p53 Oligomerization Domain Mutants: A New Class of Mutants That Retain “License to Kill”

Fifty percent of human tumors contain inactivating mutations in TP53, and the overwhelming majority of these mutations are missense mutations in the DNA binding domain of this transcription factor that eliminate the transcriptional potential of this protein. In addition, thousands of families in the world possess germline mutations in p53 and develop Li–Fraumeni syndrome (LFS), which is characterized by tumors of the brain, breast, bone, and adrenal cortex.

In unstimulated cells, p53 is maintained at very low levels, but in response to DNA damage and other stress stimuli, p53 becomes stabilized and activated as a transcription factor. Activated p53 exerts its tumor suppression through a combination of pathways, including the transcriptional pathway, wherein p53 binds to a four-pentanucleotide repeat consensus DNA binding site, and induces expression of CDKN1A, which is required for growth arrest and senescence induced by p53, and BBC3 (PUMA), PMAIP1 (NOXA) and BAX, which play a role in the induction of apoptosis; much of the transcription-dependent activity of p53 occurs in a tissue-specific manner (see ref. 1 for review). Finally, p53 also possesses transcription-independent tumor suppressor function in DNA repair and apoptosis; the latter is mediated by virtue of direct interaction with BCL2 family members (see ref. 2 for review).

p53 binds to its consensus DNA-binding site as a tetramer. Although missense mutations in the oligomerization domain of p53 account for 20% of LFS and 2% of somatic mutations, the impact of these mutants on cancer and how they may differ from DNA binding domain mutants are important areas that have long been underinvestigated. Perhaps the most famous oligomerization mutant of p53 is R337H. This mutation is present in nearly 1/300 individuals in Brazil, and previous epidemiologic and clinical studies revealed that the high rate of the R337H mutation in southern and southeastern Brazil predisposes carriers to adrenocortical tumors and other tumors, including breast carcinoma, soft-tissue sarcoma, osteosarcoma, choroid plexus carcinoma, and thyroid and lung cancers. It is of note that we now know that the severity of this mutant is exacerbated in carriers by the existence of a linked stop allele of XAF1; because XAF1 functions in a feed-forward loop with p53 to enhance its transcriptional activity, the impact of the R337H mutation alone on human cancer may be overestimated (see ref. 3 for review). In support of this, mouse models show an extremely modest impact of the R337H mutant on spontaneous cancer (4, 5). In this issue of Cancer Discovery, Choe and colleagues from the Prives group and Gencel-Augusto and colleagues from the Lozano group delve into the oligomerization domain mutant A347D (6, 7). The A347D, or AD mutant, is among the most frequent oligomerization domain mutants and proves to be significantly more consequential to cancer risk.

Choe and colleagues perform their studies in a diverse set of endogenous contexts that include AD patient samples, mesenchymal stem cells, and CRISPR-engineered knockin mutants in the U2OS osteosarcoma line (6). In an equally rigorous accompanying article, Gencel-Augusto and colleagues report on the phenotype of mice containing the AD mutant, which predominantly forms dimers, as well as the R342P mutant, which exclusively forms monomeric p53 (7). The results of both studies are complementary, compelling, and also quite surprising. First and foremost, both groups show that the AD mutant is hyperstable, constitutively dimeric, and completely defective in the stress-induced transactivation of canonical p53 target genes. Both groups also show that cells containing the AD mutant show markedly enhanced oxidative phosphorylation. Given the overwhelming defect in the transcriptional function of the AD mutant, it is surprising that Gencel-Augusto and colleagues show that AD mice show less cancer than p53 knockout mice, supporting the premise that this mutant retains tumor suppressor function. This group also shows that the cancer incidence of mice with monomeric p53 (R342P) is indistinguishable from p53 knockout mice, so tumor suppression in AD mice appears to require dimer formation.

In a fascinating series of experiments, both groups show that the mechanistic basis for tumor suppression by the AD mutant appears to lie in “neomorphic” activities of this mutant. Specifically, Choe and colleagues find that the AD dimer is capable of activating genes enriched in binding sites for ETS transcription factors. In different cell types in their model, Gencel-Augusto and colleagues find evidence that the AD dimer interacts with a noncanonical p53 binding site, containing two instead of four copies of the pentameric p53 binding site. They show that these dimer sites are enriched in genes coregulated by PPARα and PPARγ, leading to marked increases in PPAR signaling in AD cells. Notably, the findings of both groups point to potential new therapies for tumor cells containing the AD mutant. Choe and colleagues find that the direct, mitochondrial pathway of cell death by p53, which involves an interaction of this protein with BCL2 family members, is retained in tumor cells containing the AD mutant. This discovery is notable given the findings of others that the prevalent form of p53 at the mitochondria is a dimer (8) and that the oligomerization domain of p53 is required for its ability to oligomerize BAK at the mitochondria (9). Data from Choe and colleagues indicate that topoisomerase inhibitors such as etoposide will show efficacy for tumors containing the AD mutant. In a similar vein, Gencel-Augusto and colleagues show that the increased PPAR activity in AD mutant cells is lost in tumors, supporting the premise that it is tumor suppressive. These authors propose that PPAR agonists will be efficacious against AD tumors. Thus, the findings described by both groups point to neomorphic activities of mutant p53 as well as new, personalized therapeutic opportunities (Fig. 1).

Interesting questions remain. First, the precise molecular basis for the cooperation of the p53 AD dimer with PPAR proteins remains to be determined; that said, the findings of both groups that the AD mutant interacts with PPAR proteins suggest that cooperation between these proteins on chromatin seems likely. Second, how the AD dimer enhances transcription from ETS binding sites remains to be precisely determined. Third, it remains to be determined whether the increased PPAR activity underlies the increased mitochondrial metabolism in AD cells, particularly as Choe and colleagues reported increased mitochondrial metabolism but no apparent increase in PPAR activity. Fourth, whether some of the properties of AD mutants are likewise shared with more conventional DNA binding domain mutants of p53 is an interesting next question. For example, there are some interesting parallels between the AD mutants and germline DNA binding domain mutations in TP53 (LFS): The Hwang group previously showed that human LFS patients maintain increased oxidative metabolism in comparison with noncarriers through increased mitochondrial respiration (10). A similar increase in oxidative phosphorylation and even increased fitness were demonstrated by Gencel-Augusto and colleagues using comprehensive laboratory animal monitoring (CLAMS). Finally, in a fascinating extension of their findings, Gencel-Augusto and colleagues show evidence that basal, nonstressed p53, which like the AD mutant is predominantly a dimer, may also be associated with a hyperactive PPAR pathway. Therefore, the studies on the AD mutant may reveal functions held normally by the wild-type p53 dimer. This finding could offer new insight into differences in the tumor-suppressive activity between unstressed and “activated” p53 proteins.