The INK4 family includes the structurally and functionally related p15
, p16 and p18 genes. in vitro they arrest the cell in G1 phase, thus f
ar in an apparently similar manner. It is not clear yet how their func
tions relate in vivo, whether these genes have distinct or redundant f
unctions. The INK4 genes are thought to be candidates for tumor suppre
ssor genes. The p16 gene has been found to be inactivated in many tumo
r types, usually by homozygous deletion. In most tumors it was noted t
hat the deletions involve the neighboring p15 gene as well, thus it wa
s not clear whether these inactivation events targeted p16, p15, or bo
th genes. p16 was also found to be inactivated by intragenic mutations
in 40% of pancreatic carcinomas; these cancers provide a unique oppor
tunity to test whether p16 and p15 must both be inactivated during tum
origenesis. If p15 were a second target at chromosome 9p, it would be
predicted to be inactivated in at least some of these tumors. In addit
ion to studies of the p15 gene in these pancreatic cancers, we assayed
for mutations in the p18 gene, which offered a third, independent sit
e to infer possible redundant functions of the INK4 proteins. Sequence
analysis indicated that p15 and p18 were not targets of inactivation
in pancreatic carcinoma. p16, therefore, plays a different role in viv
o, since inactivation of only this gene had proven to confer a selecti
ve growth advantage to evolving clones of pancreatic tumor cells. The
preference for homozygous deletions as a means to inactivate p16 remai
ns unexplained.