Tumor-Initiating Cells in Childhood Neuroblastoma—Letter

We report here that the neuroblastoma tumor-initiating cell (TIC) cultures (NB12, NB88, and NB122R) isolated from neuroblastoma patients and established in culture by Dr. David Kaplan and colleagues [reported by Hansford et al. (1) and Pietras et al. (2)] are severely overgrown by Epstein-Barr virus (EBV)–infected B lymphocytes. Our findings advise precaution when using the TIC lines as models for neuroblastoma development, even at modest passage numbers.

Hansford and colleagues have reported the successful retrieval of metastatic TICs from bone marrow aspirates of patients with non-MYCN amplified, high-stage aggressive neuroblastoma (1). These cells were suggested to be putative neuroblastoma stem cells on the basis of their immature neuronal phenotype (2). The TICs have been distributed to more than 20 laboratories worldwide and recently, several reports have identified drugs selectively targeting these TICs. At the 2010 Advances in Neuroblastoma Research Meeting in Stockholm, we reported that the TIC cultures (NB12, NB88, and NB122R) from 3 individual patients have a principally normal genotype [single-nucleotide polymorphism (SNP) analyses]. Kaplan and colleagues also reported the expression of B-cell markers and VDJ gene rearrangements in TIC cultures. The observations prompted us to investigate the phenotype of these TICs in more detail.

We found that the neuroblastoma patient–derived TIC cultures are contaminated by EBV-infected B-lymphoblastoid cells. They express B-lymphocyte antigens while lacking markers of immature B-, myeloid, and stem cells as analyzed by extensive flow cytometry (Fig. 1A). At fairly low passages (7 to 15), 78% to 97% of cells in the cultures were positive for CD19 and CD20, two well-characterized B-cell markers (Fig. 1B and C). Furthermore, the TIC cultures expressed the EBNA2 and EBNA3A proteins and EBV-specific mRNA (Fig. 1D–F). We carried out flow cytometric and Western blot analyses on these TICs obtained from a second independent laboratory with consistent results, excluding the possibility that the contamination arose in our laboratory (Fig. 1A–F).

TIC xenograft tumors stained positive for EBNA2 as compared with completely negative patient-derived tumors and xenograft specimens of the neuroblastoma cell line SK-N-BE(2)c (Fig. 1G–J). Most cells in TIC xenografts exhibited prominent cytoplasm occasionally with immunoblast/plasma cell–like features suggestive of a B-cell phenotype. There were no neurofibrillary matrix, rosette formation, or dispersed lumped chromatin pattern typical of neuroblastoma.

EBV infects most humans, normally beginning during childhood, and establishes a latent infection in a fraction of B lymphocytes distributed mainly in blood, peripheral lymphoid tissue, and bone marrow. As the 3 TICs analyzed by us were from 3 different individuals (SNP-data), it is likely that the retrieved bone marrow aspirates with neuroblastoma cells were contaminated with patient-derived EBV-carrying lymphoblastoid cells. Our observations explain the normal SNP profile, the B-cell phenotype, and the VDJ gene rearrangementsi n TIC cultures.

See the Response, p. 823