Aging and p53: getting it straight A commentary on a recent paper by Gentry and Venkatachalam
Year: 2005
Type: letter
Abstract: Mice harboring the p53 ‘m’ allele – the result of a 0.6-Mb deletion affecting the p53 N-terminus as well as its upstream region – alongside a normal copy of the p53 gene, show decreased lifespan, premature aging and reduced cancer incidence. In the June issue of Aging Cell, Gentry and Venkatachalam provide a characterization of the entire m region, identifying 23 genes located upstream of p53. These authors propose that at least part if not all of the phenotype associated with the presence of the ‘m’ allele is caused by haploinsufficiency of one or more of the deleted genes and not by a truncated p53. We disagree with their interpretation and believe that the truncated p53 protein is the major, and likely the sole contributor, to both early aging and low cancer incidence in this mouse mutant. Tyner et al. (2002) previously described mice with a shortened lifespan, early onset of a number of normal aging symptoms and low cancer incidence. These mice were mutated for p53 by conventional gene targeting in embryonic stem cells. However, the mutation was anything but conventional. Somehow a large region of DNA was deleted such that the N-terminus of p53 was removed along with an undetermined amount of DNA upstream of the p53 gene. The mutant allele was called the ‘m’ allele and is used in reference to p53. However, much more than p53 is affected by the ‘m’ allele. Homozygosity of this mutation results in embryonic lethality, so that the premature aging phenotype was only observed in heterozygotes, called p53+/m. Tyner et al. suggested that accelerated aging in these mice is due to the expression of a truncated p53 protein. They found that mRNA specifically encoding the ‘m’ allele is expressed in almost all tissues of the p53+/m mouse and translated into a p24 protein. However, the protein is hard to detect in vivo and it is unknown if other genes within the large deletion contribute to the aging phenotype. Thus, one could question if it is the truncated p53 protein that causes accelerated aging in the mouse or rather the haploinsufficiency of any or all of the genes residing in the genomic region affected by the deletion. In the June issue of Aging Cell, Gentry and Venkatachalam present their results that define the genetic alteration of the ‘m’ allele and show an approximate 0.6-Mb deletion that contains at least 23 genes located upstream of p53 (Gentry & Venkatachalam, 2005). These authors propose that at least part if not all of the aging phenotype is caused by haploinsufficiency of one or more of the deleted genes and not by a truncated p53. We congratulate Gentry and Venkatachalam for characterizing the deleted region in the ‘m’ allele, which is important. However, we believe that their interpretation of the potential phenotypic consequences of the deletion confuse rather than clarify the issue of p53 as a putative pro-aging factor. We believe there are at least three problems with their interpretation in which they discount a truncated p53 as the culprit of the premature aging phenotype and the reduced cancer incidence in these mice. First, they fail to fully consider the ‘m’ allele in a genetic background deleted for wild-type p53 or adequately compare its effects with mice that overexpress full-length, normally regulated p53 or a natural variant of p53, i.e. p44. Second, they do not adequately compare the p53+/m mice with other mutant mice that exhibit premature aging, possibly also involving p53 activity. Third, they do not provide direct evidence that haploinsufficiency of genes in the deleted region contribute to the premature aging phenotype. After all is considered, we believe that the truncated p53 protein is the major, and likely the sole contributor to both early aging and low cancer incidence, as we explain in more detail below. Gentry and Venkatachalam believe that the apparent embryonic lethality of the p53m/m mice – in striking contrast to the p53–/– mice, which are viable – underscores the importance of the upstream deletion in causing the reduced cancer and premature aging phenotypes. This comes down to the question of whether the ‘m’ allele produces a dominant, variant form of p53 or is actually a null allele. There is some debate as to whether the ‘m’ allele actually produces truncated p53 since it is difficult to detect. However, a comparison of the phenotypes of mice with a null mutation in p53 to those with the ‘m’ allele shows that it is very likely that the ‘m’ allele produces a truncated p53 affecting p53 function. Mice that are homozygous null for p53 (p53null/null) exhibit a high incidence of early onset of cancer and live about 6–9 months (Donehower et al., 1992). Mice that are heterozygous null