fig. 5). There is evidence that a protein phosphatase type 1 can be inactivated by serine/threonine phosphorylation, and this same class of phosphatases has been implicated in the dephosphorylation of Rb (29,30). It will be interesting to determine if the phosphorylation status and activity of this type 1 phosphatase is affected by staurosporine.
In the different MEFs we studied, we did not observe any obvious differences in the ability of cells to arrest in the G 2 phase 24 h after staurosporine treatment. This indicated that p53, p21, Rb, or p16 were not essential components of the induction of this G 2 arrest response. A possible target of the action of staurosporine in G 2 has been presented before as the Cdc2 kinase (36).
In summary, our results highlight the importance of the Rb protein in the mechanism through which staurosporine induces G 1 arrest in the first cell cycle. Continued investigation of the actions of staurosporine on Rb phosphorylation status could uncover the route and eventual target of the actions of staurosporine as it relates to Rb.