Cancer Therapy: Preclinical

Synergism between and 17-AAG in Leukemia Cells: Critical Roles for , FLT3, Topoisomerase II, Chk1, and Rad51 QingYao,1Brenda Weigel,2 andJohn Kersey1, 2

Abstract Purpose: DNA-damaging agents, such as etoposide, while clinically useful in leukemia therapy, are limited by DNA repair pathways that are not well understood. 17-(Allylamino)-17- demethoxygeldanamycin (17-AAG), an inhibitor ofthe molecular 90 (Hsp90), inhibits growth and induces in FLT3+ leukemia cells. In this study, we evaluated the effects of etoposide and17-AAG in leukemia cells and the roles of Hsp90, FMS-like tyrosine kinase 3 (FLT3),checkpoint kinase1 (Chk1),Rad51,and topoisomerase IIinthis inhibition. Experimental Design: The single and combined effects of 17-AAG and etoposide and the mechanism ofthese effects were evaluated. FLT3 and the DNA repair-related proteins, Chk1and Rad51, were studied in small interfering RNA (siRNA)^ induced cell growth inhibition experi- ments in human leukemia cells with wild-type or mutated FLT3. Results: We found that etoposide and the Hsp90/FLT3 inhibitor 17-AAG, had synergistic inhib- itory effects on FLT3+ MLL-fusion gene leukemia cells. Cells with an internal tandem duplication (ITD) FLT3 (Molm13and MV4;11)were more sensitive to etoposide/17-AAG than leukemias with wild-type FLT3 (HPB-Null and RS4;11). A critical role for FLT3 was shown in experiments with FLT3 ligand and siRNA targeted to FLT3. An important role for topoisomerase II and the DNA repair-related proteins, Chk1and Rad51,in the synergistic effects was suggested from the results. Conclusions: The repair ofpotentially lethal DNA damage by etoposide in leukemia cells is dependent on intact and functioning FLT3 especially leukemias with ITD-FLT3. These data suggest a rational therapeutic strategy for FLT3+ leukemias that combines etoposide or other DNA-damaging agents with Hsp90/FLT3 inhibitors such as 17-AAG.

Combination remains the primary treatment level of the Rad51 protein positively correlates with cellular eto- modality for cancer patients. DNA-damaging agents, such as poside resistance, and overexpression of Rad51 confers etoposide etoposide, are currently in clinical use for therapy of leukemia. resistance (7). Checkpoint kinase 1 (Chk1) is an important Etoposide, a topoisomerase II inhibitor, induces double strand regulator of genome maintenance by the HRR system. Chk1 in- breaks (DSB) during DNA replication (1). DSBs can initiate a teracts with Rad51; Chk1-depleted cells fail to form Rad51 nuc- strong proapoptotic signal when damaged DNA is left unre- lear repair foci after hydroxyurea-induced DSBs and inhibition of paired as cell survival relies on the efficiency of DNA repair (2). DNA replication (8). Chk1 can also be activated by a number To prevent DNA damage–induced apoptosis, the breaks must be of replication inhibitors, including etoposide (9–13). Chk1 repaired by different components of the DNA repair machinery gene deletion sensitizes cells to DNA replication inhibitors (13, (3, 4). In proliferating cells, homologous recombination DNA 14–16). Collectively, it is possible that disrupting Chk1-Rad51 repair (HRR) predominates, whereas quiescent cells use nonho- signaling might enhance etoposide cytotoxicity in tumor cells. mologous end joining (5). The major enzymatic component of FMS-like tyrosine kinase 3 (FLT3), a receptor tyrosine kinase HRR in eukaryotic cells is Rad51, as cells deficient in Rad51 with an important role in early hematopoiesis, is expressed and accumulate DSBs after replication or at stalled replication forks often mutated in human leukemias. Patients with acute myelo- (6). Etoposide induces increased expression of Rad51 (6). The genous leukemia (AML) harboring internal tandem duplications (ITD) of the FLT3 receptor have a poor prognosis compared with patients lacking such mutations. The ITD mutations of FLT3 occur in f30% of de novo AML (17, 18), including many of the 1 Authors’Affiliations: The Cancer Center, University ofMinnesota MMC 806, 420 subgroup of AML that also have MLL fusion genes (19, 20). We Delaware St. SE, and 2 Division ofPediatric Hematology-Oncology and Bone Marrow Transplantation, University ofMinnesota, Minneapolis, Minnesota and others have previously shown that FLT3 is a client of heat Received 7/18/06; revised 10/18/06; accepted 10/19/06. shock protein 90 (Hsp90) (20, 21). The Hsp90 complex is inhi- Grant support: National Cancer Institute (CA087053) (J. Kersey) and the bited by the benzoquinone ansamycin antibiotic, Childrens Cancer Research Fund (J. Kersey). and its derivatives, including 17-(allylamino)-17-demethoxygel- The costs ofpublication ofthis article were defrayed in part by the payment ofpage danamycin (17-AAG; refs. 22, 23). Cells with ITD-FLT3 are charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. particularly sensitive to the geldanamycin family of Hsp90 inhi- Requests for reprints: John H. Kersey, MMC 806, 420 Delaware St. SE, bitors. 17-AAG has entered clinical trials in patients with solid Minneapolis, MN 55455. Phone: 612-625-4659; Fax: 612-626-3069; E-mail: tumors (23, 24) and leukemia, including those with ITD-FLT3.3 [email protected]. F 2007 American Association for Cancer Research. doi:10.1158/1078-0432.CCR-06-1750 3 Unpublished data.

