Targeting anaplastic lymphoma kinase (ALK) gene alterations in neuroblastoma by using alkylating pyrrole-imidazole polyamides

Anaplastic lymphoma kinase (ALK) aberration is related to high-risk neuroblastomas and is an important therapeutic target. As acquired resistance to ALK tyrosine kinase inhibitors is inevitable, novel anti-ALK drug development is necessary in order to overcome potential drug resistance against ATP-competitive kinase inhibitors. In this study, to overcome ALK inhibitor resistance, we examined the growth inhibition effects of newly developed ALK-targeting pyrrole-imidazole polyamide CCC-003, which was designed to directly bind and alkylate DNA within the F1174L-mutated ALK gene. CCC-003 suppressed cell proliferation in ALK-mutated neuroblastoma cells. The expression of total and phosphorylated ALK was downregulated by CCC-003 treatment but not by treatment with a mismatch polyamide without any binding motif within the ALK gene region. CCC-003 preferentially bound to the DNA sequence with the F1174L mutation and significantly suppressed tumor progression in a human neuroblastoma xenograft mouse model. Our data suggest that the specific binding of CCC-003 to mutated DNA within the ALK gene exerts its anti-tumor activity through a mode of action that is distinct from those of other ALK inhibitors. In summary, our current study provides evidence for the potential of pyrrole-imidazole polyamide ALK inhibitor CCC-003 for the treatment of neuroblastoma thus offering a possible solution to the problem of tyrosine kinase inhibitor resistance.


Introduction
Neuroblastoma is a pediatric cancer and the most frequently occurring extracranial solid tumor in children. The prognosis of neuroblastoma depends on risk group and stage. Patients with high-risk or stage IV neuroblastoma have a survival rate of less than 50%, despite extensive treatment [1]. The anaplastic lymphoma kinase (ALK) gene is known to be one of the factors related to the prognosis of neuroblastoma, and mutations of this gene are related to highrisk neuroblastoma. Point mutations and gene amplifications have been described in approximately 10% of patients with neuroblastoma and 14% of individuals considered as high-risk CCC-003 was achieved in a one-pot procedure: the coupling of the N-terminal-free polyamide backbone in the presence of a large excess (10 eq) of WSC and indole-seco-CBI in order to obtain CCC-003. The compounds PyPyPyPyβPyPyPy-γ-ImImβImImPy-indole-seco-CBI (CCC-003) and PyPyPyβImPyPyIm-γ-PyImPyβImPy-indole-seco-CBI (Mismatch polyamide) were shown in Fig 1A. The mismatch polyamide had no binding motif within the coding region of the ALK gene.

High-resolution denaturing polyacrylamide gel electrophoresis (PAGE)
The detailed method of PAGE analysis was described previously [20]. All DNA fragments were purchased from Hokkaido System Science. DNA samples were dissolved in 1 × TE buffer (pH 7.2). CCC-003 was dissolved in DMSO. ODN 1 and 2 (25 bp) were annealed in a final volume of 20 μL containing 1 μM of each strand ( Fig 1D). Annealing was performed by heating to 95˚C and cooling to the starting temperature of 25˚C at a cooling rate of 2˚C/ 2 min. For DNA alkylation, 2 μL of 1 μM or 5 μM CCC-003, 5 μM Mismatch, or DMSO was added to 4 μL of double stranded DNA fragments in a final volume of 20 μL containing 5 mM sodium phosphate buffer (pH 7.0) and incubated at 25˚C for 18 h. After heating at 95˚C for 25 min, 2 μL of 1 M piperidine was added to the alkylated DNA samples followed by heating at 95˚C for 25 min again. The samples were recovered by vacuum centrifugation at 45˚C for 15 min. The pellets were dissolved in 40 μL of denaturing buffer (formamide: 10 × loading buffer: 10 × TBE buffer: nuclease-free water = 5:1:1:3) and heated at 95˚C for 25 min. The 5 μL of samples were loaded on 20% polyacrylamide gel containing 7 M of urea and electrophoresed (room temperature, 200 V, 90 min) in 1× TBE buffer. The gel was imaged on a ChemiDoc MP Imaging System (Bio-Rad, Hercules, CA, USA).

