When LMO1 Meets MYCN, Neuroblastoma Is Metastatic

Abstract

LMO1 is a high-risk neuroblastoma susceptibility gene, but how LMO1 cooperates with MYCN in neuroblastoma tumorigenesis is unclear. In this issue of Cancer Cell, Zhu et al. develop a novel zebrafish model that elucidates a mechanism by which LMO1 and MYCN synergistically initiate neuroblastoma and contribute to metastatic disease progression.

Neuroblastoma (NB) is the most common extracranial solid pediatric tumor arising in presumptive neural crest sympatho-adrenal progenitors. It comprises more than 5% of malignancies in children and accounts for 10.2% of cancer-related deaths in childhood (Bosse and Maris, 2016; Brodeur, 2003). Approximately 70%–80% of patients over 18 months of age present with metastatic disease, usually in the bone marrow, bone, lymph nodes, liver, intracranial, and orbital boney sites. The long-term survival rates of high-risk patients remain at approximately 50% (Maris, 2010). Sporadic neuroblastoma is genetically heterogeneous. Recent genome-wide-association studies (GWAS) and next-generation sequencing (NGS) studies have identified novel germline and somatically acquired genetic mutations, as well as single-nucleotide polymorphisms (SNPs), associated with low- or high-risk disease (Bosse and Maris, 2016). Remaining to be elucidated are the mechanisms by which SNPs contribute to neuroblastoma initiation and impact tumor progression.

Among the NB high-risk GWAS loci identified in NB, the LIM domain only 1 (LMO1) locus (Wang et al., 2011) is intriguing because LMO1 is a member of a family of cysteine-rich transcription factors, LMO1-4, that are required for many developmental processes and have been implicated in the onset or the progression of different cancers (Matthews et al., 2013). Interest centered on LMO1 because DNA copy number alterations within the LMO1 locus were also identified in 12.4% of NB cases (Wang et al., 2011). A highly associated SNP (rs2168101) within the first intron of LMO1 was found to occur at a GATA transcription factor-binding motif, which created a novel super-enhancer that led to high-level LMO1 transcription (Oldridge et al., 2015). In neuroblastomas with LMO1-risk SNP or somatic copy number gains, LMO1 expression is high and is associated with disease progression. However, LMO1’s role in NB tumor initiation and progression has not been rigorously demonstrated.

The observation that risk-associated SNPs in the LMO1 locus are especially enriched in high-risk neuroblastoma patients with metastatic disease (Wang et al., 2011) led to the current study in this issue of Cancer Cell, in which Zhu et al. generated transgenic zebrafish lines that express human LMO1 (Zhu et al., 2017). LMO1, under control of the zebra-fish dopamine-β -hydroxylase gene (dβh) promoter, is expressed in the peripheral sympathetic nervous system. By co-injecting dβh:LMO1 and the dβh:mCherry constructs into zebrafish embryos at the one-cell stage, the authors are able to visualize tumor development in vivo. No tumors were observed in any fish lines with transgenic expression of LMO1 alone. Because there is a strong association between the expression of LMO1 and MYCN in the subset of high-risk neuroblastoma lacking MYCN amplification (Zhu et al., 2017), the authors asked whether the co-expression of LMO1 and MYCN could affect the onset and penetrance of neuroblastoma. By utilizing their dβh promoter-driven EGFP-MYCN transgenic fish, which develop neuroblastoma tumors (Zhu et al., 2012), they found that when LMO1 and MYCN transgenic fish interbreed, tumors develop in 80% of the MYCN;LMO1 progeny by 24 weeks of age, whereas only 20%–30% of the dβh-MYCN progeny have tumors. This indicates that high levels of LMO1 expression enhance initiation of neuroblastoma in vivo.

