In CKI-Responsive Cells Compared with CKI-Unresponsive Cells

Protection against chemotherapy-induced cytotoxicity by cyclin-dependent kinase inhibitors (CKI) in CKI-responsive cells compared with CKI-unresponsive cells

Mathias Schmidt1,2 and Zhen Fan*,1

1Department of Experimental Therapeutics, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, Texas, TX 77030, USA

Inactivation of the retinoblastoma (Rb) protein caused by Keywords: chemotherapy; chemoprotection; Rb; cell gene mutation, association with oncoproteins from small cycle; CDK DNA viruses, mutational inactivation of p16Ink4a,or overexpression of cyclin D is a common feature of many human cancer cells and is causally associated with the Introduction aberrant proliferation control of cancer cells; whereas normal cells maintain an integrated cell cycle machinery The development of neoplasia frequently involves and are subject to cell cycle checkpoint control by cyclin- inactivation of the retinoblastoma (Rb) protein and dependent kinase (CDK) inhibitors (CKIs). To determine p53 tumor-suppressor pathways and disruption of cell whether this dierence can be translated into a cycle checkpoints that monitor the integrity of therapeutic advantage to protect normal cells from replication and division of mammalian cells (Sherr, adverse cytotoxicity caused by chemotherapy, we 1994; Weinberg, 1991, 1995). These genetic alterations established cell model systems for ecdysone-inducible in neoplasia are often causally associated with expression of p16Ink4a, p21Waf1, and p27Kip1 in one CKI- uncontrolled proliferation of cancer cells. In the past responsive cell line (A431 human vulvar epidermoid decades, the design and concept of conventional cancer carcinoma cells with functional Rb) and one CKI- chemotherapy has been in principle based on defective unresponsive cell line (SiHa human cervical cancer cells control of cell cycle events in cancer cells and the with nonfunctional Rb). Expression of p16Ink4a, p21Waf1,or quantitative dierence in the proportion of fast- p27Kip1 in both SiHa and A431 cells strongly inhibited dividing cells between cancer cells and normal cells CDK2 activity, indicating functional expression of the (Darzynkiewicz, 1995; Kohn et al., 1994). While many CDK inhibitors in both cell lines. However, only in A431 chemotherapeutic drugs can successfully kill fast- cells did expression of p16Ink4a, p21Waf1, or p27Kip1 cause dividing cancer cells, the drugs unfortunately also Rb dephosphorylation, arrest cell cycle traversal, and damage several types of rapidly proliferating normal potently inhibit cell proliferation. Induction of p16Ink4a, cells such as the hematopoietic precursors, intestinal p21Waf1, or p27Kip1 in SiHa cells failed to cause Rb cells, and hair follicle cells. This nonselective killing for dephosphorylation or to arrest cell cycle traversal, and rapidly proliferating normal cells often causes serious such induction only minimally inhibited cell proliferation. adverse eects in cancer patients, such as bone marrow We then compared the chemosensitivity of clones derived suppression, anorexia (appetite loss), nausea, vomiting, from these two cell lines when the CKIs were and were diarrhea, stomatitis, and alopecia (hair loss) and thus not induced. Induction of p16Ink4a,p21Waf1, or p27Kip1 limits the doses of these drugs that can be tolerated. conferred strong resistance to paclitaxel- or cisplatin- The progression of the mammalian cell cycle is driven mediated cytotoxicity on the CKI-responsive A431 cells by periodic activation of a group of nuclear enzymes but not on the CKI-unresponsive SiHa cells. Our results called cyclin-dependent kinases (CDK) in response to support a novel chemotherapy strategy for treating various intracellular and extracellular signals (Morgan, patients with Rb pathway-impaired cancers by concur- 1995; Sherr, 1996). Cells become committed to division rent administration of chemotherapy with CKIs as at a point in mid to late G1 phase of the cell cycle (known chemoprotective agents for normal cells. Oncogene as the restriction point) following phosphorylation of the (2001) 20, 6164 ± 6171. Rb protein. The CDK inhibitors (CKIs) decelerate the cell cycle traversal by inactivating the CDKs that phosphorylate the Rb protein, and they are therefore key negative regulators of eukaryotic cell cycle traversal. *Correspondence: Z Fan, Box 36, The University of Texas MD Two families of CKIs, the INK and KIP/CIP inhibitors, Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX have been identi®ed. The INK family of CKIs includes 77030, USA; E-mail: [email protected] p16Ink4a,p15Ink4b p18Ink4c, and p14/p19Arf, which are 2Current address: Byk Gulden Lomberg Chemische Fabrik GmbH, Byk-Gulden-Str. 2, D-78467 Konstanz, Germany inactivated by gene deletion, by point mutation, or by Received 9 April 2001; revised 28 June 2001; accepted 5 July 2001 silencing through methylation in a high percentage of CDK inhibitors and chemoprotection M Schmidt and Z Fan 6165 human tumors. The KIP/CIP family consists of p21Waf1, tion of p16Ink4a,p21Waf1, or p27Kip1 in individual clones p27Kip1, and p57Kip2, which are rarely mutated in human of both