www.aacrjournals.org 1591 Clin Cancer Res 2007;13(5) March 1, 2007 Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2007 American Association for Cancer Research. Cancer Therapy: Preclinical

Table 1. Inhibition of proliferation of leukemia cell lines by 17-AAG and etoposide correlates with mutation/ expression of FLT3

Cell lines Lineage FLT3 status MLL fusion gene IC50 of 17-AAG (nmol/L) IC50 of etoposide (nmol/L) Molm13Myeloid ITD Yes 31 F 338F 9 MV4;11 Myeloid ITD Yes 40 F 246F 4 RS4;11 Pro-B Wild-type Yes 700 F 52 102 F 19 HPB-NULL Pre-B Wild-type No 470 F 3673F 8 THP-1 Myeloid Wild-type Yes 800 F 48 855 F 58 KM3Myeloid Wild-type No 2200 F 164 124 F 17 IE8 Pre-B Negative No 2100 F 196 111 F 24 U937 Myeloid Negative No 4500 F 321 11538 F 458

Abbreviation: ITD; FLT3internal tandem duplication mutation.

Previous research has shown an important role for other Materials and Methods receptor tyrosine kinases, the epidermal growth factor receptor (EGFR; refs. 25–27) and the insulin-like growth factor receptor Reagents. 17-AAG was provided by the Developmental Therapeu- (IGF-IR; refs. 28, 29) in DNA repair. In the present studies, we tics Program of the National Cancer Institute. Both 17-AAG and found that MLL-fusion gene leukemia cell lines with ITD-FLT3 etoposide were dissolved in DMSO to a 10 mmol/L stock solution. had reduced cellular levels of the DNA repair proteins Chk1 ‘‘SMARTpool’’ for human FLT3, a combination of four synthetic siRNA and Rad51 and were more sensitive to etoposide in cell growth duplexes each of which was designed to reduce mRNA for human FLT3, experiments than leukemias with wild-type FLT3 (HPB-Null human Chk1, human Rad51, and nonsilencing negative control siRNA, and RS4;11). Furthermore, etoposide, in combination with were obtained from Dharmacon (Lafayette, CO). the Hsp90/FLT3 inhibitor 17-AAG, had synergistic inhibitory Cell culture and cell proliferation assay. Human leukemia cell lines, effects in FLT3+ cells. These studies raise the possibility that MV4;11, Molm13, HPB-Null, and RS4;11 were described previously (20). MV4;11 and Molm13 have previously been found to have ITD-FLT3 DNA repair may be dependent on FLT3 in FLT3+ leukemia cells. mutations, whereas RS4;11 and HPB-Null have wild-type FLT3 as In this study, we used small interfering RNA (siRNA) targeted to shown in Table 1. The cell lines were maintained in RPMI 1640 FLT3, Chk1, and Rad51 and showed a critical role for each of containing 10% fetal bovine serum. For cell proliferation assays using these molecules in the synergistic effects of etoposide and the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) Hsp90/FLT3 inhibitor, 17-AAG. method, cells were seeded in replicates of six at a density of 1 Â 105

Fig. 1. Combination effects of 17-AAG and etoposide in human leukemia cell lines.The median effects of drug combination for MV4;11, Molm13, RS4;11and HPB-Null are shown. Cells were incubated with 17-AAG, etoposide, or both for 72 h, and proliferation was assessed using MTTassays. The experiments were repeated at least thrice.

Clin Cancer Res 2007;13(5) March 1, 2007 1592 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2007 American Association for Cancer Research. Synergistic Effects of Hsp90 Inhibitor and Etoposide

Fig. 2. The effects of FLT3 siRNA, 17-AAG, or etoposide alone and the combination oftwo agents on RS4;11cells. A, growth ofRS4;11cells transfected with controlled siRNA or FLT3 siRNA in the presence or absence ofFL. B, Western blot analysis showed that FLT3, Chk1, and Rad51expression and phosphorylation levels in RS4;11cells transfected with control siRNA or FLT3 siRNA affected by FL at 6 h. C, RS4;11cells were treated with 1000 nmol/L 17-AAG,1000nmol/Letoposide,andthe combination of1000 nmol/L 17-AAG and 1000 nmol/L etoposide in the presence of 100 ng/mL FLT3-ligand. Cell growth ofRS4; 11cells with different treatment with or without FLT3-ligand at days 0, 1, 2, and 3. D, Western blot showed the total level ofFLT3, p-Chk1, Chk1,Rad51,andPARPafterRS4;11cellswere treated for 48 h. The experiments were repeated thrice.

