Decline and fall of the tumor suppressor
Epidemiological studies show that cancer is primarily a disease of the aged. Cancer rates increase dramatically in humans beginning in the six and seventh decades of life (1). Although the complex relationship between cancer and aging has long been recognized, a clear understanding of the mechanisms behind this relationship has remained elusive. Among the genes that protect us from early cancers are the tumor-suppressor genes, a category of genes that generally encode negative growth regulators. Yet the frequent cancers in older individuals suggest that the tumor-suppressor genes somehow lose their effectiveness late in life. Are the tumor suppressors in aged individuals actively promoting cancers, are they losing their cancer-prevention function, or are they at the mercy of larger forces that circumvent or eliminate their function? In a recent issue of PNAS, the article by Feng et al. (2) provides compelling evidence that some tumor suppressors remain structurally intact but decline in functional activities during aging.
So why do tumor suppressors become less effective in preventing cancer with age? One dominant theory is that cellular mutation loads increase with age in tissues (3). It has been proposed that the emergence of a cancer cell may be the result of an unlucky accumulation of mutations in a set of cooperating oncogenes and tumor-suppressor genes (4). This proposal suggests that innate tumor-suppressor function does not have to change but its function can be eliminated as a result of mutagenic insults. However, Feng et al. (2) add a new twist by showing that cancers in the aged are not merely the result of accumulated oncogenic mutations but may be partially dependent on intrinsic nonmutational declines in tumor-suppressor function.
The tumor suppressor in question is p53, also known as the “guardian of the genome,” which responds to an array of cellular stresses and mediates cell cycle arrest and DNA repair in damaged cells (5, 6). Depending on cell type, environmental context, and/or degree of stress, p53 may instead facilitate a cell death or apoptosis response in the damaged cell. Alternatively, a permanent cell cycle arrest, or senescent state, may be induced. Regardless of whether the damaged cell is arrested and repaired, killed, or made senescent, the end result is the same: the organism is protected from propagation of mutation-bearing cells that could potentially become cancerous. Among the hundreds of tumor-suppressor genes now identified, p53 may be the most important. The p53 gene is mutated in over half of all human cancers, and it has been estimated that >80% of human cancers have dysfunctional p53 signaling (6, 7).
Feng et al. (2) took a simple but elegant approach to measure p53 activity in different tissues of young and old mice of different inbred strains. Surprisingly, they found that, in virtually all tissues examined, p53 activity was significantly decreased in old tissues compared with young tissues. They did this by γ-irradiation of the mice, followed by harvesting of the tissues 6 h later and examination of the tissues for p53 protein and various markers of p53 function. The p53 protein was stabilized and p53 protein levels were increased as expected in irradiated young mouse spleens. However, spleens from old mice showed significantly less p53 stabilization than their young mouse counterparts. To assess p53 functionality, two major assays were used. First, because p53 is a transcription factor that activates known target genes, RNA levels of six such p53 target genes (e.g., p21, Mdm2) were examined after irradiation. All tissues examined from old mice showed a less robust p53 target induction than that seen in young mouse tissues. In the second assay, apoptotic cell numbers in irradiated young and old spleens were measured, and old spleens exhibited less efficient apoptosis than young spleens, again indicating reduced p53 functionality with age.
One interesting wrinkle noted by Feng et al. (2) was that female inbred mice showed earlier reduction in p53 activities compared with male mice. At 20 months of age (late middle age for a mouse), female tissues displayed major declines in p53 activity, whereas male mice at 20 months retained youthful levels of p53 activity. However, by 28 months, the males had caught up to the females and exhibited greatly reduced p53 activity. Although the basis for this p53-associated sexual dimorphism is unclear, it does correlate with the fact that females of the C57BL/6 inbred strain have a shorter median lifespan than males (www.nia.nih.gov/ResearchInformation/ScientificResources/AgedRodentColoniesHandbook/StrainSurvivalInformation.htm) (Fig. 1). The authors suggest that the age-associated reduction in p53 activity could presage cancers in part by increasing mutation rates and inhibiting elimination of mutated cells. Importantly, Feng et al. showed that the reduced p53 activity is cell autonomous and not attributable to a systemic or hormonal effect, because isolated splenocytes from old mice retain reduced p53 activities in response to various stressors.
