4-Hydroxytamoxifen (59002) was obtained from Merck Chemicals (Darmstadt, Germany), doxycycline (631311) from Clontech Laboratories (Mountain View, CA, USA), and PP2A Immunoprecipitation Phosphatase Assay Kit (17-313) from Merck Millipore (Darmstadt, Germany). Dasatinib (S1021), dactolisib (S1009), sorafenib (S7397), erlotinib (S7786), lorlatinib (S7536), afatinib (S1011), trametinib (S2673), and SC79 (S7863) were from Selleck (Munich, Germany). OP449 was provided by Oncotide Pharmaceuticals (Durham, NC, USA).

The human NB cell lines SK-N-AS, SK-N-SH, and SK-N-BE(2)-C were purchased from ATCC (LGC Standard, Wesel, Germany). GI-M-EN, SH-SY5Y, IMR32, LAN5, and KELLY cells were from DSMZ (Braunschweig, Germany), and NB69 cells were from ECACC (Salisbury, UK). IMR5 and NLF cells were a gift from G. M. Brodeur (CHOP, Philadelphia, PA, USA). SK-N-AS and SK-N-SH cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, Paisley, UK) supplemented with 10% of heat-inactivated fetal calf serum (FCS; Gibco), SH-SY5Y cells in DMEM with 20% FCS, and SK-N-BE(2)-C cells in a 1:1 mixture of EMEM and Ham’s F12 (Gibco) with 10% FCS. GI-M-EN, SH-EP, NLF, and KELLY cells were grown in Roswell Park Memorial Institute (RPMI) 1640 medium (Gibco) with 10% FCS, NB69 in RPMI 1640 medium with 15% FCS, and IMR5, IMR32, and LAN5 cells in RPMI 1640 medium with 20% FCS. All media were supplemented with 2 mM of l-glutamine (Gibco) and 100 U/ml of penicillin/streptomycin (Gibco). The authenticity of cells was verified by short tandem repeat (STR) profiling using the Cell Line Authentication Kit (Sigma-Aldrich, Taufkirchen, Germany). Cells were tested for mycoplasma contamination using the LookOut Mycoplasma PCR Detection Kit (MP0035, Sigma-Aldrich).

GeneArt CRISPR Nuclease Vectors (Life Technologies, Darmstadt, Germany) expressing guiding RNA against PPP2CA or PTPA, and Cas9 and GFP were lipofected into KELLY NB cells. Transfected cells were sorted for green fluorescent protein (GFP) expression and deposited as single cells in 96-well plates. Clones from single cells were randomly picked and subjected to Western blotting for PPP2CA, PTPA, and control tubulin. The percentage of clones with complete ablation of PPP2CA and PTPA proteins, of all clones investigated, was calculated. Sanger sequencing was used to determine the presence of premature stop codons leading to homozygous or heterozygous gene knock-outs.

SH-EP MYCN-ER cells were used as previously described (37). Cells were cultured in the presence or absence of 300 nM of 4-hydroxytamoxifen (4-OHT). Total RNA was extracted using Direct-zol RNA Kit (R2051, Zymo Research, Freiburg, Germany) according to the manufacturer’s instructions and reverse-transcribed using SuperScipt III (18080-51, Invitrogen, Darmstadt, Germany). Gene expression of ODC1 was determined by quantitative real-time PCR 48 h after activation of MYCN using forward primer GCATGTGGTGATTGGATGGTGTT, reverse primer TGGCTCTGGATCTGCCTTCATGAGT, and Advanced Universal SYBER Green Supermix (1725274, Bio-Rad, Feldkirchen, Germany) in a CFX Connect Thermal Cycler (Bio-Rad).

SH-EP MYCN-ER cells were stably transduced with two lentiviral shRNAs targeting PPP2CA or a control shRNA with a non-silencing sequence at a multiplicity of infection (MOI) of 50 (TRIPZ, RHS4696-201904073, RHS4696-201904413, and RHS5087-EG332; Dharmacon, CO, USA) according to the manufacturer’s instructions. After selection with puromycin, cells were treated with 5 μg/ml of doxycycline to induce the expression of shRNA. Knock-down efficiency of PPP2CA was assessed by Western blotting.