Cell lines and cell culture
Human neuroblastoma SH-SY5Y, Kelly, and SK-N-AS cells were maintained in DMEM, RPMI1640, and MEM/F12, respectively. All cell lines were supplemented with 10% FBS (Thermo Fisher Scientific, Waltham, MA, USA) as well as penicillin/streptomycin (Thermo Fisher Scientific) and cultured in a humidified atmosphere at 37˚C with 5% CO 2 . SH-SY5Y cells were purchased from ATCC, while Kelly and SK-N-AS cells were obtained from the mutation. (C) SPR sensorgrams for the interaction of CCC-003-Dp with hairpin DNAs containing the F1174L mutation sequence, Mismatch-Dp with hairpin DNAs containing the F1174L mutation sequence, CCC-003-Dp with hairpin DNAs containing the wild-type sequence, and Mismatch-Dp with hairpin DNAs containing the wild-type European Collection of Authenticated Cell Cultures. Murine Ba/F3 cell line was obtained from RIKEN BioResource Center and maintained in RPMI 1640 medium supplemented with 10% FBS, penicillin/streptomycin and 10 ng/ml of recombinant mouse interleukin-3 (IL-3, Fujifilm-Wako Chemicals). Mycoplasma contamination was tested for using the Mycoplasma Detection Set (Takara Bio, Kusatsu, Japan).

Electroporation and cell selection
Expression vectors of wild-type and F1174L-mutated ALK genes were kind gifts from Dr. J. Takita (Kyoto University, Kyoto, Japan). Ba/F3 cells were electroporated with the ALK vectors or pcDNA3 empty vector using a Neon transfection system (Thermo Fisher Scientific) at 1635 V/20 ms/1 pulse. Twenty-four hours after the electroporation, the cells were subjected to G418 selection (1 mg/ml, Merck) supplemented with IL-3 (10 ng/ml) for three weeks. Ectopic expression of the ALK proteins was confirmed by western blot.

Cell proliferation assay (WST assay)
Neuroblastoma cell lines were seeded in 96-well plates at a final density of 3×10 3 cells/well and allowed to attach overnight. The cells were then treated with DMSO or experimental compounds. After 72 h, cell proliferation was determined by WST assay using the Cell Counting Kit-8 (Fuji Film-Wako Chemicals) according to the manufacturer's instructions and was quantified on an MTP-310 Microplate Reader (Corona Electric, Hitachinaka, Japan) or Multiskan Sky (Thermo Fisher Scientific). ALK-expressing Ba/F3 cells were seeded in 96-well plates at 5×10 3 cells/well with different concentrations of IL-3 and cell viability was determined by WST assay. The cells were also treated with CCC-003 without IL-3 for 72 h and subjected to WST assay.

Quantitative reverse transcriptase PCR (qRT-PCR)
Neuroblastoma cell lines were seeded in 6-well plates at a final density of 2.5×10 5 cells/well and allowed to attach overnight. The cells were then treated with DMSO or experimental compounds. After 24 h, RNA was extracted using the RNeasy Plus Mini Kit (Qiagen, Hilden, Germany) following the manufacturer's instructions. Purified RNA (300 ng-1 μg) was used for cDNA synthesis with SuperScript VILO Master Mix (Thermo Fisher Scientific), following the manufacturer's protocol. The q-PCR Master Mix and specific primer sets were used for SYBR Green-based quantitative PCR (q-PCR; Applied Biosystems, Foster, CA, USA) on an Applied Biosystems StepOnePlus Real-Time PCR System (Thermo Fisher Scientific). The PCR primer sequences used were as follows: ALK, 5'-ACTCCAGGGAAGCATGG-3' and 5'-CACTCGAA ATGGGTTGTCTG-3'; RPS18, 5'-GAGGATGAGGTGGAACGTGT-3' and 5'-TCTTCAGTC GCTCCAGGTCT-3'.

