Abstract
Neuroblastoma (NB) is the most common solid tumour in childhood and accounts for 15% of childhood cancer deaths. It is known that high-risk NB is highly correlated with MYCN amplification. The MYCN is a member of the MYC transcription factor family which controls various target genes that regulate essential cellular processes, such as proliferation, cell growth, apoptosis, and differentiation in certain tissues in children. However, overexpressed MYCN induces proliferation and cell growth and suppresses apoptosis and differentiation pathways in NB cells. Since RNA interference (RNAi) was first described, many research groups have investigated the application of RNAi with the use of short interfering RNA (siRNA). Our aim is to induce apoptosis and differentiation using RNAi as a novel therapeutic strategy for MYCN-amplified NB. Our hypothesis is that MYCN silencing by anti-MYCN siRNA induces apoptosis and differentiation at the mRNA and protein level. We are encapsulating siRNA with liposome and integrin-receptor targeting peptide to deliver MYCN siRNA into NB cells and optimising cationic and anionic polyethylene glycol (PEG)ylated receptor-targeting nanocomplexes (RTNs). In this project, we also aimed to optimise the methods to store RTNs for a long time in trehalose, which is known as a cryoprotectant. As a result, MYCN was silenced by the siRNA at both the mRNA and protein levels, and the siRNA-mediated MYCN reduction induced downstream effects, such as a neuronal differentiation marker TrkA upregulation and the morphological changes of the cells. The anti-MYCN siRNA delivered using RTNs successfully silenced MYCN mRNA in vivo as well. We used an NB cell line with non-functional p53 and resistance toward p53-pathway dependent drugs, probably induced by multiple sessions of chemotherapy and radiotherapy. Therefore, the results are promising for a novel therapy for relapse NB with MYCN amplification. In addition, we successfully demonstrated that trehalose maintains the biophysical properties and the function of RTNs, consisting of either DNA or siRNA at -80 °C. This allows us to make a large amount of RTN for many experiments, store it for the long term, and transport it to a place far from the laboratory.
1. Introduction
1.1. Neuroblastoma
1.1.1. Neuroblastoma in clinical presentation
Neuroblastoma (NB) is one of the most common solid malignant tumours in childhood, and its incidence rate is approximately 6-10 per one million infants (Domingo-Fernandez et al. 2013). Moreover, NB accounts for more than 7% of malignancies and 15% of cancer-related mortality in childhood (Nara et al. 2007; Huang et al. 2011; Maris et al. 2007). The outcomes are widely diverse: more than 90% patients with low-risk NB are cured, while high-risk NB has a poor outcome with less than a 50% survival rate (Whittle et al. 2017).
1.1.2. Origin of neuroblastoma
Most (65%) primary NBs are diagnosed in the adrenal medulla and are associated with the sympathetic ganglia (Maris 2007; 2010). It is well-accepted that the origin cells of NB arise from the sympathoadrenal lineage of the neural crest during the development stage (Fig. 1.1) (Buechner & Einvik 2012; Cheung & Dyer 2013; Marshall et al. 2014). During embryogenesis, the neural crest gives rise to diverse cells, such as enteric neurons, Schwann cells, and adrenal medulla (Pichler & Calin 2014). The rest of the primary tumour cases arise in the chest, neck, and pelvis (Maris 2010; Cheung & Dyer 2013). Patients with bilateral adrenal NBs are rare, which suggests that they may have a predisposing genetic lesion and that two independent genetic lesions in the cells of the left and right sympathoadrenal lineage induce bilateral tumours (Cheung & Dyer 2013).
1.1.3. Familial neuroblastoma
Inherited mutations in the signalling pathways, which are important for the development of the sympathoadrenal lineage, are associated with familial genetic syndromes characterised by predisposition to NB and defeats in development (Cheung & Dyer 2013) although familial NB is rare (< 1%-2%) (Mosse et al. 2008). Significant heterogeneity in the clinical presentation of NB is seen even in the same family (Mueller & Matthey 2009). The predisposition mutation first described was a paired-like homeobox 2b (PHOX2B), and the gene encodes a homeodomain transcription factor, which promotes cell-cycle exit and neuronal differentiation (ibid.). Furthermore, PHOX2B plays a crucial role in the development of the neural crest-driven autonomic neurons (ibid.). Therefore, it was hypothesised that a mutation in the differentiation pathway regulated by PHOX2B in the sympathoadrenal lineage of neural crest may contribute to NB tumorigenesis (Cheung & Dyer 2013). However, mutations in PHOX2B can explain only a small subset of familial NB (Mosse et al. 2009).
The most common heritable form of NB is caused by a lesion in the anaplastic lymphoma receptor tyrosine kinase (ALK) gene (Mosse et al. 2009). It was reported that six out of eight families with three or more affected members had an ALK mutation in the germline cells, while the other two families had PHOX2B mutations (Mueller & Matthey 2009). An ALK mutation is expressed in the developing sympathoadrenal lineage in the neural crest, and it may control the balance of proliferation and differentiation through multiple cellular pathways, such as the mitogen-activated protein kinase (MAPK) and RAS-related protein (RAP1) signal transcription pathways (Motegi et al. 2004, Schonherr et al. 2010). Furthermore, it was reported that PHOX2B regulates ALK gene expression directly (Bachetti et al. 2010). In addition, ALK signalling may be crucial for the proliferation of the sympathoadrenal lineage during development (Reiff et al. 2011).
