Cell death-based treatment of neuroblastoma

Neuroblastoma (NB) is the most common solid childhood tumor outside the brain and causes 15% of childhood cancer-related mortality. The main drivers of NB formation are neural crest cell-derived sympathoadrenal cells that undergo abnormal genetic arrangements. Moreover, NB is a complex disease that has high heterogeneity and is therefore difficult to target for successful therapy. Thus, a better understanding of NB development helps to improve treatment and increase the survival rate. One of the major causes of sporadic NB is known to be MYCN amplification and mutations in ALK (anaplastic lymphoma kinase) are responsible for familial NB. Many other genetic abnormalities can be found; however, they are not considered as driver mutations, rather they support tumor aggressiveness. Tumor cell elimination via cell death is widely accepted as a successful technique. Therefore, in this review, we provide a thorough overview of how different modes of cell death and treatment strategies, such as immunotherapy or spontaneous regression, are or can be applied for NB elimination. In addition, several currently used and innovative approaches and their suitability for clinical testing and usage will be discussed. Moreover, significant attention will be given to combined therapies that show more effective results with fewer side effects than drugs targeting only one specific protein or pathway.


Introduction
Neuroblastoma (NB) is the most common solid childhood tumor outside the brain. It originates from primitive cells of the sympathetic nervous system 1 . NB causes 15% of childhood cancer-related mortality and overall survival rate for metastatic tumors is considerably low, 40% after 5 years 2,3 . Most incidences are diagnosed during the first year of life, which also gives a better prospect for the outcome, whereas older patients have a poorer diagnosis 4,5 . In some NB cases, spontaneous regression has also been detected; however, underlying mechanisms remain unclear 6,7 . Moreover, NB is a complex disease that has high genetic, biological, clinical, and morphological heterogeneity, and is therefore difficult to target for successful therapy [8][9][10] . Thus, NB is under thorough investigation to better understand its progression and to improve the treatment to increase the survival rate.
Several classification systems have been used in order to improve risk assessment and prognosis of NB. For example, the outcome of the disease can be assessed by the presence or absence of stroma, the degree of differentiation, and the mitosis-karyorrhexis index 11 . Currently, even more parameters are used for the classification of NBs, such as stage, age, histologic category, grade of tumor differentiation, the status of the MYCN oncogene, chromosome 11q status, and DNA ploidy. These are the most statistically significant and clinically relevant factors in use to describe two stages of localized (L1 and L2) and two stages of metastatic disease (M and MS) 12 .
The main drivers of NB formation are abnormalities in sympathoadrenal cells that derive from neural crest cells (Figure 1) 13 . Several germline and sporadic genomic rearrangements have been detected in NB, for example, LIN28B (encoding lin 28 homolog B) 14 , PHOX2B (pairedlike homeobox 2b) 15 , ALK (anaplastic lymphoma kinase) 16 , GALNT14 (polypeptide N-acetylgalactosaminyltransferase 14) 17 , and MYCN 18 (Table 1). Around 2% of NB cases appear to be hereditary, with ALK being the first gene identified to be responsible for familial NB 16,19 . Furthermore, MYCN oncogene amplification is found in 20% of all NB cases, especially in patients who are resistant to therapy and have poor prognosis 18,20,21 . More than 50% of these high-risk patients relapse even after intensive treatment 22 . Whole-genome sequencing has been used to identify additional mutations and genes responsible for de novo NB development, but no other specific "NB driver mutations" have been found 23,24 . Thus, MYCN amplification seems to be the major cause of sporadic NB and other mutations support tumor aggressiveness 25 . Therefore, investigation of the MYCN gene amplification is considered to be a mandatory step for treatment specification 26 .
In this review, we provide a thorough overview of how different modes of cell death are exploited or can be employed as treatment for NB. In addition, several novel or already clinically tested drugs against NB and their mechanisms of action are discussed. A special emphasis is also placed on combined therapies that attack many pathways and have been shown to be more effective than drugs targeting only one specific protein or pathway.

