Cell Type–Dependent Effects of Polo-Like Kinase 1 Inhibition Compared with Targeted Polo Box Interference in Cancer Cell Lines

3189

Cell type–dependent effects of Polo-like kinase 1 inhibition compared with targeted polo box interference in cancer cell lines

Jenny Fink, Karl Sanders, Alexandra Rippl, certain cell lines but highly contrasting effects in others. Sylvia Finkernagel, Thomas L. Beckers, This may point to subtle differences in the molecular and Mathias Schmidt machinery of mitosis regulation in cancer cells. [Mol Cancer Ther 2007;6(12):3189–97] Nycomed GmbH, Konstanz, Germany Introduction Abstract Polo-like kinase 1 (Plk1) has been identified as key player Multiple critical roles within mitosis have been assigned for G2-M transition and mitotic progression in both normal to Polo-like kinase 1 (Plk1), making it an attractive and tumor cells (1). Multiple roles have been assigned to candidate for mitotic targeting of cancer cells. Plk1 Plk1 at the entry into M phase, mitotic spindle formation, contains two domains amenable for targeted interference: condensation and separation of chromosomes, exit from a kinase domain responsible for the enzymatic function mitosis by activation of the anaphase-promoting complex, and a polo box domain necessary for substrate recogni- and in cytokinesis (reviewed in ref. 2). Moreover, recent tion and subcellular localization. Here, we compare two reports implicated an involvement of Plk1 in the resump- approaches for targeted interference with Plk1 function, tion of cell cycle reentry after checkpoint activation through either by a Plk1 small-molecule enzyme inhibitor or by DNA-damaging agents (3). It is therefore not surprising inducible overexpression of the polo box in human cancer that targeted interference with Plk1, primarily by anti- cell lines. Inducible expression of the Plk1 polo box bodies, antisense, or small interfering RNA technology but resulted in growth inhibition of RKOp27 human colon also with small-molecule inhibitors (4, 5), has been reported adenocarcinoma cells without obvious signs of mitotic to result in a blockade in mitosis with subsequent induction abnormalities. A Plk1 kinase inhibitor in the same cell line of cell death (6–9). Some of these reports even describe a arrested cells in mitosis with subsequent onset of differential response of nontransformed cells toward apoptosis. Similarly, PC-3 human prostate cancer cells interference with Plk1 function as compared with tumor were growth inhibited on expression of the polo box. cells. Moreover, overexpression of Plk1 has been observed Prolonged expression of the polo box in these cells in a wide variety of carcinomas of different pathologic resulted in the occurrence of binucleated or multinucleat- origin, including breast (10), ovary, colon, pancreas, lung, ed cells. In contrast, U2OS human osteosarcoma cells endometrium, brain, skin, head and neck, esophagus, responded to overexpression of the polo box with a gastric tract, and prostate (reviewed in ref. 11). Some of massive mitotic accumulation coinciding with the onset these studies correlated Plk1 overexpression with a poor of apoptosis. Comparison of spindle formation revealed patient prognosis even in multivariate analyses (12). very similar mitotic abnormalities in polo box–over- Human Plk1 is composed of two functional domains: The expressing U2OS cells compared with U2OS cells treated NH2-terminal kinase domain is highly conserved among with the Plk1 kinase inhibitor. We conclude that interfer- polo-like kinase family members. The canonical consensus ence with polo box function and inhibition of Plk1 kinase sequence for protein kinases in the ATP-binding subdo- activity can exert very similar phenotypic effects in main I (Gly-X-Gly-X-X-Gly-X-Val) is changed within the Plk1 kinase domain into Gly-X-Gly-Gly-X-Ala-X-C. The COOH-terminal region contains two polo boxes, each being 60 to 70 amino acids in length. Despite only a low amino acid sequence identity (12%), both polo boxdomains (PBD) Received 1/23/07; revised 8/16/07; accepted 10/9/07. form an intramolecular dimer with identical folds of a six- The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked stranded h-sheet and an a-helix(13). Together with the advertisement in accordance with 18 U.S.C. Section 1734 solelyto region in between them and a linker between the kinase indicate this fact. domain and the first polo box, it functions as a phospho- Note: Current address for J. Fink: Institut Curie, Centre National de la Recherche Scientifique, Unite´Mixte de Recherche 144, 26 rue d’Ulm, serine/threonine binding module (13, 14). Using a phos- 75248 Paris, France. Current address for T.L. Beckers: Oncotest GmbH, pho-peptide library–based proteomic screen, the optimal am Flughafen 12-14, D-79108 Freiburg, Germany. recognition sequence has been identified as Ser-[pSer/ Requests for reprints: Mathias Schmidt, Nycomed GmbH, RPD/SO, pThr]-[Pro/X] (15). Many of the identified Plk1 substrates Byk-Gulden-Strasse 2, D-78467 Konstanz, Germany. Phone: 49-7531-842945; Fax: 49-7531-8492945. contain this consensus motif (16). The current polo box E-mail: [email protected] working model suggests that priming kinases such as Copyright C 2007 American Association for Cancer Research. mitogen-activated protein kinases, cyclin-dependent doi:10.1158/1535-7163.MCT-07-0048 kinases, and other mitotic kinases generate priming