Chr
Gene
Gene name
Status
Predisposition
Function
References
1p36.31
CHD5
Chromodomain helicase DNA-binding protein
Deletion
Sporadic
Chromatin remodelling and gene transcription
Fujita et al. 2008
1p36.3–p36.2
DFF45
DNA fragmentation factor
Mutation
Sporadic
Apoptosis
Abel et al. 2004
1q23.3
DUSP12
Dual specificity phosphatase 12
SNP
Sporadic
Negatively regulate members of the mitogen-activated protein (MAP) kinase superfamily (MAPK/ERK, SAPK/JNK, p38)
le Nguyen et al. 2011
1p36.3
UBE4B
Ubiquitination factor E4B
Mutation
Sporadic
Ubiquitination
Korona et al. 2003
2p23
ALK
Anaplastic lymphoma receptor tyrosine kinase
Mutation
Hereditable
Genesis and differentiation of the nervous system
Mosse et al. 2008
2q35
BARD1
BRCA1-assosiated RING domain
SNPs
Sporadic
Interaction with the N-terminal region of BRCA1
Capasso et al. 2009
4p12
PHOX2B
Paired-like homeobox 2B
Germline missense or frameshift
Hereditable Sporadic
Transcription factor
Mosse et al. 2004
5q11.2
DDX4
DEAD box polypeptide 4 isoform
SNP
Alteration of RNA secondary structure
le Nguyen et al. 2011
5q11.2
IL31RA
Interleukin 31 receptor A precursor
SNP
Signalling via activation of STAT-3 and STAT-5
le Nguyen et al. 2011
6p22
FLJ44180
Long intergenic non-protein coding RNA 340
SNP
–
Maris et al. 2008
6p22
FLJ22536
Long intergenic non-protein coding RNA 340
SNP
–
Maris et al. 2008
9p21
CKDNA
Cyclin-dependent kinase inhibitor 2A
Mutation, deletion
Sporadic
Cell-cycle G1 control
Ghiorzo et al. 2006
9p23–p24.3
PTPRD
Protein tyrosine phosphatase, receptor type, D
Microdeletion
Signalling molecules that regulate a variety of cellular processes including cell growth, differentiation, mitotic cycle, and oncogenic transformation
Stallings et al. 2006
11q13
CCND1
Cyclin D1
Amplification, rearrangement
Proto-oncogene, control of cell-cycle/cellular proliferation
Michels et al. 2007
11p11.2
HSD17B12
Hydroxysteroid (17-beta) dehydrogenase 12
SNP
Converts estrone into oestradiol in ovarian tissue
le Nguyen et al. 2011
6q21
LIN28B
Lin-28 homolog B
SNP
RNA-binding protein, negatively regulates let-7 processing
Molenaar et al. 2012
11p15.4
LMO1
LIM domain only 1
SNP, amplification
Sporadic
Cysteine-rich transcriptional regulator
Wang et al. 2011
12q24
PTPN11
Protein tyrosine phosphatase, non- receptor type 11
Mutation
Sporadic
Regulate a variety of cellular processes including cell growth, differentiation, mitotic cycle, and oncogenic
Martinelli et al. 2006
17q11.2
NF1
Neurofibromin 1
Deletion
Negative regulator of the RAS signal transduction pathway
Origone et al. 2003
17q13.1
P53
Tumour protein p53, TP53
Mutation
Sporadic
DNA-binding protein
Carr-Wilkinson et al. 2010
18q21.3
DCC
Netrin 1 receptor
Mutation, deletion
Sporadic
Member of the immunoglobulin superfamily of cell adhesion molecules
Kong et al. 1997
20p11
SLC24A3
Solute carrier family 24
SNP
Sporadic
Plasma membrane sodium/calcium exchangers
Maris et al. 2008
1.1.3. Sporadic neuroblastoma
In sporadic NB, approximately 6%-10% are caused by somatic ALK-activating mutations and 3%-4% carry a high frequency of ALK gene amplification (Cheung & Dyer 2013). In addition, activating ALK mutations or amplifications are associated with high-risk NB suggesting that ALK is an oncogene in NB. There are several other hallmark genes associated with high-risk NB with a poor outcome, including BARD1, LMO1, and LIN28B (Table 1.1), however, MYCN amplification is the most common genetic lesion (Brodeur & Bagatell 2014).
Risk
MYCN Amplification
Stage
Age
at Diagnosis
Overall
Survival (%)
Current Treatment Approach
Low-risk
No
4S
< 12 months
> 91 ± 2
Supportive care
No
Locoregional
≤ 21 years
> 95
Surgery ± Chemotherapy
Intermediate
risk
No
4
> 18 months
89 ± 2
Surgery and moderate intensity chemotherapy
High-risk
Yes
Locoregional
≤ 21 years
53 ± 4
Dose-intensive chemotherapy, surgical resection of residual primary tumour, radiation to primary and resistant metastatic sites, myeloablative therapy with autologous stem cell rescue, anti‑GD2 immunotherapy and 13‑cis-retinoic acid
Yes
4
< 18 months
29 ± 4
Yes or no
4
≥ 18 months and ≤ 21 years
31 ± 1
No
4
≥ 12 years
< 10
It is known that high-risk NB is strongly associated with MYCN amplification, and it has a poor prognosis, whereas tumours with low-risk retain the ability to differentiate and regress.
spontaneously (Westermark et al. 2011; Brodeur 2003). MYCN amplification occurs in approximately 20%-25% of the primary tumours of NB (Gustafson & Weiss 2010, Buechner & Einvik 2012; Huang et al. 2011) and is a marker for prognosis, and survival rates, which are low in MYCN-amplified NB (Table 1.2) (Seeger et al. 1985; Gustafson & Weiss 2010).