Genetic background
Anaplastic lymphoma kinase Changes in the ALK gene are identified as being responsible for~50% of familial and~1% of all NBs 16 (Table 1). ALK is a member of the insulin receptor superfamily of transmembrane RTKs (receptor tyrosine kinase). Mutations and amplifications of the ALK gene can lead to a constitutive activation of ALK that supports cell survival and proliferation in the peripheral neuronal   29 . For instance, crizotinib 30 and entrectinib 31 reduce the cells' proliferation rate and are currently in Phase 1/2 trials (NCT00939770, NCT01606878, and NCT02650401) for relapsed or refractory NB; however, there are problems with their off-target effects and acquired resistance. Therefore, new-generation ALK inhibitors are already been developed and tested for NB therapy, for example, lorlatinib (NCT03107988) 32 , AZD3463 (ref. 33), and ceritinib (NCT01742286) 34 . In addition to reducing the proliferation rate, clinical tests have shown that most ALK inhibitors also sensitize NB cells to conventional cytotoxic drugs and their combined use is causing more prominent cell death 35,36 . On the other hand, this approach is helpful for only ALK-positive tumors and, due to the high heterogeneity of NB, more strategies are needed for successful treatment of NBs carrying other mutations.

MYCN
MYCN is part of the MYC family of transcription factors that regulate several cellular processes including proliferation, cell cycle, glycolysis, glutaminolysis, mitochondrial function, and biogenesis [37][38][39] . MYCN expression is essential for normal prenatal development and is present until a few weeks after birth 40 . Amplifications of the MYCN gene are known to be responsible for increased tumor growth, proliferation, and NB development (Table 1) 41,42 . Deregulation of MYC induces cell proliferation and apoptosis; however, this apoptotic signal is inhibited by reducing p53 activity, overexpressing antiapoptotic proteins, or downregulating pro-apoptotic proteins 43,44 . Thus, a combined suppression of MYCinduced apoptosis and MYC-driven proliferative signals supports extensive tumor development.
MYCN usually has a very short half-life, but after amplification it is highly expressed and forms heterodimers with MAX to act as a transcriptional factor and support constant NB tumor growth 45 . Therefore, downregulation of MYCN is one possible approach to induce apoptosis, decrease NB proliferation, and/or induce neuronal differentiation 46 . For example, antisense oligonucleotides 47 and RNAi [48][49][50] have been successfully used for MYCN downregulation in NB that resulted in decreased tumor growth, cellular migration, and invasion. The described approach has proved to be effective in the laboratory; however, off-target effects and clinical delivery of these compounds to the tumor site are still problematic.
Blocking the MYCN/MAX interaction is another option for NB therapy, because unbound MAX homodimerizes and stimulates differentiation 51

Other genomic abnormalities
Overexpression and amplifications of LIN28B are very common in NB cells and can in turn lead to high MYCN expression (Table 1) 14,57 . Moreover, whole-genome sequencing revealed that 25% of the patients have rearrangements in TERT (encoding telomerase reverse transcriptase) 58,59 promoter and 10% in transcriptional regulator ATRX (encoding the RNA helicase) 23 , supporting rapid cellular proliferation (Table 1). Chromosomal copy number alterations are also represented in almost all NBs, for example, more than 50% have gain of 17q (ref. 60) and 30% have loss of 1p36 and/or 11q1 (ref. 61) ( Table 1). These arrangements have a strong correlation with MYCN amplification and poor prognosis. However, the function of these regions and how they regulate NB formation is still unclear 60,61 . Targeting NB via stimulation of various modes of cell death

Apoptosis induction in NB therapy
Apoptosis is essential for the normal growth of an organism, being involved in early embryonic and immune system development. It also has an important role in the maintenance of normal tissue homeostasis and helps to eliminate damaged and harmful cells 62 . Therefore, misregulation of apoptotic pathways has an important role in cancer development, because mutations or amplifications in the oncogenes (e.g., MYC) can compromise apoptotic pathways. On the other hand, apoptosis induction is the most prominent anticancer strategy.

Targeting p53/MDM2 interaction
The members of the p53 protein family are important regulators of cell cycle and apoptosis in normal and transformed cells 63 . In addition, p53 as well as p73 act as tumor suppressors. Mutations in the p53 gene that control cell fate occur in more than 80% of tumor cell lines and more than 40% of human cancers 64 . However, abnormalities of p53 are mostly found in relapsed NB after chemotherapy, but not at the time of the diagnosis 65,66 . Instead, overexpressed MYCN regulates p53 and MDM2 (murine double minute 2) expression to achieve stringent control over cell death ( Figure 2) 67,68 . Tumors such as NB, which generally have wild-type p53, are likely to induce the degradation of p53 and avoid cell death by overexpression or amplification of MDM2, which is a negative regulator and the primary E3 ubiquitin ligase for p53 (ref. 65,67,69). For instance, MYCN binds to the promoter of MDM2 to induce its expression and vice versa, suggesting that downregulation of MDM2 can also be used to decrease MYCN expression and stabilize p53 to induce apoptosis ( Figure 2) 67,70,71 .
