cotransport Introduction Neuroblastoma NB which arises from
Neuroblastoma (NB), which arises from embryonal neural crest cells, is the most common extra-cranial solid tumor of childhood and accounts for more than 15% of all child cancer death [1,2]. Patients with NB frequently manifest disseminated disease at first visit and prognosis of these patients is poor with 3-year survival rate less than 40% [1,2]. It has been well-known that bone is one of the most common target sites of NB spread. More than 60% of advanced NB patients manifest bone metastases that are radiologically osteolytic or a mixture of osteolytic and osteoblastic type [3,4]. Children with bone metastases suffer from complications including intolerable bone pain, pathological fractures and bone marrow failure. Currently-available anti-NB therapies are not satisfactorily effective at treating bone metastases and these complications. Accordingly, these children show miserable clinical courses and poor survival rate [5,6]. Therefore, control of bone metastases is an important goal in the management of infants with NB  and effective therapeutic interventions designed based on the understanding of the pathophysiology of bone metastasis of NB have been awaited. However, the mechanism underlying the preferential metastasis of NB to bone remains poorly understood.
Recent studies suggest that the interactions between cancer cotransport and bone microenvironment are critical to the pathophysiology of bone metastasis [8,9]. Bone provides a fertile soil for metastatic cancer cells by releasing growth factors such as insulin-like growth factors (IGFs)  and transforming growth factor β (TGFβ)  as a consequence of osteoclastic bone resorption during bone remodeling. Cancer cells stimulated by these bone-derived growth factors consequently produce increased levels of osteoclast-stimulating factors such as parathyroid hormone-related protein (PTH-rP)  and prostaglandin E2 (PGE2) , which in turn further promotes osteoclastic bone resorption, establishing a vicious cycle between bone-resorbing osteoclasts and metastatic cancer cells [8,9]. Contribution of similar vicious cycle between NB and bone microenvironment  mediated by receptor activator of NF-κB ligand (RANKL) [14,15] and IGF-1  to the development and progression of NB bone metastasis has been also suggested.
In the present study, we developed an animal model of NB bone metastasis to advance our understanding of the mechanism of bone metastasis. We showed that an inoculation of the SK-N-AS human NB cells into the left heart ventricle of nude mice caused osteolytic bone metastases with increased osteoclastogenesis. SK-N-AS cells expressed COX-2 mRNA and produced substantial amounts of PGE2. The selective COX-2 inhibitor NS-398 inhibited PGE2 production and osteolytic bone metastases with reduced osteoclastic bone resorption and subcutaneous SK-N-AS tumor enlargement with reduced angiogenesis. These results suggest that COX-2/PGE2 system plays a critical role in the pathophysiology of osteolytic bone metastases and tumor progression of NB. Inhibition of COX-2/PGE2 is a potential therapeutic intervention for bone metastases in children with NB.
Materials and methods
Discussion In the present study, we have shown that, of the three human NB cell lines examined, only the SK-N-AS human NB cell line reproducibly developed radiologically and histologically discernible osteolytic bone metastases following intracardiac inoculation in nude mice. Histological examination demonstrated that there were numerous TRAP-positive osteoclasts present along the endosteal surface of eroded bone in the SK-N-AS bone metastases. In the co-cultures with mouse BMCs, SK-N-AS cells induced TRAP (+) MNC formation in the absence of osteoclastogenic factor. The induction was RANKL-dependent, since mRANKL mRNA expression was induced in mouse BMCs and TRAP (+) MNC formation was decreased by OPG in the co-culture with SK-NAS cells. On the other hand, SK-N-DZ and SK-N-FI cells that failed to develop bone metastases did not induce TRAP (+) MNC formation in the co-cultures with mouse BMCs. These results suggest that SK-N-AS cells possess the capacity to cause osteolytic bone metastases accompanied with RANKL-dependent osteoclastogenesis. Subsequently, we attempted to understand this capacity at molecular levels. The result that SK-N-AS CM increased TRAP (+) MNC formation in mouse BMCs suggests that SK-N-AS cells produce a soluble osteoclastogenic factor. We found that SK-N-AS cells strongly expressed COX-2 and produced substantial levels of PGE2, while non-bone metastatic SK-N-DZ and SK-N-FI cells did not express COX-2 and produced little PGE2. Furthermore, the selective COX-2 inhibitor NS-398 significantly reduced bone metastases and osteoclastogenesis in vivo and in vitro. Taken together, these experimental data suggest that COX-2 expression and consequent PGE2 production are responsible for causing bone metastases with increased osteoclastogenesis in SK-N-AS human NB cells. Consistent with our results, it was reported that elevated COX-2 expression and PGE2 production and induction of apoptosis and inhibition of tumor growth in vivo by the nonsteroidal anti-inflammatory drugs in NB [28,29]. Thus, we show here an establishment of a reproducible animal model of osteolytic bone metastases of human NB in which COX-2/PGE2 plays a critical role. We believe that this model leads us to deepen our understanding of the mechanism of bone metastasis of NB.