1.2. MYCN
1.2.1. MYCN and c-Myc
Amplified MYCN was first found in DNA homologous to viral myc (v-myc) in an NB cell line in 1983 (Westermark et al. 2011, Huang & Weiss 2015). MYCN is a member of the Myc proto-oncogene family that includes c-Myc and MYCL, which are evolutionarily well-conserved transcription factors (Westermark et al. 2011; Whitfield & Soucek 2012). MYCN is normally located on chromosome 2p24.3 (Schwab et al. 1983; Bell et al. 2010), while c-Myc is on 8q24.21. In addition, c-Myc and MYCN are highly homologous and several domains, such as the DNA-binding domain, are shared between the two proteins (Fig. 1.2a) (Gherardi at al. 2014). MYCN is strictly expressed in certain tissues scuh as ?? of the developing embryos in humans and mice but is almost completely absent in adults, while c-Myc is expressed in all proliferative tissues in adults (Westermark et al. 2011). C-Myc and MYCN are likely to be complementary to each other during embryonic development. The c-Myc expression occurs in many tissues, except neuroepithelium, while MYCN is expressed in all proliferative tissues at high levels (ibid) (Fig. 1.2.b).
Mutations in MYCN in humans and mice is involved with birth defects, and mouse embryos with null MYCN die around E11.5, whereas mice with null MYC die around E10.5 (Huang & Weiss 2013). Hence, the role of MYCN in embryonic development seems to be essential. Similarly, MYCN is initially expressed in the entire cells during neural crest development, promoting migration and differentiation (Wakamatsu et al. 1997). Additionally, Olsen et al. (2017) reported that they successfully transformed wild-type neural crest cells to NB by enforced expression of MYCN in mice.
1.2.2. Functions of MYCN
The protein controls the expression of many genes involved in essential cellular processes, such as proliferation, cell growth, protein synthesis, metabolism, apoptosis, and differentiation (Huang et al. 2011; Feng et al. 2010; Nara et al. 2007; Westermark et al. 2011). MYCN consists of a DNA-binding, basic-region, helix-loop-helix/leucine-zipper (bHLHZip), MYC box (MB), and transactivation domain (TAD) (Fig. 1.3.a). The C-terminus bHLHZip domain heterodimerises with its partner protein, Myc-associated factor X (MAX), to form a transcriptional activator that binds to the E-box sequence: 5’-CACGTG-3’ (Westermark et al. 2011), while c-Myc can also bind non-canonical 5’-CANNTG-3’ (Bell et al. 2010).
The MYC/MAX heterodimer recruits transcription co-factors, such as complexes containing transcription domain-associated protein (TRRAP), with either tat-interactive protein 60 kDa (TIP60) or GCN5, which are histone acetyl transferase (HATs), or acetyl transferase, such as p300/CBP (ibid) (Fig. 1.3.B left). Then, histone acetylation, stimulated by the MYC/MAX complex with the recruitment, leads to an open chromatin structure that provides docking sites for additional proteins, promoting transcription. The MYC/MAX dimer also recruits factors that stimulate transcriptional elongation, such as P-TEFb and TFII-H, through phosphorylation, or the MYC/MAX complex can phosphorylate the C-terminal domain of RNA pol II (Fig. 1.3.B, left).
The MYCN can also bind to other transcriptional factors, such as Miz-1 and SP-1, on the initiator element and repress gene expression (Fig. 1b, right). The MYC/MAX recruits factors, such as histone deacetylases (HDACs) and DNA methylase 3a (Dnmt3a), which inhibit transcription induced by Miz-1/SP-1 (Westermark et al. 2011).
1.2.3. Stability and degradation of MYCN
Tight control of the MYCN function is essential for modulating cellular mechanisms (Beltran 2014). There are many mechanisms that regulate MYCN, and one such mechanism is to control the MYCN protein turnover, which is the balance between the cell-cycle rate and MYCN degradation (ibid.). The MYCN protein has a half-life of approximately 20 to 30 minutes in normal cells, while the MAX expression is abundant and consistent in quiescent and proliferation cells (Lu et al. 2003). On the other hand, the protein is extremely stable (about 100 times more than normal) in some amplified NB tumour cell lines which ensures the cells remain in cell cycles and do not go into the G1 phase (Bonvini et al. 1998; Beltran 2014). The extraordinary stability of the MYCN protein in NB with MYCN amplification may be correlated with co-amplification and co-expression of NCYM (Suenaga et al. 2014) and co-amplified AURKA (Otto et al. 2009).
