Synergistic induction of p53 mediated apoptosis by valproic ... - Nature

Leukemia (2012) 26, 910 - 917 & 2012 Macmillan Publishers Limited All rights reserved 0887-6924/12 www.nature.com/leu

ORIGINAL ARTICLE

Synergistic induction of p53 mediated apoptosis by valproic acid and nutlin-3 in acute myeloid leukemia E McCormack1,2, I Haaland2, G Vena˚s2, RB Forthun2, S Huseby3, G Gausdal3, S Knappskog4, DR Micklem3, JB Lorens3, Ø Bruserud1,2 and BT Gjertsen1,2 Although TP53 mutations are rare in acute myeloid leukemia (AML), wild type p53 function is habitually annulled through overexpression of MDM2 or through various mechanisms including epigenetic silencing by histone deacetylases (HDACs). We hypothesized that co-inhibition of MDM2 and HDACs, with nutlin-3 and valproic acid (VPA) would additively inhibit growth in leukemic cells expressing wild type TP53 and induce p53-mediated apoptosis. In vitro studies with the combination demonstrated synergistic induction of apoptosis in AML cell lines and patient cells. Nutlin-3 and VPA co-treatment resulted in massive induction of p53, acetylated p53 and p53 target genes in comparison with either agent alone, followed by p53 dependent cell death with autophagic features. In primary AML cells, inhibition of proliferation by the combination therapy correlated with the CD34 expression level of AML blasts. To evaluate the combination in vivo, we developed an orthotopic, NOD/SCID IL2rgnull xenograft model of MOLM-13 (AML FAB M5a; wild type TP53) expressing firefly luciferase. Survival analysis and bioluminescent imaging demonstrated the superior in vivo efficacy of the dual inhibition of MDM2 and HDAC in comparison with controls. Our results suggest the concomitant targeting of MDM2-p53 and HDAC inhibition, may be an effective therapeutic strategy for the treatment of AML. Leukemia (2012) 26, 910 -- 917; doi:10.1038/leu.2011.315; published online 8 November 2011 Keywords: nutlin-3; valproic acid; p53; in vivo imaging; acute myeloid leukemia

INTRODUCTION Elderly acute myeloid leukemia (AML) patients do not tolerate the intensive combination chemotherapy or bone marrow transplantation required to treat their disease, resulting in an overall survival rate o10%.1 - 3 Thus, the necessity for more specific targeted and less toxic therapy is critical for these patients. Although mutations in TP53 occur in o10% of patients with AML, overexpression of its main negative regulator, MDM2, is frequently observed.4 Furthermore, aberrant recruitment of histone deacetylases (HDACs) is detected in AML, leading to myeloid differentiation block and leukemic blast accumulation.5 Nutlin-3 is a small-molecule antagonist of MDM2, which has been found to bind specifically to the p53-binding pocket of MDM2, activate the p53 pathway in cancer cells with wild type p53, and inhibit tumor growth in a non-genotoxic manner in xenografted tumor mice.6 p53-mediated effects of nutlin-3, such as induction of cell cycle arrest and apoptosis, have been demonstrated in cancers characterized by non-mutated TP53, including AML.7 Valproic acid (VPA) is a well-tolerated anticonvulsant drug that has been used clinically for more than three decades.8 Recently, it has also been shown to have HDAC inhibitor activity, and to induce differentiation and apoptosis in AML progenitors and blasts, with a limited toxicity profile in elderly AML patients.9 - 12 Both VPA and nutlin-3 affect the regulation of p53; nutlin-3 by inhibiting MDM2 and VPA by inhibiting HDACs that participate in p53 deacetylation and destabilization.6,13 We therefore hypothesized that concomitant inhibition of MDM2 and HDACs would additively induce p53-mediated apoptosis and inhibit tumor

growth. The effects of the combinational treatment on apoptosis, p53 and target gene induction, and inhibition of proliferation were tested in vitro in AML cell lines and primary AML cells, and sensitivity towards the combinational treatment was correlated to clinical parameters. To evaluate the combinatorial treatment in vivo, we established a luciferase expressing MOLM-13luc AML xenograft model in NOD-scid IL2rgnull mice. Subsequent, bioluminescent imaging permitted non-invasive spatio-temporal detection of disease development and therapy response monitoring.

MATERIALS AND METHODS Cell lines, primary AML cells, normal peripheral blood lymphocytes and cord blood cells Cell lines and cell culture conditions are described in Supplementary information. AML samples were collected from patients after informed consent and approval by the regional Ethics Committee (REK Vest; http://helseforskning.etikkom.no; Norwegian Ministry of Education and Research). Normal peripheral blood lymphocytes were obtained from healthy blood donors (Blood bank, Haukeland University Hospital, Bergen, Norway), whereas normal umbilical cord blood was obtained from healthy individuals after caesarian section (Women’s Clinic, Haukeland University Hospital, Bergen, Norway). Preparation and culturing of cells is described in Supplementary Information.