cells/mL in 96-well plates in the presence or absence of various concen- Upstate Biotechnology, Inc. (Lake Placid, NY); and actin from Sigma trations of 17-AAG and etoposide. Plates were incubated for the indicated (St. Louis, MO). For detection, the blots were incubated with horseradish hours at 37jC, 5% CO2, and proliferation was measured using Cell Titer peroxidase–conjugated anti-immunoglobulin G antibody (Promega) 96 Aqueous Non-radioactive Cell Proliferation Assay (Promega, Madi- and developed using the enhanced chemiluminescence detection system son, WI). Results are reported as means F SE for each point on the curve. (Amersham Pharmacia Biotech, Arlington Heights, IL). The densitomet- Transfection with siRNA. The cells were washed twice and resus- ric analysis was done using Molecular Analyst (Bio-Rad, Hercules, CA). pended in OptiMEM medium (Invitrogen, San Diego, CA). Five million Analysis of combined effects of drugs. Each of the cell lines were cells were then mixed with 10 Ag of siRNA duplexes. Electroporation simultaneously exposed to 17-AAG and etoposide for 72 h, and was done using a BTX Electro Square Porator ECM830 (BTX, Holliston, inhibition of cell proliferation was determined by Cell Titer 96 Aqueous MA) with one pulse at 240 V. After electroporation, cells were Non-radioactive Cell Proliferation Assay. To calculate combined drug immediately mixed with RPMI 1640 containing 10% fetal bovine effects, the combination index (CI) method was used. The CI values serum. Cell growth was calculated with the viable cell numbers detected were determined at different effect levels of growth inhibition as a by trypan blue exclusion test. quantitative measure of the degree of drug interaction using the Western blotting. Cells were lysed by sample buffer [62.5 mmol/L computer software CalcuSyn (BioSoft, Ferguson, MO). Combined drugs Tris (pH, 7.4), 2% SDS, 10% glycerol, and the protease inhibitor were used at fixed molar ratios. The drug doses that produced about ‘‘Complete Mini’’ (Roche, Indianapolis, IN)], and the cell lysate was 50% of growth inhibition in single drug experiments were chosen to clarified by centrifugation. About 20 Ag of cell lysate of each sample were determine an appropriate fixed molar ratio of two combined drugs. electrophoresed by 12% SDS-PAGE, transferred to a nitrocellulose Detection of apoptosis. Cells cultured under the various conditions membrane, and immunoblotted with different antibodies. Antibodies were harvested and labeled using the CaspaTag caspase activity kit were purchased as follows: Hsp90 and Hsp70 from BD PharMingen (San (Serologicals Corporation, Norcross, GA). Briefly, harvested cells were Diego, CA); FLT3, Chk1, Rad51, and poly(ADP-ribose) polymerase incubated with FAM-VAD-FMK that irreversibly binds to caspases-1 to (PARP) from Santa Cruz Biotechnologies, Inc. (Santa Cruz, CA); p-Chk1 caspase-9. After two washes, propidium iodide (PI) was added, and cells (S317) from Cell Signaling Technology (Beverly, MA); g-H2A.X from were analyzed on a FACS Calibur flow cytometer.

www.aacrjournals.org 1593 Clin Cancer Res 2007;13(5) March 1, 2007 Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2007 American Association for Cancer Research. Cancer Therapy: Preclinical

Fig. 3. FLT3 siRNA and/or etoposide mediated inhibition on cell growth ofMV4;11cells. A, Western blot analysis showed that protein expression levels in MV4;11cells transfected with control siRNA, 50, 100, 200, and 400 pmol FLT3 siRNA after 24 h. B, cell growth ofMV4;11cells after transfected with control siRNA, 50, 100, 200, and 400 pmol FLT3 siRNA at days 0, 1, 2, and 3. C, effect of etoposide on cell growth of MV4;11cells transfected with control siRNA and 400 pmol FLT3 siRNA after 72 h. The experiments were repeated at least thrice.

Cell cycle analysis. One million cells were suspended in 1 mL FLT3 and least sensitivity of cells that lack FLT3 (20). Cell lines solution containing 50 mg/mL PI, 0.1% sodium citrate, and 0.1% expressing ITD-FLT3 also had the greatest sensitivity to Triton X-100. The PI-stained samples were analyzed within 24 h. The etoposide. The IC50s of etoposide for Molm13 and MV4;11 cell cycle distribution and apoptosis were determined by the analysis of cells were 38 and 46 nmol/L, respectively. The cell lines nuclear DNA content using CellQuest-Pro software (Becton Dickinson, expressing wild-type FLT3 showed intermediate sensitivity to Mountain View, CA). 17-AAG and etoposide. Cell lines that were FLT3 negative were very resistant to 17-AAG. Results To determine the combination effect of 17-AAG and etopo- side, we compared the cells that express ITD-FLT3 with cells that 17-AAGand etoposide are synergistic in the inhibition of cell express wild-type FLT3. Each of the cell lines were exposed to proliferation of cell lines expressing FLT3, especially in cells with varying concentration of 17-AAG and etoposide at a fixed molar mutated FLT3. Our initial studies addressed the relationship ratio for 72 h, and inhibition of cell proliferation was deter- between FLT3 mutation and sensitivity to 17-AAG and mined. CIs were plotted against fractional effects as described in etoposide. We examined human leukemia cell lines for Materials and Methods. CI values significantly <1 indicate expression of FLT3 and inhibition of cell growth by 17-AAG synergy; values close to 1 indicate an additive effect; and values and etoposide using MTT cell proliferation assay. Results are significantly >1 indicate an antagonistic effect of the two agents. shown in Table 1. Inhibition by 17-AAG as reported in our As seen in Fig. 1, 17-AAG and etoposide synergistically previous study shows greatest sensitivity of cells with mutated inhibited cell proliferation of ITD-FLT3 cells (MV4;11 and

Clin Cancer Res 2007;13(5) March 1, 2007 1594 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2007 American Association for Cancer Research. Synergistic Effects of Hsp90 Inhibitor and Etoposide