MYCN is degraded by F-box and WD repeat domain-containing 7 (Fbxw7a), ubiquitin ligase, via the ubiquitin-proteasome system (Sjostrom et al. 2005; Gustafson & Weiss 2010). This occurs during the mitotic phase after the cell-cycle kinase cyclin B/CDK1 and glycogen synthase kinase 3β (GSK3β) phosphorylate MYCN at serine 62 (S62) and threonine 58 (T58), respectively (Sjostrom et al. 2005). Additionally, PP2A dephosphorylates the S62 phosphate in the Pin1-mediated process (Gustafson & Weiss 2010). Because AKT can inactivate GSK3β, MYCN is stabilised by the PI3K/AKT pathway (Barone et al. 2013; Chesler et al. 2004) (Fig. 1.4.). The AKT/PI3K pathway is activated by receptor tyrosine kinases (RTKs), such as ALK, Trk, and IGF 1R. The RTKs are activated by binding ligands or by mutations, causing continuous activation including activated ALK (Gustafson & Weiss 2010). Furthermore, MYCN is normally located on chromosome 2p24.3 (Schwab et al. 1983; Bell et al. 2010) while ALK is linked to MYCN on chromosome 2p23 (Gustafson & Weiss 2010). There is no evidence to show the direct connection between ALK and MYCN on the chromosome. However, it is likely that activated ALK and other RTKs are involved with the stabilisation of MYCN (Barone et al. 2013, Gustafson & Weiss 2010).
AKT can also activate mammalian target of rapamycin (mTOR) indirectly through several signalling mechanisms. The mTOR forms a complex called mTORC1, and the complex inhibits PP2A by phosphorylation, which contributes to MYCN activation (Gustafson & Weiss 2010). Similarly, Aurora A kinase (AURKA), which is a mitosis kinase and is usually expressed during G1 and mitosis, inhibits Fbxw7-mediated degradation and stabilises MYCN. Overexpression and amplification of AURKA often occurs in MYCN-amplified NB. The two proteins cooperate to promote tumour proliferation and oncogenic activity (Otto et al. 2009). Brockmann et al. (2013) reported that inhibition of AURKA triggers degradation of MYCN in NB. Furthermore, when the MAPK pathway is hyperactivated by activation of H-RAS through an oncogenic RAS mutation and other oncoproteins, the pathway induces MYCN accumulation by accelerating MYCN translation (Kapeli & Hulin 2011).
Suenaga et al. (2014) reported that NCYM, which is the Cis-antisense gene of MYCN, is co-expressed and co-amplified with MYCN in some human primary NB and the gene codes a protein which inhibits MYCN phosphorylation by GSK3β, and so promotes MYCN stabilisation (Suenaga et al. 2014). However, Zhao et al. (2016) reported that they could not identify the NCYM protein, with either of the two commercially available anti-NCYM antibodies. In addition, Duffy et al. (2013) stated that GSK3β inhibitors reduce the MYCN mRNA levels and NB cell viability. The MYCN phosphorylation by GSK3β can stabilise MYCN (Duffy et al. 2013).
1.2.4. Transcription of MYCN
Even though MYCN has been investigated for decades, the main transcription factors and underlying essential mechanisms for MYCN expression in NB remain poorly understood (Zhao et al. 2016). Despite the findings of transcription factors for MYCN expression in the other cancer cells, such as specificity protein 1 (Sp1) in cervical cancer (Inge et al. 2001), they are not sufficient for the activation of MYCN transcription in NB (Kramps et al. 2004). Furthermore, Suenaga et al. (2009) reported that the MYCN promoter was activated by the MYCN protein through the direct recruitment to intron 1 on the MYCN gene containing two putative E-box sites.
Recently, evidence has shown that noncoding RNA plays a critical role in pathogenesis in NB. Yu et al. (2009), for example, reported that noncoding RNA expressed in aggressive NB (ncRAN), which is mapped on chromosome 17q25.1, is associated with a poor clinical outcome. Another example is that the loss of lncRNA NB-associated transcript-1 (NBT-1) leads to NB progression via increasing proliferation and suppressing differentiation of neuronal precursors (Pandey et al. 2014).
Importantly, Lui et al. (2016) reported that a long noncoding RNA transcribed from -14 kb upstream from the MYCN transcription start site (lncUSMycN) is overexpressed in NB tissues and cell lines with MYCN amplification, and it upregulates NCYM by activating the transcription of the NCYM gene. The NCYM RNA forms a complex with NonO, RNA-binding protein (Fig. 1.5.), and they upregulate the expression of MYCN mRNA. In addition, Suenaga et al. (2014) mentioned that MYCN activate the NCYM transcription.
1.2.5. MYCN and target genes
Many genes have been identified as downstream targets of MYCN (Fig. 1.6.), but it is still not clear how these genes are regulated directly or indirectly (Valentijn et al. 2012); for example:
• MDM2 (negative regulator of p53, upregulated by MYCN),
• p53 (tumour suppressor, upregulated),
• the neuronal-leucine rich repeat-1 (NLRR1: transmembrane protein with unknown function, upregulated),
• S-phase associated kinase 2 (SKP2; a component of the ubiquitin ligase complex, upregulated),
• DKK3/Wnt/β-catenin pathway (DKK3, supressing Wnt: downregulated, Wnt/β-catenin pathway, stimulating expression of target genes: upregulated) (Koppen et al. 2007), and
• PI3K/Akt/mTOR pathway (inducing S6 kinase and elF4E, initiation factor of eukaryotic translation) (Slack et al. 2005; Valentijn et al. 2012; Bell et al. 2010, Beltran 2014).