Understanding these peculiarities of NB and targeting the p53-MDM2 pathway may be helpful in finding better therapeutic treatments for pediatric patients with wildtype p53 (ref. 72,73). For example, small antagonistic molecules, like nutlin-3 (ref. [74][75][76], MI-773/219/63 (ref. 75), and idasanutlin (RG7388) 77 , which bind to MDM2 to block its interaction with p53, have shown promising results in NB. These inhibitors attenuate the proliferation of MYCN-expressing NB cells and some of them are being tested in clinical trials; however, the development of resistance, toxicity, MDM2 accumulation, and the need for wild-type p53 make the trials challenging 78 . In addition to the regulation of p53-MDM2, MYCN facilitates an increase in the expression of FAK (focal adhesion kinase), which interacts with p53 and causes its sequestering in the cytoplasm ( Figure 2). Interrupting this binding by small molecules or peptides enables p53 to move to the nucleus to induce apoptotic cell death of in vivo breast and colon tumors 79 .
Furthermore, MYCN-upregulated MDM2 can similarly bind with another member of the p53 family, tumor suppressor TAp73 (p73 locus encodes two isoformstumor suppressor (TAp73) and putative oncogene (ΔNp73)) ( Figure 2). MDM2 decreases TAp73 transcription and supports resistance to the treatment 80,81 . It has been discussed that besides regulating p53 and MDM2 levels, MYCN might also directly decrease TAp73 expression and support NB tumor growth 82 . In addition, there are results showing that overexpression of TAp73 can in turn reduce MYCN expression and induce differentiation of NB cell lines, indicating that the balance between TAp73 and MYCN levels can influence the outcome of the NB development and treatment ( Fig. 2) 83,84 . These new approaches have led to novel combinatorial therapeutic strategies that simultaneously reduce toxicity and enhance the outcome of the treatment and are being tested in preclinical and clinical trials for NB 75  that MYCN has many cellular targets, disrupting its interaction with one of them is probably not enough for successful treatment.

BCL-2 family
Other important apoptosis regulators are B-cell lymphoma/leukemia 2 (BCL-2) family proteins, which are divided into two groups: pro-apoptotic and anti-apoptotic proteins. The main anti-apoptotic proteins are BCL-2, BCL-xL, and myeloid cell leukemia (MCL)-1, which prevent outer mitochondrial membrane (OMM) permeabilization by binding and inhibiting pro-apoptotic proteins. Apoptosis-promoting proteins from this family can in turn be divided into two groups: BH-3 only and effector proteins. The pro-apoptotic BH-3 only proteins (Bid and Bim) respond to apoptotic stimuli and inhibit antiapoptotic BCL-2 proteins or activate the effector proteins (BAK and BCL-2-associated X protein), which form pores in the OMM to induce cytochrome c release and apoptosis. The balance between pro-and anti-apoptotic proteins determines the fate of the cells through regulation of the mitochondrial apoptotic pathway 88,89 . As with p53, mutations in BCL-2 are scarce in NB, although dysregulation and increased levels of the BCL-2 gene are frequent [90][91][92] . Moreover, in B-cell lymphomas a link between MYC and BCL-2 expression has been described, because overexpression of MYC in tumor cells is often found together with rearrangements in the BCL-2 family to support tumor growth and suppress apoptosis 93,94 . Therefore, therapies that change the balance between proand anti-apoptotic proteins are promising strategies for tumor treatment.
One possible approach might be using conventional chemotherapeutics together with inhibitors of antiapoptotic BCL-2 proteins (e.g.,   95 , although there have been problems with modest outcome, side effects, 96 and resistance in relapsed NBs 97 . This is due to the compensatory upregulation of the anti-apoptotic MCL-1 protein that rescues cells from apoptosis. However, when the MCL-1 inhibitor (e.g., A-1210477) is used in combination with ABT-199, successful induction of NB cell death has been demonstrated 98 .