So how does LMO1 accelerate neuroblastoma onset and increase penetrance in MYCN transgenic fish? Using immunohistochemistry, Zhu et al. analyzed the interrenal gland region (the presumptive source of sympathoadrenal progenitors) of 5-week-old transgenic fish. The number of mCherry⁺ sympathoadrenal cells was similar between LMO1 transgenic and control fish, indicating that expression of LMO1 alone does not affect the number of sympathoadrenal progenitors at this stage. In contrast, MYCN;LMO1 transgenic fish had ~3-fold more GFP⁺ sympathoadrenal cells compared to MYCN-only fish. To determine whether the increase in sympathoadrenal cell number is due to a faster rate of neuroblast growth, the authors performed EdU pulse-labeling experiments in the control dβh:EGFP, MYCN, LMO1, and MYCN;LMO1 transgenic fish. They found that the fraction of EdU-incorporating GFP⁺ sympathoadrenal cells was significantly increased in the transgenic fish co-expressing MYCN and LMO1, compared to those expressing either MYCN or LMO1 alone. To determine whether transgenic expression of LMO1 rescued the MYCN-induced developmentally timed apoptotic response in the interrenal gland at 5.5 weeks of age, Zhu et al. used activated Caspase-3 to mark apoptotic cell death. A similar number of fish (MYCN alone or MYCN;LMO1) with apoptotic cells in the interrenal gland expressed activated Caspase-3. This suggested that transgenic expression of LMO1 was unable to rescue the MYCN-induced developmentally timed apoptosis at 5.5 weeks of age. However, because there was a significantly higher number of GFP⁺ or mCherry⁺ sympathoadrenal cells in the fish coexpressing MYCN and LMO1 than in fish expressing MYCN alone, at this time, they reasoned that expression of LMO1 synergizes with MYCN to induce neuroblastoma by increasing the proliferation of hyperplastic sympathoadrenal cells, which overcomes the MYCN-induced developmentally timed apoptotic response at 5.5 weeks of age in the transgenic fish expressing both MYCN and LMO1 (Figure 1A).

Figure 1. MYCN and LMO1 Synergistically Impact Neuroblastoma Initiation and Metastasis.

(A) In zebrafish, the MYCN overexpression causes expansion and apoptosis of the sympathoadrenal neuroblasts at 5–5.5 weeks postfertilization (wpf), and a small group of zebrafish with MYCN overexpression developed neuroblastoma (left); the overexpression of LMO1 synergizes with MYCN to induce neuroblastoma in a large group of zebrafish by increasing the proliferation of sympathoadrenal neuroblast cells, which overcomes the MYCN-induced developmentally timed apoptotic response (right). (B) ECM-associated genes and integrins are downstream targets of LMO1 that mediate LMO1 function in neuroblastoma metastasis.

Clinically, MYCN amplification is associated with high-risk, metastatic disease in neuroblastoma patients. However, murine or fish NB models driven by MYCN overexpression alone are not associated with metastatic disease. Importantly, Zhu et al. discovered that MYCN;LMO1 transgenic fish have metastatic disease. In the MYCN, MYCN;LMO1, and MYCN;ALK transgenic fish at 5 to 7 months of age, the primary tumors arising in the interrenal glands are histologically comparable to human neuroblastomas. The authors observed widespread tumor masses in multiple regions distant from the interrenal gland and kidney in the fish co-expressing MYCN and LMO1, but not in MYCN-only or MYCN and ALK transgenic fish. The metastatic sites include orbit and spleen (analogous to lymph node), common sites of NB meta-static spread in patients, as well as gill (analogous to lung), a less-frequent yet high-risk site in humans. Next, they asked whether LMO1 affects migration and invasiveness in human NB cells. The BE2C NB cell line has a T/- rs2168101 genotype within the LMO1, in which the T nucleotide disrupts the GATA binding motif, resulting in the absence of a super-enhancer and low-level LMO1 expression. Overexpression of LMO1 increased the in vitro migratory and invasive behavior of BE2C cells.

Clues as to how LMO1 affected neuroblastoma metastatic behavior were provided by RNA sequencing of BE2C cells expressing LMO1 or the control vector. Gene set enrichment assays showed significant enrichment of a gene signature encoding “matrisome-associated proteins,” consisting of structural extracellular matrix (ECM) proteins and ECM-associated enzymes, as well as gene signatures for “ECM regulators” and “integrins” in LMO1-expressing BE2C cells. Among the upregulated ECM-associated genes, the lysyl oxidase (LOX) family encodes enzymes known to promote metastasis (Kagan and Li, 2003; Erler et al., 2006). The authors treated the LMO1-overex-pressing BE2C cells with a LOX inhibitor, β-aminopropionitrile (BAPN), and found that BAPN reduced the in vitro migratory and invasiveness of LMO1-expressing cells while having little effect on the vector-control cells. These results suggest that members of the LOX family are critical downstream targets of LMO1 that mediate LMO1 function in neuroblastoma metastasis (Figure 1B).

In this study, Zhu et al. provide a compelling link as to how a risk allele contributes to NB development and progression. Importantly, they discover that LMO1 induces metastatic neuroblastoma to sites in LMO1;MYCN transgenic fish that are in part analogous to sites in neuroblastoma patients. This MYCN;LMO1 zebrafish neuroblastoma model not only enhances our understanding of how developmental alterations lead to neuroblastoma initiation and metastasis but will also be an important model for determining critical genes involved in metastatic disease and for using unbiased screens to identify drugs that target metastatic neuroblastoma.