There are studies published by several groups that MYCN expression at higher levels also triggers not only proliferation and cell growth, but also apoptosis (Gustafson & Weiss 2010). Gustafson and Weiss (2010) pointed out that, if the hypothesis is true, then it contradicts the hypothesis that MYCN amplification is strongly correlated with high-risk NB. The inhibition mechanisms of apoptosis as a contributor to MYCN-mediated transformation are complex and poorly understood (Gustafson & Weiss 2010).
Furthermore, many groups have observed the effect of MYCN silencing in MYCN-amplified NB using antisense or RNA interference, and the results are various and appear to depend upon the experimental condition, such as the cell lines. Overall, the reduction of MYCN expression is regarded to induce cell arrest in the G1 phase of the cell cycle, differentiation, and/or apoptosis (Kang et al. 2006; Westermark et al. 2011; Bell et al. 2010). Conversely, the increase of MYCN expression triggers the re-entry of the quiescent cells into the cell cycle, which notably shorten the G1 phase and decrease the cell attachment to the extracellular matrix (Bell et al. 2010; Gherardi et al. 2014).
Additionally, high levels of MYCN help the development of NB in terms of proliferation, whereas normal cells can be differentiated. If MYCN is inhibited, then NBs have reduced functions of proliferation and cell growth and eventually cannot survive (Soucek & Evan 2010). Differentiation and apoptosis pathways are important in therapeutics for NB. The details are below.
1.2.5.1. Differentiation
NB is thought to arise originally from the neural crest cells and thereby can self-renew and maintain pluripotency. It is likely that MYCN is associated with the regulation of these stem cell-like properties because Myc can be replaced with MYCN in reprogramming fibroblasts into iPS cells (Nakagawa et al. 2010). Hence, MYCN (and Myc) promotes ‘a stem-like state’ because it is likely to block differentiation pathways and induce self-renewal and pluripotency factors (Huang & Weiss 2013). The pluripotency genes KLF2, KLF4, and LIN28B are upregulated by MYCN (Cotterman & Knoepfler 2009). In addition, endosomal and mesodermal differentiation markers (BMP4 and GATA6) are upregulated in MYCN knockout mice (Varlakhanova et al. 2010). Similarly, differentiation proteins, such as cyclin-dependent kinase-like 5 (CDKL5) (Valli et al. 2012) and tissue transglutaminase (TG5) (Liu et al. 2007), are supressed by MYCN. Valli et al. (2012) mentioned that CDKL5 stops the cell cycle and promotes differentiation after MYCN knockdown.
Similarly, expression of MYCN downregulates cell division control protein 42 (CDC42), a G-protein involved in a cytoskeleton remodelling pathway, and upregulates nm23 (the nucleotide diphosphate kinase, nm23-H1:NME1 and nm23-H2:NME2), which also downregulates CDC42. This interaction inhibits differentiation of MYCN-amplified NB (Valentijn et al. 2005; Bell et al. 2010). Additionally, MYCN upregulates PAX3, which encodes a transcription factor that is expressed in active progenitor cells during early embryogenesis; the PAX3 expression is subsequently downregulated during neural differentiation (Harris et al. 2002).
Tropomyosin receptor kinase A (TrkA) is a member of the tyrosine receptor kinase family, along with TrkB and TrkC (Westermark et al. 2011). Moreover, TrkA is a high affinity NGF receptor, while TrkB is a brain-derived neurotropic factor (BDNF), and TrkC is neurotrophin-3 (ibid.). In addition, TrkA is predominantly expressed at a later stage in development, and it is likely that it plays a crucial role in the complete differentiation of sympathetic neurons in normal cells (Dixon & McKinnon 1994). When TrkA is co-expressed with neurotrophin receptor p75NTR, differentiation is enhanced (Ho et al. 2011).
Iraci et al. (2011) reported that TrkA and p75NTR are downregulated by MYCN/SP-1/MIZ-1 repression complex recruiting HDAC1, which affects malignancy of NB by inhibiting the cell response to NGF. Hence, the expression level of TrkA and p75NTR is normally low in aggressive MYCN-amplified NB (Iraci et al. 2011).
1.2.5.2. Apoptosis
The protein p53 is widely known as a proapoptotic protein, but it is rare to find p53 mutations in primary NB tumours (Westermark et al. 2011; Huang et al. 2011). MYCN upregulates both proliferation and apoptosis (Fulda et al. 2000). Hence, the outcome is dependent upon the status of apoptotic factors cooperating with MYCN, such as BCL2 (the anti-apoptotic protein) (Strasser et al. 1990) or p53 (Chesler et al. 2008). Therefore, MYCN is likely to cooperate with suppressors of p53 signalling, such as miRNA-380-5p (Swarbrick et al. 2010), CUL7 (Kim et al. 2007), BMI1 (Huang et al. 2011), H-Twist (Valsesia-Wittmann et al. 2004), and MDM2 (Slack et al. 2005). Interestingly MDM2, an E3 ubiquitin ligase, has an important role in the apoptosis pathway in NB by promoting survival by ubiquitination and degradation of p53 (Fig. 1.5.). It is also thought that MDM2 binds the AU-rich elements of the 3’UTR of MYCN mRNA, thereby stabilising the mRNA (Gu et al. 2012). The MYCN can trigger transcription of TP53 and MDM2, and MDM2 is a target of p53-mediated transcription (Slack et al. 2005; Chen et al. 2010).
