Induced genomic instability in irradiated germ cells and in the offspring; reconciling discrepancies among the human and animal studies


Many studies confirmed that radiation induces genomic instability in whole-body systems. However, the results of the studies are not always consistent with each other. Attempts are made in the present review to resolve the discrepancies. Many of the studies in human and experimental animals utilize the length change mutation of minisatellite sequences as a marker of genomic instability. Minisatellite sequences frequently change their length, and the data obtained by conventional Southern blotting give rather qualitative information, which is sometimes difficult to scrutinize quantitatively. This is the problem inevitably associated with the study of minisatellite mutations and the source of some conflicts among studies in humans and mice. Radiation induction of genomic instability has also been assessed in whole-body experimental systems, using other markers such as the mouse pink-eyed unstable allele and the specific pigmentation loci of medaka fish (Oryzias latipes). Even though there are some contradictions, all these studies have demonstrated that genomic instability is induced in the germ cells of irradiated parents, especially of males, and in offspring born to them. Among these, transmission of genomic instability to the second generation of irradiated parents is limited to the mouse minisatellite system, and awaits further clarification in other experimental systems.

Genomic Instability And Its Biological Significance In Whole-Body Systems

Genomic integrity is crucial for the survival of the cells, and three mechanisms have evolved for the maintenance of the genome. DNA repair operates on damaged genome to restore the original sequence integrity. Cell cycle checkpoint assures the efficient elimination of errors during DNA replication and chromosome segregation. Apoptosis is the most effective, but the ultimate, way to eliminate potentially deleterious cells carrying DNA damage. Genomic information is transmitted faithfully under the supervision of these three mechanisms, and a failure of any of them leads to devastating results such as carcinogenesis in the affected generation and hereditary problems in the subsequent generation. The fidelity, however, is costly, and the mechanism of damage tolerance allows damaged cells to survive and proliferate at the cost of possible mutation (Radman, 2001). A group of translesional DNA polymerases is an example of such mechanism, in which DNA damage is bypassed without being repaired (Boudsocq et al., 2002). Downregulation of fidelity is not restricted to DNA replication. Precision of cell cycle checkpoint is downregulated in an adaptation/override phenomenon in yeast, where cells escape G2/M block in the presence of unrepaired DNA double-strand breaks (Leroy et al., 2003). Induced genomic instability can be viewed as the mechanism of damage tolerance in which the fidelity of genomic maintenance is downregulated persistently in irradiated cells.

Radiation is recognized to induce genomic instability in a variety of experimental systems (Morgan, 2003a,Morgan, 2003b). Mutagenesis associated with induced genomic instability has two important features: untargeted mutation and delayed mutation. In the former, mutation is induced in a DNA region even without damage when the fidelity of the genomic maintenance mechanism is downregulated (Friedberg et al., 1995). In the latter, mutation is induced in a delayed fashion in the descendants of irradiated cells. These two features are not shared by the classic mutagenic pathway in which mutation occurs in a targeted manner within a certain period of time after the introduction of damage to DNA.

Parental irradiation in experimental animals sometimes induces genomic instability in germ cells, which is detected in the offspring. Genomic instability may also occur in offspring born to irradiated parents. These phenomena are sometimes called transgenerational genomic instability, but the former is not transgenerational in a strict sense. Several laboratories have published a series of papers on genomic instability in humans and in mice. However, a number of discrepancies are noted among them. In the present review, attempts are made to resolve some of the issues.

Minisatellite mutation and male spermatogenesis

Minisatellite sequences, also known as VNTR (variable number of tandem repeats), were first noted in the human genome for their high length variability among individuals (Wyman and White, 1980). They are several kbp sequences composed of varying numbers of GC-rich core repeat uints of 6 to over 100 bp in length (Jeffereys et al., 1985). Minisatellite sequences are distributed widely from higher eukaryotes to protozoa, and similar sequences are even found in lower eukaryotes such as yeast (Mannazzu et al., 2002). Among those, human and mouse minisatellites are studied in detail. Some of the human and mouse minisatellites are highly unstable, and the rate of the instability varied among minisatellite loci. In addition, the rate varied by three orders of magnitude even among alleles of a human minisatellite locus (Buard et al., 1998).

A variation in the sequence motif among the core repeats within a human minisatellite makes it possible to perform a digital analysis of detailed arrangements of the repeats within a minisatellite stretch. Thus, the mechanism of hypervariability was investigated mainly in humans (Jeffereys, 1997). An analysis of the highly variable MS32 and CEB1 minisatellite loci revealed that a new germline mutation includes intra-allelic and interallelic exchanges. A unique feature of the latter was that it did not involve the exchange of the flanking sequences (Jeffereys et al., 1994). This can be explained by the synthesis-dependent strand annealing during meiotic recombination. Also, a bias was noted for the recombination breakpoint, which increased in the frequency toward one of the ends. This was thought to suggest the presence of recombination hot spot in the vicinity of the end (Jeffreys, 1994; Richard and Paques, 2000). Meiotic recombination is initiated by double-strand breaks introduced by Spo11, a meiosis-specific topoisomerase like nuclease conserved throughout the eukaryotes (Martini and Keeney, 2002). Spo11 introduces about 300–400 double-strand breaks in the pachytene stage chromosomes. These breaks are repaired by recombination of homologous, but not of sister chromatids, and meiosis-specific cohesion molecules are responsible for this substrate specificity in meiotic recombination (Schwacha and Kleickner, 1997; Haber, 2000; Kitajima et al., 2003). Thus, interallelic exchanges in meiotic mutation of human minisatellite sequences are a reflection of the mechanism specific to the pachytene stage germ cells.