Mol Cancer Ther 2007;6(12). December 2007

phosphorylation sites, which are bound by the polo boxof concentration twice as high as the final concentration in the Plk1. Thus, Plk1 is brought into close proximity to its test. RKOp27 and PC-3 cells were seeded into 96-well flat- substrate, which facilitates subsequent phosphorylation bottomed plates at a density of 5,000 per well in a volume of (17–19). It has to be noted that a priming phosphorylation 50 AL/well. To determine the effect on arrested RKOp27 of Plk1 substrates is not absolutely mandatory because cells, these cells were seeded at a density of 15,000 per well in bacterially expressed unphosphorylated substrates of Plk1 medium containing 10 Amol/L ponasterone A. Twenty-four can also be phosphorylated by Plk1 (20). Several studies hours after seeding, 50 AL of each compound dilution in have shown that this motif recognition model is also crucial DMEM were added per well with 1% DMSO–treated cells as for the targeting of Plk1 to specific subcellular locations a control. After incubation with the compounds for 72 h at (21–23). Interestingly, Sak, the fourth member of the polo- 37jC, 10 AL of Alamar Blue solution (Biosource) were added like kinase family, contains only one polo box. Crystal and the fluorescence was measured (excitation, 544 nm; structure analyses of the polo boxmotif have shown that emission, 590 nm). For calculation of cell viability, the the Sak polo boxforms a homodimer in vitro and in vivo emission value from untreated cells was set as 100% viability and localizes to centrosomes and the cleavage furrow and the emission rates of treated cells were set in relation to during cytokinesis (24). the values of untreated cells. The corresponding IC50 values Due to its structural hallmarks, its overexpression in were determined from the concentration-effect curves. various cancer types, and the intrinsic dependence of Flow Cytometric Analysis tumor cells on its activity in mitosis, Plk1 represents an After completion of the desired treatment, cells were attractive candidate for molecular targeted cancer therapy. harvested by trypsinization, and an aliquot of 1 Â 106 cells It is conceivable that specific inhibitors of polo-like kinases was washed once with cold PBS and then fixed with cold may be developed through targeting either the kinase 70% ethanol. The DNA was stained with a solution con- domain or the PBD. taining 25 Ag/mL propidium iodide and 10 Ag/mL RNase In this study, we compared the physiologic consequences A in PBS for 6 h. Cell cycle distribution was analyzed with of Plk1 kinase inhibition with a small-molecule inhibitor (4) a FACScan flow cytometer (Becton Dickinson) at an exci- and the disruption of polo boxfunction through inducible tation wavelength of 488 nm. ectopic overexpression of the polo box in three different Cell Staining cancer cell lines. We observed that interference with polo On completion of the desired treatment, cells were boxfunction and inhibition of Plk1 kinase activity can exert harvested by trypsinization. An aliquot (5 Â 103 cells) very similar phenotypic effects in certain cancer cells. Yet was spun onto glass slides using a cytospin and subse- distinct differences were observed with respect to respon- quently fixed and stained with the Hemacolor kit according siveness toward these two forms of targeted Plk1 interfer- to the manufacturer’s instructions. For the staining of ence in other cell types. microtubules, cells were fixed with 4% paraformaldehyde in PBS for 5 min, washed with PBS, and permeabilized with 0.2% Triton X-100 for 5 min. After blocking with 2% bovine Materials and Methods serum albumin for 30 min, the cells were incubated with an Materials and all other reagents were purchased from antibody against a-tubulin for 1 h at 37jC. Binding was Sigma Chemical unless otherwise specified. The Hemacolor visualized by incubation with a Cy3-labeled antimouse staining kit was obtained from Merck; antibodies were antibody (Dianova). The slides were washed with PBS obtained from various suppliers: Plk1 (Santa Cruz Biotech- between each antibody incubation step and DNA was nology or Cell Signaling Technology); h-actin and a-tubulin counterstained for 5 min with 100 ng/mL Hoechst 33342 (Sigma); Myt-1, cleaved caspase-3, and phospho-histone (Invitrogen) followed by final rinsing with deionized water. H3 (Cell Signaling Technology); and V5 (Invitrogen). The The slides were covered with Vectashield and sealed before Fugene 6 transfection reagent was obtained from Roche. analysis by confocal microscopy. Cells and Cell Cultures Immunoblot Analysis RKO human colon cancer cells, PC-3 human prostate Cells were lysed in a lysis buffer containing 50 mmol/L cancer cells, and U2OS human osteosarcoma cells were Tris (pH 7.4), 150 mmol/L NaCl, 1% NP40, 50 mmol/L obtained from LCG Promochem and maintained in DMEM NaF, 1 mmol/L Na3VO4, 1 mmol/L phenylmethylsulfonyl supplemented with 10% FCS. RKOp27 cells (25) were fluoride. The lysates were cleared by centrifugation and cultured in double-selection culture medium containing the supernatants were collected. Equal amounts of lysate 200 Ag/mL Zeocin and 500 Ag/mL neomycin (G418). protein were separated by SDS-PAGE and subsequently CellViabilityAssay blotted onto polyvinylidene difluoride membranes for Metabolic activity correlating with cell proliferation/ incubation with antibodies used for Western blot analysis viability was quantified with the Alamar Blue assay as as indicated. Specific signals were visualized by use of a described (26) using PC-3 prostate carcinoma or RKOp27 chemiluminescence detection kit (Roche). colon carcinoma cell lines (25). The compounds were Cloning of PBDExpressionVectors dissolved in DMSO (20 mmol/L) and subsequently diluted Three different expression vectors were generated to in semilogarithmic steps. DMSO dilutions were further transiently or stably express the PBD of Plk1 in cancer cells: diluted 1:100 into DMEM containing 10% FCS to a final pcDNA5/TO V5-HisA-PBD1 consists of amino acids 326 to