In contrast, mutations in the p53 pathway are found in NB at relapse, which may occur in response to cytotoxic chemotherapy. Chemotherapy initially has efficacy in the treatment of MYCN-amplified NB, partly because of MYCN-mediated p53 activation (Huang & Weiss 2013). Eventually, these tumours may acquire resistance to the therapy as a result of mutations that inactivate p53 (ibid.). The MYCN and MDM2 upregulate each other, thereby conferring a survival advantage to NB, leading to relapse (ibid.).
Similarly, several groups found that the promoter of apoptotic initiator caspase-8 is methylated (Casciano et al. 2004; Banelli et al. 2005; Lazcoz et al. 2006), and it is likely that it is a mechanism of apoptosis evasion in NB with MYCN amplification. Loss of caspase-8 contributes resistance to tumour necrosis factor-related, apoptosis-inducing, ligand-induced apoptosis in NB cells (Eggert et al. 2001).
Furthermore, prosurvival signalling cascades are constitutively activated in MYCN-amplified NB, while proapoptotic signalling is supressed. For instance, activation of tropomyosin receptor kinase (TrkB), the same group member as TrkA, is frequently seen in MYCN-amplified NB, while the expression level is low in non-MYCN-amplified NB (Nakagawa 1994). The TrkB activation is associated with resistance to chemotherapy and can upregulate MYCN mRNA. This may imply that TrkB activation is associated with MYCN-amplified NB (Ho et al. 2002; Dewitt et al. 2013).
1.2.6. Current treatment of neuroblastoma with MYCN amplification
Like other cancers, the treatment methods used for NB therapy include surgery, chemotherapy, monoclonal antibody treatment, and radiotherapy (Fig. 1.2.) (Maris et al. 2007; Macmillan Cancer Support 2015). The NB tumours vary remarkably based upon their stage and biological features (Murphy & Quaglia 2014).
Localised NB generally have favourable biological features, and surgery alone successfully treats them (Maris 2007). Even MYCN-amplified NB could achieve remission for a long time after surgery alone if it is localised (ibid.). As chemotherapy, 13-cis-retinotic acid (RA) has become a standard of therapy in high-risk NB after neuronal differentiation by retinoid was shown in vitro (Cheung & Dyer 2013). However, RA resistance can be several NB cells have obtained, which leads to relapse (Clark et al. 2013).
Similarly, anti-GD2 monoclonal antibody has become a standard of care for patients with high-risk NB since 2010. The GD2 is expressed in mature neurons and during foetal development and across NB cells with high density, membrane proximity, and homogeneity. However, the dose is constrained due to the side effects and its efficacy has been observed only in minimal residue disease and has hardly been seen in high-risk NB (Cheung & Dyer 2013).
In short, overexpression of MYCN induces proliferation and cell growth and supresses apoptosis and differentiation, which contributes to NB tumorigenesis. In addition, there is no therapy that demonstrates efficacy to cure patients with MYCN-amplified NB. Therefore, MYCN might be a promising target for treatment against high-risk NB, and gene therapy silencing of MYCN by RNA interference (RNAi) can be a novel therapy for NBs.
1.3. Gene Therapy
1.3.1. RNA interference
Fire et al. (1998) first discovered RNAi as double-stranded RNA (miRNA) that enabled gene silencing in the nematode worm Caenorhabditis elegans. After that, small interfering RNA (siRNA) was identified as 25-nucleotide antisense RNA in plants by Hamilton and Baulcombe (1999). Then, Elbashir et al. (2001) reported that 21-nt double-stranded RNA mediates RNAi in mammalian cells in vitro.
RNAi is a naturally occurring mechanism to regulate genes in most eukaryotic cells and uses a small double-strand RNA (dsRNA) molecule from endogenous or exogenous origin to homology-dependently control gene activity (Aagaad & Rossi 2007; Almeida et al. 2005). In addition, siRNA is 21-22 nt long dsRNA that is two nucleotides longer at its 3’ side, allowing it to be recognised by the enzymatic machinery of RNAi that evolutionally induces homology-independent degradation of the target mRNA (ibid.). The siRNA is generated by ribonuclease digest of dsRNA in the Dicer integrative complex (Zhang et al. 2004; Viderira et al. 2014) (Fig. 1.8.). The siRNA is subsequently integrated into the RNAi-induced silencing complex (RISC) consisting of Argonaute 2 (Ago2), Dicer, and dsRBP. It is known that Dicer has two partner, double-stranded RNA-binding proteins (dsRBP): transactivation response RNA-binding protein (TRBP) and protein activator of PKR (PATC). When Ago2 forms a complex with Dicer and TRBP or PATC, the complex can select the predicted strand, while Ago2 alone uses the guide and passenger RNAs equally (Noland & Doudna 2013).