The MS32 minisatellite mutates less frequently in somatic cells, and the breakpoint of mutations distributes randomly throughout the sequence in human blood lymphocytes (Jeffereys and Neumann, 1997). The digital analysis has indicated that somatic mutations involve simple intra-allelic events such as unequal sister chromatid exchanges, duplications and deletions within the stretch (Jeffereys and Neumann, 1997). Homologous chromatids are rarely in close association in somatic cell nuclei, so that interallelic exchanges are nearly impossible to occur. In contrast, sister chromatids are aligned by mitotic cohesion during the S to M phase, and they serve as efficient substrates of recombination (Yokomori, 2003). In addition, unequal sister chromatid exchanges, duplication and deletion occur even without double-strand breaks, and can be initiated by the stalled replication fork during DNA synthesis (Cox, 2000). The difference in the mechanism of minisatellite mutation between germ cells and somatic cells is important to understand some features of minisatellites. For example, the hypervariable human MS32 minisatellite exhibits somatic instability, but not meiotic instability, in the transgenic mice, which is well due to the lack of the recombination hot spot near the integration site (Buard et al., 2000). In any case, DNA replication is required for the minisatellite to mutate.

Core repeat units of minisatellites are generally shorter in the mouse than in humans. Two of the most variable mouse minisatellites, hm6-hm and hm-2, have a 5 bp core repeat of a GGGCA motif, and a 4 bp core repeat of a GGCA motif, respectively (Kelly et al., 1989; Mitani et al., 1990; Gibbs et al., 1993). These short repeat units have no sequence variation within a stretch of minisatellite, and this makes the digital analysis of the sequence impossible. Owing to this homogeneity of the repeat units, the mouse minisatellites are sometimes known as ESTR (expanded simple tandem repeat). Analyses at the single-molecule level revealed that the majority of mutations at the hypervariable Ms6-hm minisatellite were those of gain and loss of a few repeat units, and the same spectrum of mutations was found in somatic cells and in mouse sperm (Yauk et al., 2002). Thus, although human and mouse minisatellites share similar features such as GC richness and high rate of instability, their behaviors differ considerably. The intra-allelic recombination seems to be the sole mechanism for the length change mutation of the mouse minisatellite, and the distinction between meiotic instability and somatic instability may possibly be restricted to human minisatellite sequences. This difference between humans and mouse is likely to contribute for their sensitivity to radiation-induced instability. Destabilization of minisatellites was shown repeatedly in the latter, while that in the former is controversial. In addition, transmission of radiation-induced destabilization was reported only in the mouse.

Female mammals undergo a cycle of oogenesis once in their lifetime. In contrast, cycles of spermatogenesis persist throughout the life of sexually mature males. Therefore, males are more convenient for analysing the effect of radiation at various stages of germ cells. During spermatogenesis, stem cells and spermatogonia undertake mitotic division cycles similar to those in somatic cells, and they have a large capacity to repair DNA damage. They then differentiate into spermatocytes where four strands of replicated homologous chromosomes undergo meiotic exchanges. During the first meiotic division, two sister chromatids of a homologous chromosome move to the same daughter cell, and then sister chromatids are separated at the second meiotic division to yield spermatids. Drastic differentiation takes place during the spermatid stage in which cytoplasm is lost, and nuclear histones are replaced by protamine. After this stage of late spermatids, the capacity to repair DNA damage is lost.

A cycle of the male spermatogenesis from stem cells to ejaculating spermatozoa takes around 40 days for the mouse and 70 days for humans (Adler, 1996). In the former, epididymal and testicular spermatozoa last for 4–6 and 6 days, respectively. Similarly, the spermatid stage lasts for 9 days, the spermatocyte stage for 14 days and the differentiating spermatogonia stage for 6 days. Therefore, the effect of radiation on each stage of spermatogenesis can be studied by mating males at various periods after irradiation; offspring born from mating of males at 0 day to 1 week after irradiation corresponds to irradiation of mature spermatozoa, at 2 weeks the late spermatid stage, at 3 weeks the early spermatid to late spermatocyte stages, and at 6 weeks and later, the stem cell stage.

Radiation induction of minisatellite instability in male germ cells in mice: stage specificity

Radiation induction of minisatellite mutation was first reported in offspring born to male mice mated at 6 weeks after irradiation, which corresponds to irradiation at the stem cell stage (Dubrova et al., 1993). This work was followed by a report of another group (Sadamoto et al., 1994). Owing to the highly frequent mutation and a small target size of the sequence, it was concluded that minisatellite mutation is untargeted and radiation-induced genomic instability is likely to be responsible (Sadamoto et al., 1994; Dubrova et al., 1998). A series of works thereafter confirmed that the mouse minisatellite is sensitive to radiation induction of length change mutations, but several discrepancies are noted. The results of mouse minisatellite studies are summarized in Table 1.