Molm13). Synergism was observed, but at a lower level in wild- stimulation, phosphorylation of both FLT3 and Chk1 were type FLT3 cells (RS4;11 and HPB-Null). For MV4;11 and Molm inhibited by FLT3 siRNA (Fig. 2B). The early effects induced by 13, CI values at IC50 were 0.86 and 0.81, CI values at IC75 were FL on Chk1 suggest that the phosphorylation of Chk1 may 0.56 and 0.68, and CI values at IC90 were 0.41 and 0.62 possibly be directly downstream of phosphorylated FLT3. At respectively. However, for RS4;11 and HPB-Null cells, CI values 48 h after FL stimulation, the phosphorylation of Chk1 is at IC50 were 1.10 and 1.09, CI values at IC75 were 0.90 and enhanced more than that at 6 h (Fig. 2D). Etoposide inhibited 0.94, and CI values at IC90 were 0.81 and 0.87, respectively. In cell growth as shown in Fig. 2C, but did not significantly summary, synergism with combination of 17-AAG and etopo- affect phosphorylation of Chk1, expression of Rad51 or total side was observed in cell lines which showed sensitivity to both FLT3 as shown in Fig. 2D. In contrast, the Hsp90/FLT3 drugs as single agents, and the degree of synergism was related inhibitor, 17-AAG, which inhibited cell growth to the same to the FLT3 status of the cells. The remainder of the experiments extent as etoposide significantly reduced the total amount of investigated the potential mechanism(s) responsible for this FLT3, Chk1, and Rad51. Cell growth was most inhibited by synergism. the combination of 17-AAG and etoposide (Fig. 2C). This was The critical role of intact Hsp90 in control of Chk1 and Rad51 reflected in more apoptosis as an increase of the cleaved PARP and growth of cells with wild-type FLT3. We initiated a series of (Fig. 2D). experiments to evaluate the role of Hsp90/FLT3, Chk1, and In conclusion, in cells with wild-type FLT3, FL induced phos- Rad51 in cells that were stimulated to proliferate by FLT3 ligand phorylation of FLT3, Chk1, and expression of Rad51. In these (FL). Cells with wild-type FLT3 (RS4;11) proliferate more in the cells with wild-type FLT3, 17-AAG but not etoposide resulted presence of FL than the cells without FL. FLT3 siRNA inhibits in down-regulation of Chk1 and Rad51. The combination of cell proliferation induced by FL (Fig. 2A). At 6 h after FL 17-AAG and etoposide resulted in the greatest inhibition of cell growth and apoptosis. The role of FLT3 and Hsp90 in survival of ITD-FLT3 cells. As shown in Table 1 and Fig. 1, cells with mutated FLT3 were found to be more sensitive to 17-AAG and etoposide than cells with wild-type FLT3. Thus, we evaluated the mechanism of synergism between 17-AAG and etoposide in cells that have mutated FLT3. Cells with ITD-FLT3 have constitutive activation of FLT3. We did FLT3 knockdown to evaluate its role in cell survival. ITD-FLT3 (MV4;11) cells were transfected with FLT3 siRNA as described in Materials and Methods. Small RNAi reduced FLT3 expression by 84% by densitometry at 24 h (Fig. 3A). In addition, as shown in Fig. 3A, FLT3 siRNA resulted in a reduction of p-AKT, total and phosphorylated Chk1, and total Rad51, but had no effects on total AKT, Hsp90, or Hsp70. Cell growth experiments showed that FLT3 siRNA inhibited cell growth in a dose-dependent manner (Fig. 3B), and apoptosis was shown by increased PARP in Fig. 3A. FLT3 siRNA, combined with etoposide, induced additional inhibi- tion of cell growth in MV4;11 cells (Fig. 3C). In summary, these experiments show that tumor cell growth is critically dependent on FLT3 signaling in leukemia cells with ITD-FLT3. We next conducted experiments to further evaluate the mechanism of the synergism between the Hsp90 inhibitor 17-AAG and etoposide in ITD-FLT3 cells, with emphasis on DNA repair and cell signaling pathways. As shown in Fig. 4, a marker of DSBs, g-H2A.X (30, 31), was induced by 17-AAG or etoposide and more by the combination. Of interest, cells incu- bated with the Hsp90 inhibitor, 17-AAG, had significantly less of the critical DNA repair proteins, Chk1, p-Chk1, and Rad51. Signal transduction kinases FLT3 and AKT, known to be Hsp90 client proteins, were reduced following 17-AAG alone or combination therapy, but not by etoposide alone. The chaperone proteins, Hsp90, Hsp70, and Cdc37 were also analyzed using Western blotting. Hsp90 and Cdc37 were not significantly altered by 17-AAG alone or in combination with etoposide. As reported in earlier studies (20, 32) Hsp70 but

Fig. 4. Combinations of17-AAG and etoposide down-regulate DNA not Hsp90 was increased by 17-AAG. It is known that topoisomerase (To p o II a), DNA repair proteins (Chk1and Rad51), signal topoisomerase IIa is increased at G2-M arrest induced by transduction kinases (FLT3 and AKT) and induced more apoptosis (cleavage of etoposide treatment (33), and this was the case in our PARP) than either agent alone.Western blot analysis ofproteins in cell lysate from MV4;11cells after treatment of 200 nmol/L 17-AAG and 200 nmol/L etoposide experiments as shown in Fig. 4. In contrast, topoisomerase alone and in combination for 0, 24, and 48 h is shown. IIa was inhibited by 17-AAG alone and in combination with

www.aacrjournals.org 1595 Clin Cancer Res 2007;13(5) March 1, 2007 Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2007 American Association for Cancer Research. Cancer Therapy: Preclinical