Targeting cellular bioenergetics pathways
Considering the key role of mitochondria in various modes of cell death, they might be potential targets for tumor therapy. For instance, many anticancer drugs destabilize mitochondria to induce apoptotic cell death 99 . Rapidly proliferating tumors easily become hypoxic, which is the reason why the majority of tumors change their source of energy from mitochondrial oxidative phosphorylation (OXPHOS) to glycolysis. These cells usually have lowered amount of mitochondria and/or mutations in one or more OXPHOS complexes [100][101][102] . In contrast, relapsing cancer cells tend to have increased levels of OXPHOS [103][104][105] . The role of MYC overexpression in these processes is to increase the expression of mitochondrial complexes and hence mitochondrial respiration 38 . These metabolic changes help cells to survive in nutrient-deprived environments 106 . Therefore, to eliminate resistant tumor cells, chemotherapeutic drugs could be used in combination with electron transport chain inhibitors, such as the complex I inhibitors metformin 107 or tamoxifen 108 , to induce leakage of electrons and excessive formation of reactive oxygen species (ROS). In addition, using non-toxic doses of the complex II blockers of the respiratory chain, such as thenoyltrifluoroacetone 109 or α-tocopheryl succinate 110 together with harmless doses of cytotoxic drugs, synergistically stimulates the formation of ROS and thereby increases the effectiveness of the therapy on breast cancer and NB cell lines.
Fast growth of the tumor cells and poor vascularization leads to hypoxia, which causes the activation of transcription factors, such as hypoxia-inducible factor 1 (HIF-1), that regulate the hypoxic adaptation ( Figure 3 help cells to survive in low O 2 conditions. HIF-1 is activated in a hypoxic environment that is very common in solid tumors. HIF-1 expression leads to the activation of glycolysis and angiogenesis, and correlates with aggressive tumors and poor outcome. HIF-1 is a heterodimer consisting of the O 2 -regulated HIF-1α subunit and a constantly expressed HIF-1β subunit 114,115 . HIF-1α becomes stabile in a low O 2 environment and binds with HIF-1β to form an active HIF-1 complex that has both anti-and proapoptotic effects 116,117 . For instance, severe and continuous hypoxia will result in HIF-1 activation, p53 expression, and apoptosis. On the other hand, simultaneous stabilization of HIF-1 with activation of the PI3K/ Akt pathway, survivin, glycolytic enzymes, p21, and/or erythropoietin can inhibit apoptosis and support NB tumor growth 118,119 . Furthermore, recent data suggest that the aforementioned TAp73 also regulates the degradation of HIF-1 and the suppression of vascularization in an oxygenindependent manner (Fig. 3) 120,121 . Therefore, loss of TAp73 activity in MYCN-overexpressed tumors can be associated with increased HIF-1 activity and thereby the stimulation of angiogenesis in tumor cells 120,122 . Another isoform of p73, NH 2 terminally truncated putative oncogene ΔNp73, is also involved in angiogenesis regulation (Fig. 3). In tumor cell lines, ΔNp73 is stabilized in O 2deficient conditions and activates vascularization via vascular endothelial growth factor A expression 121 , indicating that cellular response to hypoxic conditions and HIF-1 activity is tightly regulated by MYCN and p53 family proteins. Moreover, HIF-1 activity is also associated with low responsiveness to differentiation therapy and the downregulation of HIF-1 can improve the outcome of the NB treatment 123 . Therefore, taking into account the importance of HIF-1 in NB tumor progression, the search for its inhibitors, such as topotecan 124 and acriflavine 125 , is a promising strategy. Several of these have already been shown to improve the effects of antiangiogenic drugs in vivo.
Cancer cells modify their metabolism to support their constant proliferation. Adjustments in cancer cells' metabolism result in excessive glycolytic activity to produce ATP, the Warburg effect, to support rapid cell proliferation. These changes are also seen in aerobic conditions, even though glycolysis generates less ATP than OXPHOS 126,127 . This decrease in oxygen demand helps tumor cells to survive in hypoxic conditions and continue proliferation due to excessive glycolytic activity 128 . Such a drastic metabolic change is attained by the activity of various oncogenes and regulatory proteins, such as MYC and HIF-1 (ref. 129,130).
