For the TH-MYCN mouse model, palpation is still the most common method used by researchers to examine tumor formation [22,55]. However, because tumors predominantly originate from the abdominal ganglion structures [22], it can be technically challenging for researchers to detect the early onset of tumorigenesis and track the rate of tumor progression by touch. Recently, PET or MRI scans have been used to evaluate the growth of TH-MYCN tumors in the presence or absence of different compounds, such as the Aurora A kinase or mTOR inhibitor [59,60]. These techniques seem promising and sensitive for monitoring tumor progression. However, they require expensive equipment, trained professionals, and lengthy procedures. In contrast, the onset of NB can be easily detected by fluorescent microscopy in the zebrafish model and large cohorts of animals (over 400 fish) can be monitored efficiently and regularly without harming the fish as early as 5 weeks post fertilization (wpf) [33,47,48]. Therefore, the zebrafish model allows demonstration of the cooperative contributions of multiple genetic alterations to NB tumorigenesis with high statistical power.

In addition to being useful in the early detection of tumor onset, fluorescent-tagged tumor cells can be monitored for the growth of tumor, especially tumor metastasis in real-time. In transgenic fish with overexpression of MYCN and LMO1 oncogenes, fluorescent-positive tumor masses were observed using fluorescent microscopy in the distant regions from the primary tumor site, the interrenal gland region (IRG), as early as five weeks of age [48]. Using pathological and immunohistochemical analyses with antibodies against neuroblastoma markers, such as Tyrosine Hydroxylase (TH) [61] and HuC [62], widespread metastasis were detected in the orbit, gill, spleen, distal portion of kidney and heart [48]. Additional sites of metastases, such as the bone, the liver and the pancreas, were also observed in another newly developed zebrafish model with loss of function of gas7 gene [58].

Although there are anatomical differences between zebrafish and mammals, most of the organs in mammals have their functional equivalent counterparts in zebrafish. For example, the zebrafish IRG is equivalent to the human adrenal gland, which contains the chromaffin cells (an important cell lineage of origin for NB [63]) interposed with interrenal epithelial cells in the head kidney [64]. Similarly, although the lymph nodes are absent in zebrafish, the spleen and kidney (the sites where T cells, B cells, and dendritic cells reside) serve as secondary lymphoid organs in zebrafish [65,66]. The kidney marrows also function similarly to mammalian bone marrow [65]. The gills of zebrafish have a similar structure to the mammalian airways and fulfill the same gas-exchange function, making it analogous to the mammalian lung [67,68]. With such functional, physiological and anatomical homology of zebrafish to mammals, the metastases detected in the fish model match quite well to those metastatic sites commonly seen in human NB patients, including the bone marrow (70%), bone (55%), lymph nodes (30%), liver (30%), and brain (18%) [69]. In addition, increased numbers and networks of picrosirius red-stained collagen fibers, indicating the enhanced stiffness of extracellular matrix, were observed in the fish tumors with overexpression of both MYCN and LMO1 [48], which is consistent with the concept that increased ECM stiffness contributes to enhanced metastasis [70,71]. Hence, the zebrafish model offers a unique advantage in dissecting the molecular and cellular basis and the contribution of microenvironment to NB metastasis.

It is challenging to develop effective and safe targeted therapies for cancer patients, especially for children. To do so requires using an entire organism to test the efficacy of new drugs or novel combinations thereof and evaluate the drug toxicity in order to develop new targeted therapies. Several NB-bearing zebrafish models have been used to demonstrate the effectiveness of novel inhibitors and combinations of new compounds in the treatment of NB, along with illuminating of the underlying mechanisms that may contribute to drug resistance in NB therapy (Figure 2).