SH-SY5Y and KELLY cells were seeded at 5 × 10⁴ cells per well or SH-EP MYCN-ER cells were seeded at 2 × 10⁴ cells per well into 6-well plates overnight. SH-SY5Y and KELLY cells were treated with increasing concentrations of OP449 on day 1. With the use of flow cytometry, dead, i.e., propidium iodide-positive, cells were excluded, and viable cells were counted. SH-EP MYCN-ER cells were treated with or without 4-OHT, and shRNA was induced by 1 µg/ml of doxycycline. Apoptosis was determined by quantification of DNA fragmentation using fluorescence-activated cell sorting (FACS) analysis of propidium iodide-stained nuclei. Cells were resuspended in 300 µl of hypotonic fluorochrome solution containing 0.1% sodium citrate, 0.1% Triton X-100, and 50 µg/ml of propidium iodide in distilled water. Cells were incubated at 4°C overnight, and hypodiploid nuclei were determined by flow cytometry. Specific apoptosis was calculated according to the following formula: 100 × [experimental cell death (%) − spontaneous cell death (%)]/[100 − spontaneous cell death (%)].

In supplemented RPMI 1640 medium for KELLY and SH-EP MYCN-ER cells and supplemented DMEM for SH-SY5Y cells, 0.75 × 10³ cells per well were seeded in triplicates into 6-well plates and allowed to attach overnight. SH-SY5Y and KELLY were treated with OP449 on days 1 and 3. SH-EP MYCN-ER cells were treated with vehicle or with 4-OHT, and shRNA was induced by 1 µg/ml of doxycycline. After 1 week (SH-EP MYCN-ER cells) and 2 weeks (SH-SY5Y and KELLY cells) in culture, colonies were stained with crystal violet solution in 3.7% formaldehyde.

Experiments were carried out in triplicates in 24-well plates coated with a layer of 0.6% low-gelling-temperature agarose (A0701, Sigma-Aldrich) with DMEM or RPMI medium. In 0.5% agar with DMEM or RPMI medium, 2 × 10³ cells per well were seeded as a single-cell suspension. OP449 in 1 ml of culture medium was added above the top agar on days 1 and 3 in SH-SY5Y and KELLY cells. For the SH-EP MYCN-ER cell soft agar assay, the medium contained 300 nM of 4-OHT and 1 µg/ml of doxycycline in 1 ml of culture medium, which was exchanged every 48 h. After 1 week (SH-EP MYCN-ER cells) and 2 weeks (SH-SY5Y and KELLY cells) in culture, colonies that had formed within the soft agar were stained with 1 mg/ml of MTT in phosphate-buffered saline (PBS).

Cells measuring 1 × 10⁴ were plated in 96-well plates, incubated overnight, and then exposed for 24, 48, 72, and 96 h to increasing concentrations of OP449. Cell viability was determined by MTT assay, with the viability of dimethyl sulfoxide (DMSO)-treated controls set at 100%.

Cells measuring 3 × 10⁴ were plated in 96-well plates and incubated overnight. Cells were pretreated for 0.5 h with SC79 at 10 µM for KELLY cells and 20 µM for SH-SY5Y cells. SC79 was continued for an additional 2 h in the presence of increasing doses of OP449. Cell viability was determined by MTT assay, with the viability of DMSO- or PBS-treated controls set at 100%.

The R2: Genomics and Visualization Platform (http://r2.amc.nl) was used to analyze the expression of SET mRNA (gse45545) and SET protein (38) in NB patients.

PP2A activity of SH-SY5Y and KELLY cells was determined using PP2A Immunoprecipitation Phosphatase Assay Kit (17-313, Merck) according to the manufacturer’s protocol. Cells were treated with increasing concentrations of OP449 for 4 h and lysed, and PPP2CA, the catalytic subunit of PP2A, was immunoprecipitated. The precipitate was incubated with K-R-pT-I-R-R as a substrate and with Malachite Green for the detection of released (free) phosphate. PP2A activity was measured in supernatants at 450 nm using a TECAN microplate reader (TECAN, Männedorf, Switzerland).