Table 1 Summary of the mouse minisatellite studies

Radiation-induced minisatellite instability exhibits a stage sensitivity during male spermatogenesis. Premeiotic stem cells and spermatogonia were found to be sensitive to radiation. An elegant experiment was carried out to examine the role of meiotic crossing over in the instability of minisatellites (Barber et al., 2000). Male mice irradiated with 1 Gy X-rays or 10 mg/kg cisplatin were mated at various times thereafter, and F1 mice were analysed. These treatments destabilized minisatellites when given before the pachytene stage (4 weeks after irradiation), but not after the spermatocyte stage undergoing diakinesis and second meiotic divisions (3 weeks after irradiation). In contrast to minisatellite instability, the frequency of meiotic crossing over, as assessed by polymorphic microsatellite markers, was not affected by irradiation. These data suggest that meiotic crossing over is not involved in radiation induction of minisatellite mutation in mice. It should be noted that the number of double-strand breaks induced by 1 Gy X-rays is around 30–40 per diploid genome, which is an order of magnitude smaller than those induced by Spo11 during meiosis.

A discrepancy was noted for the sensitivity of the postmeiotic stages to radiation induction of minisatellite instability. Dubrova and his group repeatedly demonstrated that the postmeiotic stage, as analysed by mating males at 3 weeks after irradiation, was insensitive to radiation induction of minisatellite mutations using three strains of mice including CBA/H (Dubrova et al., 1993, 1998a; Barber et al., 2000). In contrast, Niwa and his group reported that the postmeiotic irradiation of C3H/HeN males at 2 weeks prior to the mating is most effective in inducing minisatellite mutation (Sadamoto et al., 1994; Fan et al., 1995). Thus, the discrepancy may be due to the difference in the strain used. It may also be due to the difference in the stages of spermatids at the time of irradiation. Mating of males at 3 weeks after irradiation corresponds to the late spermatocyte to early spermatid stage (Adler, 1996; Barber et al., 2000). At this stage, germ cells still retain a capacity to repair double-strand breaks, but replication and recombination of DNA does not take place. The late spermatid stage, as analysed by mating males at 2 weeks after irradiation, has been shown to be highly sensitive to various chemical mutagens, and this was partly explained by the damage to protamin, the loss of which was later shown to cause severe DNA damage during condensation and packaging of DNA in the sperm head (Russell et al., 1990,Russell et al., 1991; Cho et al., 2003). The postmeiotic spermatids and spermatozoa were shown to be sensitive to radiation induction of specific locus mutation (Russell et al., 1998). Careful re-examination of the early and late spermatid stage irradiation would resolve at least part of the issue.

Irradiation of spermatozoa was shown to be sensitive to length change mutation of the Ms6-hm minisatellite locus (Fan et al., 1995; Niwa and Kominami, 2001). Since spermatozoa lack any activity of DNA metabolism, mutation of minisatellite sequences must have taken place in zygotes after fertilization. Induction of somatic mutation at the Ms6-hm locus was reported to occur in early embryonic development (Kelly et al., 1989; Gibbs et al., 1993). In addition, somatic mutation of the Ms6-hm sequence can be induced by radiation in highly proliferating tissues such as spleen, but not in the nonproliferating brain cells (Yauk et al., 2002). In contrast, Dubrova and his group indicated that spermatozoa are again insensitive to radiation induction of minisatellite mutation (Barber et al., 2002). Thus, the discrepancy on the spermatozoa radiosensitivity is not resolved at present.

Radiation induction of minisatellite instability in male germ cells in mice: transacting nature of the mutagenesis and transmission to the second generation

Although the issue is not fully resolved, minisatellite mutation in the offspring born to irradiated spermatozoa indicates that the destabilization events must have taken place in zygotes. In zygotes, paternal and maternal pronuclei share the same cytoplasmic environment, and, therefore, the mechanism of destabilization could operate on both alleles. Indeed, minisatellite mutation was induced at the unirradiated maternal allele as well (Niwa and Kominami, 2001). This also suggests the presence of a crosstalk between maternal and paternal genomes. A genomic crosstalk was demonstrated in mouse zygotes, where a p53-dependent lacZ reporter was activated in the microinjected female pronucleus upon fertilization by irradiated sperm (Shimura et al., 2002). This genomic crosstalk resulted in the suppression of DNA synthesis of the male pronucleus as well as the unirradiated female pronucleus. The slowing down of replication fork is likely to increase the frequency of slippage and mitotic recombination of repeat sequences. Interestingly, activation of the p53-dependent lacZ reporter and the suppression of DNA synthesis were not triggered in zygotes fertilized with sperm irradiated at the late spermatid stage (Shimura, unpublished observation). Consistent with these observations, irradiation of late spermatids did not induce mutation of the maternally derived minisatellite in offspring, although the same treatment was most effective in inducing minisatellite mutation at the paternally derived allele (Sadamoto et al., 1994). These results suggest that late spermatids retain some capacity to process DNA damage in such a way that it is no longer recognized by the damage-sensing mechanism in zygotes. The same damage processing is likely to mutate minisatellite sequences at this stage of spermatogenesis.

The elevated frequency of the minisatellite mutation was reported to persist for at least two subsequent generations after irradiation of the founder males (Barber et al., 2002). Both spermatid and spermatogonia stages were sensitive to this transgenerational destabilization of minisatellite sequences. The higher mutation frequency was passed on through either of male and female lineages, suggesting that the memory of DNA damage survives for at least two cycles of spermatogenesis or oogenesis. The mechanism of the transgenerational memory of DNA damage is of extreme interest. Since the chromatin proteins are replaced during male spermatogenesis in the male lineage, it is likely that the memory is kept in DNA, possibly by epigenetic modification of cytosine methylation. In fact, Cre-mediated introduction of strand breaks and the deletion of the loxP element during meiosis was found to result in methylation and silencing of the region which transvects to the allelic region. This silencing was shown to be stably transmitted to the subsequent generations (Rassoulzadegan et al., 2002).

