4.4. ORGAN-SPECIFIC MUTATION ACCUMULATION
Genome instability is considered a major causal factor in cancer and aging (60- 62). Somatic DNA alterations such as point mutations, deletions and rearrangements are thought to activate and/or inactivate specific genes involved in key cellular processes, such as cell cycle regulation and growth control. Age-related accumulation of such mutations in various organs and tissues is likely to be a major factor responsible for the increased incidence of cancer with age. Genome instability can also affect the control of gene expression, for example through mutations in promoter regions or in genes encoding transcription factors, but also as a consequence of large deletions and rearrangement events involving millions of base pairs. Such large structural alterations could gradually impair genome functioning, for example by causing local hemizygosities or influencing DNA looping and bending.
Even over long distances, such changes could impair existing patterns of gene regulation, which might explain the gradual appearance of dysfunctional ("senescent") cells (63). The underlying cause of genome instability is DNA damage, i.e. chemical changes in DNA induced by a variety of spontaneous agents, including oxygen free radicals (64). To preserve genome integrity, all organisms are equipped with an intricate network of DNA damage response pathways (65). These enable cells to cope with the fundamental problems arising from acute and long-term effects of DNA lesions that otherwise would lead to a quick collapse of the genetic machinery. However, genome stability systems such as DNA repair and cell cycle arrest systems are by no means error-free. Errors associated with DNA processing result in mutations, which can range from point mutations to the large deletion and rearrangement events mentioned above.
In contrast to DNA damage, mutations are irreversible, except by cell death. Hence, the mutation load of a given organ or tissue could predispose to cancer and other pathological end points of the aging process. The availability of methods to monitor genomic instability in vivo is important to study the basic mechanisms underlying tumor formation and age-related cellular degeneration and death. Below it will be discussed the development and application of a series of transgenic animal models that allow the study of these questions as a function of the animal's DNA repair status.
4.4.1. INFLUENCE OF DEFECTS IN CELL CYCLE CHECKPOINTS AND DNA REPAIR
Mouse mutants carrying defined defects in genome stability systems are now being generated at an increasing pace. As expected, most of these knockout mice are predisposed to cancer. However, several knockout mice involve genes that in humans not only cause increased cancer predisposition, but also symptoms of accelerated aging, i.e. the segmental progeroid syndromes, such as Werner's syndrome (68), Cockayne's syndrome (69) and ataxia telangiectasia (70). (The fact that heritable mutations in such genes cause accelerated aging supports the notion that genome stability systems are major longevity assurance systems.) In some of these diseases a high level of genomic instability has been demonstrated in peripheral blood lymphocytes of the patients, suggesting that an accelerated accumulation of somatic mutations might be the ultimate cause of the disease (71). It was begun to crossbreed lacZ mouse model for detecting mutations with DNA repair mouse mutants.
Eventually, each of these hybrids will harbor a specific mutation in a genome stabilization pathway as well as the lacZ reporter gene for detecting somatic mutations (detailed plasmid in 86).
126.96.36.199. THE TP53 GENE
A major category of genome stabilization systems involves cell cycle checkpoints, that is, monitoring systems for DNA damage that temporarily halt replication until the lesions are repaired (72).
Mutant frequencies in liver, spleen and tumors of 4-6 months old lacZ control and p53-/-, lacZ hybrid mice
a Mean mutant frequency ±SD.
b Mann-Whitney test.
c Tumors derived from 22- and 33- month-old mice.
Two cell cycle checkpoint genes are understood in some detail: TP53 and ATM (ataxia telangiectasia mutated). The TP53 gene is the tumor suppressor gene most frequently mutated in human tumors. Its gene product has a central role in the Gl checkpoint and is thought to act, at least partly, through the induction of p21, an inhibitor of cyclin dependent kinases and PCNA (proliferating cell nuclear antigen). Various signals of cellular distress, including DNA damage (most notably DNA double-strand breaks), hypoxia and low levels of nucleotide triphosphate pools, use p53 as the trigger to a response (73). A major component of this response clearly is cell cycle arrest. However, induction of p53 can also result in apoptosis, a form of programmed cell death (74). Interestingly, there is evidence that p53 plays a role in replicative senescence of normal mouse and human cells in culture (73).
