Small molecule inhibitors of aurora-a induce proteasomal degradation of N-myc in childhood neuroblastoma

Abstract

Amplification of MYCN is a driver mutation in a subset of human neuroendocrine tumors, including neuroblastoma. No small molecules that target N-Myc, the protein encoded by MYCN, are clinically available. N-Myc forms a complex with the Aurora-A kinase, which protects N-Myc from proteasomal degradation. Although stabilization of N-Myc does not require the catalytic activity of Aurora-A, we show here that two Aurora-A inhibitors, MLN8054 and MLN8237, disrupt the Aurora-A/N-Myc complex and promote degradation of N-Myc mediated by the Fbxw7 ubiquitin ligase. Disruption of the Aurora-A/N-Myc complex inhibits N-Myc-dependent transcription, correlating with tumor regression and prolonged survival in a mouse model of MYCN-driven neuroblastoma. We conclude that Aurora-A is an accessible target that makes destabilization of N-Myc a viable therapeutic strategy.

Figure 1

Effect of MLN8054 on N-Myc expression and stability in MYCN-amplified neuroblastoma cell lines. (A) The panels show immunoblots of IMR-32 and IMR-5 cells treated for 48 hr with the indicated concentrations of MLN8054 or solvent control (0.1% v/v DMSO). The results are representative of four independent experiments (n=4; unless otherwise indicated, n indicates the number of independent experiments in the following legends). (B) The panels show representative immunoblots of five MYCN-amplified neuroblastoma cell lines upon treatment with MLN8054 (1000 nM; 48 hr) or solvent control (n=3). Ratios of N-Myc to Vinculin are indicated below the panel. (C) The graphs show qRT-PCR assays of MYCN mRNA levels relative to β2-microglobulin and RPS14 as control. RNA was isolated from the indicated cell lines treated with MLN8054 (1000 nM, 24 hr) or solvent control. Error bars indicate Mean + standard deviation (SD) of triplicate technical assays (n=2). (D) IMR-32 cells were pre-incubated with 1000 nM MLN8054 or solvent control for 24 hr as indicated, then treated with cycloheximide (100 μg/ml), harvested at the indicated time points, and immunoblotted with the indicated antibodies. Exposures of N-Myc blots were adjusted to equalize exposure at 0 min (n=5). (E) Quantification of five independent N-Myc stability assays performed as described in panel D. Data are plotted as mean +/− SEM. Levels of N-Myc at 0 min were set to 1 for both conditions (*** p<0.0001; ** p<0.02; differences at later time points were not statistically significant). (F) IMR-32 cells were pre-treated with 1000 nM MLN8054 or solvent control for 24 hr as indicated and subsequently labeled with ³⁵S-labeled methionine. The panel shows immunoprecipitations using N-Myc antibodies or controls after the indicated chase times. (G) Quantitation of N-Myc stability assays. The panel shows the average (Mean+/−SEM) of three independent pulse-chase experiments performed as described in panel F in IMR-32 cells (** p<0.02; * p<0.05). (H) IMR-32 cells were treated for 14 hr with 1000 nM MLN8054 in the presence or absence of MG132 (5 μM) then immunoblotted for the indicated proteins (n=2). See also Figure S1.

Figure 2

Structurally related Aurora-A inhibitors dissociate Aurora-A/N-Myc complexes. (A) IMR-32 cells were treated with MLN8054 (1000 nM, 4 hr) in the presence of MG132 (5 μM). Subsequently, cell lysates were immunoprecipitated with either control or N-Myc antibodies and probed with the indicated antibodies. Input: cell lysate loading control. Relative amounts of Aurora-A and of Max recovered in N-Myc immunoprecipitates are indicated below each panel (n=4). Asterix indicates light chain of antibody. (B) The experiment was performed as described in (A) in the presence of the indicated concentrations of MLN8237 (n=2). (C) IMR-32 cells were treated with either DMSO or the indicated concentrations of MLN8237 (4 hr) in the presence of MG132. Subsequently, cell lysates were immunoprecipitated with either control or Aurora-A antibodies. Immunoblots depict amounts of the indicated proteins (n=4). Input: cell lysate loading control. (D) Proximity ligation assays analyzing N-Myc/Aurora-A and N-Myc/Max complexes in SH-EP cells stably expressing N-Myc. Blue, DAPI staining of nuclei; red dots, PCR amplification products indicating complex formation of N-Myc with Aurora-A or with Max as indicated. Scale bars indicate 10 μm. Where indicated, MLN8237 (500 nM) or MK5108 (1000 nM) was added for 4 hr (n=3). SH-EP cells not expressing N-Myc serve as negative control. Lower panel shows Mean+SD of triplicate biological replicates. (E) IMR-32 cells were treated with the indicated concentrations of CCT137690 (4 hr) in the presence of MG132. N-Myc/Aurora-A complexes were analyzed as described in panel B. (F) Quantitation of Aurora-A/N-Myc complexes in response to treatment with MLN8237, CCT137690 and MK5108. Data are shown for one experiment for each inhibitor performed as described above. See also Figure S2.