Oncogenic MYC upregulates glucose import (e.g., GLUT1), glycolytic enzymes (e.g., hexokinase 2 (HK2) and PDK1), and mitochondrial biogenesis, thereby ensuring metabolic intermediates that support cell growth 131,132 . Elevated glucose transport into the cells and glycolysis itself can be targeted for cancer cell-specific therapy 133,134 . For example, glucose analog 2-DG (2-deoxy-D-glucose) that is phosphorylated by HK2 cannot be metabolized further and accumulates in the cell, leading to the inhibition of glycolysis and tumor growth [135][136][137][138] . This approach has been successful in several NB cell lines 139 and also in xenograft models, 140 regardless of their MYCN status, indicating its potential for clinical significance. Furthermore, the clinical efficacy of 2-DG is enhanced when combined with cytotoxic drugs in breast 141 , head and neck 142 , and ovarian 143 cancer cell lines.
Another hexokinase inhibitor lonidamine was under clinical trials and revealed promising results in combination therapy for ovarian cancer clinical trial 144 and NB cell lines 145 . Furthermore, HK inhibitor 3-bromopyruvate (3-BrPA) effectively reduces cell growth of leukemia 146 , breast 147 , and colon 146 cancer cells without any significant toxicity or recurrence 146,147 . It has been efficient when used alone or in combination with other inhibitors (e.g., rapamycin 148,149 ) or cytotoxic drugs (e.g., platinum-based agents 150 and doxorubicin 151 ) for NB, leukemia, breast, lymphatic, colon, and hepatic cancers. There is also a modified version of 3-BrPA named 3-bromo-2-oxopropionate-1-propyl ester, which is a cell-permeable ester that has a strong effect on GLUT1-and MKI67expressing NB cells, but is less damaging for normal cells 152 . In addition to HK inhibitors, small-molecule PDK (pyruvate dehydrogenase kinase) inhibitors, such as dichloroacetate (DCA) 153,154 , or the downregulation of lactate dehydrogenase A (LDHA) 155 can also be used to reverse the glycolytic shift by directing pyruvate into mitochondria, to restore the characteristic phenotype of non-malignant cells. For example, DCA has successfully reduced lactate production, proliferation rate, cell viability, and increased respiration in NB cell lines 156,157 . In addition, LDHA inhibitor FX11 has successfully inhibited aerobic glycolysis and growth of NB cell lines 158 .
Besides increased glucose metabolism, many tumors, and especially NB, show signs of glutamine dependency 159 . Glutamine regulates cellular energetics, redox state, amino acid production, cell signaling, and nucleotide synthesis 160,161 . Therefore, glutamine addiction helps cancer cells to acquire substrates for rapid proliferation and to survive better in complex environments. In tumors, stimulation of glutaminolysis in low glucose and oxygen conditions is mainly induced by MYC, whereas MYC knockdown results in reduced glutamine metabolism in glioblastoma cell line 162 . Thus, removal of glutamine should lead to the death of addicted cells, whereas oxaloacetate, pyruvate, and α-ketoglutarate can rescue cells from dying, suggesting that MYC-driven glutamine metabolism is a major carbon source for the tricarboxylic acid cycle [162][163][164][165] . Therefore, targeting glutamine metabolism for MYC-driven tumors is a promising strategy for cancer therapy.
Glutamine depletion results in activating transcription factor 4 (ATF4)-dependent, but p53-independent, apoptosis as a result of the stimulation of expression of the pro-apoptotic BCL-2 family proteins PUMA and NOXA. Therefore, combinations of ATF4 agonists and glutaminolysis inhibitors have shown the induction of apoptosis and a decrease in NB tumor growth 164 . Inhibitors of glutaminase 1 by small molecules such as 986 (ref. 166) and bis-2-[5-phenylacetamido-1,2,4-thiadiazol-2-yl] ethyl sulfide [167][168][169] , suppressed cell growth, migration, invasion, and resistance to oxidative stress in MYC-overexpressing tumors. However, MYCN-amplified NB cells that predominantly express GLS2 might be less sensitive to these drugs 164,167 . Besides GLS blockers, inhibitors of glutamate dehydrogenases, such as epigallocatechin-3-gallate 170 , or aminotransferases, such as aminooxyacetate 171 , can be used to block subsequent glutamate processing. However, problems with identifying the predominant pathway in specific cancers make it difficult to predict the NB sensitivity to these drugs.