Radiation induction of minisatellite instability in male germ cells in mice: doubling dose, RBE and the low dose-rate irradiation

One discrepancy among published reports concerns the doubling dose values for the induction of minisatellite mutation. Dubrova and his group estimated the doubling dose to be around 0.5 Gy for premeiotic stage irradiation, and 0.33 Gy in later studies, which is the same as that obtained by the specific locus test in mice (Russell and Kelly, 1982; Dubrova et al., 1993,Dubrova et al., 1998a,Dubrova et al., 1998b). The doubling dose estimations were also made by Niwa and his group, and the values were 3.4, 0.83 and 4 Gy for stem cells, spermatids and spermatozoa, respectively (Fan et al., 1995; Niwa and Kominami, 2001).

Most of the minisatellite mutations are those of the length change type, in which the repeat units in a minisatellite sequence change in number. Southern blotting is used to analyse the change in length, which can give only a rough estimation of the number of repeat units in a particular minisatellite allele. The frequency of alleles plotted against their length takes a unimodal distribution pattern, in which the nonmutated allele equaling in length with the parental allele peaks at the center, and alleles differing in length gradually decrease in number on both sides. For a population with a higher than normal mutation frequency, the distribution pattern becomes wider and more flat. In order to score mutant alleles, one has to define the extent of the length change by which a particular allele is judged as a mutant. This means that the criterion of mutant alleles is rather arbitrary, and is likely to differ in different laboratories. An application of a more stringent criterion, such as scoring larger length changes, lowers the mutation frequencies of both the control and exposed populations. At the same time, however, this brings a more steep increase and a larger difference between the mutation frequencies of the two populations. Thus, the criterion influences the mutation frequency and the doubling dose estimation. In fact, doubling doses of 1.6 Gy, instead of 0.33 Gy, can be calculated from the data published by the same research group as Dubrovas' (Yauk et al., 2002). Re-examination and the adjustment of the criteria of mutant alleles are likely to resolve the differences in the doubling dose estimations.

The RBE of fission neutron for the induction of minisatellite mutation was calculated, and the values were 5.9, 2.6 and 6.5 for stem cells, spermatids and spermatozoa, respectively (Niwa et al., 1996). An independent analysis was made, which gave the RBE 3.36 for the stem cell irradiation (Dubrova et al., 2000). Ionizing radiations produce a variety of DNA damages, and double-strand break and clustered damage are the most deleterious of all (Goodhead, 1999). The yield of double-strand break and clustered damage is dependent on LET, and peaks at around 100 keV/μm (Nikjoo et al., 1999). Therefore, higher RBEs of fission neutron for minisatellite mutation suggest that minisatellite instability is triggered by double-strand breaks.

It is generally accepted that low dose-rate exposures to low LET radiations are less effective in inducing mutation, although the inverse dose-rate effect was sometimes reported (Hill et al., 1984; Crompton et al., 1990). Dose-rate effect was reported for germline mutation in the mouse, and the effect becomes the lowest at a dose rate around 10 mGy/min (UNSCEAR, 1993). At a lower dose rate, cells can have more time to repair DNA damage before the next one is delivered much later. Interestingly, doses of γ-radiation delivered over 100 h at a dose rate of 0.166 mGy/min were found to be as effective as those given at a rate of 0.5 Gy/min for the induction of minisatellite instability in the spermatogonia stem cells (Dubrova et al., 2000).

Radiation induction of minisatellite instability in humans: acute and chronic irradiations

Minisatellite analyses were conducted on the children of atomic bomb survivors (Kodaira et al., 1995). Children were born more than 10 years after the bomb, and the average gonadal dose to the exposed parents was 1.9 Sv. Comparison of the minisatellite mutation rates between the control and exposed groups using nine locus-specific probes and one finger print probe indicated no increase in the mutation rate. Minisatellite mutations were also assessed on sperm DNA samples obtained from three seminoma patients at various times after the completion of radiotherapies (May et al., 2000). Total doses to the normal testicle were 0.38–0.82 Gy, and the sperm DNA was analysed for mutation at two minisatellite loci, using the small-pool PCR method. No increase in the mutation rate was observed for irradiation of stem cells, spermatogonia, spermatocytes and spermatids. A similar study conducted on 10 chemotherapy patients also failed to detect any increase in the mutation rate of the minisatellite MS205 locus, except for a slight increase in one patient receiving high-dose treatments with mechlorethamine and procarbazine (Zheng et al., 2000). These data suggest that human minisatellite sequences are relatively insensitive to acute and fractionated exposures of moderate to high doses. It is important to remind that a somewhat similar analysis on the sperm DNA of male mice using the single-molecule PCR assay yielded a statistically significant increase in the mutation frequency by irradiation of spermatogonia stem cells (Yauk et al., 2002).