Figure 3

Effects of Aurora-A inhibitors on N-Myc in cells expressing wild-type Aurora-A and drug-resistant Aurora AT217D. (A) MYCN-amplified SMS-KCN cells were stably transduced with expression plasmids encoding a tetracyclin-dependent transactivator. They were subsequently stably transduced with control vector or vector expressing HA-tagged wild-type Aurora-A or Aurora-AT217D. Pools of cells were subsequently treated with MLN8054 (500 nM) and doxycycline (1 μg/ml) as indicated. The panels show colony assays stained with crystal violet 4 days later (n=3). (B) Immunoblot analysis of the SMS-KCN cells described in (A) depicting expression of the indicated proteins; MLN8054 was added for 48 hr (500 nM) (n=3). (C) The SMS-KCN cells described in (A) were treated with MLN8054 (2000 nM; 4 hr) in the presence of MG132 (5 μM) and doxycycline (1 μg/ml; 12 hr) as indicated. Cell lysates were prepared and immunoprecipitated with HA-antibodies. The panels show immunoblots probed with the indicated antibodies. Levels of N-Myc recovered in HA immunoprecipitates (relative to DMSO-treated cells) are indicated below each lane (n=2). (D) IMR-32 cells were incubated with the indicated series of MLN8054 derivatives (1000 nM; 24 hr). The upper panels show immunoblots of cell lysates probed with indicated antibodies (n=3). IC₅₀ data of these compounds from in vitro assays of Aurora-A kinase are shown at the bottom. See also Figure S3.

Figure 4

Role of Fbxw7 in MLN8054-induced degradation of N-Myc. (A) SH-EP cells were stably transduced with retroviruses expressing either N-Myc (WT) or N-MycT58AS62A (mut) then incubated with MLN8054 (1000 nM; 48 hr) or solvent control as shown. The panels show immunoblots of cell lysates probed with the indicated antibodies (n=3). (B) Immunoblots of IMR-32 cells that were transfected with indicated siRNAs. Cells were incubated with MLN8054 or solvent control for 24 hr and analyzed as in panel A (n=3). Ratios of N-Myc to Vinculin are indicated below the panel. (C) IMR-32 were transfected with siRNA targeting FBXW7 or control siRNA. 24 hr before harvesting, cells were treated with MLN8054 (1000 nM) or DMSO as control. CHX was added for the indicated times. Exposures of N-Myc blots were adjusted to equalize exposure at 0 min (n=3). (D) Quantification of three independent stability experiments performed as described for panel C. Levels of N-Myc at 0 min were set to 1 for all conditions. Values are shown as mean +/− SEM (* p<0.02 relative to siFbxw7-treated cells). (E) SH-EP cells engineered to stably express wild type N-Myc were incubated with MLN8054 (1000 nM) for the times shown. Immunoblots depict expression of the indicated proteins (n=2). (F) IMR-32 were incubated with MLN8054 (1000 nM) for the indicated times before lysates were prepared and blotted as before. (G) IMR-5 neuroblastoma cells were incubated with the indicated drugs for 18 hr. The panels show immunoblots of cell lysates probed with the indicated antibodies. Amounts of N-Myc relative to Cdk2 are indicated below each lane (n=2).