Cells were lysed in radioimmunoprecipitation assay (RIPA) buffer (Tris pH 8, 50 mM), NaCl (150 mM), 1% NP-40, 0.1% sodium dodecyl sulfate (SDS), 1% sodium deoxycholate (DOC), EDTA (pH 8, 1 mM), 2 mM of dithiothreitol (DTT), and Protease Inhibitor Cocktail EDTA-free (Roche, Basel, Switzerland). Whole protein lysate measuring 20 µg was resolved on 4%–12% Bis-Tris gels (NP0335BOX, P0323BOX, Invitrogen). Nitrocellulose membranes were blocked with 5% non-fat dry milk in Tris-buffered saline with Tween 20 (25 mM of Tris-HCl (pH 7.4), 137 mM of NaCl, and 0.1% Tween 20) and then incubated overnight at 4°C with the following antibodies: mouse anti-phospho-PPP2CA (sc-271903, Santa Cruz Biotechnology, Heidelberg, Germany, 1:500) and mouse anti-PPP2CA (610555, BD Transduction Laboratories, CA, USA, 1:5,000). For the detection of phosphorylated proteins, cells were lysed in RIPA buffer (Tris-HCl pH 7.5, 10 mM), NaCl (150 mM), EDTA (2 mM), 1% Triton-X 100, 10% glycerol, and DTT (2 mM), and Western blotting analyses were performed with rabbit anti-ERK1/2 (M5670, Sigma, 1:10,000), mouse anti-ERK-phospho-Thr202/Tyr204 (612358 BD Transduction Laboratories, 1:2,000), rabbit anti-AKT (9272, Cell Signaling Technologies, Frankfurt, Germany, 1:1,000), mouse anti-AKT-phospho-Thr308 (558316 BD Transduction Laboratories, 1:2,000), rabbit anti-AKT-phospho-Ser473 (9271, Cell Signaling Technologies, 1:2,000), rabbit anti-p70S6-phospho-Thr389 (9205, Cell Signaling Technologies, 1:1,000), and p70S6 (9202, Cell Signaling Technologies, 1:1,000) antibodies.

As secondary antibodies, goat anti-mouse IgG (H+L)–horseradish peroxidase (HRP) conjugate (170-6516, Bio-Rad, 1:10,000) and goat anti-rabbit IgG (H+L)-HRP conjugate (65-6120, Invitrogen, 1:10,000) were used for 1 h at room temperature.

For detection, enhanced chemiluminescence (ECL) (GE Healthcare, Buckinghamshire, UK) and WESTAR ETA ULTRA (Cyanagen, Bologna, Italy) reagents were employed.

The concentration ranges of OP449 and kinase inhibitors were selected to span concentrations below and above the IC₅₀ values for each drug. The ratio of the drugs in combinations was fixed according to the ratio of the IC₅₀ of the drugs. The combination index (CI) of survival rates in the combination studies was calculated according to the Chou–Talalay method (39) using CompuSyn software. Weighted average CI values were calculated as CI_wt = (CI₅₀ + 2CI₇₅ + 3CI₉₀ + 4CI₉₅)/10, where CI₅₀, CI₇₅, CI₉₀, and CI₉₅ represent the CI values obtained at 50%, 75%, 90%, and 95% inhibition of cell viability, respectively (39).