cell lines with high expression levels inhibited malignancies (Sherr, 2000; Sherr and Roberts, 1999). the activities of CDK2, indicating the functional A hallmark of neoplasia is the disturbance of cell expression of these inhibitors in these cells. However, cycle controls. Although overexpression of such CKIs only in A431 cells did induction of the CDK inhibitors as p21Waf1 and p27Kip1 has been shown to cause arrest cell cycle traversal and increase drug resistance. increased resistance to chemotherapy in certain types In SiHa cells, functional expression of p16Ink4a, p21Waf1, of human cancer cells (Fueyo et al., 1998; Ruan et al., or p27Kip1 was associated with inhibited CDK activity, 1998; Schmidt et al., 2000; St Croix et al., 1996; but cell proliferation and drug sensitivity were not Waldman et al., 1997), recent studies suggested that altered. Comparing the results with those for unin- CKIs may actually be used to protect normal cells duced cells, we found that inducible expressions of from unwanted toxicity from chemotherapy and we p16Ink4a,p21Waf1, and p27Kip1 all imposed strong may thus turn this disadvantage into a therapeutic resistance to chemotherapy on the Rb-functional advantage (Davis et al., 2001; Stone et al., 1996). The A431 cells but were much less eective in the Rb- rationale for such strategy is based on the hypothesis disrupted SiHa cells. Our data provide experimental that substantial CKI-mediated chemoresistance can be support for the novel therapeutic strategy of combining found only in cells with normal Rb. In most cancer chemotherapy with inhibition of CDK as a means of cells, the Rb pathway is inactivated. (Weinberg, 1995; chemoprotection of normal cells from chemotherapy- Wiman, 1993). This disturbance is often caused by induced nonspeci®c cytotoxicity in patients with Rb- infection from small DNA viruses such as the human pathway impaired cancers. papilloma virus (HPV), by mutational inactivation of p16Ink4a or overexpression of cyclin D and, in a limited set of malignancies (such as retinoblastoma and Results osteosarcoma), by genetic mutation of the Rb gene. The inactivation of Rb allows cancer cells to evade the Establishment of p16Ink4a-, p21Waf1-, and p27Kip1-inducible normal cell cycle checkpoint control, whereas normal expression clones in A431 and SiHa cells cells have functional Rb and thus are liable to checkpoint regulation. This CKI-mediated resistance p16Ink4a, p21Waf1,orp27Kip1 cDNA was subcloned into to chemotherapy might therefore be an ideal approach the ecdysone-inducible mammalian cell expression to protect normal cells from unwanted chemotherapy- vectors, and the expression vectors were transfected induced toxicity in patients with Rb pathway-impaired in A431 and SiHa cells for stable transfectant clones. cancer. The approach may justify a novel therapeutic Over 10 clones for each cell line were obtained with strategy by concurrently treating Rb pathway-impaired high expression levels of p16Ink4a, p21Waf1, or p27Kip1 human cancer with chemotherapy in combination with upon induction. Figure 1 shows Western blot analyses inhibiting CDK activity to protect normal cells from of p16Ink4a, p21Waf1, and p27Kip1 proteins in these chemotherapy-induced cytotoxicity. representative clones (termed SiHa-p16, SiHa-p21, In this study, we employed the ecdysone-inducible and SiHa-p27, and A431-p16, A431-p21, and A431- expression system to ectopically express three repre- p27, respectively) 24 h after induction of gene expres- sentative CDK inhibitors, p16Ink4a, p21Waf1,orp27Kip1, sion with muristerone A. Induction of p16Ink4a, p21Waf1, each of which inhibits the progression of the cell cycle with a dierential timing, in selected cell lines that contain functional or nonfunctional Rb as a `proof-of- concept' experimental model. Because of technical diculties in obtaining normal human cell transfectant for inducible expression of p16Ink4a, p21Waf1, or p27Kip1 in the cells, we instead used A431 human vulvar epithelial carcinoma cells, which contain functional Rb but mutated p53 (Kwok et al., 1994). For Rb-nonfunctional cells, we used SiHa human cervical carcinoma cells that are known to be infected with HPV (Baker et al., 1987); the HPV virus produces two oncoproteins, E6 and E7, that inactivate p53 and Rb, respectively (Dellas et al., 1998; Dyson et al., 1989; Funk and Galloway, 1998; Jones et al., 1997; Morozov et al., 1997; Schener et al., 1990; Verlhac et al., 1994). HPV infection is one of the major causes of Rb inactivation Figure 1 Inducible expression of p16Ink4a, p21Waf1, and p27Kip1 in human cancer (Jones and Munger, 1996; Ludlow in SiHa and A431 cells. The cells were incubated for 24 h in and Skuse, 1995; Munger et al., 1992; Vousden, 1995; culture medium in the absence (7) or presence (+) of 3 mM Ink4a Waf1 zur Hausen, 1996). We examined how the CKIs muristerone A. Equal amounts of lysate from p16 -, p21 -, and p27Kip1-induced cells were loaded in each lane of an SDS ± p16Ink4a,p21Waf1, and p27Kip1 modulated the drug- PAGE, followed by Western blot analysis with speci®c antibodies mediated cytotoxicity in A431 and SiHa cells. Induc- against p16Ink4a, p21Waf1, or p27Kip1