For instance, using the dβh-MYCN zebrafish NB model (designated the MYCN line), Radic-Sarikas et al. demonstrated that the epidermal growth factor receptor (EGFR) kinase inhibitor, lapatinib, can prolong and enhance the cytotoxicity of YM155, an anti-cancer drug, by inhibiting the multidrug-resistance efflux transporter ABCB1 [49]. This led to the synergistic inhibition of the growth of MYCN-overexpressing NB in vivo. Similarly, using the transgenic fish with overexpression of both MYCN and proliferation associated 2G4 (PA2G4), Koach et al. demonstrated that a small molecule, WS6, can competitively bind to PA2G4 to prevent its interaction with MYCN, leading to destabilization and reduced expression of MYCN and in turn suppressed growth of NB [55]. Using the same zebrafish MYCN model together with the transgenic fish overexpressing both MYCN and GAB2, Zhang et al. demonstrated that the MEK inhibitor trametinib can enhance the sensitivity of MYCN and GAB2-overexpressing NB to the treatment of CBL0137, a histone chaperone FACT inhibitor [47]. In addition, He et al. applied the MEK inhibitor (trametinib) and retinoid (isotretinoin) on the juvenile compound fish with loss of nf1 in the context of MYCN overexpression and demonstrated synergistic killing of tumor cells by this combination treatment [46]. Taken together, these zebrafish NB models showed considerable translational potential for investigating new strategies to improve the treatment of this devastating childhood tumor.

Retrovirus-mediated insertional mutagenesis in zebrafish was established in 1996 [73]. It is a powerful forward-genetic approach for identifying genes that are critical for embryonic development [74].

In 2009, led by Drs. Nancy Hopkins and Jacqueline Lees, a team of researchers performed a comprehensive insertional mutagenesis screen in zebrafish with a goal to identify genes susceptible to cancer. Four mutant fish lines with viral insertions in the fbxw4 gene were identified. Interestingly, the mis-regulated fibroblast growth factor 8 (fgf8) in these mutants was found to contribute to neuroblastomagenesis [42].

Although the forward-genetic approach was successful, it was quite challenging to screen for fish with spontaneous development of NB and to effectively identify disease-driving genes. Therefore, the reverse-genetic strategy became a more popular approach to model NB using the zebrafish in the past decade.

Using a reverse-genetic approach, the first transgenic zebrafish model of neuroblastoma was developed in Dr. Thomas Look’s laboratory in 2012. Two stable transgenic lines with overexpression of MYCN or ALK^F1174L were generated using meganuclease (I-SceI)-mediated transgenic strategy [33]. It is important to note that human genes, rather than zebrafish genes, are commonly chosen for developing a transgenic zebrafish NB model. Doing so allows researchers to avoid the potential issues of targeting transgenes or validating the efficacy of drugs designed to interact with human proteins.

The I-SceI-mediated transgenesis is highly efficient in zebrafish and frogs, although the transgene integration rate seems low in other species [75,76]. The strategy to build transgene constructs for a NB zebrafish model is similar to that of the TH-MYCN mouse model. However, the TH promoter did not work well in zebrafish. Therefore, the dβh promoter was used instead to drive expression of genes of interest in the PSNS [33]. The dβh-EGFP-MYCN construct was used to generate the first zebrafish NB model (designated as MYCN line). Due to the instability of MYCN protein, EGFP-MYCN expression was visible under fluorescence microscopy but was not strong [33]. Hence, researchers applied a strategy of co-injection of dβh-EGFP or dβh-mCherry DNA with dβh promoter-driven transgene constructs containing genes of interest for the subsequent development of multiple transgenic lines.

The co-injected DNA has been demonstrated to be co-integrated into the fish genome, leading to their co-expression in the PSNS cells [33]. The EGFP expression is detectable in the embryos as early as one-day post fertilization [33]. Hence, the EGFP-positive embryos can be easily identified and sorted using fluorescent microscopy, which eliminates additional genotyping procedures after the transgenic line is confirmed. This approach is cost-efficient for conducting large-scale experiments (which require screening hundreds of transgenic fish) at a time when lab space is at a premium.

