The rate of homozygous deletions of CDKN2A/p16 is variable between differen
t tumor entities, and in addition it is higher in established cell lines in
comparison with primary tumors. Such incongruencies may reflect statistica
l sampling errors, true differences depending on tissue derivatisation and
CDKN2A/p16 loss under selective pressure in tissue culture. Clarification o
f these issues is warranted in the context of defining tumor suppressor gen
es such as CDKN2A/p16 as targets for gene replacement therapies. We therefo
re compared established cell lines derived from human glioblastomas and the
ir corresponding primary tumors by multiplex PCR methodology. Archival earl
y passages were included to determine the time point at which the p16 statu
s of a cell line changes if it is different from the original tumor. It was
found that in 2 of 11 cases (18%) the primary tumor had no p16 alteration
whereas the corresponding cell lines had a homozygous p16 deletion. Trackin
g the in vitro evolution of these two cell lines we found that CDKN2A/p16 w
as lost already in the earliest passages. This suggests a clonal outgrowth
advantage of a subpopulation of p16 deleted tumor cells rather than instabi
lity of the CDKN2A/p16 genotype in vitro. Including 20 additional glioblast
oma-derived cell lines we detected that in 19 of the total 31 lines at leas
t one exon was lost bringing the rate of p16 loss in the whole panel to 61%
. This compares to a rate of 49% which was found in original glioma tissue
from 47 unselected other patients. It is concluded, that in cell culture se
lective pressure favours the outgrowth of pre-existing CDKN2A/p16 negative
clones, which account for the difference of CDKN2A/pl6 status between cell
lines and tumors. In no case did we see a change of the CDKN2A/p16 status d
uring prolonged tissue culture periods of up to 8 years.