Oncogene CDK inhibitors and chemoprotection M Schmidt and Z Fan 6166 or p27Kip1 was detectable approximately 2 h after inhibited CDK2 activity in both SiHa and A431 cells, induction, peaked 24 h after induction, and persisted as measured by an in vitro kinase assay with GST-Rb for at least 96 h without further supplementation with as substrate (Figure 3a). However, Western blot muristerone A (data not shown). Expression of p16Ink4a, analysis with an anti-Rb antibody that recognized p21Waf1, or p27Kip1 alone was sucient to substantially both hyperphosphorylated (inactive) and hypophos- inhibit the proliferation of A431 cells; in contrast, phorylated (active) Rb within the cells gave sharply induction of these inhibitors in the HPV-positive SiHa dierent results (Figure 3b). Proliferating A431 cells cells only minimally inhibited proliferation (Figure 2), (including A431-p16, A431-p21, and A431-p27 clones) presumably because of the HPV E7 protein that displayed a high proportion of hyperphosphorylated sequesters Rb from E2F and thereby renders the cells Rb, which was markedly reduced when p16Ink4a, p21Waf1, capable of bypassing the inhibitory eects of p16Ink4a, or p27Kip1 was induced in these A431 clones. In p21Waf1, or p27Kip1 on CDK. contrast, we found that the SiHa cells (including SiHa-p16, SiHa-p21, and SiHa-p27 clones) exhibited an unchanged ratio of hyperphosphorylated and Effects of p16Ink4a,p21Waf1, and p27Kip1 on CDK activity hypophosphorylated Rb regardless of whether p16Ink4a, and Rb phosphorylation p21Waf1, or p27Kip1 was induced in these cells. We therefore examined the eects of p16Ink4a, p21Waf1, Flow cytometric analysis of both SiHa and A431 and p27Kip1 on CDK activity and Rb phosphorylation cells with and without induction of p16Ink4a, p21Waf1,or in these two cell clones. The marked dierence in p27Kip1 further con®rmed the results obtained with the inhibition of the proliferation by p16Ink4a,p21Waf1,and cell proliferation assay (Figure 2). Induction of p16Ink4a, p27Kip1 in SiHa and A431 cells was not correlated with p21Waf1, or p27Kip1 in A431-p16, A431-p21, or A431-p27 their eects on the activities of CDKs within the cells. cells signi®cantly increased the proportion of cells in Induction of p16Ink4a,p21Waf1, or p27Kip1 strongly G1 phase of the cell cycle and decreased the proportion of cells in S phase. In contrast, induction of p16Ink4a, p21Waf1, or p27Kip1 expression in SiHa cells only modestly increased the proportion of the cells in G1 phase and decreased the proportion of the cells in S phase (Figure 4).

