Regulation of p16^CDKN2 Expression and Its Implications for Cell Immortalization and Senescence

Abstract

p16CDKN2 specifically binds to and inhibits the cyclin-dependent kinases CDK4 and CDK6, which function as regulators of cell cycle progression in G 1 by contributing to the phosphorylation of the retinoblastoma protein (pRB). Human cell lines lacking functional pRB contain high levels of p16 RNA and protein, suggesting a negative feedback loop by which pRB might regulate p16 expression in late G 1 . By a combination of nuclear run-on assays and promoter analyses in human fibroblasts expressing a temperature-sensitive simian virus 40 T antigen, we show that p16 transcription is affected by the status of pRB and define a region in the p16 promoter that is required for this response. However, the effect is not sufficient to account for the differences in p16 RNA levels between pRB-positive and -negative cells. Moreover, p16 RNA is extremely stable, and the levels do not change appreciably during the cell cycle. Primary human fibroblasts express very low levels of p16, but the RNA and protein accumulate in late-passage, senescent cells. The apparent overexpression of p16 in pRB-negative cell lines is therefore caused by at least two factors: loss of repression by pRB and an increase in the number of population doublings.CDKN2, also referred to as INK4A, CDK4I, and MTS1 (23,36,49), is a putative tumor suppressor gene on human chromosome 9p21 that is genetically linked with familial inheritance of malignant melanoma and is inactivated by a variety of mechanisms in a broad spectrum of human cancers (reviewed in references 22 and 51). CDKN2 encodes a 156-amino-acid protein, designated p16, that specifically binds to and inactivates the cyclin-dependent kinases (CDKs) CDK4 and CDK6 (16,19,41,49). These kinases are the major catalytic partners for cyclins D1, D2, and D3 and collaborate with cyclin E-CDK2 in controlling the G 1 /S transition in mammalian cells (reviewed in references 9, 43, and 53).One of the critical substrates of the G 1 -specific cyclin-dependent kinases is the product of the retinoblastoma gene (pRB) which, in its hypophosphorylated state, exerts a negative influence on G 1 progression (42,63). Since phosphorylation of pRB equates with its functional inactivation, a relatively robust model has emerged in which the role of the cyclin D-dependent kinases is to initiate the phosphorylation of pRB (9,52,63). Consistent with such a scheme, antibody neutralization studies have indicated that D-type cyclin function is required up to a point in G 1 coincident with the block imposed by pRB (3,44). This requirement is lost in cells lacking functional pRB (29,30). Ectopic expression of p16, which specifically interferes with the cyclin D-dependent kinases CDK4 and CDK6, imposes an analogous G 1 block that is also dependent on functional pRB (14,24,31,34,48). Conversely, ectopic expression of D cyclins can accelerate G 1 progression, particularly when coupled with elevated expression of cyclin E (2,20,35,(44)(45)(46). The model can readily explain the role of p16 as a tumor suppressor, since loss of function will...

Tissue-specific alternative splicing in the human INK4a/ARF cell cycle regulatory locus

“…Probing of the Northern blot with a beta-actin cDNA probe indicated that mRNA was of good quality and equally loaded ( Figure 3a). Probing the same blot with an oligonucleotide probe speci®c for the p12 alternative transcript (derived entirely from intron 1 sequence) revealed a single band of the expected size for initiation at the known INK4a promoter only in pancreas (Figure 3b) (Hara et al, 1996;Serrano et al, 1993). The blot was then stripped and re-probed with an exon 3-speci®c oligonucleotide probe capable of detecting ARF, INK4a, and p12 simultaneously.…”

“…We transfected the p12 expression vector into the PANC-1 cell line which is deleted for the INK4a/ARF locus but is wild-type for pRb and into the C-33A cell line which retains expression of INK4a but has lost pRb (Hara et al, 1996), and placed the cells in G418 selection for 12 days after which time the number of viable cells was determined. Figure 6 shows that expression of p12 suppressed growth by nearly 40 and 60% in C-33A and PANC-1 cells, respectively, relative to empty expression vector (pcDNA3.1(+)).…”

“…All numbering of CAT constructs is relative to the transcription start site for ARF (Mao et al, 1995) and the 5'-most start site for INK4a (Hara et al, 1996). Transfection and CAT assays were performed essentially as described (Robertson and Jones, 1998).…”

“…We therefore determined if the p12 cDNA could code for the Figure 5a shows that the antibody was speci®c for INK4a expressed from the endogenous locus in C-33A cells (a high level INK4a-expressing cervical cancer cell line (Hara et al, 1996)), transfected with empty expression vector (pcDNA3.1, right panel) and from an INK4a expression vector transfected into the INK4a-/ARF-pancreatic cell line PANC-1 (p16senseNeo, left panel). PANC-1 and C-33A cells transfected with the p12 expression vector (p12senseNeo) produced a protein of approximately 12 kD reactive with the INK4a antibody.…”

“…Detection of INK4a was Figure 6 Growth suppression assay. PANC-1 cells (wild-type pRb , dark bars) and C-33A cells (mutant pRb (Hara et al, 1996), light bars) were transfected in triplicate with the indicated expression vectors and placed in G418 selective media for 12 days. The number of viable cells was counted and expressed relative to the empty expression vector control (pcDNA3.1, set at 100%).…”