for p53 (p53+/null) also develop tumors more rapidly than the homozygous normal mice (median tumor incidence of 18 months), resulting in their premature death at around 2 years (Venkatachalam et al., 1998). From the comparison of p53+/m with p53+/null mice we can deduce that the ‘m’ allele affects p53 since the p53+/m mice do not show increased cancer as compared to their p53+/null counterparts. On the contrary, p53+/m mice are profoundly cancer resistant. It is therefore likely that the truncated version of p53 in the ‘m’ allele is expressed and interacts with full-length p53 to exert its effect, unless one would assume that haploinsufficiency of any of the genes deleted in the ‘m’ allele overrides the pro-cancer effect of haploinsufficiency of p53 itself. A comparison of the ‘m’ allele in the presence or absence of a full-length p53 allele is important to understand the potential phenotypic consequences of the truncated p53 gene and haploinsufficiency of genes deleted in the ‘m’ allele. The p53null/null mice do not exhibit any accelerated aging phenotype other than cancer. Interestingly, while we know that the ‘m’ allele exerts an effect, either through p53 or any of the 23 deleted genes, p53null/m mice exhibit the exact same phenotype as p53null/null mice (Tyner et al., 2002). There are two conclusions that can be drawn from this comparison. First, if truncated p53 is the culprit, then it does not work alone but requires full-length p53 to cause premature aging (without full length p53 the ‘m’ allele appears to be null for p53 activity). Since p53 functions as a tetramer, a possible mode of action for the truncated p53 protein is to function as part of this tetramer, together with full-length p53, thereby modifying its function to favor early aging and low cancer incidence. Second, haploinsufficiency of the genes deleted in the m region has apparently no phenotypic consequence in the p53null/m mice, at least for the duration of their 6- to 9-month lifespan. Hence, if haploinsufficiency of any of these deleted genes would be responsible for the increased cancer resistance in the p53+/m mice, why would it do nothing to alleviate the cancer-prone phenotype of the p53null/m mice? Indeed, these results indicate that cancer incidence is a p53-dependent trait, which is independent of the genes deleted in the ‘m’ allele. Again, this comparison argues in favor of a truncated p53 protein that functions in association with full-length p53. Finally, Gentry and Venkatachalam argue that p53 activity is unlikely to play a role in the premature aging phenotype of p53+/m mice since overexpression of full-length p53 from a bacterial artificial chromosome (BAC) clone fails to induce aging but succeeds in reducing cancer (Garcia-Cao et al., 2002). However, in these ‘super’ p53 mice, p53 activity is not constitutively increased but under normal regulatory control. Indeed, the basal levels of p53 in these mice are not affected. This strengthens our conclusion from the above that it is the combination of truncated with full-length p53 that results in premature aging. Indeed, there is no reason to expect that increased p53 gene dose under normal regulatory control would lead to premature aging, although it clearly can lead to reduced cancer (since more p53 protein can be mobilized when it is needed). At this point, one must consider the p44 transgenic mice that overexpress a natural, truncated version of p53 (Maier et al., 2004). Gentry and Venkatachalam argue that overexpression of p44 does not result in an aging phenotype with reduced cancer incidence, which would contradict a similar role as the ‘m’ allele in p53+/m mice. We believe that their interpretation of the p44 transgenic phenotype is not correct. Indeed, in their paper, Maier et al. (2004) explicitly state that these animals have a very low incidence of cancer, a result similar to the situation in the p53+/m mice. This is ascribed to impairment of normal cell proliferation, which is in keeping with the proliferation suppressive capability and lack of immortalization potential of p44 noted as early as 1987 by Rovinski and Benchimol (1988). Decreased tumor incidence in the p44 mice may be less obvious in view of its much shorter lifespan, it is nevertheless present. Gentry and Venkatachalam also argue that the p53+/m mice and the p44 transgenic mice do not exhibit similar phenotypes, with the p44 transgenic mice showing a much shorter lifespan and more severe symptoms. However, a much higher level of expression of the p44 allele as compared to the ‘m’ allele could easily explain this discrepancy. Indeed, the ratio of p44 to p53 may be critical for eliciting the premature aging phenotype. Meanwhile, apart from reduced longevity, the p44 mice show early lordokyphosis, osteoporosis and testicular atrophy. Hence, there