Western blot
Neuroblastoma cell lines were seeded in a 10-cm dish at a final density of 1.5×10 6 cells and allowed to attach overnight. The cells were then treated with DMSO or experimental sequence. Oligonucleotides were immobilized on the surface of an SA sensor chip. Five curves of the lowest, mid-low, middle, mid-high, and highest concentration of PI polyamide indicate 100, 200, 300, 400, and 500 nM, respectively. (D) Oligonucleotides used in PAGE analysis. (E) PAGE analysis confirmed the alkylation site on target DNA fragment by CCC-003. Annealed DNA fragment (dsDNA) using 5 0 -TAMRA-labeled DNA oligonucleotides (ODN1 and ODN2) was incubated with DMSO, Mismatch polyamide or CCC-003 and heated to visualize alkylated bands by PAGE analysis. The alkylated DNA fragments were analyzed with reference DNAs (ODN3 and ODN4).

Annexin V staining
Neuroblastoma cells were exposed to experimental compounds for 48 h before the experiment. Annexin V staining was performed using the MEBCYTO Apoptosis Kit (MBL, Nagoya, Tokyo) according to the manufacturer's instructions. Cells were analyzed on a BD FACSVerse flow cytometer using FlowJo Software (Becton Dickinson, Franklin Lakes, NJ, USA).

In vivo study
SH-SY5Y cells (5×10 6 cells) were mixed 1:1 with Matrigel (Corning, New York, NY, USA) and injected subcutaneously into the flanks of total 30 female BALB/c nu/nu mice (5-6 weeks old) purchased from Charles River Laboratories (Yokohama, Japan). After the tumor size reached 500 mm 3 , the mice were randomized into two groups and intravenously injected with DMSO or CCC-003 (0.3 mg/kg/week) once a week for 2 weeks (n = 8 in each group). Tumor volume (W×W×L/2) and body weights were measured every 3 or 4 days. Animal health was monitored by body weight and general assessment of animal behavioral activity. Adverse events were judged by body weight change and a humane endpoint was determined by 20% reduction in body weight. None of mice reached the endpoint during the experiments. When necessary, mice were anesthetized with 2% isoflurane. The mice were sacrificed when the tumor volume reached 2,000 mm 3 by administration of anesthesia followed by cervical dislocation.

Hematoxylin & eosin (H&E) and immunohistochemical staining
Compounds (0.3 mg/kg) or DMSO were injected intravenously into female BALB/c nu/nu mice harboring SH-SY5Y xenografts (n = 3 mice per group) when the tumor volume surpassed 500 mm 3 . After 24 h, tumors were resected and fixed in 4% paraformaldehyde. Sections (4 μm) were deparaffinized by immersion in xylene and were then rehydrated, followed by H&E staining according to standard procedures and immunostaining using an anti-cleaved caspase-3 antibody (Cell Signaling Technology). Cleaved caspase-3-positive cells were counted in 20 high-power fields of each tumor at ×400 magnification.

Statistical analysis
Statistical analyses were performed using GraphPad Prism 6.0 (GraphPad Software, San Diego, CA, USA). For in vitro studies, one-way analysis of variance (ANOVA), followed by a Bonferroni the post-hoc test, was used for multiple group comparisons. For in vivo studies, two-way repeated measures ANOVA, followed by the Bonferroni post-hoc test, was used to evaluate differences in tumor volume and body weight between the groups. For cleaved caspase 3 immunostaining, unpaired t-test was used. P-values below 0.05 were considered statistically significant.

Ethical statement
This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the Ministry of Education, Culture, Sports, Science and Technology of Japan. The protocol was approved by the Committee on the Ethics of Animal Experiments of Chiba Cancer Center (IRB ID #18-2). All efforts were made to minimize their suffering.

