The MYC family of transcription factors is a major driver of human cancer and potential therapeutic target. However, no clinically viable drugs have been yet developed that are able to directly tackle MYC oncoproteins. In our laboratory, we are exploring alternative approaches aiming to disturb signalling downstream of MYC. MYCN is frequently activated in neuroblastoma, a paediatric solid malignancy that, in its metastatic form, has a very poor prognosis. An important pathway regulated by MYC is the CKS1/SKP2/p27kip1 axis. In this study, we have repurposed the anti-psychotic drug Prozac to disrupt CKS1/SKP2/p27Kip1 signalling and assess its potential as an anti-neuroblastoma agent in vitro and in vivo. Using DNA editing technology, we show that stabilisation of p27Kip1 operated by Prozac in MYC-activated cells is essential for the anti-neuroblastoma activity of the drug. Furthermore, dosing mice with a concentration of Prozac equivalent to that used in long-term clinical trials in children with psychiatric disorders caused a significant reduction of metastatic disease in two models of high-risk neuroblastoma. The favourable toxicity profile of Prozac suggests that long-term treatments might be implemented in children with MYC/CKS1high neuroblastomas.
Neuroblastoma is a rare cancer of the sympathetic nervous system that occurs during early childhood and infancy. Although it only accounts for 7% of cancer diagnosis of children under 15 year, it is responsible for most paediatric mortalities owing to solid tumours and is the commonest cancer diagnosed within the first year of life1. Neuroblastomas arise from sympathoadrenal lineage neural crest cells and primary tumours form in the sympathetic nervous system, usually within the adrenal medulla or paraspinal ganglia2. A major molecular alteration in neuroblastoma is the amplification of the locus encoding the oncogenic transcription factor MYCN. MYCN belongs to a small gene family that includes c-MYC and l-MYC, and encodes for a 60–64 kDa, nuclear, phosphoprotein, which is one of the BHLH transcription factors responsible for assisting in the differentiation of neuronal progenitor cells transitioning from the neural crest to the sympathetic nervous system3. In the nervous system, MYCN has been shown to play a significant role in the development of tissues and the differentiation pathways of neuronal progenitor cells3. In neuroblastoma, MYCN amplification is used as a prognostic marker and is indicative of a high-risk disease1,2. One of the signalling pathways downstream of MYC proteins is the CKS1/SKP2/p27Kip1 axis. When p27Kip1 (encoded by CDKN1B) is expressed, the complexes formed by the cyclin directed kinase CDK2 cannot be established, and the cell cannot enter the cell cycle. CKS1 (encoded by CKS1B) facilitates S phase entry through the degradation of the CDK2 inhibitor, p27Kip14. CKS1 acts as a co-factor in the SKP2-cullin-F-box complex for p27Kip1 ubiquitination. Although SKP2 binding and subsequent ubiquitination of p27Kip1 can occur without CKS1, it is drastically reduced without the presence of CKS14,5. Without CKS1 to bridge the leucine-rich-repeats, SKP2 cannot obtain optimal substrate positioning6,7. Although CKS2 shares with CKS1 a similar CDK interaction domain and binding sites for p27Kip1, it cannot interact with SKP2 and thereby cannot help reduce p27Kip1 levels but, instead, protects it from degradation by SKP26,8. CKS1 recognises phosphorylated p27Kip1 through an anion pocket and will bind to promote SKP2 binding, as CKS1 contains specific residues in the N-terminus that allows association with SKP2. CKS2 also contains the anion pocket to recognise and bind to phosphorylated p27Kip1 but lacks the association region for SKP2. Therefore, CKS2 protects p27Kip1 from degradation, whereas CKS1 promotes degradation, and p27Kip1 levels are determined, in part, by the CKS which has the highest concentration8. Neuroblastoma tumours have been shown to express high levels of SKP29 and c-MYC upregulates SKP2 expression10. Further studies have additionally shown that in neuroblastoma cells MYCN targets a separate E-box location and, similarly to c-MYC, augments transcription of SKP211. Furthermore, c-MYC and MYCN induce CKS1B transcription10,12. Targeting SKP2 causes p53-independent apoptosis in non-amplified neuroblastoma cells, whereas in MYCN-amplified cells it was noted a decrease in growth but not apoptosis11. Likewise, when CKS1B was inhibited in tumour cells, apoptosis and growth arrest followed stabilisation of p27Kip113,14. A genome-wide, drop-out shRNA screen carried out in our laboratory has identified CKS1B as a potential therapeutic target gene in MYCN-amplified neuroblastoma by inducing synthetic lethality12. Although pharmacological inhibitors of SKP2 are not currently available15, CKS1 can be inhibited by a small molecule that is safe and available. Fluoxetine, also known as Prozac, is a serotonin uptake inhibitor originally developed to treat depression. However, Prozac also has been shown to induce G1 arrest through inhibition of the CKS1–SKP2 binding interaction site, resulting in elevated p27Kip1 levels and differentiation of neuronal stem cells13,16.
