MDM2 and p53: a question of balance.

Since the initial discovery of mutations in the p53 (also known as TP53) gene in colorectal carcinoma, notable progress has been made in elucidating the role of p53 in the regulation of gene expression, cycle progression, and apoptosis (7). It is now recognized that p53 is induced in response to some types of DNA damage and that it is a key component of the physiologic checkpoint pathway that governs the transition of cells from G[ to S phase in the cell cycle. In the absence of normal p53 function, cells may progress through the cell cycle despite genomic injury, producing daughter cells with ever increasing levels of genetic aberrations. In model systems, mutations in the p53 gene are, in fact, associated with an increased rate of aneuploidy and gene amplification (2). Biochemically, p53 exerts its regulatory effects as a transcription factor via binding to a specific recognition sequence in target genes. An unexpected connection between p53 and a previously described protein was made when immunoprecipitation studies (3,4) demonstrated that p53 interacts with the MDM2 oncoprotein. MDM2 (murine double minute-2) was originally identified as an amplified gene carried on double minutes in a transformed murine cell model (5). Overexpression of MDM2 was shown to produce a transformed phenotype, but its biochemical function remained a mystery until it was recognized that the 90-kd protein that coprecipitated with p53 was identical with MDM2. The p53/MDM2 interaction negatively regulates the transcriptional activating function of p53. Thus, overexpression of MDM2 results in an effect similar to mutational inactivation of p53. This finding has led to the proposition that MDM2 amplification may consititute an alternative pathway to mutational inactivation of p53 (3). This model would provide at least a partial solution to the puzzle of why p53 mutations only occur in a subset of otherwise apparently similar tumors and predicts that tumors with MDM2 amplification should lack p53 mutation and vice versa. The connection between these genes runs to another level, since p53 can induce expression of MDM2 via a p53 binding site in the MDM2 gene, suggesting the existence of an autoregulatory feedback loop in p53 function (6,7). The precise balance between the free and complexed forms of these proteins is probably critical in regulating p53 function. Complementing these functional studies and strengthening the argument that MDM2 has a critical role in regulating p53, descriptive studies (3,8-10) have documented the amplification and overexpression of the MDM2 gene in human sarcomas and glial tumors. In addition, MDM2 amplification occurred only in the absence of p53 mutation, supporting the concept that amplification and p53 mutation are alternative mechanisms of p53 dysfunction (11). These observations are now extended by three studies reported in this issue of the Journal. Habuchi et al. (12) and Lianes et al. (13) add urothelial tumors to the list of cancers that carry MDM2 amplification and overexpression, while Florenes et al. (14) further associate MDM2 amplification and expression with p53 status in sarcomas. Amplification is, as expected, associated with high levels of MDM2 expression, and this association may be expected to result in altered p53 function in tumors bearing MDM2 amplification. Nothing in cancer biology, including MDM2 amplification, is ever quite as simple as it first appears, and the genetic complexity of MDM2 amplification deserves comment. MDM2 maps to chromosome 12ql4.3-15 (15), and the amplification unit is frequently quite large, including numerous flanking genes. Genes located centromeric to MDM2 that are variably coamplified with it include SAS (76), CDK4 (17), CHOP (18), and GLI (79) as well as other less fully characterized transcripts (20). Each of these genes can be plausibly related to the regulation of cell growth. GLI has been demonstrated to have transforming activity. CHOP, normally a growth arrest-associated transcription factor, is the target of the tumor-specific t(12; 16) in myxoid liposarcoma in which it forms a fusion protein with the FUS/TLS gene on chromosome 16 (21). SAS belongs to a large family of integral membrane proteins that may be related to signal transduction. CDK4 encodes a cyclin kinase with specificity for D cyclins. MDM2, SAS, and CDK4 are most frequently included in the amplification unit. A substantial proportion of sarcomas with amplification of chromosome 12 sequences have both CDK4 and MDM2 amplification, while only one of these genes is amplified in other tumors. This finding is of some importance in interpreting descriptive studies of tumors and is particularly interesting because of the role of CDK4 in cell cycle progression. CDK4 is known to be inhibited by WAF1/CIP1, a gene that was identified as a target of p53 induction (22). Recently, mutational inactivation of another CDK4 inhibitor, pi6, has been observed in several cancer cell