The ATM gene encodes a protein similar to several yeast and mammalian PIK-related kinases (PIK=phospholipid phosphatidylinositol kinases) involved in cell cycle control (75). Like p53, ATM is part of a DNA damage-sensitive checkpoint, but also plays a role in normally developing or undamaged cells. In addition, ATM acts as an upregulator of p53 and a functional link of ATM with p53 has been discussed (76). Crosses were made between the lacZ-plasmid mouse model and a TP53 knockout mouse (77). Subsequently, several organs and tissues of the homozygous knockouts (TP-/-) as well as of the heterozygous animals (TP53+/-), with the lacZ-plasmid vector also present, were analyzed for mutation frequencies. The TP53-/- animals were sacrificed as late as possible, i.e. between 4 and 6 months, which is about the time they usually die as a consequence of cancer (77).
Based on the involvement of the TP53 product in preventing cells with severely damaged DNA to continue cycling, the expectation was that the TP53-/- animals would have a much higher spontaneous mutation frequency. Surprisingly, this was hardly the case. While both in liver and spleen a significantly higher mutation frequency was observed, the difference with the control animals was less than twofold (Table 4.III). (The heterozygous TP53 knockouts were not different from the lacZ controls.) In some of the tumors, which were also analyzed (mainly lymphomas and sarcomas), significantly higher mutation frequencies than the normally found average (in tissues from young animals) were observed (Table 4.III).
However, the same was true for a lymphoma and a sarcoma obtained from an aged animal (Table 4.III) and can be explained by various factors, including a mutator phenotype. Thus, based on these results, TP53 does not seem to play such a major role in mutation induction as its function as "guardian of the genome" might have suggested. This is in spite of its important role in carcinogenesis. This seeming discrepancy can be explained by assuming that TP53's early role as a cell cycle checkpoint and in apoptosis of cells with severe DNA damage can be taken over by alternative pathways. Increased predisposition to cancer can then be explained by the absence of its gene product at a later stage, i.e. when the tumor cells actually have to be removed. The lack of influence of the TP53 gene product on mutation induction is underscored by the identical amounts of mutations induced by ethyl nitrosourea in its target organ, the spleen, in the TP53-/- and control animals (results not shown).
Studies with another cell cycle checkpoint, i.e. ATM, are currently in progress.
188.8.131.52. THE XPA NUCLEOTIDE EXCISION REPAIR GENE
A major DNA repair mechanism in mammalian cells is nucleotide excision repair (NER). This system entails multiple steps that employ a number of proteins to eliminate a broad spectrum of structurally unrelated lesions such as UV-induced photoproducts, chemical adducts, as well as intra-strand crosslinks and some forms of oxidative damage (78). Deficiency in NER has been shown to be associated with human inheritable disorders, such as xeroderma pigmentosum (XP), Cockayne's syndrome (CS) and trichothiodystrophy (TTD). These disorders are characterized by UV-sensitivity, genomic instability and various signs of premature aging (78). XP patients can be classified into at least seven complementation groups (XPA-XPG). One of the most frequently found groups in human is XPA. Its deficiency causes a complete block in NER and severe symptoms, including a >2000-fold increased frequency of UVB-induced skin cancer and accelerated neurodegeneration (79, 80).
Mutant frequencies in liver and brain of lacZ control and XPA-/-, lac Z hybrid mice
a Mean mutant frequency ±SD.
|2 months||lacZ control||5.7±1.4||4||4.6±1.8||4|
|4-6 months||lacZ control||5.4±2.0||18||4.8±1.6||15|
|9-12 months||lacZ control||9.3±1.6||10||5.4±1.2||9|
|25-34 months||lacZ control||12.2±5.8||19||5.0±1.4||17|
Only liver showed a significant difference between genotype at age groups "4-6" and "9-12"; p=0.0004 and p=0.008, respectively, as compared by Mann-Whitney test.