CCC-003 preferentially bound to the DNA sequence of mutant ALK
To disrupt the expression of the F1174L mutant ALK gene, we designed and synthesized PI polyamide CCC-003, which specifically recognizes the DNA sequence of mutated ALK, as well as a mismatch polyamide (Mismatch), which had no binding motif within the coding region of the ALK gene ( Fig 1A, S1 and S2 Figs). CCC-003 possesses an alkylating moiety, seco-CBI, which can alkylate the adenine residue adjacent to a C to A mutation within the ALK gene ( Fig 1B).
To validate the specific binding of CCC-003 to mutated ALK, we examined the affinity of PI polyamide to DNA fragments containing the F1174L-mutated sequence using an SPR assay. The non-alkylating variant CCC-003-Dp exhibited high affinity for target F1174Lmutated oligonucleotides when compared to oligonucleotides with the wild-type ALK sequence (Fig 1C and S3-S5 Figs). The dissociation constant (K D ) values of CCC-003-Dp for the mutation and wild-type oligonucleotides were 6.37(± 0.41)×10 −8 and 58.20(± 1.62)×10 −8 M, respectively, and the binding affinity of CCC-003-Dp for the mutation sequence was 9.14-fold higher than for the wild-type sequence (Table 1). Preferential binding of PI polyamide to the mutation and wild-type sequence was not observed for the Mismatch.
To determine the alkylation sites of target DNA fragments by CCC-003, we used high-resolution denaturing PAGE analysis. All DNA oligonucleotides used in this study are listed in Fig  1D. ODN3 and ODN4 were prepared as reference DNAs. As shown in Fig 1E, CCC-003 treatment demonstrated a fragment band corresponding to ODN3, indicating a specific alkylation by CCC-003 and heat-induced DNA cleavage at the targeted adenine adjacent to F1174L mutated sequence of ODN2. On the other hand, no obvious fragmented DNA was observed by Mismatch polyamide treatment. These data suggest that CCC-003 selectively alkylates the adenine nucleotide of F1174L-mutated ALK gene.

CCC-003 suppressed cell viability and ALK expression in ALK-mutated neuroblastoma cells
We assessed anti-proliferative activity of CCC-003 on ALK F1174L-mutant cells using IL-3-dependent murine hematopoietic cell line, Ba/F3. The cells were electroporated with expression vectors for wild-type and F1174L-mutated ALK or empty pcDNA3 vector (Fig 2A). G418 selected cells that expressed ALK proteins showed sustained viability after IL-3 removal ( Fig  2B). We treated these ALK-expressing Ba/F3 cells with CCC-003 and found that CCC-003 suppressed IL-3-independent cell viability in the cells with F1174L-mutated ALK expression but not in those expressing wild-type ALK (Fig 2C). We next performed WST assays in human neuroblastoma cells to examine the effect of CCC-003 on cell proliferation. The IC 50 values of CCC-003 in ALK-mutated SH-SY5Y and Kelly cells (1.8 and 7.3 nM) were lower than those for wild-type ALK-harboring SK-N-AS and SK-N-FI cells (14.3 and >100 nM, respectively, Fig 2D). The expression level of ALK in neuroblastoma cell lines was examined by qRT-PCR. SH-SY5Y and Kelly cells expressed F1174L-mutated ALK, while SK-N-AS cells expressed wild-type ALK at low levels ( Fig 2E). In ALK-mutated neuroblastoma cells, CCC-003 significantly reduced the mRNA and protein expression levels of ALK in a dose-dependent manner (Fig 2F and 2G). These results suggest that CCC-003 downregulates F1174L-mutated ALK expression and suppresses the proliferation of neuroblastoma cells.

CCC-003 affected the F1174L-mutated ALK gene in a different manner from ALK inhibitors
We compared CCC-003 to two ALK inhibitors, crizotinib and alectinib. The WST assay demonstrated that cell growth inhibition in ALK-mutated cells by CCC-003 was more pronounced when compared to crizotinib and alectinib (Fig 3A-3D). As expected, ALK inhibitors suppressed the phosphorylated form of ALK without changing ALK protein expression in ALKmutated cells (Fig 3E). In contrast, CCC-003 treatment suppressed both the phosphorylated form and total ALK expression. These data suggest that CCC-003 inhibits ALK activation by suppressing its gene expression through a mode of action that is different from that of other small-molecule ALK inhibitors.

