Multiple mechanisms of p16INK4A inactivation in non-small cell lung cancer cell lines

p16INK4A, a specific inhibitor of cyclin-dependent kinase (cdk)4 and cdk6, is a candidate tumor suppressor in malignancies with wild-type retinoblastoma (Rb). Loss of p16INK4A frees these cdks from inhibition, permitting constitutive phosphorylation of Rb and inactivation of its growth suppressive properties. Consistent with this model, Rb-positive non-small cell lung cancers (NSCLCs) have little or no detectable p16INK4A protein, whereas Rb-negative lung cancers have abundant p16INK4A. However, only some NSCLCs have homozygous deletions or nonsense mutations in a remaining p16INK4A allele, suggesting that other mechanisms must account for absent or low levels of p16INK4A protein. Here, we analyzed 9 Rb-positive NSCLC cell lines for the controls governing p16INK4A activity. Four lines had homozygous deletions of p16INK4A (SK-LU-1, SK-MES-1, A-427, and SW900), and three had a point mutation in a single allele. First, in H520 cells, the previously reported deletion at codon 45 results in a frameshift that produces no detectable protein. Second, in Calu-3 cells, a His to Tyr substitution at codon 83 produced a variant with a shortened half-life that was unable to form complexes with cdk4 or cdk6. Third, in H661 cells, the previously reported point mutation in the second intron splice donor site resulted in a smaller p16INK4A protein. Although this variant formed complexes with cdk4 and cdk6, it had a profoundly reduced half-life, producing low steady-state levels of p16INK4A and abundant levels of free cdks. Finally, Calu-1 and Calu-6 cells transcribed no detectable mRNA encoding authentic p16INK4A. These cell lines displayed methylation of the CpG island surrounding the first exon of p16INK4A and expressed abundant levels of a nontranslated mRNA containing an alternative first exon (E1 beta), as did all other cell lines in which the p16INK4A locus was not deleted. These data indicate that Rb-positive NSCLC cells have evolved a variety of pathways to suppress p16INK4A expression. Reintroduction of p16INK4A into these cell lines by retroviral transfer resulted in a reduced growth rate, increased abundance of hypophosphorylated Rb, accumulation of cells in G1, and a less transformed morphology in Rb-positive, but not Rb-negative cells, suggesting that loss of p16INK4A is essential for maintenance of the transformed phenotype.