To assess whether PPP2CA may be amenable to therapeutic inhibition, we ablated its gene in NB cells using CRISPR/Cas9. The KELLY cell line was selected, as this is a paradigmatic MYCN-amplified NB cell line with mutated ALK, p53, and ARF. In addition, the copy number of 9q34 is increased, where both PTPA, the regulatory subunit of PP2A that activates PPP2CA, and SET, the major endogenous inhibitor of PP2A, are located. We reasoned that if PPP2CA were essential for NB viability, then few, if any, knock-out clones lacking PPP2CA should be generated, as such clones would die out after knock-out. Indeed, none of the many randomly picked clones that grew out had lost PPP2CA protein ( Figure 1B ). The lack of homozygous knock-out in the surviving clones was confirmed on the genomic level by sequencing ( Figure 1B ). Only two clones had a homozygous knock-out, and these two clones still harbored PPP2C protein, possibly an alternative PPP2C isoform. Of note, in the heterozygous knock-outs, the remaining allele apparently upregulated PPP2CA protein expression, as protein levels were not decreased ( Figure 1B ). These data show that KELLY NB cells do not tolerate a complete lack of PPP2CA. In silico data mining of genome-scale CRISPR/Cas9 knock-out screens confirmed the dependency of KELLY cells on PPP2CA and extended this conclusion to other NB cell lines and additional cancer entities ( Figure 1C ). In contrast to ablation of PPP2CA, knock-out of PTPA in KELLY cells resulted in the outgrowth of many clones with a homozygous knock-out (46% of randomly picked clones), all of them completely devoid of PTPA protein ( Figure 1D ). Thus, these data strongly suggest that PPP2CA, but not PTPA, is an essential survival gene in NB and thus a suitable target for its therapy.

Being unlikely that therapeutic targeting of PPP2CA can achieve complete inactivation of PPP2CA, we investigated the effects of incomplete depletion of PPP2CA by knocking down PPP2CA in SH-EP MYCN-ER NB cells. As we wanted to know how MYCN influences the effects of PP2A inhibition, SH-EP MYCN-ER cells, where MYCN can be activated by the addition of 4-hydroxytamoxifen (4-OHT), were used. Induction of PPP2CA lentiviral shRNAs in SH-EP MYCN-ER repressed expression of PPP2CA in MYCN-activated and non-activated cells ( Figure 2A , left panel). MYCN activation was verified by induction of ODC1, a bona fide transcriptional target gene of MYCN ( Figure 2A , right panel).

Knock-down of PPP2CA decreased cell growth ( Figure 2B ), increased apoptosis ( Figure 2C ), and decreased clonogenicity ( Figure 2D ) and anchorage-independent growth ( Figure 2E ). All these effects were more pronounced when MYCN was activated ( Figures 2B–E ). Together, these results show that depletion of PPP2CA decreases the survival of NB cells, augmented by MYCN.

As the role of SET in NB patients is unknown, we investigated the association of SET expression with patient survival and MYCN copy number in a large patient cohort. High mRNA expression of SET was associated with markedly decreased overall survival ( Figure 3A , left panel) and with amplification of MYCN ( Figure 3A , middle panel). Along this line, high protein expression of SET was associated with MYCN amplification in a different patient cohort ( Figure 3A , right panel). Thus, SET is overexpressed in poor-prognosis NB patients.

Given the strong expression of SET in poor-prognosis NB, including MYCN-amplified NB, we asked whether inhibition of SET by OP449 would decrease the viability of NB cells. NB cell lines with non-amplified MYCN (SK-N-AS, SH-SY5Y, NB69), amplified MYCN (KELLY, SK-N-BE(2)-C, LAN5), 1p36 deletion (NB69, SK-N-AS, SK-N-BE(2)-C), activating ALK mutations (SH-SY5Y, KELLY, LAN5), inactivating p53 mutations (SK-N-AS, KELLY, SK-N-BE(2)-C), and ARF deletions (SK-N-AS, KELLY) were treated with OP449 ( Figure 3B ). In these NB cells with diverse molecular alterations, OP449 markedly decreased viability, independent of their underlying molecular alterations ( Figure 3B ).

Next, we investigated the effects of OP449 in NB cells in more detail in the paradigmatic MYCN non-amplified SH-SY5Y and MYCN-amplified KELLY cells. OP449 attenuated cell growth, induced apoptosis, and inhibited clonogenicity in a dose-dependent manner ( Figures 4A–C ). Anchorage-independent growth was inhibited at higher doses ( Figure 4D ). Taken together, OP449 decreases the aggressiveness of SH-SY5Y cells and, more pronounced, of the MYCN-amplified KELLY NB cells.

Phosphorylation of Y307 inhibits PPP2CA (40). Treatment of SH-SY5Y and KELLY with OP449 decreased phosphorylation of PPP2CA at Y307 ( Figure 5A ), suggesting activation of PP2A. Determination of the enzymatic activity of PP2A proved reactivation of PP2A by OP449 in NB cells ( Figure 5B ).