Mol Cancer Ther 2007;6(12). December 2007

603 of human Plk1 fused to a COOH-terminal V5 antibody epitope and a 6Â His cluster (Invitrogen). The sequence corresponding to amino acids 326 to 603 of human Plk1 was amplified using PCR techniques from RKOp27 cDNA and inserted into pIND/V5-HisA via HindIII/XhoI restriction sites included in the PCR primers. The 5¶ primer also contained a 5¶-gccgccacc-3¶ Kozak sequence and a start codon. The sequence encoding Plk1 amino acids 326 to 603 fused to the V5-His epitope was subsequently excised via HindIII/PmeI restriction sites and inserted into pcDNA5/ TO via the same restriction sites. pcDNA5/TO V5-HisA- PBD2 is the same as pcDNA5/TO V5-HisA-PBD1 except that the sequences encoding amino acids 337 and 340 of human Plk1 were exchanged into alanines by PCR-directed mutagenesis. For the generation of pCDNA5/TO-PBD3, pcDNA5/TO V5-HisA-PBD2 was used as a template for a two-step PCR in which a stop codon has been inserted directly after the XhoI restriction site to omit the fusion of the polo boxwith the V5 epitope and the histidine tag. Additionally, a nuclear localization sequence derived from the SV40 large T antigen (PKKKRKV) has been inserted at the 5¶ end directly following the ATG start codon via the 5¶ PCR primer. Generation of Plk1 PBD ^ Overexpressing Cell Lines Master cell lines of RKOp27, PC-3, and U2OS were generated by transfecting them with pcDNA6/TR (Invi- trogen) using the Fugene 6 kit (Roche Diagnostics). Stable clones were selected and cultured in medium containing 10 Ag/mL blasticidin S (Invitrogen) and expression of the tet repressor in single clones was analyzed by Western blotting with a tet repressor specific antibody (Mobitec). Subsequently, clones of each cell line stably expressing the Figure 1. Interference with Plk1 function in RKOp27 cells byinducible tet repressor were transfected with pcDNA5/TO V5-HisA- overexpression of the Plk1 PBD or bypharmacologic inhibition of Plk1 PBD1 (RKOp27), pcDNA5/TO V5-HisA-PBD2 (PC-3), or kinase activity. A, growth rates of two representative RKOp27 PBD clones pcDNA5/TO-PBD3 (U2OS). Stable clones were selected in with (filled circles) or without (open triangles) induced expression of the Â 4 medium containing 10 Ag/mL blasticidin S and 50 Ag/mL PBD. Cells were seeded out at a densityof 2 10 per six-well plate. Cell numbers were determined with a Vicell counter at the indicated time hygromycin B (Roche). Inducible expression of the polo box points. The PBD expression was induced with 2 Ag/mL doxycycline (Dox) was examined by Western blot analysis with a V5 epitope– and detected byWestern blotting of uninduced ( À) and induced (+) specific antibody (RKOp27 and PC-3) or with a Plk1- lysates with an antibody against the V5 epitope (inset). B, methylene blue/eosin staining 6 d after induction of the Plk1 PBD expression (top) specific antibody (U2OS) of lysates from cells induced with and confocal analysis of the mitotic spindles after 48 h of induced PBD 2 Ag/mL doxycycline for 24 h. Control lysates were taken expression (bottom). The cells were fixed and stained with antibodies from uninduced cells. against a-tubulin to visualize mitotic spindles byconfocal microscopy. Counterstaining of the DNA was conducted with Hoechst 33342. C, cell cycle – dependent cytotoxicity of compound 1, an inhibitor of Plk1 kinase activity. RKOp27 cells were seeded out as described in Materials and Results Methods. After 24 h, different concentrations (as indicated) were added to Interference with Plk1 Function in RKOp27 Human arrested (filled boxes) or proliferating (open triangles) RKOp27 cells. Cell viabilitywas determined 72 h later and calculated as percentage of Colon Adenocarcinoma Cells untreated control cells. We used the TRexsystem to generate a system for the inducible expression of the human Plk1 PBD (amino acids 326–603) in RKOp27 cells. The tet-induced expression in read-outs turned out to be altered on induced expression of two representative clones is shown in Fig. 1A. Several the polo box, we observed an f50% decreased growth rate morphologic, molecular, and cellular markers were ana- in cells with induced expression of the PBD (Fig. 1A). Cell lyzed to investigate the effects of Plk1 PBD overexpression cycle analyses showed that the S-phase content was in RKOp27 cells. These include altered phospho-histone H3 consistently decreased in induced clones compared with content, confocal analyses of mitotic spindles, methylene that in uninduced cells. To compare the effect of induced blue/eosin staining to reveal a potentially increased mitotic overexpression of the Plk1 PBD with that of pharmacologic indexor the occurrence of multinucleated cells, and inhibition of the kinase domain, we used a small-molecule induction of apoptosis. Whereas neither of the mentioned inhibitor of Plk1 (compound 1; refs. 4, 5). RKOp27 cells