Model showing the correlation between cellular p53 and ataxia–telangiectasia mutated (ATM) activity levels and longevity in mice. p53 and ATM protein levels and activity remain high during the first half of the mouse lifespan but then decline significantly in aged mice (red line). Male p53 and ATM levels are indicated by the solid red line, and the dashed red line represents female p53/ATM activity levels, which decline sooner than in males. Survival (blue lines) correlates well with p53/ATM levels, although males (solid blue line) show extended median longevity compared with females (dashed blue line). Cancer rates (green line) increase dramatically in aged mice and begin to appear after the decline in ATM and p53 activity levels.
The mechanism by which p53 activity declines with age remains uncertain, but one important clue comes from the discovery by Feng et al. (2) that stress-activated ataxia–telangiectasia mutated (ATM) protein levels decline with age in parallel with declining p53 activity. Moreover, levels of ATM autophosphorylation in response to radiation (a marker of ATM activation) are also reduced in older mice. This observation is important because ATM is a pivotal DNA damage sensor kinase that phosphorylates p53 and facilitates p53 activation (8, 9). In addition, ATM activates a host of DNA damage response, DNA repair, and cell cycle checkpoint proteins that are critical for the maintenance of genomic DNA integrity (8). The absence of a functional ATM gene is associated with lymphoid tumors in both humans and mice (8, 10). The reduction in ATM and p53 signaling was shown not to be merely a result of a generalized reduction in all signaling pathways because S6 kinase and AKT protein and phosphorylation were not reduced at the same rate as ATM and p53 phosphorylation in aged spleen.
Feng et al. (2) thus show that two proteins critical for DNA damage response, DNA repair, cell cycle control, and apoptosis decline with age (Fig. 1). The age-associated declines in ATM and p53 signaling functions in mice, if applicable to other mammals, provide an important addition to our understanding of how cancer rates increase with age. The importance of p53 in aging and cancer in mice has recently been underlined by several studies showing that additional copies of p53 in the mouse germ line can provide augmented cancer resistance, thereby counteracting the p53 decline seen in nontransgenic mice (11–13). When the additional p53 alleles were normally regulated, no effects on longevity were noted, except that in one exciting study, additional copies of p19ARF (a p53 regulator) and p53 added to the germ line resulted in extended median longevity relative to normal mice (14).
Like all exciting discoveries, the one described by Feng et al. (2) raises more questions than it answers. Chief among
Two proteins critical for DNA damage response,DNA repair, cell cycle control, and apoptosis decline with age.
these is the question of whether ATM and p53 functional decline occurs in aging humans. We have found that spontaneous tumors
in normal mice rarely exhibit p53 mutations (L.A.D., unpublished data). This finding is consistent with a scenario in which
lower age-associated p53 levels produce less selective pressure for mutational inactivation of p53. Yet, in human tumors,
p53 mutations are frequent, raising the question of whether p53 activity is retained at higher levels longer in aging humans.
Oncogenic conversion of human cells is known to require more genetic alterations than conversion in mouse cells (15), so perhaps complete inactivation of p53 rather than mere reduction of p53 activity is a prerequisite for human tumor cell
progression.
Another key question is the mechanism by which p53 and ATM decline may contribute to increased tumorigenesis. It is possible that increasing genomic instability is part of the reason, because studies by Vijg and colleagues (3) have found that many tissues show moderate increases in mutational load with age. However, these mutation rates tend not to accelerate with age as might be expected from the p53/ATM activity profile. One alternative mechanism for cancer induction is the reduced clearance of damaged and defective cells in aged organisms. Reduced p53/ATM levels could result in accumulation of senescent and dysfunctional cells. In particular, senescent cells could actively promote tumor cell formation through abnormal secretion of growth factors and proteases that stimulate adjacent cells to proliferate, as shown in elegant experiments by Campisi and colleagues (16). This scenario could be coupled with increased cellular proliferation and a lower threshold for oncogenic conversion in cells with reduced p53 activity.
The study by Feng et al. (2) provides intriguing possibilities for our understanding of the aging/cancer connection, but we caution that the findings are correlative and confined to one species. Nevertheless, they provide a framework for reconsideration of our current models and the impetus for designing further experiments to resolve many remaining questions.
Footnotes
- *To whom correspondence should be addressed. E-mail: larryd{at}bcm.tmc.edu
-
Author contributions: G.H. and L.A.D. wrote the paper.
-
The authors declare no conflict of interest.
-
See companion article on page 16633 in issue 42 of volume 104.
- © 2007 by The National Academy of Sciences of the USA