Figure 3 Eect of p16Ink4a, p21Waf1, and p27Kip1 on the inhibition of CDK activity and the status of Rb phosphorylation. (a) Inhibition of CDK2 activity. SiHa and A431 cell clones were incubated as described in the legend for Figure 1. CDK2 was immunoprecipitated from the lysates with the corresponding speci®c antibodies. The immunoprecipitates were subjected to an Rb kinase assay as described in the text. (b) Status of Rb phosphorylation upon induction of p16Ink4a, p21Waf1, or p27Kip1 in Figure 2 Eect of p16Ink4a, p21Waf1, and p27Kip1 on SiHa and SiHa and A431 cells. SiHa and A431 cells were incubated as A431 cell proliferation. SiHa and A431 cell clones were described in the legend for Figure 1. Two hundred micrograms of individually seeded onto six-well tissue culture plates at a density each lysate from p16Ink4a, p21Waf1, or p27Kip1 induced or of 26104 cells/well in the absence (&) or presence (&)of3mM uninduced cells was loaded in each lane for an SDS ± PAGE, muristerone A in the culture medium. The cells were cultured for followed by Western blot analysis with an Rb-speci®c antibody. 5 days. Relative cell number was determined using an MTT assay GST-Rb: glutathione S-transferase-Rb fusion protein; ppRb: as described in the text. Values are means with standard deviation hyperphosphorylated Rb protein; pRb: hypophosphorylated Rb for triplicate wells for each condition protein

Oncogene CDK inhibitors and chemoprotection M Schmidt and Z Fan 6167 cisplatin-mediated cytotoxicity was not aected upon Differential response of SiHa and A431 cells to p16Ink4a, p21Waf1, or p27Kip1 induction. chemotherapy-induced cytotoxicity following induction of p16Ink4a, p21Waf1, or p27Kip1 SiHa cells (including SiHa-p16, SiHa-p21, and SiHa- Discussion p27 clones) and A431 cells (including A431-p16, A431- p21, and A431-p27 clones) were then compared for In this study, we examined the eects of inducible their sensitivities to conventional chemotherapy when expression of p16Ink4a, p21Waf1, and p27Kip1 on chemo- p16Ink4a,p21Waf1,orp27Kip1 was or was not induced in therapy-induced cytotoxicity in the Rb-functional A431 these cells. Paclitaxel, which acts primarily by stabiliz- cell line and in the Rb-disrupted SiHa cell line. We ing microtubules in cells in M phase and therefore is an found that the Rb-functional A431 cells were very M phase-speci®c drug, was used to evaluate the responsive to inducible expression of CKI, whereas the chemosensitivity of the two cell lines when p16Ink4a, Rb-disrupted SiHa cells were not. SiHa cells contain a p21Waf1, or p27Kip1 was or was not induced in the cells. fully integrated HPV 16 genome, including E7 (Baker Induction of p16Ink4a, p21Waf1,orp27Kip1 conferred et al., 1987); the latter abrogates the Rb-mediated strong resistance on A431 cells to paclitaxel-mediated checkpoint of the cell cycle by sequestering Rb from cytotoxicity (Figure 5). In contrast, induction of E2F through a well-de®ned motif, LxCxE (Huang et p16Ink4a, p21Waf1, or p27Kip1 did not produce any al., 1990; Jones et al., 1990) and by promoting Rb signi®cant dierence in the sensitivity to paclitaxel- degradation through a ubiquitin-proteasome-dependent mediated cytotoxicity in the SiHa clones. Compared pathway (Berezutskaya et al., 1997; Berezutskaya and with results for A431 cells, in which the IC50 for Bagchi, 1997; Boyer et al., 1996). In our current study, paclitaxel-mediated cytotoxicity increased from 510 to we used this HPV-mediated Rb-disrupted cell line, 4104 mM when p16Ink4a, p21Waf1, or p27Kip1 was induced, which bypasses the eects of p16Ink4a, p21Waf1,orp27Kip1, the IC50 of SiHa cells was not signi®cantly altered to represent a variety of cancer cells that have following p16Ink4a, p21Waf1,orp27Kip1 induction. inactivated Rb pathway due to various known causes. Similar results were obtained with another antineo- Our results indicate that patients with Rb pathway- plastic drug, cisplatin, which acts to induce DNA impaired cancers may bene®t from receiving both crosslinking and therefore is primarily a cell cycle- chemotherapy and CKIs to protect normal cells from nonspeci®c drug (Figure 6). The IC50 for cisplatin- chemotherapy-induced cytotoxicity. mediated cytotoxicity increased from 10 to 4104 mM The intention in our current study was to establish when p16Ink4a,p21Waf1, or p27Kip1 was induced in the cell models to prove the potential utility of a novel A431 cell clones, whereas in SiHa cells the IC50 for therapeutic strategy by which chemotherapy may be

Figure 4 Cell cycle distribution analysis upon induction of p16Ink4a, p21Waf1, or p27Kip1 in SiHa and A431 cells. SiHa and A431 cell clones were cultured for 24 h in culture medium in the absence (7) or presence (+) of 3 mM muristerone A. Cell cycle distribution was analysed by ¯ow cytometry after the DNA was stained with propidium iodide. The analysis was repeated, and average values from three independent experiments are presented