Mol Cancer Ther 2007;6(12). December 2007

with inducible expression of the cell division cycle kinase resistance to antimitotic agents like paclitaxel. We used the inhibitor p27kip1 (25) have successfully been used to RKOp27 system to test the cell cycle–dependent cytotox- investigate cell cycle–dependent cytotoxicity of known icity of compound 1. As expected for inhibitors of Plk1, and investigative chemotherapeutic agents (27, 28). Induc- compound 1 is only cytotoxic for proliferating RKOp27 kip1 tion of p27 arrests proliferating RKO cells in the G1 cells (Fig. 1B, open triangles), but not for RKO cells that phase of the cell division cycle, thereby inducing complete have been arrested in the cell cycle by the induced expression of p27kip1 (Fig. 1B, filled squares). We have previously shown that compound 1 is able to induce apoptosis in these cells (4), and therefore the reduced viability observed after treatment with the Plk1 inhibitor is attributable to induction of apoptosis rather than to a profound growth delay. From these data, we conclude that interference with Plk1 function by overexpression of the PBD has a milder effect on the cell physiology of RKOp27 cells (growth retardation) than pharmacologic inhibition of Plk1 kinase activity, which exerts cytotoxicity to proliferating, but not arrested, cells. Interference with Plk1 Function in PC-3 Human Prostate Cancer Cells We extended our studies to PC-3 human prostate cancer cells in which we established a different inducible expression system for the Plk1 PBD. In contrast to RKOp27 cells, we mutated amino acids 337 and 340 of the D-box sequence into alanines to prevent degradation of the ectopically expressed PBD at the end of mitosis (29) and thereby to potentially obtain more sustained effects. The doxycycline-induced expression in four representative clones is shown in Fig. 2A and their growth properties in Fig. 2B. One clone (clone 4) with extremely low inducible expression of the PBD was also included in the studies. A 60% to 70% growth reduction was observed in all clones except clone 4 after 10 days of induced ectopic overexpression of the Plk1 PBD. Methylene blue/eosin staining of a representative clone (clone 9) after 7 days of induced PBD expression revealed the occurrence of a large proportion of binucleated cells (>40%) and some cells with abnormal mitoses (Fig. 2C). No other cell cycle alterations or signs of mitotic accumulation were observed (data not shown). In contrast, treatment of PC-3 cells for 24 h with

Figure 2. Interference with Plk1 function in PC-3 cells byinducible overexpression of the Plk1 PBD or bypharmacologic inhibition of Plk1 kinase activity. A, Western blot analysis of lysates from four uninduced (À) and induced (+) clones with an antibodyagainst Plk1 (Santa Cruz Biotechnology). Note the almost undetectable inducible expression in clone 4. B, growth inhibition of the clones as used in A on inducible expression of the Plk1 PBD. Cells were seeded out at a densityof 2 Â 104 per six-well plate. Cell numbers were determined with a Vicell counter at the indicated time points and percentage of growth reduction was calculated compared with the uninduced clones. The PBD expression was induced as described in Fig. 1. C, methylene blue/eosin staining of clone 9 seven days after induction of the Plk1 PBD expression. Note the occurrence of a large proportion of cells with two nuclei. D, methylene blue/eosin staining of PC-3 cells after 24 h of treatment with the Plk1 kinase inhibitor compound 1. More than one third of the cells were accumulated in mitosis with randomlydispersed condensated chromo- somal material. E, cytotoxicity of compound 1, an inhibitor of Plk1 kinase activityon PC-3 cells. The cells were seeded out as described in Materials and Methods. After 24 h, different concentrations of compound 1 (as indicated) were added to the cells. Cell viabilitywas determined 72 h later and calculated as percentage of untreated control cells.

Mol Cancer Ther 2007;6(12). December 2007

when incubated for 72 h (Fig. 2E). These data again reveal a relatively mild phenotypic effect of induced expression of the Plk1 PBD in PC-3 cells, in contrast to inhibition of the Plk1 kinase activity that leads to mitotic accumulation and subsequent onset of cell death. Mitotic Phenotype of Plk1 PBD Overexpression in U2OS Osteosarcoma Cells Our results obtained with inducible expression of the PBD in RKO and PC-3 cells are in stark contrast to observations made by Hanisch et al. (30) and Seong et al. (23), who reported a significant mitotic accumulation of HeLa and U2OS cells, respectively, on inducible expression of the Plk1 PBD. To investigate whether cell type– specific factors may account for these differences, we also examined U2OS cells. The expression vector chosen was almost identical to the one used in PC-3 cells, with the exception that a stop codon was inserted after amino acid 603 to omit fusion with the V5/His epitope included in the vector and a nuclear localization sequence was inserted. Two independent clones were investigated in detail (Fig. 3). Induction of Plk1 PBD expression coincided in both clones with a strong increase in the reactivity of lysates with phospho-histone H3 antibodies being indic- Figure 3. Western blot analysis of U2OS osteosarcoma cell lysates with ative of mitotic accumulation. Phospho-histone H3 reac- inducible expression of the Plk1 PBD. Top, first row, time course of PBD tivity was increased as soon as 18 h after induction of PBD expression in clone 5 (left) and clone 6 (right). Expression of the PBD was expression and sustained for 48 h. The subsequent detected in lysates of cells induced for different time points as indicated with antibodies against Plk1. Second row, changes in phospho-histone decrease in the phospho-histone H3 signal correlated with H3 reactivityof induced lysates. Fourth row, induction of apoptosis on an onset of apoptosis as detected with antibodies against prolonged induced expression of the PBD as detected with antibodies cleaved caspase-3. Moreover, we observed a strongly specific for cleaved caspase-3. Bottom, differential electrophoretic mobilityof Myt-1after induction of PBD expression. N., lysates of cells increased electrophoretic mobility of Myt-1 in lysates of treated with 80 ng/mL nocodazole for 24 h. Note the different time points induced cells similar to the one in nocodazole-treated of cell harvesting as compared with the top blots. Control cells were left U2OS cells. Obviously, induced expression of the Plk1 uninduced and were harvested 96 h after seeding. Western blots were PBD led to an accumulation of the cells in mitosis with reprobed with antibodies against h-actin or a-tubulin as indicated to check z loading of equal amounts of protein. subsequent onset of apoptosis on prolonged ( 48 h) induced PBD expression. Clones 5 and 6 were subjected to cell cycle analyses to compound 1 led to accumulation of cells in mitosis (mitotic determine the degree of G2-M accumulation and to index, 34%) with condensed chromatin randomly spread potentially observe additional alterations in the cell cycle throughout the cells (Fig. 2D). Moreover, compound 1 was distribution pattern (Fig. 4A). A significant G2-M accumu- f cytotoxic to PC-3 cells with an IC50 value of 500 nmol/L lation was observed in both clones after 24 and 48 h of PBD