are enough similarities between the phenotypes of the p44 transgenic mice and the p53+/m mice to conclude that both models display increased aging and reduced cancer. A major argument in favor of a role of constitutively activated p53 as a pro-aging factor is its likely involvement as a culprit in several other mouse models displaying defects in genome maintenance in combination with premature aging. Gentry and Venkatachalam refer to two such models, i.e. the Ku80 knockout (Vogel et al., 1999) and the Brca1 hypomorph (Cao et al., 2003). It is not clear as to why the authors believe that the DNA repair phenotype in the Ku80 or Brca1 mutant mice ‘confound a comparative analysis of these models with the p53+/m mouse’. This section of their paper is not clearly worded. We believe Gentry and Venkatachalam argue that mutations in p53 along with mutations in either Ku80 or Brca1 cause aging. Whether or not this is what they mean to say, it is not correct. On the contrary, deletion of p53 in the Brca1 and Ku80 mutant mice has been reported to alleviate symptoms of accelerated aging, at least at the level of replicative senescence. It is true that the aging phenotype in Brca1-mutant mice is only observed in a p53+/– background, but that is due to the fact that this Brca1 mutation is embryonically lethal in a p53+/+ background. The fact that these mice can be rescued by diminishing p53 dose indicates a causal role of p53 responses in the lethality of the Brca1 mutation. Indeed, according to the interpretation by Cao et al. (2003) themselves, it is the activation of p53 that causes the premature aging phenotype. In the case of the Brca1 mutant (and probably the Ku80 null mice) p53 activation is caused by chronic DNA damage as a consequence of inefficient DNA repair. This situation is very similar in other premature aging mutants in the mouse. For example, the Terc–/–/Atm–/– or Terc–/–/Wrn–/– mouse mutants display an array of premature aging phenotypes, most likely as a consequence of increased genomic instability eliciting p53 responses including apoptosis (Wong et al., 2003; Chang et al., 2004; Vijg & Hasty, 2006). The final weakness in the claim that a 50% reduction of gene dose in the deleted region in the ‘m’ allele contributes to the anticancer and pro-aging phenotype of the p53+/m mice is that it is not supported by direct evidence. To obtain such direct evidence it would be necessary to study the effect of the deletion in the presence of a fully intact p53 gene. If the hypothesis proposed by Gentry and Venkatachalam will be received favorably by the scientific community, such work will undoubtedly be initiated. Meanwhile, we believe that in light of the various comparisons described above activation of p53 is the most likely explanation for the premature aging symptoms which so often accompany diminished or defective DNA repair. As schematically depicted in Fig. 1, the ‘m’ allele or p44 stabilize p53 in the absence of increased genotoxic stress, similar to the continuous stabilization of p53 by ongoing oxidative stress as a consequence of certain defects in genome maintenance. This involves a gain-of-function rather than a loss-of-function as would be the expected result of haploinsufficiency of a gene deleted in m. Such continuous activation should be distinguished from the situation in the super p53 mice, in which p53 is more rapidly stabilized only when needed. Our next focus should be to further unravel the details of the p53 responses, including the role of p44, why they seem to accelerate only a part, but not all normal aging symptoms, if similar mechanisms can also be the cause of normal aging and, finally, if this novel insight will offer us new targets for prevention or treatment of aging-related diseases. Schematic representation of the putative role of p53 as a pro-aging factor. Increased copy number of the p53 gene in the context of its normal regulatory sequences (as part of a BAC clone) causes its exacerbated response (green), but not constitutive activation. Constitutive activation of p53 (red) can be accomplished through chronic DNA damage (as in various mouse mutants harboring genetic defects in genome maintenance), or through the m or p44 allele in the context of the normal p53 gene product. Note that in the p44 transgenic mice the ratio of p53 vs. its variant is altered in favor of the latter. We are grateful to Dr Larry Donehower for critically reading the manuscript, an anonymous reviewer for useful comments, and John David Garza for preparing the figure. The authors’ research is supported by grants from the NIA, NCI and NIEHS.
Source: Aging Cell
Authors: Jan Vijg, Paul Hasty
Language: English
Subfield: Oncology
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
Funder: National Institute on Aging
OpenAlex ID: W2030228737