We reasoned that the SET inhibitor OP449 may synergize with rationally chosen kinase inhibitors known to be efficacious in NB, as both act on phospho-signaling pathways important for the maintenance of NB. The efficacy of OP449 and the kinase inhibitors dasatinib, dactolisib, sorafenib, erlotinib, lorlatinib, afatinib, and trametinib was determined in SH-SY5Y and KELLY cells. These cell lines were differentially sensitive to OP449 ( Figure 6A and Supplementary Figure 1 ). The various kinase inhibitors differed in their IC₅₀ within and between the two cell lines ( Figure 6A and Supplementary Figure 1 ). The synergy between OP449 and the kinase inhibitors was evaluated. As determined by the weighted average CI (CI_wt), which emphasizes the more relevant higher inhibitory efficacies, OP449 synergized with dasatinib and dactolisib in SH-SY5Y cells and with dasatinib, sorafenib, erlotinib, and lorlatinib in KELLY cells ( Figure 6A and Supplementary Figure 1 ). Thus, dasatinib, but not the other kinase inhibitors, exhibited relevant synergy with OP449 in both cell lines.

Since OP449 reactivated the phosphatase PP2A, known to inhibit the AKT pathway, and because AKT is important for the growth of NB, we first determined the impact of OP449 on AKT phospho-signaling in SH-SY5Y and KELLY cells. OP449 decreased activating phosphorylation of S473 and T308 in AKT and of p70S6K ( Figure 6B ). Thus, OP449 inhibits activating phospho-signaling of the AKT pathway in NB.

Since OP449 synergized with the kinase inhibitor dasatinib, we investigated their combined effect on the AKT pathway. Dasatinib reduced activating phosphorylations of AKT and p70S6K ( Figure 6B ). Of note, activating phosphorylations of AKT were more reduced by the combination of OP449 with dasatinib compared to single drug use ( Figure 6C ). No synergy was observed in ERK1/2 phosphorylation ( Figure 6C ). In summary, OP449 and dasatinib synergize in inhibiting activating phosphorylations in the AKT pathway of NB cells.

Given that OP449 synergized with dasatinib in SH-SY5Y and KELLY cells, we wanted to know whether OP449 and dasatinib synergize in a broader panel of MYCN non-amplified and MYCN-amplified NB cell lines. OP449 and dasatinib reduced cell viability at 72 h with IC₅₀ ranging at 1.2–7.1 and 0.03–25.7 µM, respectively ( Figure 7 ). The CI_wt values indicate that OP449 and dasatinib synergistically reduced the viability of SH-EP, SK-N-SH, NB69, and GI-M-EN but not of SK-N-AS MYCN non-amplified NB cells ( Figure 7A ) and synergized in NLF and SK-N-BE(2)-C but not in IMR5, LAN5, and IMR32 MYCN-amplified NB cells ( Figure 7B ). Together, these data show that OP449 synergizes with dasatinib in many MYCN-amplified and non-amplified NB cell lines.

Given that OP449 inhibited activating phospho-signaling of AKT and that OP449 and dasatinib synergized in this process, we asked whether cellular inhibition by OP449 could be rescued by activating AKT. SC79, an activator of AKT, was used, and activation was determined by phosphorylation of AKT and the downstream p70S6 kinase. For SH-SY5Y and KELLY cells, 20 and 10 µM of SC79 were found to be suitable doses, respectively ( Figure 8A ). AKT activation occurred at 0.5 h and subsided at 2 h ( Figure 8A ). Therefore, these doses and a duration of 2 h were chosen for subsequent rescue experiments. As sensitivity for OP449 was higher in KELLY than in SY5Y cells, lower doses of OP449 were used in KELLY cells. Of note, at higher OP449 concentrations, activation of AKT led to significant, albeit partial rescue, of OP449-mediated viability inhibition, more pronounced in KELLY cells ( Figure 8B ). Thus, inhibition of NB cells by OP449 is in part mediated by AKT inhibition.