Clinical parameters of AML patients Clinical parameters including FAB classification, cell surface markers, karyotype, resistance, survival and FLT3/NPM1 mutational status were

1 Department of Medicine, Hematology Section, Haukeland University Hospital, Bergen, Norway; 2Institute of Medicine, Hematology Section, University of Bergen, Bergen, Norway; 3Department of Biomedicine, University of Bergen, Bergen, Norway and 4Institute of Medicine, Section of Oncology, University of Bergen, Bergen, Norway. Correspondence: Professor BT Gjertsen, Institute of Medicine, Hematology Section, Haukeland University Hospital, University of Bergen, N-5021 Bergen, Norway. E-mail: [email protected] Received 8 March 2011; revised 30 August 2011; accepted 16 September 2011; published online 8 November 2011

Synergism of valproic acid and nutlin-3 in AML E McCormack et al routinely analyzed and collected. RNA extraction and p53-mutational analysis was performed as previously described.14

Compounds Nutlin-3 and -3a (Cayman Chemical Company, Ann Arbor, MI, USA) was dissolved in DMSO and stored at 80 1C. When used in cell culture work, the final concentration of DMSO did not exceed 0.1%. For animal experiments, nutlins were administered orally (p.o.) as a suspension in vehicle consisting of 2% hydroxypropyl cellulose and 0.5% tween 80 (Sigma, St Louis, MO, USA) in sterile water. VPA (Orfiril, Desitin Arzneimittel GmbH, Hamburg, Germany) (100 mg/ml in solution) was stored at 80 1C for cell culture work. For animal experiments, VPA was administered by intraperitoneal injection.

Western blotting Western blotting was performed as previously described.15 For details, see Supplementary Information.

Flow cytometry p53 flow cytometric analysis, detection of apoptosis by annexin V-propidium iodide (Annexin-PI) and detection of Lysotracker Red stained cells are described in detail in Supplementary Information.

Transmission electron microscopy (TEM) For evaluation of autophagic features, cells were prepared for and analyzed by transmission electron microscopy as previously described.16

Cell viability and proliferation assays Evaluation of apoptosis and inhibition of proliferation in cell lines and primary AML cells after drug treatment was performed using different assays; Hoechst 33342, Annexin-PI, 3H-thymidine incorporation, ATP and Alamarblue assay, for details see Supplementary Information.

Transfections of luciferase expressing construct and sorting of cells The luciferase expressing construct L192 and the TetActivator were transfected into Phoenix cells, and retroviral infection was performed as previously described.17 Sorting of highly bioluminescent MOLM-13luc cells is described in Supplementary Information.

MOLM-13 xenograft model and optical imaging The animal experiments were approved by The Norwegian Animal Research Authority and performed in accordance with The European Convention for the Protection of Vertebrates Used for Scientific Purposes. For all experimental details, see Supplementary Information.

Statistical analysis For details on statistical analysis, see Supplementary Information.

RESULTS Combinational therapy of nutlin-3 and VPA induces apoptosis synergistically in the wild type TP53 AML cell line MOLM-13 We were particularly interested in whether nutlin-3, an MDM2 antagonist,6 and VPA, a known class I and II HDAC inhibitor affecting p53 acetylation,13 could work in tandem to elicit enhanced apoptotic effect on AML cells expressing wild type p53. To determine synergistic effects of VPA and nutlin-3 on cell viability, MOLM-13 cells were treated with increasing doses of VPA (0 -- 2 mM) or nutlin-3 (0 -- 20 mM) alone or in combination at a fixed ratio (1:100) for 72 h (nutlin-3 for the last 24 h) and analyzed by Annexin-PI. The dose-effect curve for each drug was determined and combination index values were calculated according to the Chou -- Talalay method,18 whereas effect on cell viability was expressed as fraction of cells affected. The results are presented in Figure 1a, indicating strong synergism between the two drugs along the entire dose-response curve. The apoptotic effect of VPA (50 -- 1000 mM, 24, 48 and 72 h) and nutlin-3 & 2012 Macmillan Publishers Limited