Oncogene CDK inhibitors and chemoprotection M Schmidt and Z Fan 6168

Figure 5 Dierential sensitivity of SiHa and A431 cells to paclitaxel-mediated cytotoxicity upon induction of p16Ink4a, p21Waf1,or p27Kip1. SiHa or A431 cell clones were seeded onto 24-well plates in normal growth medium with (*) and without (*) induction of p16Ink4a (left), p21Waf1 (middle), or p27Kip1 (right) by 3 mM muristerone A for 24 h. The cells were then exposed for 3 h to various doses of paclitaxel, followed by an additional 72 h culture period in normal growth medium with (*) and without (*)3mM muristerone A. Relative cell number was determined with an MTT assay as described in the text

Figure 6 Dierential sensitivity of SiHa and A431 cells to cisplatin-mediated cytotoxicity upon induction of p16Ink4a, p21Waf1,or p27Kip1. SiHa or A431 cell clones were seeded onto 24-well plates in normal growth medium with (*) and without (*) induction of p16Ink4a (left), p21Waf1 (middle), or p27Kip1 (right) by 3 mM muristerone A for 24 h. The cells were then exposed for 3 h to various doses of cisplatin, followed by an additional 72 h culture period in normal growth medium with (*) and without (*)3mM muristerone A. Relative cell number was determined with an MTT assay as described in the text

co-administered with CKIs in cancer cells, as a means or without inducible expression of p16Ink4a,p21Waf1,or of protecting normal cells from chemotherapy-induced p27Kip1, ectopic expression of these inhibitors signi®- cytotoxicity. We found that, in contrast to results for cantly protected A431 cells from paclitaxel- or SiHa cells, which showed similar cytotoxic responses to cisplatin-induced cytotoxicity. It is interesting to note both paclitaxel- and cisplatin-induced cytotoxicity with that expression of these cell cycle inhibitors protected

