To the best of our knowledge, this is the first study to successfully generate a model fish carrying mutant rad50. In this study, to ensure a null mutation of medaka rad50, we used genome editing to target the position of the coiled-coil region preceding the hook construct to generate a stop codon close to the back of that position. With the loss of the hook construct, rad50 stopped expressing functional RAD50 [16, 28–30]. Therefore, the 2 base-deletion in the rad50 of rad50Δ2/+ medaka would have resulted in the loss of functional rad50. In contrast, a 9-base deletion in rad50 that caused a small in-frame deletion of rad50 retained some rad50 function in rad50Δ9/+ medaka. Furthermore, STIII medaka is a transparent fish in which morphological abnormalities, including organ defects and tumorigenesis, can be easily observed. Therefore, rad50Δ2/+ and rad50Δ9/+ STIII medaka were deemed to be appropriate models that could be used to evaluate the phenotypes affected by rad50 mutations.
In humans, mutation of RAD50 has been shown to be associated with microcephaly and growth retardation, but not immunodeficiency [12, 13]. Mutant mice with hypomorphic RAD50 showed hydrocephalus and defects in primitive hematopoietic and gametogenic cells, along with liver tumorigenesis, but did not show any evidence of histological defects in the cerebellum or telangiectasia [16, 31]. In the present study, morphological changes were observed in the eyes, thyroid gland, liver, and kidneys of rad50Δ2/+ medaka. Telangiectasia was observed in the medaka model. Telangiectasia, which are found on the surface of the skin or the edges of eyeballs [9], constitute a part of the findings of A-T patients [8]. This finding has not been previously observed in experimental animal models of A-T [16, 31]. To the best of our knowledge, rad50Δ2/+ medaka is the first experimental animal model exhibiting telangiectasia, similar to that found in A-T patients. Histiocyte and lymphocyte infiltration in rad50Δ2/+ medaka may be caused by hematopoietic and immune system disorders. Some A-T patients show immune disorders, and mutant mice with MRN dysfunction show defects in primitive hematopoietic and gametogenic cells [16, 31]. A-T is strongly associated with abnormal proliferation of cells involved in the lymphatic system, which weakens the immune system [32]. Histiocyte and lymphocyte infiltration in rad50Δ2/+ medaka is different from that seen in mutant mice [16, 31] but may mimic findings in humans. However, the possibility remains that histiocytes and lymphocytes may have increased and infiltrated after eye disruption and not before. No histological changes were observed in reproductive tissues, and no abnormal mating behavior was observed in rad50Δ2/+ medaka, and embryogenesis of offspring was normal.
The hindbrain of fish, which corresponds to the cerebellum of mammals, integrates processes such as vestibular motor reflexes, gaze retention, and lateral line signal detection to maintain posture against flow velocity [33–35]. Several experiments using zebrafish and medaka have demonstrated that deficiencies in the genes required for hindbrain homeostasis may reduce rheotaxis ability [36]. Mice with a mutation in the gene regulating the arrangement of cells in the cerebellum showed disordered Purkinje cells and granular cells in the cerebellum, along with poor performance in the rotarod test, which measures coordination and motor learning [37]. Therefore, disordered cells in rad50Δ2/+ medaka hindbrain may lead to deterioration of rheotaxis ability. In addition, spindle-shaped Purkinje cells, that were not observed in the wildtype, were characteristically observed in rad50Δ2/+ medaka, which may also be associated with defects in their rheotaxis ability. The proximal axons of Purkinje cells in the cerebellum swell into spindle-shaped cells (torpedo) in A-T patients [38].
No malignant tumors were observed in rad50Δ2/+ medaka, but capillary hemangiomas of the retina and nodular thyroid hyperplasia were. These results may substantiate reported carcinogenic properties of RAD50, as suggested by the prevalence of RAD50 mutations in cancer patients. Of 317 Finnish breast and ovarian cancer patients, 8 carried RAD50 heterozygous mutations, indicating a significant association with cancer morbidity and RAD50 polymorphism in Finnish people [39]. On the other hand, a study of 7657 Chinese breast cancer patients demonstrated a significant association between the RAD50 mutation and the recurrence and mortality of breast cancer; however, no association with morbidity was noted [40]. Further studies may be required to elucidate the clinical significance of RAD50 mutation in human cancers, but its tumorigenic properties, as shown by other defective MRN-ATM pathway molecules, could be safely demonstrated using studies based on animal models such as ours.