Using a similar transgenesis approach, researchers have developed a panel of transgenic lines with overexpression of control reporters or genes of interest. These include transgenic fish lines overexpressing EGFP [33] or mCherry [48] alone as a control group, or NB-relevant genes including ALKwt [33], PTPN11 [47], GAB2 [47], LIN28B WT [56], LIN28B_MU [56], LMO1 [48], and DEF [50]. Interestingly, most of the aforementioned genes have been shown to cooperate with MYCN to contribute to the pathogenesis of NB through different cellular mechanisms (Figure 2). However, none of these genes can sufficiently drive NB tumorigenesis when they were overexpressed alone, suggesting that MYCN is a potent oncogene and a key driver for the NB initiation.

Recently, a newer stable transgenic zebrafish line with overexpression of MYCN under control of the dβh promoter was developed in Dr. Thomas Look’s lab [50]. Instead of using the DNA construct carrying a dβh-EGFP-MYCN fusion gene, two transgenic constructs, dβh-MYCN (cDNA+3′UTR) and dβh-EGFP, were coinjected into the one-cell stage of wild-type embryos in developing this line (designated as TgMYCN_TT). The advantage of the coinjection approach has been illustrated above. The major difference of the TgMYCN_TT line to the MYCN line is the inclusion of the 3′ UTR of MYCN gene, containing microRNA recognition sites [77], in the transgene construct. Thus, the MYCN expression in the TgMYCN_TT line could be regulated at the post-transcriptional level, which might link to the increased penetrance of NB to ~70% by 29 wpf [50] in the TgMYCN_TT line as compared to ~30% in the MYCN line by 25 wpf [48]. Moreover, overexpression of c-MYC—a highly expressed MYC-family gene in the MYCN non-amplified high-risk NBs—in the PSNS of transgenic fish could also induce NB tumorigenesis [51], thus implicating the key oncogenic role of MYC family genes in NB pathogenesis.

CRISPR can effectively target genes of interest and achieve up to 75–99% modification rate in zebrafish [78,79]. This high-throughput targeting can generate close to 30% germline mutations [80] with as low as 1% of off-target effects [81]. Shi et al. generated arid1aa and arid1ab knockout zebrafish using this robust technology in the NB model with MYCN overexpression to examine the relevance of ARID1A as a NB suppressor gene in vivo and found that both arid1aa and arid1ab deficiency increases penetration of MYCN-driven NB in zebrafish [57]. Alternatively, targeted deletion in the zebrafish genome can be achieved efficiently using the TALENs system [82]. With this method, a gas7 knockout fish line was generated and bred with the MYCN transgenic line, which led to development of widespread metastasis in the compound fish [58].

Given that zebrafish modeling for NB research has only become common in the last decade, there are still a plethora of research methods in the zebrafish that have not been applied to the NB study, but that have potential to be used in the near future. For example, the Cre-Lox recombination, a site-specific recombinase technology, has been widely used in different model systems. Briefly, a ubiquitously or tissue-specifically expressed tyrosine recombinase enzyme (Cre) can be used to recombine a pair of short target sequences, called the Lox sequences, leading to manipulation of DNA sequence at specific sites. Such Cre-LoxP conditional expression approach has been used to target the oncogenic Kras/Ras pathway under control of the nestin promoter [83]. Given that many lineage-specific promoters are available in zebrafish, this system is highly applicable when studies require specific activation of gene of interest in a given subset of cell population. For example, Langenau et al. was able to express mouse c-Myc in lymphoblasts of zebrafish using this experimental approach [84]. Heat shock inducible promoters also appear to be applicable which has been used to track individual retina-neuron axon pathways [85]. Additionally, tamoxifen-inducible Cre recombinase can conditionally activate gene expression when needed, such as CreERT2 in zebrafish [86]. The whole system can be easily integrated using CRISPR/Cas9 in zebrafish [87]; a detailed review on the zebrafish Cre-LoxP system has previously been published [88].