Oncogene CDK inhibitors and chemoprotection M Schmidt and Z Fan 6169 A431 cells from the cytotoxic eects of two very synthetic CDK inhibitor prevented chemotherapy- dierent drugs. Paclitaxel is a microtubule stabilizer induced alopecia in rats (Davis et al., 2001). The data that acts in an M phase-dependent fashion; cisplatin is from our experiments suggest that the strategy of a DNA alkylating agent whose activity is cell cycle- inhibiting CDK activation with natural or synthetic independent. A431 cells contain a functional Rb but inhibitors may also be used systemically to protect mutated p53 (Kwok et al., 1994). In our other study, normal cells from chemotherapy-induced adverse we transfected RKO human colon cancer cells, which eects, provided that the Rb status in the cancers can contain both wild-type Rb and wild-type p53 (Kapoor be determined before the treatment. For patients with and Lozano, 1998) with the same ecdysone-inducible Rb-functional cancers, administration of CDK inhibi- expression vectors of p16Ink4a,p21Waf1, or p27Kip1 used in tors would compromise the eect of chemotherapy and current study and we found that the expression of the thus is not recommended for these cancer patients. cell cycle inhibitors conferred greater resistance to cell In summary, our data support the idea of a novel cycle-dependent drugs (5-¯uorouracil, camptothecin, therapeutic strategy by using CKIs such as p16Ink4a, and etoposide) than to cell cycle-independent drugs p21Waf1, or p27Kip1 or synthetic small-molecule CDK (cisplatin, doxorubicin, and melphalan) (Schmidt et al., inhibitors as chemoprotective agents to protect normal 2001). These results are apparently dierent from the cells from the toxic eects of chemotherapy. This observation seen in A431 cells in current study. The strategy may also apply to other conventional similar protective response to paclitaxel- and cisplatin- therapies, such as radiation therapy, and apparently mediated cytotoxicity seen in p16Ink4a-, p21Waf1-, or applies only to Rb-de®cient cancer cells. Given that Rb p27Kip1-arrested A431 cells may be due to the mutated mutation/inactivation is one of most frequent genetic p53 status in A431 cells. alterations found in human cancers, this novel An early study showed that an IPTG-inducible chemotherapy strategy may hold great promise for expression of p16Ink4a in H294T melanoma cells arrested improving the eectiveness of conventional chemo- cells and conferred on the cells resistance to cytotoxi- therapy for patients with cancer. city induced by methotrexate, vinblastine, and cisplatin (Stone et al., 1996). Our current study expanded the candidates of the cell cycle inhibitors into p16Ink4a, p21Waf1, or p27Kip1 on paclitaxel- and cisplatin-induced Materials and methods cytotoxicity. To better justify this novel chemotherapy strategy, we employed the ecdysone-inducible expres- Materials sion system to obtain stable cell clones expressing Anti-p16Ink4a and p27Kip1 antibodies were from Santa Cruz p16Ink4a, p21Waf1, or p27Kip1, and we compared the Biotechnology Inc. (Santa Cruz, CA, USA); anti-p21Waf1 sensitivity of the cell clones with and without induction antibody was from Neo Markers Biotechnology Inc. (Union of these CDK inhibitors. This approach allowed an City, CA, USA). Antibody against Rb was from PharMingen unbiased comparison between two isogenic cell popula- Biotechnology Inc. (San Diego, CA, USA). Anti-CDK2 antibody and glutathione S-transferase-Rb fusion protein Ink4a tions that dier only in the expression of p16 , (GST-Rb) were from Santa Cruz Biotechnology Inc.. All p21Waf1, or p27Kip1. We compared the results in one other reagents were purchased from Sigma Chemical Co. (St. CKI-responsive cell line and one CKI-unresponsive cell Louis, MO, USA) unless otherwise speci®ed. line. It would be ideal to also perform this study in nonmalignant cells; however, as mentioned in the Cells and cell culture introduction, we encountered technical diculty in obtaining nonmalignant cell clones for inducible SiHa human cervical carcinoma cells and A431 human expression of the cell cycle inhibitors, and we therefore epidermoid carcinoma cells were obtained from American used A431 cells instead. A caveat is that, although Type Culture Collection (Manassas, VA, USA) and were A431 cells contain normal Rb, the cells are supposed to maintained in DMEM supplemented with 10% fetal bovine serum, 100 units/ml penicillin G, 100 mg/ml streptomycin, dier from normal cells in some other aspects and thus and 0.25 mg/ml amphotericin B. may not completely represent human normal cells. Similarly, a cautious note should be made that, although HPV infection is one of the major causes of Construction of the inducible pIND vectors containing p16Ink4a, Waf1 Kip1 Rb inactivation in human malignancies, HPV may p21 , or p27 have other eects on the SiHa cells that we used as the The construction of the pINDp21Waf1 and pINDp27Kip1 vectors Rb-nonfunctional cell model in the current study. For was previously described (Schmidt et al., 2000). The cDNA future studies, cell lines that contain Rb mutation such encoding p16Ink4a was ampli®ed by PCR using the primers 5' at as Saos-2 may be considered. aag ctt gcc gcc atg gag ccg gcg gcg g and 3' t ata tct aga act atc One important study that needs to be carried out is ggg gat gtc tg. The PCR fragment was subcloned into the pIND vector via the restriction sites HindIII and XbaI. to evaluate this therapeutic strategy in an in vivo model. Although it is possible for p16Ink4a, p21Waf1,orp27Kip1 to be delivered to mice in a viral or nonviral way, an Establishment of p16Ink4a-, p21Waf1-, and p27Kip1-inducible alternative approach is to use synthetic small-molecule expression SiHa and A431 cell clones inhibitors of CDK. In a recent study that reported very Transfection was performed with the Fugene-6 transfection promising and interesting results, topic application of a kit (Roche Diagnostic, Indianapolis, IN, USA). SiHa and