Figure 4. Mitotic accumulation of U2OS cells after inducible expression of the Plk1 PBD. A, cells were seeded out and induced with 2 Ag/mL doxycycline for the time periods as indicated. Control cells were left untreated. Cell cycle analysis was conducted as described in Materials and Methods. B, methylene blue/eosin staining. Cells were treated as described in A and aliquots were subjected to methylene blue/eosin staining as described in Materials and Methods.

Mol Cancer Ther 2007;6(12). December 2007

expression. Induction of apoptosis became obvious at the (mitotic indexof 44%, with 62–67% monopolar spindles; latter time point as indicated by the increase in the sub-G1 Fig. 6B); however, we observed far less bipolar spindles. fraction of the samples. Hardly any viable cells were Overexpression of the PBD plus treatment with the Plk1 observed at 72 h of PBD expression. A considerable amount inhibitor did not further increase the mitotic index(39%); of cells with a DNA content of >4N (being indicative of cells were arrested almost exclusively with monopolar endoreplication) was also observed in clone 6, but only to a spindles (74–83%). minor degree in clone 5. Methylene blue/eosin staining Because Plk1 involvement in spindle checkpoint activa- (Fig. 4B) again confirmed that both clones displayed an tion is well described (31–34), we investigated whether increased mitotic indexon induced expression of the induction of the polo boxor inhibition of Plk1 kinase Plk1 PBD. activity would result in BubR1 phosphorylation, which Mitotic markers were also analyzed after inhibition of the occurs during spindle checkpoint activation (35). As shown kinase function by compound 1 (Fig. 5). There was a similar in Fig. 7, neither inducible expression of the PBD nor increase of phospho-histone H3 reactivity, mitotic shift of inhibition of Plk1 kinase activity resulted in phosphoryla- Myt-1, or activation of caspase-3 in cells induced to tion of BubR1, which is detectable by a lower electropho- overexpress the PBD compared with cells treated with the retic mobility, as observed on treatment with nocodazole, Plk1 inhibitor compound 1. Moreover, a combination of despite significant mitotic accumulation of the cells (see PBD overexpression plus treatment with compound 1 did Fig. 6B). Kinase assays with Plk1 and BubR1 as substrate not have any obvious additive effect. To compare both were conducted to exclude that BubR1 is a direct substrate treatment modalities in greater detail, confocal images were of Plk1 (data not shown). From these data, we conclude that taken in which the spindle apparatus was stained with both Plk1 kinase inhibition and PBD overexpression anti–a-tubulin antibodies (Fig. 6A). PBD overexpression similarly fail to induce phosphorylation of BubR1 despite increased the mitotic indexfrom 5% to 40% with the their capacity to induce a strong mitotic accumulation. occurrence of both monopolar (29–50%) and bipolar spindles. Incubation of the same cells with 2 Amol/L of the Plk1 inhibitor induced very similar mitotic phenotypes Discussion In this study, we investigated the molecular alterations of Plk1 inhibition in cancer cell lines either with inducible overexpression of the Plk1 PBD or with pharmacologic inhibition of Plk1 kinase activity. Strikingly similar phenotypes were observed in U2OS human osteosarcoma cells: both types of treatment led to a significant mitotic accumulation accompanied with subsequent onset of programmed cell death. A large proportion of the cells displayed monopolar spindles on kinase inhibition, whereas overexpression of the PBD resulted in the occurrence mitotic cells displaying monopolar or bipolar spindles. Although the role of Plk1 in the activation of the spindle checkpoint is well described (31–34), neither treatment, however, led to the induction of BubR1 phosphorylation, which is considered as one marker of spindle checkpoint activation. In kinase assays with human BubR1, we excluded that BubR1 is a direct substrate of Plk1. It has to be noted that phosphorylation of BubR1 alone may not be a reliable marker of spindle checkpoint activation and that investigation of other markers may be necessary to thoroughly investigate Figure 5. Western blot analysis of U2OS osteosarcoma cell lysates with whether or not Plk1 inhibition activates the spindle inducible expression of the Plk1 PBD after treatment with 2 Amol/L of checkpoint. The cellular phenotype observed after ectopic compound 1. The cells were treated for 24 h (Dox 24 and 38h)as expression of the PBD in U2OS cells contrasts the effects indicated. First row, Western blot analysis of induced PBD expression in clone 5 (left) and clone 6 (right) with an antibodyagainst the COOH observed in RKO and PC-3 cells. There, inhibition of Plk1 terminus of Plk1. Second row, increased phospho-histone H3 reactivityof kinase activity with compound 1 clearly induced a mitotic lysates with induced PBD expression or treated with compound 1 (Cpd. 1; phenotype with subsequent onset of cell death. Inducible Plk1 inhibitor). Third row, differential electrophoretic mobilityof Myt-1 after induction of PBD expression or treatment with compound 1. Last expression of the Plk1 PBD, however, only resulted in a row, induction of apoptosis on induced expression of the PBD or treatment growth delay in RKO and a growth delay plus the with compound 1 as detected with antibodies specific for cleaved occurrence of an increased number of binucleated cells in caspase-3. Noc., lysates of cells treated with 80 ng/mL nocodazole for PC-3. Although we have not experimentally shown that 24 h served as positive control for mitotic cells. Western blots were reprobed with antibodies against h-actin as indicated to check loading of overexpression of the PBD in these cells actually titers equal amounts of protein. away Plk1 from its subcellular localizations and substrates,