Autophagy and NB therapy
Autophagy is a catabolic survival mechanism that is activated in somatic cells under metabolic stress, to provide the cell with metabolites and to eliminate damaged organelles, protein aggregates, and infecting organisms. Extensive autophagy can also lead to cell death, but its function is not yet fully understood [172][173][174][175] . In many solid tumors, including NB, the outcome of the chemotherapeutic agents is also affected by the cellular stimulation/ activation of autophagy, which can lead to unexpected consequences and autophagy-mediated cell survival or death 176 . However, there are ongoing discussions and research to better understand whether extensive activation of autophagy could be used to induce cell death or whether it should be blocked, because it helps cells to survive in extreme environments and therefore support tumor growth.
For example, one of the reasons why previously discussed ALK inhibitors may cause resistance is due to their ability to activate autophagy-mediated cell survival. This can be avoided by using ALK inhibitors together with autophagy inhibitors, such as chloroquine, which have been shown to increase cell death of ALK-positive lung cancer 177,178 In addition, research on histone deacetylase 10 has shown its role in autophagy-mediated cell survival and poor outcomes in high-risk NB 179 . Moreover, BCL-2, a regulator of apoptosis, also controls and inhibits autophagy, which is why it seems to be one of the key factors and a potential target in balancing autophagy and apoptosis 180 . Therefore, inhibition of autophagy in combination with other apoptosis-inducing drugs is a potential strategy to induce apoptotic cell death of NB cells, especially in resistant tumors 181,182 .
Targeting PI3K/AKT/mTOR pathway The PI3K/AKT/mTOR (mechanistic target of rapamycin) signaling pathway is an important regulator of autophagy. In NB, it correlates with a poor outcome and is shown to be upregulated by constitutively activated ALK and MYCN genes [183][184][185] . The PI3K/AKT/mTOR pathway is regulated by the aforementioned RTKs, which are shown to be involved in malignant NB cell transformation, when mutated and/or amplified. Therefore, several inhibitors of RTK and PI3K/AKT/mTOR pathways have also been tested for NB therapy 186,187 . However, there are also problems with resistance, as these inhibitors cause secondary mutations and autophagy activation that supports cell survival 188,189 .
Protein kinase mTOR is considered to be the main inhibitor of autophagy and controller of cellular metabolism [190][191][192] . Deregulation of mTOR expression is very common in tumor cells and it is targeted in many NB studies, as its inhibition destabilizes MYCN, reduces NB growth, and induces excessive autophagy activation that will result in the stimulation of cell death 36,184,193 . Although clinical benefits from mTOR inhibitors, when used alone, have been modest, their effectiveness for NB in combination therapies is under investigation [194][195][196][197] . For example, the mTOR inhibitor temsirolimus (rapamycin analog) has been tested for NB in clinical trials, in combination with standard chemotherapy and monoclonal antibodies (NCT01767194) 195 . In addition, the combination of mTOR inhibitors, such as dactolisib 198 ,or INK128 (ref. 199), with ALK inhibitors or other conventional chemotherapeutics has shown the ability of the treatment to overcome drug resistance and to prevent NB tumor growth. Moreover, elevated levels of AKT are also very common in NBs 185 . Studies on combined AKT targeting have shown even more successful results, for example, the combination of AKT inhibitor perifosine and mTOR inhibitor temsirolimus is in clinical testing for pediatric solid tumors (NCT01049841) 200 . Furthermore, AKT inhibitor MK2206 in combination with etoposide or rapamycin has shown promising results in NB cell lines 201 . Taken together, targeting the PI3K/AKT/mTOR pathway and thereby inducing excessive autophagy can be used as a strategy for cancer therapy; however, targeting several pathways simultaneously should be used to avoid resistance to treatment.