Oncogene CDK inhibitors and chemoprotection M Schmidt and Z Fan 6170 A431 cells were ®rst transfected with the pIND-regulatory Flow cytometric analysis vector pVgRXR (Invitrogen, Carlsbad, CA, USA) and were selected in medium containing 200 mg/ml Zeocin. Stable Cell cycle distribution analysis with propidium iodide was master clones were evaluated by a transient transfection of performed as previously described (Fan et al., 1995). Brie¯y, the cells with pIND-LacZ (Invitrogen) for 24 h, followed by upon the completion of the desired treatment, cells were cell staining with X-Gal solution (fresh 1 mg/ml X-Gal in a harvested by trypsinization, and an aliquot (16106 cells) was solution containing 5 mM K3Fe(CN)6,5mM K4Fe(CN)63- washed once with ice-cold PBS, followed by ®xation with 2% H2O and 1 mM MgCl2 in PBS) for over 6 h at 378C. A formaldehyde in PBS for 15 min. The cells were then kept in satisfactory blue-staining clone was chosen for second-round 70% ethanol at 7208C until DNA staining. DNA was stable transfection with pINDp16Ink4a, pINDp21Waf1,or stained with a solution containing 25 mg/ml propidium iodide pINDp27Kip1 vectors. Cell clones expressing both pIND and and 10 mg/ml RNAse A in PBS for 6 h at 378C. Cell cycle pVgRXR vectors were selected using a culture medium distribution was analysed with a ¯ow cytometer at an containing 200 mg/ml Zeocin and 500 mg/ml neomycin excitation wavelength of 488 nm. (G418). Expression of p16Ink4a, p21Waf1, or p27Kip1 was induced by 3 m muristerone A for 24 h and was examined M Rb kinase assay by Western blot analysis with speci®c antibodies. The kinase assay with GST-Rb as a substrate was performed as previously described (Chou et al., 1999; Wu et al., 1996). Cell proliferation assay Brie¯y, after cell lysis, CDK2 was immunoprecipitated. The A431 and SiHa cells were seeded onto six-well tissue culture immunoprecipitates were subjected to an in vitro kinase plates at a density of 26104 cells/well in medium reaction with 30 ml of reaction mixture containing 10 mCi of supplemented with 10% fetal bovine serum in the absence [g732P]ATP, 20 mM unlabeled lithium ATP, and 0.2 mgof or presence of 3 mM muristerone A. Relative cell numbers puri®ed GST-Rb protein (Santa Cruz Biotechnology, Inc.) in were obtained using an MTT assay as described below. a kinase buer (50 mM HEPES ± KOH, pH 7.5, 10 mM MgCl2,1mM dithiothreitol, 2.5 mM EGTA, 10 mM b- glycerophosphate, 0.1 m sodium orthovanadate, 1 mM Chemotherapy cytotoxicity assay (MTT assay) M NaF). The reaction was allowed to proceed at 308C for Cells were seeded onto 24-well culture plates. After exposure 30 min and was then stopped by boiling the samples in SDS of the cells to various doses of chemotherapeutic drugs and sample buer for 5 min; the products were separated by 10% incubation for various times, relative cell number was assayed SDS ± PAGE and then subject to autoradiography. by adding 50 ml of 10 mg/ml MTT (3-[4,5-dimethylthiazol-2- yl]-2,5-diphenyltetrazolium bromide) to 0.5 ml of culture medium and incubating the cells for 3 h at 378CinaCO2 incubator, followed by cell lysis with 500 ml of lysis buer Abbreviations Rb, retinoblastoma protein; CDK, cyclin-dependent ki- containing 20% SDS in dimethyl formamide/H2O (1 : 1, v/v), pH 4.7, at 378C for more than 6 h. Relative cell number was nase; HPV, human papilloma virus; GST-Rb, glutathione then determined by measuring the optical absorbance of cell S-transferase-Rb fusion protein. lysates at a wavelength of 595 nm and normalizing the value with a corresponding control (untreated cells).

Immunoblot analysis Acknowledgments Cells were lysed in a lysis buer containing 50 mM Tris, Mathias Schmidt was supported by a fellowship from the pH 7.4, 150 mM NaCl, 0.5% NP-40, 50 mM NaF, 1 mM Ernst Schering Research Foundation, Berlin, Germany. Na3VO4,1mM phenylmethylsulfonyl ¯uoride, 25 mg/ml The authors are grateful to Mr Michael Worley of the leupeptin, and 25 mg/ml aprotinin. The lysates were cleared Department of Scienti®c Publications for editorial assis- by centrifugation, and the supernatants were collected. Equal tance with the manuscript. This study was supported in amounts of lysate protein were used for Western blot analysis part by the NCI cancer center core grant (CA16672), a with the indicated antibodies (Fan et al., 1995). Speci®c research award from Bristol-Myers Squibb Company, and signals were visualized by use of the ECL chemoluminescence a start-up fund to Z Fan by The University of Texas MD detection kit (Amersham, Arlington Heights, IL, USA). Anderson Cancer Center.

References

Baker CC, Phelps WC, Lindgren V, Braun MJ, Gonda MA Davis ST, Benson BG, Bramson HN, Chapman DE, and Howley PM. (1987). J. Virol., 61, 962 ± 971. Dickerson SH, Dold KM, Eberwein DJ, Edelstein M, Berezutskaya E and Bagchi S. (1997). J. Biol. Chem., 272, Frye SV, Gampe Jr RT, Grin RJ, Harris PA, Hassell 30135 ± 30140. AM, Holmes WD, Hunter RN, Knick VB, Lackey K, Berezutskaya E, Yu B, Morozov A, Raychaudhuri P and Lovejoy B, Luzzio MJ, Murray D, Parker P, Rocque WJ, Bagchi S. (1997). Cell Growth Dier., 8, 1277 ± 1286. Shewchuk L, Veal JM, Walker DH and Kuyper LF. Boyer SN, Wazer DE and Band V. (1996). Cancer Res., 56, (2001). Science, 291, 134 ± 137. 4620 ± 4624. Dellas A, Schultheiss E, Leivas MR, Moch H and Torhorst J. Chou JL, Fan Z, DeBlasio T, Ko A, Rosen N and (1998). Anticancer Res., 18, 3991 ± 3998. Mendelsohn J. (1999). Breast Cancer Res. Treat., 55, Dyson N, Howley PM, Munger K and Harlow E. (1989). 267 ± 283. Science, 243, 934 ± 937. Darzynkiewicz Z. (1995). J. Cell Biochem., 58, 151 ± 159.