Fig. 5. Effects of17-AAG or etoposide alone and the combination of two agents on cell cycle distribution and apoptosis for ITD-FLT3 leukemia cell lines. A, MV4;11cells were treated with 200 nmol/L 17-AAG or 200 nmol/L etoposide alone and the combination oftwo agents. Cell cycle arrest was determined for0, 1, and 2 days with PI staining and fluorescence-activated cell sorting analysis. <2N, apoptotic and necrotic cells. B, MV4;11cells were treated with 200 nmol/L17-AAG or 200 nmol/L etoposide alone and the combination oftwo agents for48 h. Apoptosis was determined with PI/Caspatag staining and fluorescence-activated cell sorting analysis. Dot pl ots were divided into four quadrants.The percentage of cells: bottom left, viable cells; top right, caspase/PI-positive dead cells; bottom right, caspase-positive cells.The experiments were repeated at least thrice. etoposide. Thus, clearly, 17-AAG abolished the increase of both important in DNA repair following etoposide-induced topoisomerase IIa induced by etoposide. DNA damage. As an end result, the combination of 17-AAG and In summary, these results from Western blot analysis show etoposide resulted in more unrepaired DNA damage, and this that 17-AAG results in reduced total levels of Chk1 and Rad51, combined with reduced topoisomerase IIa most likely explains

Clin Cancer Res 2007;13(5) March 1, 2007 1596 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2007 American Association for Cancer Research. Synergistic Effects of Hsp90 Inhibitor and Etoposide the inhibition of cell growth and increased apoptosis resulting increased, late apoptotic (CaspaTag/PI double positive) cells from the combination of the two agents. increased and necrotic (CaspaTag negative/PI positive) cells We determined in more detail the effect of 17-AAG and increased compared with control cells. With etoposide treat- etoposide on cell cycle and apoptosis of ITD-FLT3 cells ment, early apoptotic cells increased, late apoptotic cells in- using MV4;11 cells. As shown in Fig. 5A, 17-AAG resulted in creased, and necrotic cells were increased. Combination the accumulation of cells in G0-G1. After treatment with treatment resulted in increased early apoptotic cells, increased 200 nmol/L 17-AAG for 48 h, MV4;11 cells were arrested in late apoptotic cells, and increased necrotic cells compared with G0-G1 phase from a baseline of 71% to 88%. S-phase cells control cells. Thus, more apoptotic cells were observed with the decreased from a baseline of 22% to 10%, and <2N portion combination of 17-AAG and etoposide than either agent (DNA fragmentation) cells were increased from 24% to 33%. alone. A second ITD-FLT3 cell line, Molm 13, showed very In contrast to 17-AAG, etoposide (200 nmol/L) induced more similar effects of 17-AAG and etoposide in combination (data cell arrest in G2-M phase (8% to 34%), decreased DNA synthesis not shown). (S phase) from 22% to 12%, and induced more <2N portion Etoposide increases the inhibition of cell growth induced by cells. Interestingly, the combined treatment had the greatest Chk1 or Rad51 depletion. Next, we tested the effect of Chk1 effect in decreasing S phase from cells and induced more <2N inhibition on cell growth and Rad51 expression. Transfection cells than those treated with 17-AAG or etoposide alone. We of Chk1 siRNA into ITD-FLT3 MV4;11 cells resulted in a 79% also used CaspaTag/PI assay to compare the induction of apop- reduction of Chk 1 based on densitometry (Fig. 6A). Chk1 tosis by the individual agents and the 17-AAG–etoposide com- siRNA reduced the level of Rad51 (Fig. 6A), inhibited cell bination. Consistent with the PARP and <2N studies described growth (Fig. 6B), and induced cleavage of PARP (Fig. 6A) in a earlier, more caspase activity was detected in cells treated with dosage-dependent manner. More inhibition of cell growth the drug combination than in cells treated with either drug was induced by the combination of Chk1 siRNA and alone. As shown in Fig. 5B, after treatment with 17-AAG for etoposide than etoposide alone in MV4;11 cells as shown 48 h, early apoptotic (CaspaTag positive/PI negative) cells in Fig. 6C.

Fig. 6. Chk1siRNA and/or etoposide mediated inhibition on cell growth ofMV4;11cells. A, Western blot analysis showed that protein expression levels in MV4;11cells transfected with control siRNA, 100, 200, 400, and 600 pmol Chk1siRNA after 24 h. B, cell growth ofMV4;11cells after transfected with control siRNA, 100, 200, 400, and 600 pmol Chk1siRNA at days 0, 1, 2, and 3. C, effect of etoposide on cell growth of MV4;11cells transfected with control siRNA and 600 pmol Chk1siRNA after 72 h. The experiments were repeated at least thrice.

www.aacrjournals.org 1597 Clin Cancer Res 2007;13(5) March 1, 2007 Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2007 American Association for Cancer Research. Cancer Therapy: Preclinical

Fig. 7. Rad51siRNA and/or etoposide mediated inhibition on cell growth ofMV4;11cells. A, Western blot analysis showed that protein expression levels in MV4;11cells transfected with control siRNA,100, 200, 400, and 800 pmol Rad51siRNA after 24 h. B, cell growth ofMV4;11cells after transfected with control siRNA,100, 200, 400, and 800 pmol Rad51siRNA at days 0, 1, 2, and 3. C, effect of etoposide on cell growth of MV4;11cells transfected with control siRNA and 800 pmol Rad51siRNA after 72 h. D, expression levels ofChk1and Rad51in Molm13, MV4;11, HPB-Null, and RS4;11leukemia cell lines.The experiments were conducted in triplicates.