In contrast to the negative results of these studies, elevation of the minisatellite mutation rate was noted among children in the radiocontaminated area after the Chernobyl accident of 1986 (Dubrova et al., 1996). Analyses were made on children born 8 years after the accident in a radio-contaminated area of Belarus. Using two locus-specific probes and one finger print probe, a 1.7-fold increase in the minisatellite mutation rate was noted over that of an unexposed English population. In addition, the extent of the increase was linearly related to the level of soil contamination. The positive correlation of minisatellite mutation and low-dose/low dose-rate exposures was further confirmed by subsequent studies (Dubrova et al., 1997,Dubrova et al., 1998b). Radiation in the contaminated areas is mainly due to 137Cs, and the dose rate was less than 5 mSv per year, which is about twice the level of the natural background radiations in the world (UNSCEAR, 2000). The cumulative dose to the parents from the time of the accident to the conception of the children ranged from 3.3 to 28 mSv over the 8-year period. It was suggested that the background minisatellite mutation rate may well differ between English and Belarus populations, and this could have contributed to the higher mutation frequency in the latter (Satoh and Kodaira, 1996).

Minisatellite analyses using seven locus-specific probes were conducted on the children of Chernobyl clean-up workers, whose gonadal doses were estimated to be no more than 150 mSv (Livshits et al., 2001). The mutation rate was not elevated among the children of the workers, except that the children conceived while their fathers were working at the facility or within 2 months after fathers stopped working exhibited a higher but statistically nonsignificant increase. Based on these observations, a radiosensitive window was proposed in which certain stages of the spermatogenesis are sensitive to low-dose and low dose-rate radiations for the induction of minisatellite mutation. Another study on the children of Estonian clean-up workers suggested a slight but nonsignificant elevation of the mutation frequency only when the dose to the fathers was above 0.2 Sv (Kiuru et al., 2003).

In the recent study of the inhabitants of the radiocontaminated area of Ukraine, a statistically significant 1.6-fold increase in the mutation rate was observed (Dubrova et al., 2002b). This increase in the mutation rate was found in the area, regardless of the year of birth of the children from 1986 to 1995. This suggested that the cumulative dose to the parents is not relevant for the induction of minisatellite mutation, and, therefore, the initial acute exposure from the short-lived radionuclide was proposed to be responsible for the increase. The acute dose by the fallout was estimated to be 0.2–0.4 Gy. Since spontaneous minisatellite mutation in humans occurs mainly by meiotic recombination, the authors suggested that the exposure to the stem cells induces later destabilization of minisatellites during meiosis, which persists from the time of the accident until 1995. In the mouse study, induction of minisatellite mutations by irradiation of spermatogonia stem cells was not associated with higher rates of crossing over at the other region of the genome (Yauk et al., 2002). Digital analysis of the new mutation in Ukraine children would clarify whether or not the meiotic destabilization is involved in the observed increase in the minisatellite mutation. In any case, the discrepancy among the human studies still remains as an enigma, and future crossexaminations by the involved groups are highly desirable.

Radiation induction of minisatellite instability in humans: doubling dose and the lack of transmission to the next generation

The study of the children of atomic bomb survivors detected no increase in the mutation rate (Kodaira et al., 1995). However, the data, within its limitation of the statistic power, are still consistent with the doubling dose of 1.7–2.2 Sv, which was estimated from the analyses of five markers on the same population (Neel et al., 1990). The negative data of three radiotherapy patients can also be consistent with this doubling dose estimation (May et al., 2000). In contrast, the doubling dose estimated from the studies of the inhabitants of radiocontaminated areas of the Chernobyl accident is extremely high. A dose less than 50 mSv delivered over the 8-year period resulted in a 1.7-fold increase (Dubrova et al., 1996), and the doubling dose could be as small as 60 mSv even without correcting for the dose and dose-rate effectiveness factor of 2 (BEIRR V, 1990). The recent study implicating the acute exposures lowered the doubling dose, but still gave the value of 0.25–0.5 Gy, which is closer to the 1 Gy estimate for humans (UNSCEAR, 1993; Sankaranarayanan and Chakraborty, 2000; Dubrova et al., 2002b). As has been discussed in the section ‘Radiation induction of minisatellite instability in male germ cells in mice: doubling dose, RBE and the low dose-rate irradiation’, the doubling dose estimation for the minisatellite mutation can be affected greatly by the criterion of mutant alleles. Accurate doubling dose estimation is in theory difficult to make, unless precise digital analysis is applied.

A study of minisatellite mutation was conducted in the area near the nuclear test sites of Kazakhstan, where the inhabitants received as high as over 1 Sv of radiations due to surface explosions of the bombs (Dubrova et al., 2002a). Analyses were made on 40 three-generation families. A 1.5-fold increase in the minisatellite mutation frequency was restricted to the generations directly exposed to radiations. The mutation frequency decreased to a normal level in the generation born after the cessation of the surface bomb tests. Since the analyses are made in the same family, the above decrease suggests that the minisatellite destabilization does not persist in the germ cells of the subsequent generations. This contrasts to the mouse studies in which the effect persisted in the nonexposed generations born to the exposed founder mice.

Radiation induction of transgenerational genomic instability detected by the reversion of the mouse pink-eyed unstable allele

Transmission of the genomic instability across the generations has been tested in experimental systems other than minisatellites. The mouse pink-eyed dilution locus has several mutant alleles including the pink-eyed unstable (p-un) and the pink-eyed Jackson (p-J) alleles (Brilliant et al., 1994). The former mutation is due to a partial tandem duplication of a 71 kb region which disrupts the wild-type function of the gene (Gondo et al., 1993). The p-J allele carries a simple deletion mutation. These mutations confer a light yellowish color to the coat and the retinal pigment epithelium (RPE). Homologous recombination at the duplication restores the wild function of the p-un allele, which can easily be identified by black pigmented spots in the coat and in the RPE. The somatic reversion can further be induced by treatment of the developing fetus with various mutagens including radiations (Schiestl et al., 1994,Schiestl et al., 1997; Bishop et al., 2000). The increase in the frequency after irradiation is too high to be due to the direct damage to the locus, and the induced genomic instability was therefore proposed to be the mechanism. In addition, the reversion of the allele in the RPE occurs predominantly by the intra-allelic recombination rather than the interallelic recombination.