Oncogene CDK inhibitors and chemoprotection M Schmidt and Z Fan 6171 FanZ,LuY,WuX,DeBlasioA,KoAandMendelsohnJ. Schener M, Werness BA, Huibregtse JM, Levine AJ and (1995). J. Cell Biol., 131, 235 ± 242. Howley PM. (1990). Cell, 63, 1129 ± 1136. Fueyo J, Gomez-Manzano C, Puduvalli VK, Martin-Duque Schmidt M, Lu Y, Liu B, Fang M, Mendelsohn J and Fan Z. P, Perez-Soler R, Levin VA, Yung WK and Kyritsis AP. (2000). Oncogene, 19, 2423 ± 2429. (1998). Int. J. Oncol., 12, 665 ± 669. Schmidt M, Lu Y, Parant JM, Lozano G, Bacher G, Beckers Funk JO and Galloway DA. (1998). Trends Biochem. Sci., T and Fan Z. (2001). Mol. Pharmacol., (in press). 23, 337 ± 341. Sherr CJ. (1994). Trends Cell Biol., 4, 15 ± 18. Huang S, Wang NP, Tseng BY, Lee WH and Lee EH. (1990). Sherr CJ. (1996). Science, 274, 1672 ± 1677. EMBO J., 9, 1815 ± 1822. Sherr CJ. (2000). Cancer Res., 60, 3689 ± 3695. Jones DL, Alani RM and Munger K. (1997). Genes Dev., 11, Sherr CJ and Roberts JM. (1999). Genes Dev., 13, 1501 ± 2101 ± 2111. 1512. Jones DL and Munger K. (1996). Semin. Cancer Biol., 7, St Croix B, Florenes VA, Rak JW, Flanagan M, Bhatta- 327 ± 337. charya N, Slingerland JM and Kerbel RS. (1996). Nat. Jones RE, Wegrzyn RJ, Patrick DR, Balishin NL, Vuocolo Med., 2, 1204 ± 1210. GA, Riemen MW, Defeo-Jones D, Garsky VM, Heim- Stone S, Dayananth P and Kamb A. (1996). Cancer Res., 56, brook DC and Oli A. (1990). J. Biol. Chem., 265, 12782 ± 3199 ± 3202. 12785. VerlhacMH,KubiakJZ,ClarkeHJandMaroB.(1994). Kapoor M and Lozano G. (1998). Proc. Natl. Acad. Sci. Development, 120, 1017 ± 1025. USA, 95, 2834 ± 2837. Vousden KH. (1995). Semin. Cancer Biol., 6, 109 ± 116. Kohn KW, Jackman J and O'Connor PM. (1994). J. Cell Waldman T, Zhang Y, Dillehay L, Yu J, Kinzler K, Biochem., 54, 440 ± 452. Vogelstein B and Williams J. (1997). Nat. Med., 3, Kwok TT, Mok CH and Menton-Brennan L. (1994). Cancer 1034 ± 1036. Res., 54, 2834 ± 2836. Weinberg RA. (1991). Science, 254, 1138 ± 1146. Ludlow JW and Skuse GR. (1995). Virus Res., 35, 113 ± 121. Weinberg RA. (1995). Cell, 81, 323 ± 330. Morgan DO. (1995). Nature, 374, 131 ± 134. Wiman KG. (1993). FASEB J., 7, 841 ± 845. Morozov A, Shiyanov P, Barr E, Leiden JM and Raychaud- Wu X, Rubin M, Fan Z, DeBlasio T, Soos T, Ko A and huri P. (1997). J. Virol., 71, 3451 ± 3457. Mendelsohn J. (1996). Oncogene, 12, 1397 ± 1403. Munger K, Schener M, Huibregtse JM and Howley PM. zur Hausen H. (1996). Biochim. Biophys. Acta, 1288, F55 ± (1992). Cancer Surv., 12, 197 ± 217. F78. Ruan S, Okcu MF, Ren JP, Chiao P, Andree M, Levin V and Zhang W. (1998). Cancer Res., 58, 1538 ± 1543.

Oncogene