Figure 5

Role of N-Myc in cellular responses of MYCN-amplified neuroblastoma cells to MLN8054. (A) The panel on the left shows crystal violet-stained culture dishes of SMS-KCN cells grown in the presence of MLN8054 (250 nM) or solvent control for four days. The graph in the middle documents proliferation of SMS-KCN cells under the same conditions. Error bars indicate SD of triplicate biological replicates. Immunoblots on the right depict levels of the indicated proteins 48 hr after addition of MLN8054. (B) FACS analysis showing cell cycle phases of SMS-KCN cells in response to MLN8054 (250 nM; 48 hr) (left). The data were combined with the growth curve shown in panel (A) to calculate the length of each cell cycle phase (right). Error bars indicate SD of triplicate biological replicates. (C) The graph shows levels of CDKN1A mRNA relative to β2-microglobulin and RPS14 as control determined by qRT-PCR. The indicated cell lines were incubated with MLN8054 (1000 nM) or DMSO for 48 hr before RNA was prepared. Error bars indicate SD of triplicate technical replicates from one of two experiments with identical results. (D) LAN-1 cells were stably transduced with lentiviruses expressing either wild-type N-Myc or N-Myc_mut and the experiment performed as described in panel C. Error bars indicate SD of triplicate technical replicates. (E) Immunoblots documenting expression of N-Myc or N-Myc_mut in LAN-1 cells exposed to MLN8054 (1000 nM; 48 hr) or solvent control. Note that the ectopically expressed N-Myc is not completely degraded in response to MLN8054. (F) SMS-KCN cells were infected with either control viruses (empty) or retroviruses expressing N-Myc or N-Myc_mut together with a GFP marker. Cells were mixed with uninfected cells such that 20% of cells expressed GFP at the start of the experiment. The graph shows the percentage of GFP-positive cells during culture in the absence or presence of MLN8054 (100 nM). Error bars represent SD of triplicate biological samples. From the data, the following doubling times can be estimated: Control cells: DMSO: 26hr; MLN8054: 36hr; N-Myc and N-Myc_mut expressing cells: DMSO: 25hr; MLN8054: 32 hr (n=2). (G) Immunoblots documenting expression of the indicated proteins at the end of the experiment shown in panel F. Note that cells growing in the presence of MLN8054 have selected for elevated levels of N-Myc. See also Figure S4.

Figure 6

Effects of MLN8054 on N-Myc function in TH-MYCN neuroblastoma mice. (A) Tumor-bearing TH-MYCN mice were treated with for 3 days with 40 mg/kg MLN8054 or vehicle. The panels on the left show immunoblots of lysates of neuroblastomas probed with the indicated antibodies. Spleen tissue is included as negative control for N-Myc antibody. The panel on the right shows a quantification of the immunoblot (data are plotted as mean + SEM). (B) Results of ChIP assays using N-Myc or control antibodies at a genomic region surrounding the E-Box of the APEX gene or a negative control region as indicated. Error bars indicate SD of triplicate technical replicates. (C) RNA was isolated from four MLN8054-treated and from four vehicle-treated neuroblastomas and subjected to microarray analysis. The panels show expression of well-characterized gene sets of MYC-regulated genes and their response to MLN8054 treatment. (D) The graphs show Kaplan-Meier diagrams of human patients stratified by the expression levels of the gene sets shown in (C). Graphs and p-values are shown for overall survival. See also Figure S5.

Figure 7

In vivo trials of MLN8054 and MLN8237 in TH-MYCN neuroblastoma mice. (A, B) Representative day 0 and 7 magnetic resonance image sections of mice treated with vehicle or with 40 mg/kg MLN8054 (A) or 30 mg/kg MLN8237 (B). Dashed white lines indicates tumor circumference. (C) H&E staining of sections of tumors from mice treated with vehicle or 40 mg/kg MLN8054 at day 3, high magnification views are shown on the right). Black arrows (top row) indicate neuronal differentiation with the presence of variably maturing ganglion-like cells. Green arrows (middle row) indicate mitotic cells. Orange arrows (bottom row) show a multivacuolated histiocytic component around tumors. Scale bars: 50 μm (left and middle panels) 10 μm (right panels). (D) Immunohistochemistry of tumors treated as in A, using indicated antibodies. Scale bar: 50 μm. (E) Kaplan-Meier plot documenting survival of mice treated twice daily with 15 mg/kg MLN8054 vs. vehicle. (F) Kaplan-Meier plot documenting survival of mice treated with 30 mg/kg MLN8237 vs. vehicle. See also Figure S6.

Figure 8

Model to explain our findings. In normal neuroblasts, degradation of N-Myc is initiated by phosphorylation at S62 of N-Myc by cyclin B/CDK1; this primes N-Myc for phosphorylation by GSK3 at T58. The pool of doubly-phosphorylated N-Myc in cells in G2/M is recognized by Fbxw7. Association of N-Myc with Aurora-A in neuroblastoma cells blocks degradation by Fbxw7 and allows N-Myc to persist into the subsequent cell cycle. MLN8237 and MLN8054 disrupt the association, thereby restoring a more physiological pattern of N-Myc degradation.