A variety of histologic/molecular characteristics portend a poor prognosis in neuroblastoma. These include 
aggressive histology, high mitotic index and N-myc amplification. However, one of the single worst 
prognostic determinants is the ability to establish a cell line in vitro. Hence, the availability of a cell line is 
predictive of a poor outcome (Figure 12). Therefore we are likely to indirectly select for patients most likely to 
benefit from any effective adjuvant therapy. 
Initially we will try to identify HLA A2 haplotype cell lines, as this may be the most common haplotype. 
Since some neuroblastoma cells are known to express MAGE-1, the recently identified melanoma associated 
antigen recognized by cytolytic T-cells in patients with HLA-Al haplotype (79), we will also try to identify 
cell lines from patients with A1 haplotype, particularly those that are MAGE-I expressors. Using molecular 
techniques, we have typed two of seven cell lines (LAN-1 and CHLA-127) as HLA-A2 haplotype, in 
collabaation with Dr. P. T^asaki (UCLA). These are presently undergoing y-EFN transduction. 
3.3.2 TRANSDUCTION OF NEUROBLASTOMA CELL LINES: A number of neuroblastoma cell lines have 
been transduced recently using the retroviral vector DA Huy-IFN (Viagene, San Diego) or an alternative 
construct using the same vector backbone, but with y-IFN sequences replaced by P-galactosidase coding 
sequences to assay for transduction efficiency. These are indicated below: 
Cell line 
LAN-1 : 
LAN-5: 
LAN-6: 
SK-N-AS: 
SK-N-RA: 
CHLA-125 
CHLA-127: 
Phenotype 
N-myc amplified, MHC-I negative, A2 haplotype. 
N-myc amplified, MHC-I negative 
N-myc overexpressing, not amplified, weakly positive for MHC-I. 
N-myc non amplified, non expressing, MHC-I strongly positive. 
N-myc non amplified, MHC-I strongly positive 
Living patient, N-myc status not yet characterized, MHC-I not yet characterized. 
Living patient, N-myc status unknown, MHC-I not yet characterized, A2 
haplotype. 
Conditions for G418 selection have been optimized using the above cell lines. G418 levels of 100 - 2000 
p/ml were tested for multiple lines. For most neuroblastoma lines, 500 pg/ml of G418 results in acceptable 
levels of cell death within 10-12 days. Optimal concentration and types of polycations for transduction have 
also been determined (Fig. 13). 
Levels of P galactosidase expression have ranged from less than 1% for LAN-6 to 12% for SKNRA without 
■p 
G418 selection (Table V). However, stable y-IFN, neo transduced lines were readily derived from both cell 
lines. Levels of y-IFN production for LAN-6, SK-N-AS and SK-N-RA are presented in Table V. Levels in 
transduced neuroblastoma cells have ranged from 12U/10^/day to as high as 596U/10^/day. 
y-EFN may produce some slowing of growth of some neuroblastoma cell lines. In addition, we have observed 
at least one cell line LAN-5, which was readily transduced using the P-gal vector (20% transduction efficiency 
at an MOI of 2) but for which stably transduced y-IFN expressing lines could not be derived. This may be 
representative of the subset of patients for which allogeneic rather than autologous cell lines may need to be 
used, due to difficulties in outgrowth of y-IFN producing cells, in some instances. 
3.3.3. INDUCTION OF MHC CLASS I AND CLASS II EXPRESSION ON NEUROBLASTOMA 
CELL LINES: As had been hoped, y-IFN transduction produced a variety of phenotypic changes in 
neuroblastoma cells, which may enhance immunogenicity and/or antigen presentation. Induction of MHC-I 
and class II expression was observed in all cell lines following transduction (Fig. 14, 15). Although similar 
MHC-I/II induction is observed with in vitro incubation with exogenous y-IFN, rapid decline in MHC-I/II is 
observed following withdrawal of y-IFN (Fig. 16).. 
As opposed to incubation with exogenous y-IFN, MHC-I/II induction persisted over a one-two month period 
(Table V). Representative MHC-I induction in SK-N-AS, SK-N-RA and LAN-6 cells is shown in Table VI 
and Figure 14. Similar high grade induction of MHC-II was observed in SK-N-AS and SK-N-RA y-IFN 
transduced cells (Figure 15). Levels of induction of MHC-I are comparable to those observed during 
incubation with exogenous y-IFN at 100-10(X)U/ml. In addition to persistent MHC-I/II expression, the added 
effects of local y-IFN production with respect to immune activation are not known. In contrast to MHC-I/II, 
we did not observe an increase in expression of the costimulatory ligand B7/BB 1 as determined by flow 
cytometry in transduced SK-N-AS, and SK-NRA cells (data not shown). Whether any of the more recently 
[96] 
Recombinant DNA Research, Volume 19 