(0.5 -- 10 mM, last 24 h) alone and in combination MOLM-13 cells, was also investigated by DNA-specific staining with Hoechst 33342 (Supplementary Figure 1A). Although induction of apoptosis by VPA alone was minimal (P40.05), nutlin-3 mediated effective apoptosis (450%) at 10 mM. At lower concentrations, particularly the combination of VPA at 500 mM and nutlin-3 at 5 mM demonstrated super-additive induction of apoptosis in comparison with single treatments. Repeat studies employing nutlin-3 (5 mM) and VPA (500 mM) exhibited significantly better effect of the combination on viability (Hoechst 33342) compared with either agent alone (Po0.001) at 24, 48 and 72 h (Figure 1b). Furthermore, Bliss Independence analysis19 of the data revealed synergistic apoptosis induction with higher actual response than expected response for the combination of nutlin-3 (5 mM) and VPA (500 mM) at all time points (Po0.0001; Supplememtary Figure 1B). The effect of the combinational therapy on apoptosis induction in MOLM-13 cells is clearly visualized by staining with Hoechst 33342 and microscopic imaging of fragmented/condensed cell nuclei, whereas no effects are seen in normal peripheral blood lymphocytes (Supplementary Figure 1C). Finally, the remarkable efficacy and synergy of the combination upon apoptosis in wild type TP53 AML cells was reaffirmed by Alamar Blue and ATP cell viability assays, with Bliss Independence analysis also confirming synergy with these assays (Supplementary Figure 1D). Synergistic apoptosis induction by nutlin-3 and VPA involves super additive induction of p53, acetylated p53 and p53 target genes As VPA’s effect increases with time and concentration,20 we incubated MOLM-13 cells for 72 h (500 mM) to visualize protein induction. Nutlin-3 was added for the last 24 h at a lower concentration than for cell viability assays (2.5 mM) to avoid massive apoptosis. As expected, nutlin-3 induced p53, acetylated p53 and subsequently MDM2 and p21, whereas VPA, as observed in cell viability studies, demonstrated negligible effect. The combination, however, super additively induced p53 and acetylated p53 and most strikingly p21 and MDM2 (Figure 1c; western blots). Also primary AML cells showed increased induction of p53 and target genes with the same treatment conditions (significant increase in p21 and MDM2 induction; Po0.05), as analyzed by flow cytometric analysis of p53, p21 and MDM2 (Figure 1d). To determine the role of p53 in the combinational therapy, we transfected MOLM-13 cells with shp53 (providing a 70% knockdown of p53; data not shown) and compared the effects of the combinational treatment between MOLM-13 cell lines (wild type p53 and shp53) with Annexin-PI co-staining and flow cytometric analysis. Although MOLM-13 cells with wild type p53 responded to treatment with nutlin-3 (5 mM), VPA (500 mM) or the combination, those transfected with shp53 demonstrated only diminutive effects upon viability (Figure 1e). These results suggest that regulation of p53 is central to the effect of combinational therapy of nutlin-3 and VPA. Combinational therapy of nutlin-3 and VPA induces autophagy in MOLM-13 cells We assessed modulation of autophagy markers LC3B and p62 in MOLM-13 cells treated with VPA (500 mM, 72 h), nutlin-3 (2.5 mM, last 24 h) or the combination of both (Figure 1c), demonstrating decreased expression of the full-length LC3B isoform and p62. In addition, flow cytometric analysis of MOLM-13 cells treated with VPA (500 mM, 48 h), nutlin-3 (5 mM, last 24 h) or the combination showed significantly increased number of acidic organelles (Lysotracker Red) with the combination compared with either agent alone (Po0.001 and o0.01, respectively) (Figure 1f). Corresponding transmission electron microscopy micrographs confirmed autophagy induction in cells treated with nutlin-3 and the combination by detection of autophagosomes. Leukemia (2012) 910 - 917

911

Synergism of valproic acid and nutlin-3 in AML E McCormack et al

912

Figure 1. Synergistic apoptosis induction, p53 dependency and mechanisms of action for the combinational therapy of nutlin-3 and VPA. (a) The AML cell line MOLM-13 (wild type TP53) was treated with increasing doses of nutlin-3 (0 -- 20 mM) and VPA (0 -- 2000 mM) alone or in combination at a fixed ratio (1:100) for 72 h (nutlin-3 added for the final 24 h) and analyzed by Annexin-PI. Combination index (CI) values were calculated according to the Chou -- Talalay method, and cell viability is expressed as fraction affected (fa). CI-values for ED50, ED75 and ED90 for the combination are shown, and the overall CI is given as mean±s.e. of mean. (b) MOLM-13 cells were treated with nutlin-3 (5 mM), VPA (500 mM) or the combination of both for 24, 48 and 72 h (nutlin-3 added for the final 24 h) and cell viability detected by Hoechst 33342. Reduction of viability by the combinational treatment was compared with either compound alone (***Po0.001). Viability data were analyzed in triplicate and results given as mean±s.d. (c) MOLM-13 cells were incubated with nutlin-3 (2.5 mM), VPA (500 mM) or combination of both for 72 h (nutlin-3 added for the final 24 h). Western blotting was performed using antibodies against p53, acetylated p53 (lys382), MDM2, p21, LC3B and p62. Protein bands were quantified using region of interest analysis on Kodak Molecular Imaging Software version 5.0.1(Carestream Health, Rochester, NY, USA) and results given as fold induction of control. (d) Primary AML cells from four different AML patients were incubated with nutlin-3 (2.5 mM), VPA (500 mM) or combination of both for 72 h (nutlin-3 added for the final 24 h), and levels of p53, p21 and MDM2 were analyzed by flow cytometry. Results are given as median fluorescence intensity (fold induction of ctr) ±s.e. (*Po0.05, n ¼ 4). (e) The effect of nutlin-3 (5 mM) and VPA (500 mM) and the combination of both on viability (Annexin-PI) after 48 h in MOLM-13 cells with wild type p53 was compared with the effect in MOLM-13 cells stably transfected with shp53 (70% knockdown of p53) Results are given as mean±s.d. (***Po0.001). (f ) MOLM-13 cells treated with VPA (500 mM), nutlin-3 (5 mM) or combination of both for 48 h (nutlin-3 for the last 24 h) were incubated with Lysotracker Red for 30 min and analyzed by flow cytometry. Results are shown as median fluorescence intensity relative to control (**Po0.01, ***Po0.001). Micrographs of cells analyzed by transmission electron microscopy (TEM) are included for each treatment. Arrows indicate autophagosomes. Leukemia (2012) 910 - 917