Rad51 siRNA resulted in inhibition of cellular Rad51 siRNA were inhibited more by etoposide than the control cells (Fig. 7A shows an 87% reduction based on densitometry). (Fig. 7C). Cell growth was inhibited in a dosage-dependent manner by Expression levels of Chk1 and Rad51 differ in ITD-FLT3 and Rad51 siRNA (Fig. 7B). As shown in Fig. 7A, Rad51 siRNA had wild-type FLT3 leukemia cells. Because the experiments above no effects on the expression of FLT3, Chk1, Hsp90, or Hsp70. showed differences between ITD-FLT3 and wild-type FLT3 in There was an increase in the cleavage of PARP by Rad51 siRNA. effects of the Hsp90 inhibitor, 17-AAG, and etoposide, we In addition, the growth of the cells transfected with Rad51 looked for baseline differences between wild-type FLT3 and

Clin Cancer Res 2007;13(5) March 1, 2007 1598 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2007 American Association for Cancer Research. Synergistic Effects of Hsp90 Inhibitor and Etoposide

ITD-FLT3 cells. In these experiments, Western blotting was down FLT3 gene expression. FLT3 siRNA significantly inhi- used to determine the quantity of the DNA repair-associated bited cell proliferation in cells with wild-type FLT3 proteins Chk1 and Rad51 in cells. As shown in Fig. 7D, Chk1 (Fig. 2A). and Rad51 were found at higher levels in etoposide/17-AAG– Etoposide is a potent inducer of DSBs and is clinically insensitive wild-type FLT3 (HPB-Null and RS4;11) cells than highly effective in AML therapy. ITD mutations of FLT3 are in etoposide/17-AAG–sensitive ITD-FLT3 cells (Molm13 and found in about 30% of cases of de novo AML. The MV4;11). In addition, both wild-type FLT3 HPB-Null and mechanisms of ITD-FLT3 and FLT3 have been explored, RS4;11 showed fewer <2N apoptotic or necrotic cells (3.6% and agents capable of inhibiting FLT3 kinase have been and 5.2%, respectively) than ITD-FLT3 Molm13 and MV4;11 developed and display promising preclinical activity (36–44). cells, (17.3% and 22.1%, respectively) in a representative The combinations of FLT3 inhibitor with other conventional experiment. chemotherapy are likely to improve the outcomes of patients with AML (45, 46). Discussion Chk1 is an essential checkpoint kinase that is required for cell cycle delays after DNA damage or blocked DNA Results from this study provide evidence that leukemia cell replication (47, 48). Chk1 can directly interact with Rad51, death produced by the topoisomerase II inhibitor, etoposide, and Rad51 is phosphorylated in a Chk1-dependent manner. and the specific inhibitor of Hsp90, 17-AAG, are synergistic. Depletion of Chk1 results in failed Rad51 nuclear foci Importantly, the sensitivity and synergism between these two formation (8). We observed that depletion of Chk1 also agents is greatest in cells with mutated FLT3. Previously, we and reduced Rad51 expression. In this study, we also found that others reported that both ITD-FLT3 and wild-type FLT3 are leukemia cells with wild-type FLT3 express higher levels of client proteins of Hsp90, and treatment with Hsp90 inhibitors, Chk1 and Rad51 than ITD-FLT3 cells. This might partially 17-AAG or herbimycin, resulted in depletion of both proteins explain why wild-type FLT3 cells have less sensitivity to (20, 21, 32). The FLT3 inhibitor, GTP14564, was previously etoposide because Chk1 and Rad51 are proteins involved in found to have a synergistic effect in both ITD-FLT3 and wild- DNA repair during DNA damage induced by etoposide. Our type FLT3 cells (34). These and other data suggest that FLT3 is studies showed that Chk1 and Rad51 are down-regulated potentially an important target for leukemia therapy. following FLT3 siRNA. These results are consistent with the In the current research, we first studied the mechanisms of previous reports that FLT3 inhibitor, CEP-701, induced a the combined effects of etoposide and inhibition of FLT3 in synergistic cytotoxicity with etoposide in ITD-FLT3 leukemia cells with wild-type FLT3. We show for the first time that wild- cells (45). type FLT3 RS4;11 leukemia cells are dependent on In conclusion, results in the present study show that the FLT3 ligand for optimal growth and survival. We also used combination of 17-AAG and etoposide have synergistic siRNA to specifically examine the role of wild-type FLT3 in inhibitory effects in ITD-FLT3 cells in vitro. A critical role for RS4;11 cells (Fig. 2A and B). Small RNAi uses double-stranded FLT3, topoisomerase II, Chk1, and Rad51 were shown in this RNA complexes to target specific genes for silencing. A short study. The synergistic activity between the Hsp90 inhibitor, 17- synthetic duplex RNA long enough to induce gene-specific AAG, and the DNA-damaging agent, etoposide, provides useful silencing but short enough to evade host responses was used leads for the design of trials that combine these two distinct (35). We used transient transfection with siRNA to knock classes of agents in vivo.