In this study, we investigated whether Prozac could be used to induce stabilisation of p27Kip1 and growth arrest/apoptosis of MYC-expressing neuroblastoma cells in vitro and in vivo.
Results and discussion
The CKS1 inhibitor Prozac increases p27Kip1 expression in neuroblastoma cell lines
We monitored CKS1 protein levels in a panel of neuroblastoma cell lines with or without activated MYC. As expected, CKS1 levels were higher in MYCN amplified (Kelly, Lan5, LU-NB-1, LU-NB-2) than non-MYCN amplified (hNB, SHEP) neuroblastoma cell lines or normal human fibroblasts (BJ, HDF) (Fig. 1a). It must be noted that non-MYCN-amplified SK-NA-S cells have a mutation that results in activation of c-MYC, which explains the elevated CKS1 levels17.
Inhibition of CKS1B is synthetically lethal with MYCN amplification/overexpression in neuroblastoma cells, suggesting that it may be used to target specifically MYChigh tumours12. As RNA interference is not yet a viable option in cancer therapy, we used Prozac to disrupt the CKS1–SKP2 interaction, with the aim of causing stabilisation of the product of the tumour suppressor gene CDKN1B -P27Kip1-, in neuroblastoma cell lines. Thus, we exposed MYC-expressing, or non-expressing, neuroblastoma cell lines to Prozac and quantified p27Kip1 levels by western blotting. The MYC-positive cell lines showed a dose-dependent increase of p27Kip1 levels after exposure to the drug. Interestingly, p27Kip1 levels did not change in the non-MYCN-amplified cell line hNB (Fig. 1b). This result is consistent with the low levels of CKS1 expression in hNB (Fig. 1a). Prozac is likely to act at the protein level; in the presence of the drug, p27Kip1 was stabilised after block of protein synthesis operated by cycloheximide (Supplementary Fig. 1).
Prozac inhibits short and long-term growth of neuroblastoma cell lines in vitro
To investigate the therapeutic potential of Prozac in neuroblastoma, we exposed multiple cell lines to increasing concentrations of Prozac for 72 h and assess metabolic activity by MTT assays. As controls, we used normal and immortalised, but non-tumourigenic, human fibroblasts. All neuroblastoma cell lines were inhibited by Prozac, with the majority of the MYC-positive cell lines being slightly more sensitive to the drug. As expected, normal fibroblasts were completely resistant up to the maximum concentration of Prozac used (Fig. 2a). To determine the long-term effects of Prozac, we implemented colony assays in which the neuroblastoma, or normal, cells were exposed to a relatively low concentration of Prozac (10 μM) for 2–3 weeks, after which colonies were scored. There was a clear-cut reduction in the number of colonies formed by the CKS1/MYChigh cell lines, whereas the CKS1/MYClow cells (SHEP, hNB, and human fibroblasts) formed colonies normally in the presence of Prozac, suggesting that CKS1 and MYC expression could be used as a biomarker to determine susceptibility to Prozac inhibition (Fig. 2b). To further confirm the role of MYC in determining sensitivity of neuroblastoma cells to Prozac, we used a switchable TET-OFF system in which transgenic expression of MYCN can be turned off in the presence of doxycyclin. Proliferation of MYCN expressing SHEP (TET21N) is markedly inhibited by a low concentration of Prozac (10 μM); however, when MYCN expression was ablated they became resistant, confirming the hypothesis that Prozac triggers synthetic lethality in MYC-expressing cells (Fig. 2c).