The RISC complex is activated to recognise homologous mRNA sequences when Dicer cleaves the sense strand (the passenger strand), allowing the remaining antisense strand to become the guide strand within the RISC complex. The RISC comprising the guide siRNA binds the complementary sequence of the target mRNA, allowing the mRNA to be degraded or cleaved (Viderira et al. 2014; Singh et al. 2009; Aagaad & Rossi 2007). The RISC comprising the guide siRNA binds mRNA and degrades or cleaves the mRNA again (Whiteheat et al. 2009).
1.3.2. miRNA and siRNA
During the last 20 years, ncRNA including siRNA and microRNA (miRNA) appears to play an important role in regulation of cellular processes (Buechner & Einvik 2012). It is observed that dysregulation of ncRNA is associated with several diseases, including cancer and cardiovascular and developmental disorders (Esteller et al. 2011; Taft et al. 2010). Moreover, miRNA and siRNA have been widely investigated because of the therapeutic potential to regulate target genes and proteins, predisposing diseases including cancers and infections (Lam et al. 2015). The physicochemical properties of siRNA and miRNA are similar, while their functions are distinct (Table 1.3.).
Synthetic single RNAs (miRNA) can inhibit the activity of the endogenous miRNA (miRNA antagonist). In addition, synthetic miRNA can mimic endogenous miRNA functions, which leads to mRNA degradation and gene silencing (miRNA mimic) (Bader et al. 2010). The main difference between siRNA and miRNA for therapeutic purpose is that siRNA targets one mRNA and binds mRNA, whose sequence is fully complementary, while miRNA targets several mRNAs. The sequence is partially complementary to the targeted mRNA (ibid.). Therefore, siRNA more specifically silences the targeted mRNA.
1.3.3. Off-target effects of siRNA
However, siRNA off-target effects have been observed. There are two types of siRNA off-target effects: miRNA-like off-target effects and innate immune response (Jackson & Linsley 2010). In miRNA-like off-target effects, the guide stand of siRNA imperfectly matches the region of 3’ UTRs of these transcripts with 5’ of the siRNA, which leads to translation arrest or mRNA cleavage (ibid.). It was revealed that siRNA downregulated a set of transcripts which are enriched for transcripts with 3’ UTR complementary to the 5’ end of the corresponding siRNA guide strand (Jackson et al. 2006; Birmingham et al. 2005). Therefore, the sequence of 5’ end of siRNA is important to avoid off-target effect silencing. In addition, it is likely that it is correlated to the siRNA concentration; reducing siRNA concentration minimises the miRNA-like off-target effects (Dharmacon). Chemically modified siRNA is also available in the market.
The other off-target effect is the innate immune response. The siRNA or the vehicles delivering siRNA, such as cationic lipids, are sensed by toll-like receptors (TLRs) expressed by mammalian immune cells, which detect pathogen-associated molecular patterns (Judge & MacLachlan 2008; Schlee et al. 2006). There are several types of TLRs which can detect RNAs: TLR3, TLR7, and TRL8. They move between the endoplasmic reticulum and intracellular compartments. The TLRs stimulate interferons, tumour necrosis factor alpha (TNFα), and interlukin-6 (IL-6) and reduce gene expression through the recognition of viral infection (Jackson & Linsley 2010). It is likely that siRNA can activate TLR3 signalling, but TLR3 is not a major mechanism of siRNA-activated immune cells (Kariko et al. 2004; Sledz et al. 2003). The siRNA seems to activate TLR7 and TLR8 and to trigger the production of pro-inflammatory cytokines in monocytes and myeloid dendritic cells or the production of interferon α (INFα) in plasmacytoid dendritic cells activated by siRNA.
The immune response toward siRNAs is dependent upon the cell types due to the selective expression of TLRs (Hornung et al. 2005; Judge et al. 2005; Sioud et al. 2005). Because not all sequences are detected by TLRs, the sequence of siRNA is important to avoid the off-target effect of the innate immune response as well (Jackson & Linsley 2010).
1.4. Receptor-targeting Nanoparticles (RTNs)
1.4.1. The siRNA delivery by RTNs
Due to its negative charge, naked siRNA cannot be taken into negatively charged cell plasma membranes. Xu et al. (2016) mentioned that siRNA delivery using nanoparticles requires: 1) protection of siRNA in the blood and no interaction with proteins in the blood for prolonged circulation, 2) cellular internalisation, and 3) siRNA release to cytoplasm and protection of siRNA from endosomal degradation.
In addition, in vivo, there is a size limitation due to the size of the normal endothelial structure; therefore, RTNs should be smaller than 150 nm to pass through the vascular endothelial barrier (Fig. 1.10. Lower). On the other hand, the structure of blood vessels is altered at the inflammation and solid tumour sites; the vasculature and endothelial structure become ‘leaky’. Particles less than 500 nm can cross the wall and have enhanced permeation and retention (called the EPR effect). This effect permits RTNs delivered into the bloodstream to be extravasated into targeted tumour cells (Li & Szoka 2007).
Previous studies on gene and drug delivery have focused on accumulation and penetration via the EPR effects. However, recent studies have revealed that accumulation in the solid tumour does not enhance its therapeutic efficacy (Xu et al. 2016) probably because of the poor drug release (Zhao et al. 2013).
We have developed cationic and anionic receptor-targeting nanocomplexes (RTNs) consisting of lipids, receptor-targeting peptide, and nucleic acid (i.e. plasmid DNA, mRNA, or siRNA) to deliver siRNA into NB cells (Fig. 1.9.). Nanoparticles, consisting of lipids and receptor-targeting peptide, allow siRNA to be delivered into targeted cells.