Experiments were conducted in which p-un females were mated with X-irradiated p-J males, and F1 mice were scored for the reverted black spots in the RPE (Shiraishi et al., 2002). The number of spots per RPE was increased by twofold when the male parents were exposed to 6 Gy at the spermatozoa stage. Since the reverted allele was that derived from the unirradiated mother, this increase in the mutation frequency was due to the transacting genomic instability triggered by fertilization by the irradiated sperm. The reciprocal cross yielded a similar increase in the frequency of reversion at the paternally derived p-un allele, confirming that the induced genomic instability operates in cis as well. The time gap between the introduction of DNA damage at fertilization and the reversion in RPE at the fetal development clearly demonstrates the delayed nature of the induced genomic instability. In the mouse p-unsystem, the number of spots per RPE represents the mutation frequency, while the size of an eye spot, or the number of pigmented cells in a spot, represents the number of cell replication after the reversion. Analyses of the frequency and the size of the spots indicated that the rate of mutation was elevated throughout the development of RPE, suggesting that the somatic expression of genomic instability does not decay at least during organogenesis of the eye.

The somatic instability of the p-un allele was not observed when radiation was delivered either to spermatogonia stem cells or to late spermatids, which is consistent with the data of the maternally derived allele mutation of minisatellites (Sadamoto et al., 1994; Shiraishi et al., 2002). Therefore, induction of somatic instability at the p-unallele and at the minisatellite locus is likely to require the presence of DNA damage at the time of fertilization. Interestingly, the somatic instability of the p-unallele by irradiated spermatozoa in F1 mice was not observed in the F2 generation (Shiraishi, unpublished observation). This suggests that the memory of DNA damage for the pink-eyed system is lost once cells undergo meiosis.

Radiation induction of transgenerational genomic instability in Drosophila and medaka fish

Several experimental systems are well suited for the study of transgenerational genomic instability. The classic mutation analysis in Drosophila first demonstrated the transmission of genomic instability to subsequent generations of males treated with X-rays and mustard gas (Auerbach, 1976). Unfortunately however, the marker of mutation used in these classic studies was the lethality which tells little about the nature of mutation. The white ivory mutation of Drosophila is due to a tandem duplication of a region of the white locus, and its high-frequency reversion can easily be identified as red spots in the ivory color eyes. Mating of irradiated males carrying a deletion mutation at the locus would make an ideal system to test whether or not the reversion is induced at the maternally derived white ivory allele.

A specific locus test was developed in the medaka fish (Oryzias latipes), where a tester strain carries three recessive visible mutant alleles (Shimada and Shima, 1991,Shimada and Shima, 1998). Radiation was successfully shown to induce mutations at these loci, and the frequencies were highest for irradiation of spermatozoa and spermatids, while spermatogonia was least sensitive. Surprisingly, the doubling dose for the induction of these mutations was as small as 30 mGy. Transparency of the eggs makes the medaka system particularly advantageous in that mosaic mutations can easily be scored at early developmental stages even in embryos carrying lethal mutations. Mosaic mutation was noted at high frequencies when males were irradiated at spermatozoa/spermatid stages, suggesting that somatic mutation does occur in a delayed fashion (Shimada and Shima, 2001). In addition, the reciprocal cross of irradiated tester males with unirradiated wild-type females also produced the mosaic, demonstrating that the delayed mutagenesis is not restricted to the irradiated male genome, but extends to the unirradiated female genome (A Shimada, personal communication).


Recent studies on several experimental systems and on human populations have established that genomic instability is induced by radiations. It is expressed as an untargeted and delayed mutation in the germ cells of irradiated individuals, and in somatic cells of the offspring born to them. In the mouse studies, the discrepancy on stage sensitivity, especially of the effectiveness of postmeiotic irradiation, in inducing minisatellite instability can be resolved by more detailed analyses of the stages. A large difference in the doubling dose estimation is likely because of the use of different criteria for the mutant minisatellite allele. However, the issue of radiation-induced minisatellite mutation in humans remains highly obscure. Crossexamination of the same population by the involved research groups may be necessary to resolve the problem. Many of the discrepancies are likely to be derived from the use of Southern blotting to analyse minisatellite mutations, which is subject to a variety of difficulties concerning how to judge mutant alleles. One important issue awaiting future clarification concerns the persistent transmission of the germline instability through generations. Minisatellite study indicated the transmission in mice but not in humans. The study on the mouse pink-eyed unstable allele indicated that the effect of radiation is erased once the germ cells experience the gamatogenesis.

Although some details do not agree, most of the works reviewed in the present report do demonstrate that genomic instability operates in whole-body systems. Radiation-induced genomic instability is a new and previously unexpected mechanism of mutagenesis (Bridges, 2001).