& 2012 Macmillan Publishers Limited


Similar results are shown for Lysotracker Red at 72 h (Supplementary Figure 2A), and in more detailed transmission electron microscopy micrographs, demonstrating accumulation of autophagosomes in the combination therapy compared with either agent alone (Supplementary Figure 2B). Combinational therapy of nutlin-3 and VPA induces apoptosis synergistically in primary AML cells Primary AML cells from 31 heterogeneous AML patient samples (Supplementary Table 1), peripheral blood lymphocytes, and cell

lines MOLM-13 (wild type TP53, length mutated FLT3), OCI-AML3 (wild type TP53, wild type FLT3), HL-60 (deleted TP53, wild type FLT3) and NB4 (mutated TP53, wild type FLT3), exhibited various sensitivities to the combination treatment in vitro as determined by Annexin-PI staining (Figure 2a). MOLM-13 cells with wild type TP53 and mutated FLT3 demonstrated synergistic effects of the combination treatment also with this assay, whereas the AML cell lines HL-60, NB4 and OCI-AML3 were more resistant. Of the 10 most sensitive patient samples, 50% were FLT3 wild type, demonstrating the potential of the combinational treatment also for patients irrespective of FLT3 status. The frequency of TP53

Figure 2. Effect of the combinational therapy of nutlin-3 and VPA on viability of primary AML cells. (a) Sensitivity for the treatment with nutlin-3 (5 mM), VPA (500 mM) or combination of both for 48 h (nutlin-3 added for the final 24 h) analyzed by Annexin-PI is shown for the AML cell lines MOLM-13, HL60 and OCI-AML3, NB4, normal peripheral blood lymphocytes (PBL) and 31 primary AML samples, with corresponding clinical parameters for each of the AML patients. Results are visualized using TMEV microarray software suite version 4.3.01 (Dana-Farber Cancer Institute, Boston, MA, USA). Samples are ranked by sensitivity towards the combinational treatment. (b) Differences in means of viability between nutlin-3 (5 mM), VPA (500 mM) or combination of both for 48 h (nutlin-3 added for the final 24 h) for the pooled patient data analyzed by Annexin-PI. Results are given as means±s.e. of mean (***Po0.001, n ¼ 31). (c) Synergism calculated by Bliss Independence for the pooled patient data for apoptosis induction of the combination of nutlin-3 (5 mM) and VPA (500 mM) for 48 h (nutlin-3 added for the final 24 h) analyzed by Annexin-PI (n ¼ 31) and Hoechst 33342 (n ¼ 34) (*Po0.05). Error bars represent s.e. of mean. (d) Bliss Independence analysis of expected and actual response for the combinational therapy of nutlin-3 (5 mM) and VPA (500 mM) at 48 h (nutlin-3 added for the final 24 h) for each of the individual responding AML patient samples analyzed by Annexin-PI (n ¼ 28; samples with no response not shown). ND, not determined; WT, wild type. & 2012 Macmillan Publishers Limited

Leukemia (2012) 910 - 917

913


914

mutation in AML is exceptionally low and in this patient set only 2 of the 31 (6.5%) samples harbored TP53 mutations, which were among the most resistant samples. However, no significant correlations were found between the patient clinical parameters and sensitivity towards the combination. Pooling of the patient data (Annexin-PI) demonstrated significant reduction in mean viability for the combinational treatment versus nutlin-3 or VPA alone (Po0.001; Figure 2b) and synergism, as calculated by Bliss Independence (Po0.05; Figure 2c). Analysis of the 28 responding AML patient samples, showed higher actual than expected response for 24 of the 28 patient samples (Figure 2d), reflecting synergistic effect of the combinational treatment on viability in most of the patient samples. The superior effect of the combinational treatment compared with either agent alone in primary AML cells is exemplified in Supplementary Figure 3A. Analysis of viability by Hoechst 33342 showed similar results (Po0.05; Supplementary Figure 3B). Inhibition of proliferation by combinational therapy of nutlin-3 and VPA correlates with CD34 expression of AML blasts Primary AML cells (n ¼ 56) exhibited various sensitivities to the combination treatment of nutlin-3 (5 mM) and VPA (500 mM) for 48 h (nutlin-3 added for the final 24 h) in vitro also in 3H-thymidine incorporation proliferation assay (Figure 3a). Further examination of sensitivity to the combination and patient clinical parameters revealed a significant Pearson correlation between sensitivity towards the combination treatment (defined by 3H-thymidine incorporation) and percentage CD34 expression (Pearson r ¼ 0.3211, P ¼ 0.02). Additionally, when patients were divided

into populations of CD34low or CD34high (high as defined by 420% of blast cells stained CD34 positive by flow cytometry), the combination of nutlin-3 and VPA resulted in a significant increase in sensitivity in CD34low leukemic cells (Po0.05, n ¼ 53; Figure 3b). Effects of the combination therapy on normal CD34 þ cord blood cells and on clonogenicity of leukemic progenitors CD34 þ cells isolated from umbilical cord blood from three different healthy donors were assayed (Annexin-PI) for their sensitivity to nutlin-3 (5 mM), VPA (500 mM) or the combination, for 48 h (nutlin-3 added for the final 24 h). Contrasted against similar results from sensitive AML patient samples (n ¼ 10) and the whole cohort of AML patient samples (n ¼ 31), where significant differences in means for viability between the combination of nutlin-3 and VPA versus either agent alone (Po0.001; Supplementary Figure 4) were determined, CD34 þ cells were only found to be more sensitive to the combination when compared with nutlin-3 alone (Po0.5), suggesting VPA as the main mediator of toxicity in normal hematopoietic progenitors. In clonogenicity assay, AML patient cells sensitive to the combination demonstrated decreased numbers of clonogenic progenitors (mean ¼ 31.9% of control; n ¼ 3) in comparison with nutlin-3 (54.8%) and VPA (60.9%) cultures (data not shown). The efficacy of nutlin-3 in insensitive patient cells (n ¼ 4) was also relatively high (62.4%), whereas VPA stimulated growth of clonogenic progenitors in these samples (129.1%), annulling the effect of nutlin-3 in the combination therapy (92.7%). These data suggest that VPA may have different effects in leukemic stem and progenitor cells compared with the major bulk of AML blasts.21