Once the targeting peptides of RTNs bind to specific receptors on the targeted cell surface, they are internalised via receptor-mediated endocytosis or clathrin-coated pits. The particles start to be degraded within the endosome and most lipids of the RTNs fuse with the lipid of the endosome. Then, peptide/siRNA complexes are released to the cytoplasm and RNAi by siRNA processes (Hart 2010) (Fig. 1.10. Upper).
Baetlett and Davis (2006) mentioned that the duration of gene silencing by RNAi is dependent upon the doubling time of the cells presumably because the siRNA is diluted. Hence, gene silencing in non-dividing cells lasts for a month or more both in vitro and vivo (Zuckerman & Davis 2015).
There are several advantages to using non-viral nanoparticles; for instance, a wide variety of formulation can be created. Moreover, RTNs can package any type of nucleic acid and drug and can deliver several nucleic acids and drugs at the same time (Hart 2010).
1.4.2. The siRNA for therapeutics for cancers in clinical trials
Koldehoff et al. (2007) first reported that their research group systemically treated a single patient with chronic myeloid leukaemia (CML) with anti-bcr-abl (a fusion gene found in most CML) siRNA using anionic lipid nanoparticles. As a result, the siRNA remarkably achieved more than 90% gene silencing and induced apoptosis in CML cells; however, after the second dose, the silencing efficiency was not as significant as the first dose, even though the dose was increased (Koldehoff et al. 2007). Numerous studies using siRNA delivered by lipid-based nanocomplexes have been brought to clinical trials. The most advanced study using siRNA encapsulated by lipid nanoparticles is inhibiting transthyretin in transthyretin amyloidosis, and it is now in Phase III (Zuckerman & Davis 2015).
In cancer studies on clinical trials, anti-KRAS siRNA delivered by biodegradable polymer matrix (siG12D LODER) for pancreatic cancer is now in phase II, which is the most advanced siRNA study for cancer (NCT0676259, not yet recruiting). The siG12D LODER is for locally advanced pancreatic cancer and is combined with chemotherapy (Golan et al. 2015). In systemic therapy using liposomes, the most advanced clinical trials are anti-PLK1 siRNA for neuroendocrine tumours (TKM 080301) and adrenocortical carcinoma and anti-PKN3 for advanced or metastasis pancreatic cancer (Atu027) on Phase I/II. Non-targeting lipid nanoparticles (Atu027, consisting of three cationic lipids) were employed and have been completed in 2016 (Lam et al. 2015). Their silencing efficiency was not published. A summary is shown in Table 1.4.
1.4.3. Toxicity of lipid-based vectors
The final goal in gene therapy using RNAi is to achieve efficient gene silencing in the targeted tissues in clinical use without immune activation and toxicity. Currently, viruses, liposomes, and polycationic polythylenimine (PEI)-based nanoparticles are primarily used in therapeutic RNAi (van den Boorn et al. 2011). These three vehicles showed some success; however, there are issues concerning the safety and toxicity in each type of vector. For example, there are many advantages in RNAi using viral vectors, such as long-term silencing by integrating small hairpin RNA (shRNA) cassettes into the genome; however, there are major drawbacks. Viral vehicles may activate complement or coagulation factors (Waehler et al. 2007), and trigger neutralisation of antibody response, which prevents multiple administration. The main issue is the dysregulation of gene expression by insertional mutagenesis and oncogenesis (van den Boorn et al. 2011).
Name
Indication
siRNA target
Phase
Delivery system
Route of administration
Trial ID
ALN-VSP02
Advanced solid tumour with liver involvement
KSP and VEGF
I, completed
Lipid nanoparticles
Intravenous
NCT01158079 NCT00882180
Atu027
Advanced solid tumour
PKN3
I, completed
Lipid nanoparticles
Intravenous
NCT00938574
Atu027
Pancreatic ductal carcinoma
PKN3
I/II, completed
Lipid nanoparticles
Intravenous
NCT0180638
CALAA-01
Solid tumour
RRM2
I, terminated
Polymer-based targeted nanoparticles
Intravenous
NCT00689065
DCR-MYCN
Solid tumour, multiple myeloma non-Hodgkin’s lymphoma
MYC
I, terminated
Lipid nanoparticles (EnCore)
Intravenous
NCT02110563
DCR-MYCN
Hepatocellular carcinoma
MYC
I/II, terminated
Lipid nanoparticles (EnCore)
Intravenous
NCT02314052
siG12D LODER
Advanced pancreatic cancer
Mutated KRAS oncogene
I, completed
II, not yet recruiting
Biodegradable polymer-based scaffold
Local implantation
NCT01188785
NCT01676259
siRNA-EphA2-DOPC
Advanced cancer
EphA2
I, recruiting
Natural liposomes
Intravenous
NCT01591356
TKM-080301
(TKM-PLK1)
Primary or secondary liver cancer
PLK1
I, completed
Lipid nanoparticles
Intravenous
NCT01437007
TKM-080301
(TKM-PLK1)
Neuroendocrine tumour and adrenocortical carcinoma
PLK1
I/II, completed
Lipid nanoparticles
Intravenous
NCT01260035