  1. Adler I . (1996). Mutat. Res., 352, 169–172.

  2. Auerbach C . (1976). Mutation Research. Chapman & Hall: London, pp. 270–278.

  3. Barber R, Plumb M, Smith AG, Cesar CE, Boulton E, Jeffreys AJ and Dubrova YE . (2000). Mutat. Res., 457, 79–91.

  4. Barber R, Plumb MA, Boulton E, Roux I and Dubrova YE . (2002). Proc. Natl. Acad. Sci. USA, 99, 6877–6882.

  5. Biological Effects of Ionizing Radiation Report V (BEIRR V). (1990). National Academy of Science. National Research Council, Washington DC, USA.

  6. Bishop AJ, Kosaras B, Sidman RL and Schiestl RH . (2000). Mutat. Res., 457, 31–40.

  7. Boudsocq F, Ling H, Yang W and Woodgate R . (2002). DNA Repair (Amst.), 1, 343–358.

  8. Bridges B . (2001). Radiat. Res., 156, 631–641.

  9. Brilliant MH, Williams RW, Conti CJ, Angel JM, Oakey RJ and Holdner BC . (1994). Mamm. Genome, 5, S104–S123.

  10. Buard J, Bourdet A, Yardley J, Dubrova Y and Jeffreys AJ . (1998). EMBO J., 17, 3495–3502.

  11. Buard J, Collick A, Brown J and Jeffreys AJ . (2000). Genomics, 65, 95–103.

  12. Cho C, Jung-Ha H, Willis WD, Goulding EH, Stein P, Xu Z, Schultz RM, Hecht NB and Eddy EM . (2003). Biol. Reprod. (e-publication ahead of print).

  13. Cox M . (2000). Proc. Natl. Acad. Sci. USA, 98l, 8173–8180.

  14. Crompton NEA, Barth B and Kiefer J . (1990). Radiat. Res., 124, 300–308.

  15. Dubrova YE, Bersimbaev RI, Djansugurova LB, Tankimanova MK, Mamyrbaeva ZZ, Mustonen R, Lindholm C, Hulten M and Salomaa S . (2002b). Science, 295, 103.

  16. Dubrova YE, Grant G, Chumak AA, Stezhka VA and Karakasian AN . (2002b). Am J Hum Genet., 71, 801–809.

  17. Dubrova YE, Jeffreys AJ and Malashenko AM . (1993). Nature Genet., 5, 92–94.

  18. Dubrova YE, Plumb M, Brown J, Boulton E, Goodhead D and Jeffreys AJ . (2000). Mutat. Res., 453, 17–24.

  19. Dubrova YE, Plumb M, Brown J, Fennelly J, Bois P, Goodhead D and Jeffreys AJ . (1998a). Proc. Natl. Acad. Sci. USA, 95, 6251–6255.

  20. Dubrova YE, Plumb M, Brown J and Jeffreys AJ . (1998b). Int. J. Radiat. Biol., 74, 689–696.

  21. Dubrova YE, Nesterov VN, Krouchinsky NG, Ostapenko VA, Neumann R, Neil DL and Jeffreys AJ . (1996). Nature, 380, 683–686.

  22. Dubrova YE, Nesterov VN, Krouchinsky NG, Ostapenko VA, Vergnaud G, Giraudeau F, Buard J and Jeffreys AJ . (1997). Mutat. Res., 381, 267–278.

  23. Fan YJ, Wang Z, Sadamoto S, Ninomiya Y, Kotomura N, Kamiya K, Dohi K, Kominami R and Niwa O . (1995). Int. J. Radiat. Biol., 68, 177–183.

  24. Friedberg EC, Walker GC and Siede W . (1995). DNA Repair and Mugatenesis. ASM Press: Washington, DC, pp. 59–61.

  25. Gibbs M, Collick A, Kelly RG and Jeffreys A . (1993). Genomics, 17, 121–128.

  26. Gondo Y, Gardner JM, Nakatsu Y, Durham-Pierre D, Deveau SA, Kuper C and Brilliant MH . (1993). Proc. Natl. Acad. Sci. USA, 90, 297–301.

  27. Goodhead DT . (1999). J. Radiat. Res., 40 (Suppl.), 1–13.

  28. Haber JE . (2000). Trends Genet., 16, 259–264.

  29. Hill CK, Han A and Elkind MM . (1984). Int. J. Radiat. Biol., 46, 11–15.

  30. Jeffereys AJ . (1997). Clin. Sci., 93, 383–390.

  31. Jeffereys AJ and Neumann R . (1997). Hum. Mol. Genet., 136–145.

  32. Jeffereys AJ, Tamaki K, MacLeod A, Monkton DG, Neil DL and Armour JA . (1994). Nature Genet., 6, 136–145.

  33. Jeffereys AJ, Wilson V and Thein SL . (1985). Nature, 314, 67–74.

  34. Kelly R, Bulfield G, Collick A, Gibbs M and Jeffreys A . (1989). Genomics, 5, 844–856.

  35. Kitajima TS, Yokobayashi S, Yamamoto M and Watanabe Y . (2003). Science, 300, 1152–1155.

  36. Kiuru A, Auvinen A, Luokkamaki M, Kakkonen K, Veidebaum T, tekkel M, Rahu M, Kakulinen T, Servomaa K, Rytomaa T and Mustonen R . (2003). Radiat. Res., 159, 651–655.