Necroptosis induction in NB therapy
Cellular stress can activate various caspase-and p53independent forms of cell death in normal and transformed cells. One of them is necroptosis, which is morphologically similar to inflammation and immune response caused by necrosis 202 . It is mediated by necrotic death receptors, their ligands, interferons, Toll-like receptors, and the necrosome complex, consisting of receptor-interacting protein kinases 1/3 (RIPK1/3) and mixed lineage kinase domain-like [203][204][205][206] . Necrosome formation induces mitochondrial ROS production and the release of apoptosis-inducing factor, which are thought to be important executors of necroptosis 206,207 . Normal cell survival is supported by the inhibition of apoptosis and necroptosis, where apoptosis induction is suppressed by FLICE-inhibitory protein inhibiting caspase- 8 (ref. 208) and necroptosis induction is blocked by caspase-8mediated cleavage of RIPK1/3 (ref. 209). Therefore, the balance between these proteins will determine whether the cell will survive or die and through which pathway. Thus, it is expected that necroptosis has an important role in several human disorders, such as neurodegenerative and inflammatory diseases 210 . Moreover, necroptotic cell death can be used as a novel approach to modulate antitumor immunity and apoptosis in the treatment of resistant cells 211 .
As many aggressive NBs do not express caspase-8 and are resistant to apoptosis, inducing necroptotic cell death to eliminate these cells is another strategy to increase the efficiency of treatments 212 . One way to trigger necroptosis in NB cells is through the increase of cytoplasmic Ca 2+ that activates calcium-calmodulin kinase II, which in turn activates RIPK1 (ref. 213). Other agents inducing necroptosis in RIPK3-expressing NB cells are polyphyllin D 214 and D-gal 215 . On the other hand, many NBs have a decreased expression of caspase-8 and low level of proteins involved in necroptosis, especially in the advanced stages, making them also resistant to necroptosis induction 216 . It is not clear why these genes are downregulated in NB, but epigenetic modifications may be the reason of this outcome. Thus, demethylating drugs and/or histone deacetylase inhibitors 217,218 can be used to overcome this issue and support the use of necroptosis as a new approach for NB therapy.

Immunotherapy in NB treatment
Owing to the limitations of current therapies, many immunotherapeutic approaches can be used to induce NB cell death through redirecting the immune system to eliminate the malignant cells and to achieve long-term immunity and protection against relapse. One way is through targeting ALK-positive NBs with antibodies, to inhibit cell growth and induce cytotoxicity 219,220 . Antibodies can also be used to deliver immunotoxins, radioisotopes, liposomes, or nanoparticles 221 . This new method of drug delivery has a high potential for very specific onthe-spot effects on tumor cells, at the same time avoiding toxicity on healthy cells. This approach is also used for other surface epitopes, because NB is derived from embryonic tissue and it expresses surface antigens that are not widespread in nonembryonic tissues, such as L1-cell adhesion molecule (L1-CAM), GD2/3 (disialoganglioside), and B7H3 (ref. [222][223][224]. These antigens can be used as biomarkers to target advanced and chemotherapy-resistant NB cells with immunotherapeutic antibodies. The described strategy has shown promising results in preclinical and clinical trials with monoclonal antibodies, such as Hu3F8 (ref. [225][226][227][228] and dinutuximab [229][230][231][232] , on GD2-positive NB tumors. It has been shown that treatment with these antibodies will lead to cytotoxicity mediated by monocytes, macrophages, granulocytes, the complement system, and natural killer (NK) cells. As anti-GD2 antibodies act via cell-mediated cytotoxicity and NK cell reactivity, NB patients with higher immune activity have better outcomes from this treatment. [233][234][235][236][237][238] This method seems to be even more effective when used in combination with cytotoxic chemotherapy, cytokines, adoptive NK cell therapy, and 13-cis-retinoic acid 232,[239][240][241][242][243][244][245] . However, there have been problems with treatment efficiency, pain toxicity, and relapse; attempts to eliminate these issues have not yet been fully successful 244 . Another problem with this kind of treatment is that, generally, it does not induce immunological memory and other parts of the immune system should be used to achieve long-term effects.
For instance, there is evidence for "natural immunity" against ALK-positive NB cells. This is due to NB's peculiarity in presenting ALK peptides on human leukocyte antigen I, which is then recognized by T cells 246,247 . This led to a novel strategy that uses designed and/or activated T cells to induce bio-distributed, long-term, and direct cytotoxicity, which is free of the immunosuppressive influences of the tumor. These designed T cells have a chimeric antigen receptor against GD2, L1-CAM, or ALK, and they have demonstrated safety and no pain toxicity in relapsed NB [248][249][250][251][252][253] . Another similar approach is to use a peptide vaccine, such as ganglidiximab 254 , made from the tumor proteins, to activate T cells against the NB [255][256][257] . These strategies are already in clinical trials and demonstrating high efficiency. However, there are several potential drawbacks with these therapies, starting with the low or altered expression of HLA and its co-stimulatory molecules on the cells, complex and expensive standardization processes, and its requirement to use disease compromised immune system 246,258 .