ALK-mutated neuroblastoma cells were more sensitive to CCC-003 than to the mismatch polyamide
To further assess the targeting specificity, we compared CCC-003 with Mismatch polyamide in WST assays. Mismatch polyamide demonstrated 10.0-and 15.6-fold higher IC 50 values than CCC-003 in SH-SY5Y and Kelly cells, respectively. In contrast, a comparable IC 50 value of Mismatch polyamide was observed in SK-N-AS cells (68.8 nM; Fig 4A). qRT-PCR and western blotting analyses revealed that the mismatch polyamide failed to suppress the expression of ALK at both the mRNA and protein levels (Fig 4B and 4C). Further, western blot analyses revealed the downregulation of Shc adaptor protein phosphorylation.

CCC-003 induced apoptotic cell death in ALK-mutated neuroblastoma cells
To clarify whether CCC-003 induces cell death in ALK-mutated neuroblastoma cells, we performed Annexin V staining. At 48 h after treatment, CCC-003 significantly increased the number of Annexin V-positive SH-SY5Y and Kelly cells in a dose-dependent manner, while the mismatch polyamide had little influence on the viability of ALK-mutated neuroblastoma cells (Fig 5A and 5B and S6 Fig). The number of Annexin V-positive SK-N-AS cells did not change and ALK F1174L-mutated SH-SY5Y and Kelly cells were treated with DMSO for 24 h. ALK mRNA expression was analyzed using qRT-PCR. RPS18 was used as an internal control. P-values were determined using one-way ANOVA followed by a Bonferroni post-hoc test ( ��� , P < 0.001). Data are presented as the mean ± SD from three independent experiments. (F) SK-N-AS, SH-SY5Y, and Kelly cells were treated with 3-30, 1-10, or 10-100 nM CCC-003 for 24 h, respectively. ALK mRNA expression was analyzed using qRT-PCR. DMSO was used as a control treatment. RPS18 expression was used as a housekeeping gene. P-values were determined using one-way ANOVA followed by a Bonferroni post-hoc test ( �� , P < 0.01; ��� , P < 0.001). Data are presented as the mean ± SD from three independent experiments. (G) SK-N-AS and SH-SY5Y cells were treated with 1-10 nM CCC-003, and Kelly cells were treated with 30-100 nM CCC-003 for 48 h. ALK protein levels were analyzed using western blotting.

CCC-003 suppressed tumor progression in an SH-SY5Y-xenograft mouse model
We evaluated the anti-tumor potential of CCC-003 in a neuroblastoma xenograft model established using SH-SY5Y cells. Tumor-bearing mice that received an intravenous injection of CCC-003 at 0.3 mg/kg BW displayed significant suppression of tumor growth compared with vehicle-treated mice, without weight loss (Fig 6A and S7A Fig). We also performed histological

PLOS ONE
analysis of SH-SY5Y xenograft-derived tumor tissue sections 24 h after administration of CCC-003 (Fig 6B and 6C). HE staining revealed chromatin condensation and nuclear fragmentation in tumors treated with CCC-003 (Fig 6C). Immunohistochemical staining using the cleaved caspase-3 antibody revealed an elevated level of apoptotic cell death in tumor tissues after CCC-003 administration (Fig 6C and 6D). There were no histological changes in tissues of the lung, liver, and kidney after CCC-003 treatment, suggestive of minimal side effects in mice (S7B Fig).