Mol Cancer Ther 2007;6(12). December 2007

Figure 6. Confocal images of U2OS cells after induced expression of the Plk1 PBD or treatment with the Plk1 kinase inhibitor compound 1. A, cells were treated for 24 h with 2 Ag/mL doxycycline to induce PBD expression, 2 Amol/L compound 1 to inhibit Plk1 kinase activity, or both as indicated. The cells were then fixed and stained with antibodies against a-tubulin to visualize mitotic spindles byconfocal microscopy.Counterstaining of the DNA was conducted with Hoechst 33342. B, determination of the mitotic index of clone 5 after the treatments as described in A.

we favor the hypothesis that cell type differences account ered as spindle checkpoint proficient, subtle defects in the for the differential response toward overexpression of the mitotic spindle checkpoint in U2OS cells, as discussed by PBD. To exclude that the subtle differences in the Schmidt and Medema (37), may account for these cell expression constructs used may be responsible of the cell type–specific differences. type–specific responses, we tested the functionality of all The molecular function of the polo boxhas been studied three constructs by transient transfections: all three in great detail through crystallographic analyses (13, 14), expression constructs were able to induce a significant substrate recognition experiments (15), or cell biological mitotic accumulation in U2OS cells but not in RKO or PC-3 studies (23, 29, 30). According to the current working model cells (data not shown). HeLa and U2OS cells seem to be (16), the PBD of Plk1 recognizes specific priming phosphor- especially well suited to study the phenotypic effects of ylation sites on proteins for subsequent binding, which PBD overexpression of both Plk1 (23, 30) and Plk3 (36). It results in a precise subcellular localization of Plk1. In a is therefore tempting to speculate that cell type–specific processive model, the kinase domain of Plk1 then phos- factors may codetermine the functional outcome of ectopic phorylates the same substrate the PBD binds to. In a PBD overexpression. Although U2OS is generally consid- distributive model, PBD binding only results in the correct localization, and additional substrates at these locations may then be phosphorylated by the kinase domain of Plk1. It has to be noted that both models are not mutually exclusive and examples exist for both scenarios. A most recent example for the phosphorylation-dependent regulation of the Plk1-substrate interaction is its association with polo box–interacting protein 1, which is greatly stabilized by Plk1-mediated phosphorylation (38). One conclusion from this model is that both domains of Plk1, the kinase domain and the PBD, are required for Plk1 to exert its function as master regulator of mitosis. Numerous experiments in the literature support the relevance of both domains. Because targeted interference with Plk1 is incompatible with faithful progression through mitosis Figure 7. Analysis of BubR1 phosphorylation after targeted interference (1, 2, 4, 6–9) and Plk1 overexpression has been observed with Plk1. U2OS cells with inducible expression of the Plk1 PBD were in various types of human cancers (10–12), this protein is of treated for 24 h with 80 ng/mL nocodazole, 2 Ag/mL doxycycline, or outstanding interest for molecular directed tumor therapy. A 2 mol/L compound 1. Lysates of cells were subjected to Western blotting With the recent publication of specific kinase inhibitors with antibodies against BubR1 (top)orh-actin (bottom) as control for equal loading. Left, clone 5; right, clone 6. Arrows, different electropho- of Plk1 (4, 5), it has now become feasible to compare retic mobilities of BubR1. the molecular and cellular differences between targeted