Figure 3. Effect of the combinational therapy of nutlin-3 and VPA on proliferation of primary AML cells. (a) Sensitivity for the treatment with nutlin-3 (5 mM), VPA (500 mM) or combination of both for 48 h (nutlin-3 added for the final 24 h) analyzed by 3H-thymidine incorporation is shown for 56 primary AML samples, with corresponding clinical parameters for each of the AML patients. Samples were analyzed in triplicate; results presented as means and visualized using TMEV microarray software suite version 4.3.01 (Dana-Farber Cancer Institute). Samples are ranked by sensitivity towards the combinational treatment. Sensitivity towards the combinational treatment of nutlin-3 and VPA analyzed by 3H-thymidine incorporation was correlated to clinical parameters of the patients, demonstrating a significant correlation between sensitivity and CD34 expression of the AML blasts (r ¼ 0.3211, P ¼ 0.02). This correlation is visualized in (b), showing differences in means of 3H-thymidine incorporation for CD34low and CD34high (high when 420% of blast cells are positive) patients treated with the combination of nutlin-3 and VPA. Results are given as means±s.e. of mean (*Po0.05, n ¼ 53). ND, not determined; WT, wild type. Leukemia (2012) 910 - 917



Combination of nutlin-3 and VPA significantly inhibits disease development in an in vivo MOLM-13 AML xenograft model We next asked the question if the combination of nutlin-3 and VPA could be effective in an in vivo model of AML with wild type TP53. We thus evaluated the combination in a bioluminescent orthotopic (Supplementary Figure 5),22 and also in a subcutaneous, xenograft model of MOLM-13, employing either nutlin-3 or nultin-3a respectively (details of regimes outlined in Supplementary Materials and methods). Although the limited treatment regime used in the orthotopic model abrogated survival studies, quantification of bioluminescence allowed for direct comparison between treatment groups, and perhaps more significantly, permitted visualization of systemic anti-leukemic efficacy (Figure 4). Although demonstrating minimal in vitro efficacy in MOLM-13 cells, VPA treated animals showed averaged lower bioluminescence (P ¼ 0.1, 0.14 and 0.22 for weeks 1 -- 3 respectively; Figure 4b). Nutlin-3 demonstrated a significant cytoreductive effect upon MOLM-13luc xenografts, particularly on day 14 (P ¼ 0.014 versus control; Figures 4a and b). Remarkably, co-treatment of MOLM-13luc leukemic mice with the combination resulted in considerable inhibition of systemic, leukemic progression as evidenced by significantly reduced whole-body bioluminescence at all time points (P ¼ 0.02, 0.0002 and 0.028 versus controls respectively; Figures 4a and b).

A second subcutaneous MOLM-13 model in NOD-scid mice (n ¼ 5 per group) that did not require a myeloreductiveconditioning regime for xenograft, employing enatiomerically pure nutlin-3a (125 mg/kg) and higher doses of VPA (350 mg/kg; details of regimes outlined in Supplementary Materials and methods) was performed (Figure 4c). Although single agent treatments reduced tumor growth appreciatively in comparison with controls (Figure 4c) the combination group significantly impaired tumor growth in comparison with single agent arms (Figure 4c) and vehicle controls. Sacrifice of animals upon reaching tumor volumes of ethical limit revealed significant increase in survival of the combination animals in comparison with controls (P ¼ 0.007) and nutlin-3a (P ¼ 0.013) or VPA (P ¼ 0.041) treated mice (Figure 4d). Upon termination of the study, (28 days) three of five animals receiving the combination treatment still had tumors below ethical limits when all other animals in the study had been euthanized. DISCUSSION We hypothesized that judicious co-treatment of VPA and nutlin-3 would additively enhance p53 acetylation and induce apoptosis in wild type p53 AML cells.12,13,23,24 The synergy observed in MOLM-13 wild type p53 cells in three independent assays (Figure 1 and Supplementary Figure 1) was remarkable.

Figure 4. Evaluation of combinational therapy of nutlin-3 and VPA in vivo. (a) Dorsal aspect, bioluminescent images of representative animals (i.e. longest surviving animal per group) with color bar illustrating photon counts per raster scan point (1 mm2) per second. (b) Quantification of total body (dorsal and ventral) photon counts (n ¼ 5 per group; except day 28 when only surviving animals were imaged), with error bars representing standard deviation. Nutlin-3 demonstrated appreciably lower photon counts to controls on day 14 (*Po0.05), the combination treatment of nutlin-3 and VPA demonstrated statistically significant depletion of photon counts on days 7 (*P ¼ 0.012), 14 (***P ¼ 0.0002) and 21 (*P ¼ 0.028) with statistical comparison on day 28 impossible because of insufficient surviving animals. (c) Plot of relative mean tumor volumes vs time for NOD-scid mice inoculated with MOLM-13 cells (5 106 per flank) and treated with VPA (350 mg/kg), nutlin-3a (125 mg/kg), the combination of both or control vehicle at the times indicated. The combination exhibited significant inhibition of tumor growth over all control groups. (d) Survival data presented in Kaplan -- Meyer curve, illustrating the efficacy of nutlin-3a and VPA (log-rank P ¼ 0.027) and increased survival of the combination of nutlin-3a and VPA (log-rank P ¼ 0.007 vs controls, P ¼ 0.013 vs nutlin-3a and P ¼ 0.041 vs VPA) in this xenograft model of wild-type p53 AML. & 2012 Macmillan Publishers Limited