To gain mechanistic insights into the nature of the inhibition, we assessed cell cycle phases and quantified DNA fragmentation, diagnostic of apoptosis. Prozac did not cause significant changes in cell cycling activity, but it induced a significant increase of DNA fragmentation in the CKS1/MYChigh cell lines Kelly and SKNAS compared to CKS1/MYClow hNB and BJ (Supplementary figure 2). The hypothesis that Prozac might cause apoptosis of neuroblastoma cells is corroborated by a recent study showing that the MYCN-amplified SK-N-BE2 cell line undergoes apoptosis in the presence of the drug18.
Expression of CDKN1B is essential for the antiproliferative effect of Prozac
To confirm that stabilisation of p27Kip1 levels operated by inhibition of CKS1 is required for the growth inhibiting effect of Prozac, we deleted the CDKN1B gene using crispr/Cas9 technology. After clonal expansion of crispr/cas9-treated Kelly cells, we selected a clone (clone 10) with disrupted CDKN1B alleles (Supplementary figure 3). As a further control, we expanded another clone (clone 16) that had undergone the crispr/cas9 procedure but retained expression of p27Kip1. Next, we implemented colony formation assays in which we exposed the crispr/cas9-treated clones, or the parental Kelly cell line, to Prozac. Notably, the p27Kip1-null clone 10 regained the ability to form colonies in the presence of Prozac, whereas the parental cell line Kelly or clone 16, which had retained expression of p27Kip1, were sensitive to the drug (Fig. 3a). Western blot analysis confirmed that expression of p27Kip1 in clone 10 was reduced almost to zero, whereas clone 16 and parental cells presented similar p27Kip1 levels (Fig. 3b). To confirm that Prozac-induced p27Kip1 is key for the inhibitory effect of the drug, we ectopically overexpressed p27Kip1 in SKNAS cells, which express low basal levels of p27Kip1 (Fig. 3c, left panel). Overexpression of p27Kip1 caused a marked inhibition of colony formation (Fig. 3c, right panel). Overall, these experiments clearly establish that Prozac inhibits neuroblastoma proliferation via stabilisation of the tumour suppressor p27Kip1.
Prozac inhibits metastatic growth of high-risk neuroblastoma cell lines in vivo
Tumour relapse is common in children with high-risk neuroblastoma and finding a drug with a favourable toxicity profile to prolong remission would be a major clinical advance. Prozac could be particularly useful in the context of minimal residual disease, where chemotherapy-resistant neuroblastoma cells in the bone marrow and other organs eventually cause cancer to relapse. To emulate this situation, we have implemented a pseudo-metastatic model in which Kelly (MYCN amplified) and SKNAS (c-MYC mutant) neuroblastoma cells are injected in the tail vein of immunocompromised mice. In these models, tumour cells colonise preferably the liver and kidneys. Kelly cells have also a tropism for the bone marrow, whereas SKNAS cells are unable to infiltrate the bone marrow but can invade lungs. To allow organ seeding and establishment of micro-metastases, treatments with a pharmacological concentration of Prozac was started 1 week after tumour cell injections and continued for 3 weeks (a scheme of the experiment is shown in Fig. 4a). After mice killing, gross metastases in the liver, kidneys, and lungs were quantified by microscopy, and neuroblastoma cells in the bone marrow were enumerated by flow cytometry with a GD2 antibody. In the Kelly model, Prozac treatments caused a sharp, statistically significant reduction of the number of metastases and neuroblastoma cell proliferation, as indicated by reduced to positivity to the ki67 marker, suggesting that the drug is antimetastatic (Fig. 4b, c). The anticancer activity of Prozac was largely confirmed in the SKNAS model, although the decrease in metastases formation in the liver did not reach statistical significance (Fig. 4d). Prozac treatments were very well tolerated by mice, as indicated by stable body weights (Fig. 4e).