  37. Kodaira M, Satoh C, Hiyama K and Toyama K . (1995). Am. J. Hum. Genet., 57, 1275–1283.

  38. Leroy C, Lee SE, Vaze MB, Ochsenbien F, Guerois R, Haber JE and Marsolier-Kergoat MC . (2003). Mol Cell, 11, 827–835.

  39. Livshits LA, Malyarchuk SG, Kravchenko SA, Matsuka GH, Lukyanova EM, Antipkin YG, Arabskaya LP, Petit E, Giraudeau F, Gourmelon P, Vergnaud G and Le Guen B . (2001). Radiat. Res., 155, 74–80.

  40. Mannazzu I, Simonetti E, Marinangeli P, Guerra E, Budroni M, Thangavelu M and Clementi F . (2002). Appl. Environ. Microbiol., 68, 5437–5444.

  41. Martini E and Keeney S . (2002). Mol. Cell, 9, 700–702.

  42. May CA, Tamaki K, Neumann R, Wilson G, Zagars G, Pollack A, Dubrovva YE, Jeffreys AJ and Meistrich ML . (2000). Mutat. Res., 453, 67–75.

  43. Mitani K, Takahashi Y and Kominami R . (1990). J. Biol. Chem., 265, 15203–15210.

  44. Morgan WF . (2003a). Radiat Res., 159, 567–580.

  45. Morgan WF . (2003b). Radiat Res., 159, 581–596.

  46. Neel JV, Schull WJ, Awa AA, Satoh C, Kato H, Otake M and Yoshimoto Y . (1990). Am. J. Hum. Genet., 46, 1053–1072.

  47. Nikjoo H, Munson RJ and Bridges BA . (1999). J. Radiat. Res., 40 (Suppl.), 85–105.

  48. Niwa O, Fan Y, Numoto M, Kamiya K and Kominami R . (1996). J. Radiat. Res., 37, 217–224.

  49. Niwa O and Kominami R . (2001). Proc.Natl. Acad. Sci. USA, 98, 1705–1710.

  50. Radman M . (2001). Nature, 413, 115.

  51. Rassoulzadegan M, Magliano M and Cuzin F . (2002). EMBO J., 21, 440–450.

  52. Richard G-F and Paques F . (2000). EMBO Rep., 1, 122–126.

  53. Russell WL, Bangham JW and Russell LB . (1998). Genetics, 148, 1567–1578.

  54. Russell LB, Hunsicker PR, Nestor LA, Cacheiro LA and Generoso WM . (1991). Mutat. Res., 262, 101–107.

  55. Russell WL and Kelly EM . (1982). Proc. Natl. Acad. Sci. USA, 79, 542–544.

  56. Russell LB, Russell WL, Rinchik EM and Hunsicker PR . (1990). Bumbury Report 34, Cold Spring Harbor Laboratory. Cold Spring Harbor, NY, pp 271–289.

  57. Sadamoto S, Suzuki S, Kamiya K, Kominami R, Dohi K and Niwa O . (1994). Int. J. Radiat. Biol., 65, 549–557.

  58. Sankaranarayanan K and CHakraborty R . (2000). Mutat. Res., 453, 107–127.

  59. Satoh C and Kodaira M . (1996). Nature, 383, 226.

  60. Schiestl RH, Aubrecht J, Khogali F and Carls N . (1997). Proc. Natl. Acad. Sci. USA, 94, 4576–4581.

  61. Schiestl RH, Khogali F and Carls N . (1994). Science, 266, 1573–1576.

  62. Schwacha A and Kleickner N . (1997). Cell, 90, 1123–1135.

  63. Shimada A and Shima A . (1991). Proc. Natl. Acad. Sci. USA, 88, 2545–2549.

  64. Shimada A and Shima A . (1998). Mutat. Res., 399, 149–165.

  65. Shimada A and Shima A . (2001). Mutat. Res., 495, 33–42.

  66. Shimura T, Inoue M, Taga M, Shiraishi K, Uematsu N, Takei N, Yuan Z, Shinohara T and Niwa O . (2002). Mol. Cell. Biol., 22, 2220–2228.

  67. Shiraishi K, Shimura T, Taga M, Uematsu N, Gondo Y, Ohtaki M, Kominami R and Niwa O . (2002). Radiat. Res., 157, 661–667.

  68. United Nations Scientific Committee on the Effect of Atomic Radiation. (1993). United Nations, New York.

  69. United Nations Scientific Committee on the Effect of Atomic Radiation. (2000). United Nations, New York.

  70. Wyman AR and White R . (1980). Proc. Natl. Acad. Sci. USA, 77, 6754–6758.

  71. Yauk CL, Dubrova YE, Grant GR and Jeffreys AJ . (2002). Mutat. Res., 500, 147–156.

  72. Yokomori K . (2003). Curr. Top. Microbiol. Immunol., 274, 79–112.

  73. Zheng N, Moncton DG, Wilson G, Hagemeister F, Chakraborty R, Conor TH, Siciliano MJ and Meistrich ML . (2000). Environ. Mol. Mutagen., 36, 134–145.

Download references


This work is supported by a Grant-in-aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan, and by a grant from the Nuclear Safety Research Association.

Author information

Correspondence to Ohtsura Niwa.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Niwa, O. Induced genomic instability in irradiated germ cells and in the offspring; reconciling discrepancies among the human and animal studies. Oncogene 22, 7078–7086 (2003).

Download citation


  • radiation
  • genomic instability
  • minisatellite
  • transgenerational

Further reading