Spontaneous regression and TrKA pathway
NB is known for its spontaneous regression by differentiation or reactivated apoptosis, which can be considered as a possible strategy for improved therapy 259,260 . Experiments with differentiation supporting vorinostat 261 , a histone deacetylase inhibitor, and didymin 262 , a citrus-derived compound, have resulted in regression of NB in xenograft models and differentiation in relapsed NB 261,262 . There are also several other simple compounds, such as all-trans retinoic acid [263][264][265][266][267][268][269] , nitric oxide 270 , and phenylacetate 267 that trigger the induction of differentiation and inhibition of NB growth by inducing the expression of neural differentiation genes. However, this mechanism is not clear, but there is evidence that NB spontaneous regression caused by retinoids is associated with increased expression of tropomyosin receptor kinase A (TrkA) receptors 269,271 .
Furthermore, spontaneous regression of NB is correlated with high expression of TrkA and its ligand nerve growth factor (NGF), which protects cells from apoptosis and directs them to differentiation, whereas NGF alone promotes apoptosis [272][273][274][275][276][277] . Therefore, changing the balance between TrkA and NGF expression can be used for the activation of NB differentiation and apoptosis. For example, re-expression of exogenous TrkA in NB cells guides cells to NGF-induced differentiation. 274,[277][278][279] Apoptotic cell death can be induced by TrkA inhibitors, like K252a (ref. 280), and GTx-186 (ref. 281) or by downregulating TrkA with miRNA-92a (ref. 282), however, these strategies are not yet clinically tested for NB. NGF can also sensitize TrkA-expressing cells for TRAILinduced apoptosis and this effect can be further increased by using inhibitors of NF-κB and/or Mcl-1 (ref. 283). However, this approach may work better for the primary NB, but not relapsed NB, which often has mutations in this regulatory pathway.
Another Trk family protein kinase is TrkB, whose expression is correlated with poor NB prognosis and MYCN amplification. For example, TrkB ligands, such as BDNF and NT-4/5, are distributed via autocrine or paracrine signaling to support overall NB viability, drug resistance, and angiogenesis of TrkB-positive tumors [284][285][286] . Therefore, targeting TrkB may reduce the malignancy of NB with dysregulated TrkB, which can be achieved by the TrkB inhibitors GNF-4256 (ref. 287) or AZD6918 (ref. 288), which have shown promising results alone and in combination in a xenograft mouse model.
Moreover, expression of a homeobox gene HOXC9 is associated with a favorable prognostic outcome and is known as a marker of spontaneous regression in infant NBs, whereas its downregulation is present in advancedstage NBs. Therefore, re-expression of HOXC9 can be used to induce NB regression or activation of apoptotic cell death in NB cell lines 289,290 . Based on all of the aforementioned information on spontaneous regression in NB, it is not clear how it is regulated. Regression seems to be as complex mechanism as all the other cellular pathways and it can include a variety of cross-talking cell death mechanisms.

Conclusion
Therapeutics inducing different modes of cell death, mainly apoptosis, have been proved to be successful, but sometimes they demonstrate a modest efficiency and side effects. The main problem with stimulating apoptosis in tumor cells is their ability to compensate for pro-apoptotic signals via upregulating anti-apoptotic agents. Therefore, searching new strategies is crucial to achieve improved outcome of NB therapy. One way to enhance the treatment is to understand better the genetic and metabolic background of NB. This in turn can be used for more specific and even personalized therapy, thereby improving the outcome of the treatment. Moreover, recent developments in NB treatment are directed towards combined therapies that target many pathways, not just different sites of one pathway. Another promising and clinically tested approach is immunotherapy, which can be used to induce NB cell death through redirecting the immune system to eliminate the malignant cells and to achieve long-term immunity and avoid relapse. However, there are several potential drawbacks, starting with the requirement to use healthy and functional immune system, as well as difficult and expensive standardization processes. Thus, there is no easy way to overcome this complex and heterogeneous disease, but step-by-step improvements are bringing us closer to prolonged survival and gain in life quality.