In spite of important progresses in the understanding of the molecular alterations underpinning the aggressive behaviour of neuroblastoma, this cancer still poses a formidable challenge to clinicians. Based on the recent discovery of mutations at the level of genes that activate neuroblastoma oncogenic pathways, such as MYCN, ALK, ATRX, TERT1,2, some investigational, molecularly targeted drugs are being experimented in clinical trials. However, as of today, children are still treated with highly toxic mixtures of chemotherapeutic drugs, radiation and potentially unsafe compounds such as retinoids. Although immunotherapy with the GD2 antibody has significantly improved the outcome of high-risk patients, more than half of the treated children do not show a clinically relevant response19.
We have recently implemented a genome-wide shRNA screen to detect MYCN-dependent vulnerabilities in neuroblastoma. In this study, we identified, among other genes, CKS1B as a synthetic lethal partner of MYCN. MYC and CKS1 are involved in a crosstalk with p27Kip1; MYC promotes degradation of p27Kip1 via activation of CKS1 and the ubiquitin ligase skp2. The latter modulates MYC protein stability, also acting as a transcriptional co-factor and enhancer of MYC-dependent activation of cell cycle-related genes20. In the present study, we show that Prozac induces stabilisation of p27Kip1 in MYChigh neuroblastoma cells, resulting in reduced proliferation and increased apoptosis in vitro. In vivo, Prozac displays a potent antimetastatic activity, suggesting that it may be repurposed to treat cancer patients.
It is important to note that the range of Prozac doses used in clinical trials in children with psychoses is 0.8–2 mg/kg21,22, equivalent to 10–24.6 mg/kg in the mouse system, after dose conversion23. The drug dosage used in our study, 20 mg/kg, is well within this range. Notably, duration of treatments in one of the human trials were up to 32 months, with manageable neurologic side effects, indicating that long-term treatment of children with anticancer concentrations of Prozac is feasible21. As high expression of MYCN and CKS1 mark cells that are susceptible to Prozac in long-term assays in vitro and in vivo, MYCN amplification and elevated CKS1 protein levels in neuroblastoma biopsies could be used as biomarkers to identify patients that might respond to the drug. We conclude that the results presented in this study warrant the opening of clinical trials in which long-term Prozac treatments could be included in consolidation or post-consolidation therapies in patients who are at high risk of disease relapse.
Maris, J. M., Hogarty, M. D., Bagatell, R. & Cohn, S. L. Neuroblastoma. Lancet 369, 2106–2120 (2007).
Brodeur, G. M. Neuroblastoma: biological insights into a clinical enigma. Nat. Rev. Cancer 3, 203–216 (2003).
Knoepfler, P. S., Cheng, P. F. & Eisenman, R. N. N-myc is essential during neurogenesis for the rapid expansion of progenitor cell populations and the inhibition of neuronal differentiation. Genes Dev. 16, 2699–2712 (2002).
Spruck, C. et al. A CDK-independent function of mammalian Cks1: targeting of SCF(Skp2) to the CDK inhibitor p27Kip1. Mol. Cell 7, 639–650 (2001).
Ganoth, D. et al. The cell-cycle regulatory protein Cks1 is required for SCF(Skp2)-mediated ubiquitinylation of p27. Nat. Cell Biol. 3, 321–324 (2001).
Hao, B. et al. Structural basis of the Cks1-dependent recognition of p27(Kip1) by the SCF(Skp2) ubiquitin ligase. Mol. Cell 20, 9–19 (2005).
Yao, Z. P. et al. Activation of ubiquitin ligase SCF(Skp2) by Cks1: insights from hydrogen exchange mass spectrometry. J. Mol. Biol. 363, 673–686 (2006).
Frontini, M. et al. The CDK subunit CKS2 counteracts CKS1 to control cyclin A/CDK2 activity in maintaining replicative fidelity and neurodevelopment. Dev. Cell 23, 356–370 (2012).
Westermann, F. et al. High Skp2 expression characterizes high-risk neuroblastomas independent of MYCN status. Clin. Cancer Res. 13, 4695–4703 (2007).
Keller, U. B. et al. Myc targets Cks1 to provoke the suppression of p27Kip1, proliferation and lymphomagenesis. EMBO J. 26, 2562–2574 (2007).
