Analysis of a family of cyclin-dependent kinase inhibitors: p15/MTS2/INK4B, p16/MTS1/INK4A, and p18 genes in acute lymphoblastic leukemia of childhood.

A newly recognized family of proteins that inhibit cyclin-dependent kinases (CDKs) termed cyclin-dependent kinase inhibitors (CDKI) have an important role in regulation of cell-cycle progression. A subfamily of these CDKIs (p15INK4B/MTS2, p16INK4/MTS1, and p18) have a high degree of structural and functional homology and are candidate tumor-suppressor genes. We evaluated the mutational status of the p15, p16, and p18 genes in 103 childhood acute lymphoblastic leukemia (ALL) samples and correlated these results with both their clinical data and additional results concerning their loss of heterozygosity in the region of the p15/p16 genes. Homozygous deletions of the p16 gene occurred extremely frequently in T-ALLs (17/22; 77%), and it was also frequent in precursor-B ALLs (12/81; 15%). Homozygous deletions of the p15 gene were also very frequent in T-ALLs (9/22; 41%), and it occurred in 5 of 81 (6%) precursor-B ALL samples. No deletions of p18 was found in any of the 103 ALL samples. Also, no point mutations of the p15, p16, and p18 genes were detected. We correlated p15/p16 alterations at diagnosis with their clinical characteristics as compared with 2,927 other patients treated similarly. Those with p15/p16 alterations were older; had higher white blood cell counts, often with T-cell ALL phenotype; and more frequently had a mediastinal mass at presentation; but they had the same nonremission, relapse, and survival rates at 5 years as did those patients whose blast cells did not have a p15/p16 deletion. To better understand the extent of alterations affecting chromosome 9p21 (location of the p15/p16 genes), loss of heterozygosity (LOH) was examined at D9S171, which is about 1 megabase proximal to the p15/p16 genes. LOH was detected in 15 of 37 (41%) informative samples. Interestingly, of the 24 informative samples that had no detectable alteration of the p15/p16 genes, 7 samples (29%) had LOH at D9S171. In summary, we show in a very large study that p15 and p16, but not p18, CDKI genes are very frequently altered in ALL; those with p15/p16 alterations are more frequently older children, have higher white blood cells at presentation, and often have a T-cell ALL phenotype. The LOH analysis suggests that another tumor-suppressor gene important in ALL also is present on chromosome 9p21.