Mol Cancer Ther 2007;6(12). December 2007

interference with kinase inhibitors and interference with the 6. Lane HA, Nigg EA. Antibodymicroinjection reveals an essential role for human polo-like kinase 1 (Plk1) in the functional maturation of mitotic PBD domain. Because no cell-permeable Plk1 PBD–specific centrosomes. J Cell Biol 1996;135:1701 – 13. inhibitors are described yet, we overexpressed the PBD to 7. Liu X, Erikson RL. Polo-like kinase (Plk)1 depletion induces apoptosis in tether away endogenous Plk1 from its original subcellular cancer cells. Proc Natl Acad Sci U S A 2003;100:5789 – 94. localization sites and substrates. Because this approach has 8. Reagan-Shaw S, Ahmad N. Silencing of polo-like kinase (Plk) 1 via siRNA causes induction of apoptosis and impairment of mitosis machinery successfully been used to validate the PBD function (30), in human prostate cancer cells: implications for the treatment of prostate we also relied in our studies on this mode of Plk1 PBD cancer. FASEB J 2005;19:611 – 3. interference. As we saw clear cell type–dependent differ- 9. Guan R, Tapang P, Leverson JD, Albert D, Giranda VL, Luo Y. Small ences in the effects of kinase inhibition versus PBD interfering RNA-mediated Polo-like kinase 1 depletion preferentially reduces the survival of p53-defective, oncogenic transformed cells and interference, it is tempting to speculate that the two inhibits tumor growth in animals. Cancer Res 2005;65:2698 – 704. approaches are not mutually compatible in terms of 10. Weichert W, Kristiansen G, Winzer KJ, et al. Polo-like kinase isoforms targeted interference with Plk1 function. Apart from all in breast cancer: expression patterns and prognostic implications. potential experimental limitations (e.g., differential extent of Virchows Arch 2005;446:442 – 50. ectopic PBD expression in the three cell lines; see Supple- 11. Takai N, Hamanaka R, Yoshimatsu J, Miyakawa I. Polo-like kinases (Plks) and cancer. Oncogene 2005;24:287 – 91. mentary Fig. S1)1, at least in the cell systems we explored, 12. Weichert W, Denkert C, Schmidt M, et al. Polo-like kinase isoform inhibitors of the Plk1 kinase domain uniformly induced expression is a prognostic factor in ovarian carcinoma. Br J Cancer 2004; a mitotic phenotype, whereas ectopic overexpression of 90:815 – 21. the PBD induced such a phenotype only in U2OS cells. 13. Cheng KY, Lowe ED, Sinclair J, Nigg EA, Johnson LN. The crystal structure of the human polo-like kinase-1 polo box domain and its Interestingly, thus far, the only cell systems in which phospho-peptide complex. EMBO J 2003;22:5757 – 68. overexpression of the Plk1 PBD induces a mitotic phenotype 14. Elia AE, Rellos P, Haire LF, et al. The molecular basis for are U2OS and HeLa cells (16, 22, 30, 36, 38). To expand the phosphodependent substrate targeting and regulation of Plks bythe basis of investigated cell lines, we transiently transfected the Polo-box domain. Cell 2003;115:83 – 95. PBD expression construct used for the U2OS cell line into 15. Elia AE, CantleyLC, Yaffe MB. Proteomic screen finds pSer/pThr- binding domain localizing Plk1 to mitotic substrates. Science 2003;299: three more cell lines: A2780 ovarian cancer cells, HCT116, 1228 – 31. and HT29 human colon cancer cell lines. Whereas in the 16. LoweryDM, Lim D, Yaffe MB. Structure and function of Polo-like former two cell lines we did not observe any increased kinases. Oncogene 2005;24:248 – 59. mitotic index(Supplementary Fig. S2), 1 HT29 indeed 17. Nakajima H, Toyoshima-Morimoto F, Taniguchi E, Nishida E. Identification of a consensus motif for Plk (Polo-like kinase) phosphoryla- reacted with a statistically significant increase in the mitotic tion reveals Myt1 as a Plk1 substrate. J Biol Chem 2003;278:25277 – 80. index. Intriguingly, HT29 cells display subtle defects in the 18. Yarm FR. Plk phosphorylation regulates the microtubule-stabilizing mitotic spindle checkpoint and respond with an extremely protein TCTP. Mol Cell Biol 2002;2:6209 – 21. high mitotic accumulation on treatment with antimitotic 19. Zhou T, Aumais JP, Liu X, Yu-Lee LY, Erikson RL. A role for Plk1 agents such as Eg5 kinesin inhibitors (39). Whereas the phosphorylation of NudC in cytokinesis. Dev Cell 2003;5:127 – 38. precise reasons for the higher sensitivities of HT29 cells to 20. Kumagai A, DunphyWG. Purification and molecular cloning of Plx1, a Cdc25-regulatorykinase from Xenopus egg extracts. Science 1996;273: PBD overexpression remain to be elucidated, their deficien- 1377 – 80. cies in the mitotic checkpoint may provide first clues toward 21. Jang YJ, Lin CY, Ma S, Erikson RL. Functional studies on the role of the mechanistic causality. In summary, our results imply the C-terminal domain of mammalian polo-like kinase. Proc Natl Acad Sci that targeted interference with Plk1 function through kinase U S A 2002;99:1984 – 9. 22. Lee KS, Grenfell TZ, Yarm FR, Erikson RL. Mutation of the polo-box inhibition may be more broadly applicable than interference disrupts localization and mitotic functions of the mammalian polo kinase through the PBD. Plk. Proc Natl Acad Sci U S A 1998;95:9301 – 6. 23. Seong YS, Kamijo K, Lee JS, et al. A spindle checkpoint arrest and a Acknowledgments cytokinesis failure by the dominant-negative polo-box domain of Plk1 in We thank Birgitta Riede for excellent technical assistance. U2OS cells. J Biol Chem 2002;277:32282 – 93. 24. Leung GC, Hudson JW, Kozarova A, Davidson A, Dennis JW, Sicheri F. The Sak polo-box comprises a structural domain sufficient for mitotic References subcellular localization. Nat Struct Biol 2002;9:719 – 24. 1. Barr FA, Sillje HH, Nigg EA. Polo-like kinases and the orchestration of 25. Schmidt M, Lu Y, Liu B, Fang M, Mendelsohn J, Fan Z. Differential cell division. Nat Rev Mol Cell Biol 2004;5:429 – 40. modulation of paclitaxel-mediated apoptosis byp21waf1 and p27kip1. 2. van Vugt MA, Medema RH. Getting in and out of mitosis with Polo-like Oncogene 2000;19:2423 – 9. kinase-1. Oncogene 2005;24:2844 – 59. 26. O’Brian J, Wilson I, Orton T, Pognan F. Investigation of the Alamar 3. van Vugt MA, Bras A, Medema RH. Polo-like kinase-1 controls recovery Blue (resazurin) fluorescent dye for the assessment of mammalian cell from a G2 DNA damage-induced arrest in mammalian cells. Mol Cell 2004; toxicity. Eur J Biochem 2000;267:5421 – 8. 15:799 – 811. 27. Schmidt M, Lu Y, Parant JM, et al. Differential roles of p21(Waf1) and 4. Schmidt M, Hofmann HP, Sanders K, Sczakiel G, Beckers TL, Gekeler V. p27(Kip1) in modulating chemosensitivityand their possible application in Molecular alterations after Polo-like kinase 1 mRNA suppression versus drug discoverystudies. Mol Pharmacol 2001;60:900 – 6. pharmacologic inhibition in cancer cells. Mol Cancer Ther 2006;5:809 – 17. 28. Mahboobi S, Sellmer A, Hocher H, et al. [4-(Imidazol-1-yl)thiazol-2- 5. McInnes C, Mezna M, Fischer PM. Progress in the discoveryof polo-like yl]phenylamines. A novel class of highly potent colchicine site binding kinase inhibitors. Curr Top Med Chem 2005;5:181 – 97. tubulin inhibitors: synthesis and cytotoxic activity on selected human cancer cell lines. J Med Chem 2006;49:5769 – 76. 29. Lindon C, Pines J. Ordered proteolysis in anaphase inactivates Plk1 to contribute to proper mitotic exit in human cells. J Cell Biol 2004;164: 1 233 – 41. Supplementary material for this article is available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/). 30. Hanisch A, Wehner A, Nigg EA, Sillje HH. Different Plk1 functions