Leukemia (2012) 910 - 917

915


916

The importance of p53 modulation and p53 response in this particular treatment was demonstrated by super-additive induction of p53, acetylated p53 and p53 target genes, in addition to abrogated effect of the combination by shRNA against p53 (Figures 1c -- e). Mechanisms involving hyperacetylation of p53 by co-treatment of HDAC inhibitors with nutlin-3 are supported by Palani et al.25 Increased induction of autophagy by the combination therapy was also noted (Figures 1c and f, Supplementary Figure 2), possibly as a result of increased p53. However, p53 has been shown to both induce and inhibit autophagy depending on location in the cell.26 In addition, autophagy may represent both a survival and a cell death mechanism;27 the role of autophagy in the response to the treatment therefore needs to be further explored. Although these results strongly indicate a central role for p53 in this combination they do not rule out the possibility of p53 independent effects. High levels of MDM2 expression have also been associated with higher sensitivity towards nutlin-3.7,28 As MDM2 has been shown to recruit HDAC1 to p53,29 a possible hypothesis that MDM2 inhibition by nutlin-3 may further prevent HDAC activity upon p53 by reducing the recruitment of HDAC1 to p53 is therefore plausible. Also in p53 independent models, the two drugs may have similar effects, as both compounds have been shown to enhance the levels of p73 and p21.12,30,31 In our study, the combination had no effect in cell lines with deleted or mutated p53, whereas there was an effect in patient samples with mutated p53, although these were mainly among the less sensitive samples (Figures 2 and 3). Further examination of mechanisms underlying the synergistic interaction of the combination in both p53 dependent and independent models is warranted. In vitro results demonstrated an overall synergistic induction of apoptosis in primary AML cells (n ¼ 31) (Figure 2). Given the heterogeneity of AML and our cohort, it is not surprising that some of the patients responded to the therapy, whereas some were more resistant. As for any targeted therapy, patient stratification is critical to identifying those that would benefit best from the combination treatment. 3H-thymidine proliferation assay revealed a significant correlation of drug efficacy to CD34low cellular expression (Figure 3), suggesting that patients of a more differentiated phenotype may respond best to the combination of nutlin-3 and VPA. Also patient samples with poor risk cytogenetics seemed to respond well to the treatment (4 of 10 most sensitive patient samples), suggesting that the combination therapy may be of benefit to both standard risk patients and poor prognosis patients. In contrast to the recent findings of Long et al.,32 we did not find any significant correlation between FLT3 mutational status and sensitivity for either nutlin-3 or VPA alone, or for the combined treatment. The role of these and other molecular response predictors33,34 need to be examined in a larger number of samples, ideally in future clinical trials where these therapeutics are evaluated. Although the combination therapy elicited a toxic effect on normal CD34 þ progenitors in vitro (Supplementary Figure 4), the concentration of VPA used in our experiments is equivalent to serum levels normally well tolerated in vivo.8 Furthermore, nutlin-3 exerted minimal toxicity on these cells in vitro, also observed by Kojima et al.7 Ongoing (Nutlin; ClinicalTrials.gov NCT00623870) and future clinical trials are however needed in order to determine the clinical safety of the therapy. To truly evaluate the combination of nutlin-3 and VPA we developed an orthotopic model of wild type TP53 AML that would permit spatio-temporal monitoring of disease pathology and systemic drug efficacy by optical imaging.22 Thus, this orthotopic model represents a more clinically relevant preclinical paradigm. Irradiated NOD-scid and NOD-scid b2mnull models for the xenotransplantation of human AML cell lines,35,36 resulted in a variable or non-leukemic state when transplanted with MOLM-13 (Supplemental Figure 4A). Subsequent xenotransplantation of luciferase expressing MOLM-13 cells in NOD-scid IL2rgnull,37 Leukemia (2012) 910 - 917

resulted in systemic leukemia with reproducible disease latency (Supplemental Figure 4). Optical imaging demonstrated significantly reduced photon counts for mice receiving the combination of nutlin-3 (200 mg/kg) and VPA (50 mg/kg), and to a lesser extent nutlin-3, for the duration of this regime (Figures 4a and b), with relapse to higher photon counts following treatment stop. Interestingly, combination treated mice showed statistically lower numbers of photon counts than controls even at day 21, 9 days after last treatment, which was not observed for nutlin-3 only treated animals. This is in contrast to resumed tumor growth following cessation of nutlin-3a treatment reported in a responsive lymphoma xenograft model, suggesting that continuous dosing is requisite for efficacy.38 In a further subcutaneous in vivo model of MOLM-13 (Figures 4c and d), mice were treated with enatiomerically pure nutlin-3a (125 mg/kg) and VPA at high dose (350 mg/kg). In this study the combination significantly inhibited tumor growth in comparison with controls and the nutlin-3a and VPA only treatment arms (Figure 4c). Furthermore, upon termination of the study on day 28, 60% of the combination treated animals were still alive and within ethical limits, whereas all animals in the other groups had been humanely euthanized (Figure 4d). These results suggest that combination of HDAC inhibitors with MDM2 antagonist is beneficial in the treatment of AML expressing wild type p53, and is to the best of our knowledge the first report of nutlin-3 efficacy in an AML preclinical model. In conclusion, our results demonstrate that combined therapy of VPA and nutlin-3 activates p53 mediated apoptosis synergistically in vitro and in vivo in AML, and propose concomitant inhibition of HDACs and MDM2 as a novel promising nongenotoxic therapeutic option in AML. CONFLICT OF INTEREST The authors declare no conflict of interest.