Microsatellite instability was observed in non-tumor cells of rad50Δ2/+ and rad50Δ2/Δ2 medaka. Our results suggest that the rad50 Δ2 mutation confers cancer predisposition through genetic instability, probably by inducing a deficiency in the formation of the MRN complex. Tsyusko et al. analyzed nine microsatellite sites (including the sites analyzed in our study) of medaka offspring that had been irradiated for 45 d [41]. The report showed that only the Medµ45 locus showed a significant difference in the mutation rate, while the other eight sites did not. Our study indicated that although the Medµ45 and Medµ58 sequences of medaka carrying rad50 Δ2 mutation were similar to the wildtype sequences, these mutants carried mutations in the Medµ52 (1 of 16 fish) and Medµ60 (2 of 16 fish) sequences. The mutation rate in the Medµ60 mutant was significantly higher than that of the offspring of irradiated medaka (2 of 16 fish vs. 1 of 187 fish [41]; p < 0.05; Fisher’s exact test). Thus, the Δ2 mutations in rad50 may have exerted an adverse effect on the stability of Medµ60 in medaka, and this effect may have been greater than that exerted by irradiation. However, accurate interpretation of our results on microsatellite instability may be difficult owing to the scarcity of reports on microsatellite sites and genomic instability in medaka.
Epidemiological studies have indicated that mutations in DNA repair genes are involved in lifespan and aging [42]. However, evidence regarding the effect of RAD50 mutations on lifespan is inadequate. Mouse early embryonic stem cells homozygous for mutated alleles of Rad50 are nonviable [43]. The study showed no distinctive features regarding growth rate, viability, and fertility in mice heterozygous for the mutant allele compared to wildtype. In contrast, Bender et al. [31] demonstrated a decrease in the growth rate compounded by the depletion of bone marrow owing to a lack of hematopoietic stem cell proliferation in a mouse hypomorphic RAD50 model [31].
The results of the current study indicated that the survival time of heterozygous mutants was shorter than that of wildtype medaka, even in environments free of exposure to mutagens, such as those involving UV light or carcinogens. The short lifespan could be owing to the rad50 mutation and/or mutation-associated histological changes occurring around 40 weeks of age after hatching. The hatched larvae resulting from a cross between heterozygotes exhibited a Mendelian genotypic ratio. As opposed to zebrafish studies which demonstrated that embryonic development was affected by rad50 [26], our results suggest that homozygous mutations in rad50 may not affect embryogenesis and hatching in a stable environment. However, normal rad50 appears to be essential for survival in the early stages of development following hatching, since the survival time in homozygous mutants was found to be much shorter after hatching.
Our study was affected by several limitations. First, the size of analyzed cohorts was relatively small, as a result of which the results of some analyses (e.g., microsatellite analysis) could not be statistically confirmed. Thus, further studies with larger sample sizes are warranted. Secondly, whereas A-T are typically autosomal recessive disorders associated with biallelic mutations of a relevant gene, most results of the present study are based on an analysis of heterozygotes of rad50 mutation. However, a subset of A-T patients carry a heterozygous ATM mutation, suggesting that some type of heterozygous ATM deficiency may give rise to the A-T phenotype. Similarly, RAD50 mutants may exert haploinsufficiency or dominant-negative effects on A-T phenotype development. A previous study also implicated the dose-dependent dominance of a heterozygous rad50 mutant in a mouse model. In our medaka model, the homozygous mutant of rad50 was semi-lethal and thus, inappropriate for precise analysis of several data sets. Regardless of the zygosity of the mutation, our medaka model exhibiting the A-T phenotype will be useful in future experiments investigating the pathogenesis of the disease.
In conclusion, to the best of our knowledge, this study is the first to demonstrate that most A-T phenotypes may be concurrently reproduced via a vertebrate rad50 germline mutation model. The rad50Δ2/+ STIII medaka developed ataxia represented by rheotaxis disability, Purkinje cell disorder in the hindbrain, telangiectasia, immunological abnormalities, and tumor formation accompanied by genetic instability, all of which are typical characteristics of A-T patients. Furthermore, heterozygous mutation of rad50 results in a short lifespan, a condition which is also seen in classical A-T patients. In addition, this fish model revealed that RAD50 mutations could cause tumorigenesis. Thus, the A-T fish model could contribute to a deeper understanding of the molecular mechanisms underlying A-T and RAD50 mutation-induced tumor, as well as help develop novel therapeutic strategies based on improved insight into molecular disorders.