References 1. Robert J, Larsen AK. Drug resistance to topoisomer- dent chromatin-binding protein required for the DNA 16. Wang S,Wang Z, Grant S. Bryostatin 1and UCN-01 ase II inhibitors. Biochimie 1998;80:247 ^ 54. replication checkpoint. Curr Biol 2000;10:1565^73. potentiate 1-h-D-arabinofuranosylcytosine-induced 2. Khanna KK, Jackson SP. DNA double-strand breaks: 10. Feijoo C, Hall-Jackson C, Wu R, et al. Activation apoptosis in human myeloid leukemia cells through signaling, repair and the cancer connection. Nat Genet ofmammalian Chk1 during DNA replication arrest: a disparate mechanisms. Mol Pharmacol 2003;63: 2001;27:247 ^ 54. role for Chk1 in the intra-S phase checkpoint moni- 232^ 42. 3. PaullTT,RogakouEP,YamazakiV,etal. Acriticalrole for toring replication origin firing. J Cell Biol 2001;154: 17. Gilliland DG, Griffin JD. The roles of FLT3 in hemato- histone H2AX inrecruitmentofrepair factors tonuclear 913^ 23. poiesis and leukemia. Blood 2002;100:1532^ 42. fociafterDNAdamage.CurrBiol2000;10:886^95. 11. Zhao H, Piwnica-Worms H. ATR-mediated check- 18. Levis M, Small D. FLT3: ITDoes matter in leukemia. 4. Burma S, Chen BP, Murphy M, Kurimasa A, Chen DJ. point pathways regulate phosphorylation and acti- Leukemia 2003;17:1738 ^ 52. ATM phosphorylates histone H2AX in response to vation ofhuman Chk1. Mol Cell Biol 2001;21: 19. Carnicer MJ, Nomdedeu JF, Lasa A, et al. FLT3 DNA double-strand breaks. J Biol Chem 2001;276: 4129 ^ 39. mutations are associated with other molecular lesions 42462^ 7. 12. LiuQ,GuntukuS,CuiXS,etal.Chk1isanessential in AML. Leuk Res 2004;28:19^ 23. 5. Hoeijmakers JH. Genome maintenance mechanisms kinase that is regulated by Atr and required for the 20. Yao Q, Nishiuchi R, Li Q, et al. FLT3 expressing for preventing cancer. Nature 2001;411:366 ^ 74. G(2)/M DNA damage checkpoint. Genes Dev 2000; leukemias are selectively sensitive to inhibitors ofthe 6. Sonoda E, Sasaki MS, Buerstedde JM, et al. 14:1448^59. molecular chaperone heat shock protein 90 through Rad51-deficient vertebrate cells accumulate chromo- 13. Zachos G, Rainey MD, Gillespie DA. Chk1-defi- destabilization ofsignal transduction-associated somal breaks prior to cell death. EMBO J 1998;17: cient tumour cells are viable but exhibit multiple kinases. Clin Cancer Res 2003;9:4483 ^ 93. 598^608. checkpoint and survival defects. EMBO J 2003;22: 21. Minami Y, Kiyoi H,YamamotoY, et al. Selective apo- 7. Hansen LT, Lundin C, Spang-Thomsen M, Petersen 713 ^ 23. ptosis oftandemly duplicated FLT3-transformed leu- LN, Helleday T. The role ofRAD51 in etoposide 14. Shi Z, Azuma A, Sampath D, et al. S-phase arrest by kemia cells by Hsp90 inhibitors. Leukemia 2002;16: (VP16) resistance in small cell lung cancer. Int J Can- nucleoside analogues and abrogation ofsurvival with- 1535^40. cer 2003;105:472 ^ 9. out cell cycle progression by 7-hydroxystaurosporine. 22. Isaacs JS, Xu W, Neckers L. Heat shock protein 90 8. Sorensen CS, Hansen LT, Dziegielewski J, et al. The Cancer Res 2001;61:1065 ^ 72. as a molecular target for cancer therapeutics. Cancer cell-cycle checkpoint kinase Chk1is required for mam- 15. Sampath D, Shi Z, Plunkett W. Inhibition ofcyclin- Cell 2003;3:213^ 7. malian homologous recombination repair. Nat Cell Biol dependent kinase 2 by the Chk1-25A pathway during 23. Goetz MP,Toft D, Reid J, et al. Phase I trial of 17- 2005;7:195 ^ 201. the S-phase checkpoint activated by fludarabine: dys- allylamino-17-demethoxygeldanamycin in patients 9. Hekmat-Nejad M, You Z, Yee MC, Newport JW, regulation by 7-hydroxystaurosporine. Mol Pharmacol withadvancedcancer.JClinOncol2005;23:1078 ^ 87. Cimprich KA. Xenopus ATR is a replication-depen- 2002;62:680 ^ 8. 24.Workman P. Auditing the pharmacological accounts

www.aacrjournals.org 159 9 Clin Cancer Res 2007;13(5) March 1, 2007 Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2007 American Association for Cancer Research. Cancer Therapy: Preclinical

for Hsp90 molecular chaperone inhibitors: unfolding kinase inhibitor PKC412 is highly effective against SU11248 is a novel FLT3 tyrosine kinase inhibitor the relationship between pharmacokinetics and phar- human acute myelogenous leukemia cells with mutant with potent activity in vitro and in vivo.Blood macodynamics. Mol CancerTher 2003;2:131 ^ 8. FLT-3. Cancer Res 2004;64:3645 ^ 52. 2003;101:3597 ^ 605.