ACKNOWLEDGEMENTS This study was supported by The Norwegian Cancer Society (Kreftforeningen), The Western Norway Regional Health Authority (E.Mc.C. and B.T.G.) and Bergen Research Foundation. We thank Lena F Hansen, Lene M Vikebø, Michaela Popa, Maren Boge, Kjetil Jacobsen, Jørn Skavland, Bjarte S Erikstein, Andre Sulen, Paulina Ruurs, Lasse Evensen, Marianne Enger, Randi Hovland, Harald Valen, Lars Helgeland, Edith Fick, Line Bjørge, Liv Cecilie Vestrheim and Reidar Myklebust for discussion, expert advice and technical assistance. The optical imaging and transmission electron microscopy was performed at the Molecular Imaging Center (FUGE, Norwegian Research Council), University of Bergen.

REFERENCES 1 Burnett A, Wetzler M, Lowenberg B. Therapeutic advances in acute myeloid leukemia. J Clin Oncol 2011; 29: 487 - 494. 2 Estey E, Dohner H. Acute myeloid leukaemia. Lancet 2006; 368: 1894 - 1907. 3 Laubach J, Rao AV. Current and emerging strategies for the management of acute myeloid leukemia in the elderly. Oncologist 2008; 13: 1097 - 1108. 4 Wojcik I, Szybka M, Golanska E, Rieske P, Blonski JZ, Robak T et al. Abnormalities of the P53, MDM2, BCL2 and BAX genes in acute leukemias. Neoplasma 2005; 52: 318 - 324. 5 Claus R, Lubbert M. Epigenetic targets in hematopoietic malignancies. Oncogene 2003; 22: 6489 - 6496. 6 Vassilev LT, Vu BT, Graves B, Carvajal D, Podlaski F, Filipovic Z et al. In vivo activation of the p53 pathway by small-molecule antagonists of MDM2. Science 2004; 303: 844 - 848. 7 Kojima K, Konopleva M, Samudio IJ, Shikami M, Cabreira-Hansen M, McQueen T et al. MDM2 antagonists induce p53-dependent apoptosis in AML: implications for leukemia therapy. Blood 2005; 106: 3150 - 3159. 8 Gerstner T, Bell N, Konig S. Oral valproic acid for epilepsy--long-term experience in therapy and side effects. Expert Opin Pharmacother 2008; 9: 285 - 292. 9 Bruserud O, Stapnes C, Ersvaer E, Gjertsen BT, Ryningen A. Histone deacetylase inhibitors in cancer treatment: a review of the clinical toxicity and the modulation of gene expression in cancer cell. Curr Pharm Biotechnol 2007; 8: 388 - 400.



917 10 Ryningen A, Stapnes C, Lassalle P, Corbascio M, Gjertsen BT, Bruserud O. A subset of patients with high-risk acute myelogenous leukemia shows improved peripheral blood cell counts when treated with the combination of valproic acid, theophylline and all-trans retinoic acid. Leuk Res 2009; 33: 779 - 787. 11 Bellos F, Mahlknecht U. Valproic acid and all-trans retinoic acid: meta-analysis of a palliative treatment regimen in AML and MDS patients. Onkologie 2008; 31: 629 - 633. 12 Quintas-Cardama A, Santos FP, Garcia-Manero G. Histone deacetylase inhibitors for the treatment of myelodysplastic syndrome and acute myeloid leukemia. Leukemia 2011; 25: 226 - 235. 13 Xu WS, Parmigiani RB, Marks PA. Histone deacetylase inhibitors: molecular mechanisms of action. Oncogene 2007; 26: 5541 - 5552. 14 Chrisanthar R, Knappskog S, Lokkevik E, Anker G, Ostenstad B, Lundgren S et al. CHEK2 mutations affecting kinase activity together with mutations in TP53 indicate a functional pathway associated with resistance to epirubicin in primary breast cancer. PLoS One 2008; 3: e3062. 15 Wergeland L, Sjoholt G, Haaland I, Hovland R, Bruserud O, Gjertsen BT. Pre-apoptotic response to therapeutic DNA damage involves protein modulation of Mcl-1, Hdm2 and Flt3 in acute myeloid leukemia cells. Mol Cancer 2007; 6: 33. 16 Bredholt T, Dimba EA, Hagland HR, Wergeland L, Skavland J, Fossan KO et al. Camptothecin and khat (Catha edulis Forsk.) induced distinct cell death phenotypes involving modulation of c-FLIPL, Mcl-1, procaspase-8 and mitochondrial function in acute myeloid leukemia cell lines. Mol Cancer 2009; 8: 101. 17 Lorens JB, Jang Y, Rossi AB, Payan DG, Bogenberger JM. Optimization of regulated LTR-mediated expression. Virology 2000; 272: 7 - 15. 18 Chou TC. Drug combination studies and their synergy quantification using the Chou-Talalay method. Cancer Res 2010; 70: 440 - 446. 19 Keith CT, Borisy AA, Stockwell BR. Multicomponent therapeutics for networked systems. Nat Rev Drug Discov 2005; 4: 71 - 78. 20 Kawagoe R, Kawagoe H, Sano K. Valproic acid induces apoptosis in human leukemia cells by stimulating both caspase-dependent and -independent apoptotic signaling pathways. Leuk Res 2002; 26: 495 - 502. 21 Bug G, Schwarz K, Schoch C, Kampfmann M, Henschler R, Hoelzer D et al. Effect of histone deacetylase inhibitor valproic acid on progenitor cells of acute myeloid leukemia. Haematologica 2007; 92: 542 - 545. 22 McCormack E, Micklem DR, Pindard LE, Silden E, Gallant P, Belenkov A et al. In vivo optical imaging of acute myeloid leukemia by green fluorescent protein: time-domain autofluorescence decoupling, fluorophore quantification, and localization. Mol Imaging 2007; 6: 193 - 204. 23 Kumamoto K, Spillare EA, Fujita K, Horikawa I, Yamashita T, Appella E et al. Nutlin-3a activates p53 to both down-regulate inhibitor of growth 2 and up-regulate mir-34a, mir-34b, and mir-34c expression, and induce senescence. Cancer Res 2008; 68: 3193 - 3203.