Discussion
ALK TKIs have been evaluated in ALK-driven neuroblastoma cell lines, mouse models, and ongoing clinical trials [2,21,22]. Currently, alectinib and crizotinib are considered effective in the treatment of ALK-positive lung cancer [23]. However, relapse due to acquired ALK inhibitor resistance caused by secondary ALK mutations often occurs in lung cancer patients [24,25]. A similar resistance development event may occur in neuroblastoma patients after ALK inhibitor treatment. To overcome the issue of ALK inhibitor resistance in neuroblastoma, we used a novel PI polyamide designed to target the mutant ALK gene in a manner different from that of ATP-competitive inhibitors. ALK F1174L mutation is a major aberration observed in both primary and relapsed neuroblastomas [2,26]. Herein, we developed CCC-003, an alkylating PI polyamide, which preferentially binds the genomic DNA of the ALK F1174L mutation, suppressing its expression and retarding the growth of ALK F1174L-positive xenograft tumors. F1174L-mutated ALK expression was previously demonstrated to have an anti-apoptotic effect during neuroblastoma genesis [27,28]. CCC-003 treatment upregulated apoptotic cell death in neuroblastoma cells and tumor tissues, suggesting that the suppression of oncogenic mutated ALK by CCC-003 was sufficient to upregulate pro-apoptotic signaling and thus exert antitumor effects in ALK-positive neuroblastoma.
In accordance with previous studies [29,30], treatment with ALK inhibitors suppressed the phosphorylated form of ALK without affecting total ALK levels in SH-SY5Y and Kelly cells. In contrast, CCC-003 treatment downregulated ALK expression and resulted in a reduced level of its phosphorylated form, suggesting a distinct mode of action for CCC-003 compared to other small molecule TKIs. According to previous reports [17,31,32], covalent alkylation of the coding DNA strand by PI polyamide with seco-CBI specifically inhibits the transcription of target genes. In the case of CCC-003, the expression of ALK mRNA was suppressed at 24 h after treatment, suggesting that CCC-003 rapidly translocated into the cell nucleus, and selective alkylation occurred at the mutation site of ALK. ALK expression was slightly reduced by mismatch polyamide treatment. The mismatch polyamide was designed to have no binding sequence in the region of ALK and thus exhibited low affinity to the wild-type sequence oligonucleotides in SPR assays, indicative of the sequence-specific manner of ALK inhibition by CCC-003. Even though CCC-003 has the unique property of a gene silencer, potential off-target effects on other functional genes are a concern because CCC-003 recognizes a relatively short DNA motif of 10 bp. Further examination of CCC-003 binding sites in the human genome is necessary to clarify its genome-wide impact and possible side effects in neuroblastoma patients. Alectinib has been shown to have high efficacy against crizotinib-resistant F1174L ALKpositive cancer cells [19,33,34]. Nevertheless, grade 3-4 adverse events are observed in response to alectinib in ALK-rearranged lung cancer [33,35]. In the present study, CCC-003 had a 10-fold lower IC 50 value compared to these two inhibitors, suggesting that it is more cytotoxic in neuroblastoma cells than other ALK inhibitors. CCC-003 treatment resulted in decreased phosphorylation of the Shc adaptor protein, indicating that proliferative signaling via ALK may be downregulated by CCC-003 at a dose low enough to avoid possible side effects. Of note, intratumoral accumulation has been reported for the in vivo dynamics of PI polyamides [36,37]. These unique characteristics of PI polyamides may have led to the negligible adverse effects observed in the xenograft mouse models treated with CCC-003.
Comparable IC 50 values for alectinib were observed among neuroblastoma cell lines, regardless of the MYCN amplification status of cells [34]. In the present study, when compared to SH-SY5Y cells, CCC-003 exhibited weaker cytotoxic activity with a higher IC 50 value in Kelly cells. There are several possible reasons for the different sensitivity of these two cell lines. According to a previous report, MYCN expression determines anti-cancer drug sensitivity in neuroblastoma cells [38]. Our results indicated that the amplified MYCN in Kelly cells might affect their sensitivity to CCC-003. Moreover, a study reports that the sensitivity to apoptosis of MYCN-overexpressing neuroblastomas is increased through the DNA damage response [39]. To clarify the association between MYCN status and sensitivity to CCC-003 in neuroblastoma cells, further examination of the DNA damage repair system in SH-SY5Y and Kelly cells would be necessary.
In the present study, we successfully developed CCC-003, a novel PI polyamide conjugated with a DNA-alkylating agent. CCC-003 exhibited high cytotoxicity in ALK-mutated neuroblastoma cells. The current work highlighted a novel strategy for the development of therapeutic agents for neuroblastoma capable of direct alkylation of target mutant gene loci. The specific binding to mutated DNA is a unique mode of action of PI polyamide leading to the induction of tumor cell death. Owing to its specific binding, CCC-003 could be a novel therapeutic agent for ALK-positive neuroblastoma without significant side effects. Since PI polyamides can be designed to recognize and target the sequences of aberrant genes, these unique compounds could serve as novel personalized agents against different types of malignant tumors harboring various genetic aberrations, including point mutations, translocations, as well as amplifications.