The XPA-protein, in combination with replication protein A, is proposed to be involved in DNA damage recognition, i.e. the preincision step of NER. Mice deficient in the NER gene XPA have been recently generated by gene targetting in embryonic stem cells (81). These NER-deficient mice were demonstrated to mimic the phenotype of humans with xeroderma pigmentosum, that is, increased sensitivity to UVB and dimethylbenz(a)anthracene (DMBA) induced skin cancer (81). This was found to be associated with an almost complete lack of UV-induced excision repair, measured as unscheduled DNA synthesis in cultured fibroblasts, and a much lower survival of fibroblasts after UV-irradiation or treatment with DMBA (81). However, at early age NER-deficient mice do not show spontaneous abnormalities. These mice develop normally, indistinguishable from wild type mice (81).
At older age, i.e. from about 15 months onwards, XPA-/- mice show an increased frequency of hepatocellular adenomas (82), which is a common pathological lesion in older mice (67). The lack of spontaneous abnormalities in young mice might be due to the fact that under normal conditions mice have only limited exposure to NER-mediated DNA damage. Hence, at early age NER appears to be dispensable. This is in contrast to base excision repair, the inactivation of which is lethal (83).
In order to test the hypothesis that loss of NER causes accelerated mutation accumulation preceding the onset of accelerated tumor formation and aging, XPA-deficient mice were crossed with the lacZ transgenic mice previously used to monitor mutation accumulation in liver and brain during aging (66). In the hybrid XPA-/-, lacZ mice, mutant frequencies were analyzed in liver and brain at early ages: 2 months, 4 months and 9-12 months. The data in Table 4.II indicate the effect of the XPA deficiency on the in vivo mutant frequency as compared to the control value, i.e., the mean mutant frequency of re-analyzed historical controls (66) and XPA+/- animals. (No difference was observed between XPA+/- and XPA+/+ animals; results not shown). In liver of 2-month-old mice, mutant frequencies were still comparable to those for the lacZ control animals. In 4-month-old XPA-/-, lacZ mice, mutant frequencies were significantly increased by a factor of 2.
A further, albeit smaller, increase was observed between 4 and 9-12 months. At the latter age level, mutant frequencies were in the range of the maximum level reported earlier for 25-34-month-old lacZ mice, with a similar high individual variation (Table 4.IV). In brain, mutant frequencies were not found to increase over the age levels studied, which is in keeping with the lack of an age-related increase reported previously for the lacZ mice (Table 4.IV). Mutant spectra were analyzed from liver and brain of 2-month and 9-12-month-old XPA-/-, lacZ mice (data not shown). The results indicated a higher level of genome rearrangements in the older animals, which almost reached the level of the earlier reported values for mice of over 30 months of age (66). The results of this study indicate that a deficiency in the NER gene XPA causes an early-accelerated accumulation of somatic mutations in liver, but not in brain.
The increase in mutant frequency in liver is in keeping with the higher incidence of spontaneous liver tumors several months later (82). Therefore it is tempting to conclude that the higher mutant frequencies in liver predict an organ-specific predisposition to cancer. In the corresponding human syndrome, xeroderma pigmentosum, internal tumor development is rare, but neurological abnormalities are frequently observed (80). Such abnormalities have thus far not been found in XPA-deficient mice. Phenotypical differences between XPA-deficient mice and human XPA might be due to tissue-dependent variation in the levels of endogenous DNA damaging species. This may lead in humans to deleterious effects in brain, whereas in mice it may cause a higher frequency of spontaneous liver tumors. Humans with XP rarely survive beyond the third decade of life, a consequence of the dramatic increase in sunlight-induced skin cancer (80).
Skin cancer does not occur in rodents, which have a fur that cannot be penetrated by UV and are also kept under conditions not permitting exposure to sunlight. This explains the lack of such a phenotype of the XPA mutation at early ages. In the young animals, NER could be essentially redundant, but become increasingly important at later ages. Indeed, it has been repeatedly argued that loss of redundancy in cell number, gene copy number, functional pathways, etc., could be responsible for the gradual loss of individual stability and increased incidence of disease associated with aging (84).