25. Schmidt-Ullrich RK, Contessa JN, Lammering G, 33. Nakada S, Katsuki Y, Imoto I, et al. Early G2/M 41. Zheng R, Friedman AD, Small D. Targeted inhibition Amorino G, Lin PS. ERBB receptor tyrosine kinases checkpoint failure as a molecular mechanism under- ofFLT3 overcomes the block to myeloid differentia- and cellular radiation responses. Oncogene 2003;22: lying etoposide-induced chromosomal aberrations. tion in 32Dcl3 cells caused by expression ofFLT3/ 5855 ^ 65. J Clin Invest 2006;116:80 ^9. ITD mutations. Blood 2002;100:4154 ^ 61. 26. FriedmannB,CaplinM,HartleyJA,HochhauserD. 34. Yao Q, Nishiuchi R, Kitamura T, Kersey JH. Human 42. Levis M, Allebach J,Tse KF, et al. A FLT3-targeted Modulation ofDNA repair in vitro after treatment with leukemias with mutated FLT3 kinase are synergisti- tyrosine kinase inhibitor is cytotoxic to leukemia cells chemotherapeutic agents by the epidermal growth cally sensitive to FLT3 and Hsp90 inhibitors: the key in vitro and in vivo. Blood 2002;99:3885 ^ 91. factor receptor inhibitor gefitinib (ZD1839). Clin Can- role ofthe STAT5 signal transduction pathway.Leuke- 43. Smith BD, Levis M, Beran M, et al. Single-agent cer Res 2004;10:6476 ^ 86. mia 2005;19:1605 ^ 12. CEP-701, a novel FLT3 inhibitor, shows biologic and 27. DittmannK,MayerC,FehrenbacherB,etal.Radia- 35. Dykxhoorn DM, Lieberman J.The silent revolution: clinical activity in patients with relapsed or refractory tion-induced epidermal growth factor receptor nuclear RNA interference as basic biology, research tool, and . Blood 2004;103:3669 ^ 76. import is linked to activation ofDNA-dependent pro- therapeutic. Annu Rev Med 2005;56:401 ^ 23. 44. Bali P, George P, Cohen P, et al. Superior activity of tein kinase. J Biol Chem 2005;280:31182^ 9. 36. Kelly LM,YuJC, Boulton CL, et al. CT53518, a novel the combination ofhistone deacetylase inhibitor 28.Trojanek J, HoT, Del Valle L, et al. Role ofthe insulin- selective FLT3 antagonist for the treatment of acute LAQ824 and the FLT-3 kinase inhibitor PKC412 like growth factor I/insulin receptor substrate 1axis in myelogenous leukemia (AML). Cancer Cell 2002;1: against human acute myelogenous leukemia cells with Rad51 trafficking and DNA repair by homologous 421 ^ 32. mutant FLT-3. Clin Cancer Res 2004;10:4991 ^ 7. recombination. Mol Cell Biol 2003;23:7510 ^ 24. 37. Levis M, Tse KF, Smith BD, Garrett E, Small D. A 45. Levis M, Pham R, Smith BD, Small D. In vitro 29. Yang S, Chintapalli J, Sodagum L, et al. Activated FLT3 tyrosine kinase inhibitor is selectively cytotoxic studies ofa FLT3 inhibitor combined with chemo- IGF-1R inhibits hyperglycemia-induced DNA damage to acute myeloid leukemia blasts harboring FLT3 inter- therapy: sequence ofadministration is important to and promotes DNA repair by homologous recombi- nal tandem duplication mutations. Blood 2001;98: achieve synergistic cytotoxic effects. Blood 2004; nation.AmJPhysiolRenalPhysiol2005;289: 885^7. 10 4:1145 ^ 5 0. F114 4 ^ 52. 38.Weisberg E, Boulton C, Kelly LM, et al. Inhibition of 46.YeeKW, Schittenhelm M, O’Farrell AM, et al. Syner- 30. Pilch DR, Sedelnikova OA, Redon C, et al. Charac- mutant FLT3 receptors in leukemia cells by the small gistic effect of SU11248 with or daunorubi- teristics of g-H2AX foci at DNA double-strand breaks molecule tyrosine kinase inhibitor PKC412. Cancer cin on FLT3 ITD-positive leukemic cells. Blood 2004; sites. Biochem Cell Biol 2003;81:123^ 9. Cell 2002;1:433 ^ 43. 104:4202^ 9. 31. Takahashi A, Ohnishi T. Does gH2AX foci formation 39. Yee KW, O’Farrell AM, Smolich BD, et al. SU5416 47. Abraham RT. Cell cycle checkpoint signaling depend on the presence ofDNA double strand and SU5614 inhibit kinase activity ofwild-type and through the ATM and ATR kinases. Genes Dev 2001; breaks? Cancer Lett 2005;229:171 ^ 9. mutant FLT3 receptor tyrosine kinase. Blood 2002; 15 :2177 ^ 9 6. 32. George P, Bali P, Cohen P, et al. Cotreatment with 100:2941 ^ 9. 48. Bartek J, Lukas J. Chk1and Chk2 kinases in check- 17-allylamino-demethoxygeldanamycin and FLT-3 40. O’Farrell AM, Abrams TJ, Yuen HA, et al. point control and cancer. Cancer Cell 2003;3:421 ^ 9.

Clin Cancer Res 2007;13(5) March 1, 2007 1600 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2007 American Association for Cancer Research. Synergism between Etoposide and 17-AAG in Leukemia Cells: Critical Roles for Hsp90, FLT3, Topoisomerase II, Chk1, and Rad51

Qing Yao, Brenda Weigel and John Kersey

Clin Cancer Res 2007;13:1591-1600.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/13/5/1591

Cited articles This article cites 48 articles, 27 of which you can access for free at: http://clincancerres.aacrjournals.org/content/13/5/1591.full#ref-list-1

Citing articles This article has been cited by 5 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/13/5/1591.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://clincancerres.aacrjournals.org/content/13/5/1591. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.

Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2007 American Association for Cancer Research.