Evans, L. et al. SKP2 is a direct transcriptional target of MYCN and a potential therapeutic target in neuroblastoma. Cancer Lett. 363, 37–45 (2015).
Chayka, O., D'Acunto, C. W., Middleton, O., Arab, M. & Sala, A. Identification and pharmacological inactivation of the MYCN gene network as a therapeutic strategy for neuroblastic tumor cells. J. Biol. Chem. 290, 2198–2212 (2015).
Krishnan, A., Hariharan, R., Nair, S. A. & Pillai, M. R. Fluoxetine mediates G0/G1 arrest by inducing functional inhibition of cyclin dependent kinase subunit (CKS)1. Biochem. Pharm. 75, 1924–1934 (2008).
La, T. et al. A p53-responsive miRNA network promotes cancer cell quiescence. Cancer Res. 78, 6666–6679 (2018).
Skaar, J. R., Pagan, J. K. & Pagano, M. SCF ubiquitin ligase-targeted therapies. Nat. Rev. Drug Discov. 13, 889–903 (2014).
Lupu, D. et al. Fluoxetine affects differentiation of midbrain dopaminergic neurons in vitro. Mol. Pharm. 94, 1220–1231 (2018).
Zimmerman, M. W. et al. MYC drives a subset of high-risk pediatric neuroblastomas and is activated through mechanisms including enhancer Hijacking and focal enhancer amplification. Cancer Discov. 8, 320–335 (2018).
Choi, J. H. et al. Fluoxetine induces apoptosis through endoplasmic reticulum stress via mitogen-activated protein kinase activation and histone hyperacetylation in SK-N-BE(2)-M17 human neuroblastoma cells. Apoptosis 22, 1079–1097 (2017).
Hoy, S. M. Dinutuximab: a review in high-risk neuroblastoma. Target Oncol. 11, 247–253 (2016).
Hydbring P., Castell A. & Larsson L. G. MYC modulation around the CDK2/p27/SKP2 Axis. Genes (Basel). 8, pii: E174 (2017).
Fatemi, S. H., Realmuto, G. M., Khan, L. & Thuras, P. Fluoxetine in treatment of adolescent patients with autism: a longitudinal open trial. J. Autism Dev. Disord. 28, 303–307 (1998).
Hollander, E. et al. A placebo controlled crossover trial of liquid fluoxetine on repetitive behaviors in childhood and adolescent autism. Neuropsychopharmacology 30, 582–589 (2005).
Nair, A. B. & Jacob, S. A simple practice guide for dose conversion between animals and human. J. Basic Clin. Pharm. 7, 27–31 (2016).
Braekeveldt, N. et al. Neuroblastoma patient-derived orthotopic xenografts retain metastatic patterns and geno- and phenotypes of patient tumours. Int. J. cancer 136, E252–E261 (2015).
Persson, C. U. et al. Neuroblastoma patient-derived xenograft cells cultured in stem-cell promoting medium retain tumorigenic and metastatic capacities but differentiate in serum. Sci. Rep. 7, 10274 (2017).
Chaiwatanasirikul, K. A. & Sala, A. The tumour-suppressive function of CLU is explained by its localisation and interaction with HSP60. Cell Death Dis. 2, e219 (2011).
Xiao, A. et al. CasOT: a genome-wide Cas9/gRNA off-target searching tool. Bioinformatics 30, 1180–1182 (2014).
This project was funded with grants from the Associazione Italiana per la Lotta al Neuroblastoma and Neuroblastoma-UK to A.S. G.S. and V.D.L are supported by AIRC (IG:18467; IG 15196). E.C. is recipient of an AIRC fellowship (IG:18467). Annalisa Nespoli is kindly acknowledged for helping with the immunohistochemistry assays. We thank Dr. Angelo Peschiaroli (CNR, Italy) for the kind gift of the pcDNA3-p27 vector.
Conflict of interest
The authors declare that they have no conflict of interest.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Bibbo’, S., Lamolinara, A., Capone, E. et al. Repurposing a psychoactive drug for children with cancer: p27Kip1-dependent inhibition of metastatic neuroblastomas by Prozac. Oncogenesis 9, 3 (2020). https://doi.org/10.1038/s41389-019-0186-3