24 Li M, Luo J, Brooks CL, Gu W. Acetylation of p53 inhibits its ubiquitination by Mdm2. J Biol Chem 2002; 277: 50607 - 50611. 25 Palani C, Beck J, Sonnemann J. Histone deacetylase inhibitors enhance the anticancer activity of nutlin-3 and induce p53 hyperacetylation and downregulation of MDM2 and MDM4 gene expression. Invest New Drugs 2010; e-pub ahead of print 3 August 2010; doi:10.007/s10637-010-9510-7. 26 Tasdemir E, Chiara Maiuri M, Morselli E, Criollo A, D’Amelio M, Djavaheri-Mergny M et al. A dual role of p53 in the control of autophagy. Autophagy 2008; 4: 810 - 814. 27 Baehrecke EH. Autophagy: dual roles in life and death? Nat Rev Mol Cell Biol 2005; 6: 505 - 510. 28 Gu L, Zhu N, Findley HW, Zhou M. MDM2 antagonist nutlin-3 is a potent inducer of apoptosis in pediatric acute lymphoblastic leukemia cells with wild-type p53 and overexpression of MDM2. Leukemia 2008; 22: 730 - 739. 29 Ito A, Kawaguchi Y, Lai CH, Kovacs JJ, Higashimoto Y, Appella E et al. MDM2-HDAC1-mediated deacetylation of p53 is required for its degradation. Embo J 2002; 21: 6236 - 6245. 30 Lau LM, Nugent JK, Zhao X, Irwin MS. HDM2 antagonist Nutlin-3 disrupts p73-HDM2 binding and enhances p73 function. Oncogene 2008; 27: 997 - 1003. 31 Blaheta RA, Michaelis M, Natsheh I, Hasenberg C, Weich E, Relja B et al. Valproic acid inhibits adhesion of vincristine- and cisplatin-resistant neuroblastoma tumour cells to endothelium. Br J Cancer 2007; 96: 1699 - 1706. 32 Long J, Parkin B, Ouillette P, Bixby D, Shedden K, Erba H et al. Multiple distinct molecular mechanisms influence sensitivity and resistance to MDM2 inhibitors in adult acute myelogenous leukemia. Blood 2010; 116: 71 - 80. 33 Hu B, Gilkes DM, Farooqi B, Sebti SM, Chen J. MDMX overexpression prevents p53 activation by the MDM2 inhibitor Nutlin. J Biol Chem 2006; 281: 33030 - 33035. 34 Khanim FL, Bradbury CA, Arrazi J, Hayden RE, Rye A, Basu S et al. Elevated FOSBexpression; a potential marker of valproate sensitivity in AML. Br J Haematol 2009; 144: 332 - 341. 35 McCormack E, Bruserud O, Gjertsen BT. Animal models of acute myelogenous leukaemia-- development, application and future perspectives. Leukemia 2005; 19: 687 - 706. 36 Pearce DJ, Taussig D, Zibara K, Smith LL, Ridler CM, Preudhomme C et al. AML engraftment in the NOD/SCID assay reflects the outcome of AML: implications for our understanding of the heterogeneity of AML. Blood 2006; 107: 1166 - 1173. 37 Agliano A, Martin-Padura I, Mancuso P, Marighetti P, Rabascio C, Pruneri G et al. Human acute leukemia cells injected in NOD/LtSz-scid/IL-2Rgamma null mice generate a faster and more efficient disease compared to other NOD/scid-related strains. Int J Cancer 2008; 123: 2222 - 2227. 38 Sarek G, Ojala PM. p53 reactivation kills KSHV lymphomas efficiently in vitro and in vivo: new hope for treating aggressive viral lymphomas. Cell Cycle 2007; 6: 2205 - 2209.

Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)


Leukemia (2012) 910 - 917