Mol Cancer Ther 2007;6(12). December 2007

show distinct dependencies on Polo-Box domain-mediated targeting. Mol 35. Chan GK, Jablonski SA, Sudakin V, Hittle JC, Yen TJ. Human BUBR1 is Biol Cell 2006;17:448 – 59. a mitotic checkpoint kinase that monitors CENP-E functions at kinetochores and binds the cyclosome/APC. J Cell Biol 1999;146:941 – 54. 31. van Vugt MA, van de Weerdt BC, Vader G, et al. Polo-like kinase-1 is required for bipolar spindle formation but is dispensable for anaphase 36. Jiang N, Wang X, Jhanwar-Uniyal M, Darzynkiewicz Z, Dai W. Polo promoting complex/Cdc20 activation and initiation of cytokinesis. J Biol box domain of Plk3 functions as a centrosome localization signal, Chem 2004;279:36841 – 54. overexpression of which causes mitotic arrest, cytokinesis defects, and apoptosis. J Biol Chem 2006;281:10577 – 82. 32. Ahonen LJ, Kallio MJ, Daum JR, et al. Polo-like kinase 1 creates the tension-sensing 3F3/2 phosphoepitope and modulates the association of 37. Schmidt M, Medema RH. Exploiting the compromised spindle spindle-checkpoint proteins at kinetochores. Curr Biol 2005;15:1078 – 89. assemblycheckpoint function of tumor cells: dawn on the horizon? Cell Cycle 2006;5:159 – 63. 33. Sumara I, Gimenez-Abian JF, Gerlich D, et al. Roles of polo-like kinase 1 in the assemblyof functional mitotic spindles. Curr Biol 2004;14: 38. Kang YH, Park JE, Yu LR, et al. Self-regulated Plk1 recruitment to 1712 – 22. kinetochores bythe Plk1-PBIP1 interaction is critical for proper chromo- some segregation. Mol Cell 2006;24:409 – 22. 34. Moshe Y, Boulaire J, Pagano M, Hershko A. Role of Polo-like kinase in the degradation of earlymitotic inhibitor 1, a regulator of the anaphase 39. Tao W, South VJ, Zhang Y, et al. Induction of apoptosis byan promoting complex/cyclosome. Proc Natl Acad Sci U S A 2004;101: inhibitor of the mitotic kinesin KSP requires both activation of the spindle 7937 – 42. assemblycheckpoint and mitotic slippage. Cancer Cell 2005;8:49 – 59.

Mol Cancer Ther 2007;6(12). December 2007

Downloaded from mct.aacrjournals.org on September 24, 2021. © 2007 American Association for Cancer Research. Cell type−dependent effects of Polo-like kinase 1 inhibition compared with targeted polo box interference in cancer cell lines

Jenny Fink, Karl Sanders, Alexandra Rippl, et al.

Mol Cancer Ther 2007;6:3189-3197.

Updated version Access the most recent version of this article at: http://mct.aacrjournals.org/content/6/12/3189

Cited articles This article cites 39 articles, 17 of which you can access for free at: http://mct.aacrjournals.org/content/6/12/3189.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/6/12/3189.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/6/12/3189. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.