Oxidative stress on the male reproductive organs of wild mice collected from an area contaminated by radioactive materials in Fukushima

The Fukushima Daiichi Nuclear Power Plant (FDNPP) accident on March 11, 2011, released approximately 7,500 PBq of radioactive materials into the environment¹ . The Japanese government designated the areas where the annual cumulative radiation dose estimated from the air dose rate exceeded 50 mSv and the annual accumulated dose was not likely to decline below 20 mSv even after five years as ‘difficult-to-return zones’ in 2012. Although humans evacuated these difficult-to-return zones, wild flora and fauna remained and were exposed to radiation. Previous studies have attempted to estimate radiation exposure levels in terrestrial organisms in difficult-to-return zones near the FDNPP^{2,3,4,5,6,7,8,9,10} . Garnier-Laplace et al.⁸ showed that forest rodents were exposed to radiation at a rate of 3.9 mGy/d during the first 30 d after the accident. Radiation exposure can reduce reproductive success, according to the International Commission on Radiological Protection (ICRP). Strand et al.⁹ estimated that terrestrial wild organisms inhabiting areas near the FDNPP were exposed to 100–300 μGy/h (2.4–7.2 mGy/d) for the first three months after the accident. Almost all radiation is emitted by short-lived radioisotopes such as iodine-131 (T_1/2 = 8 d)⁹ . For at least three months after the accident, wild organisms were exposed to just 10–70 μGy/h (0.24–1.68 mGy/d) due to the radioactive decay of short-half-life contaminants⁹ . Thereafter, Cs-134 and Cs-137 became the main sources of radiation to which wild organisms were exposed in Fukushima. These isotopes have longer half-lives (T_1/2 = 2 and 30 years, respectively).

The ICRP defines a set of reference species and recommends dosages for assessing and managing radiation exposure in wild organisms¹¹ . According to an ICRP report, mammals are the most sensitive of all organisms to radiation and show symptoms when exposed to > 0.1 mGy/d. Currently, more data on exposure are available for rodents than for other mammals¹¹ . In Japan, a candidate reference rodent species is the large Japanese field mouse (Apodemus speciosus). It is a medium-sized mouse with a long tail, large eyes, and a brown or orange-brown coat¹² , that dwells primarily on the ground but occasionally climbs trees¹² . Therefore, the species may be more exposed to radiation accumulated on the ground than other animals that are arboreal or have a longer distance between the soil surface and the body. The Japanese field mouse inhabits a wide range of habitats ranging from open fields to secondary lowland forests. Its mean home range is 304–1,853 m², smaller than that of other terrestrial wild mammals^12,13 . This characteristic makes it possible to evaluate the effects of contaminants, such as ionising radiation, in the local environment.

The biological effects of ionising radiation are well-known and can be categorised into two types. The first category includes deterministic effects that appear when the radiation dose exceeds a threshold; responses include ulcerations, permanent infertility, and death. The other category includes stochastic effects such as carcinogenesis. These effects occur in a dose-dependent manner, and resulting traits are inherited by the progeny of affected parents in the form of DNA lesions.

Stochastic effects occur either directly or indirectly due to radiation exposure. Recent studies on the absorbed radiation dose rate in wild rodents living in radiation-contaminated areas have shown that the contribution of external exposure, which is composed of gamma rays, is greater than that of the internal dose rate^4,10 . Gamma rays, a type of low-linear-energy transfer radiation, produce more indirect effects than direct effects¹⁴ . For example, radiation generates reactive oxygen species (ROS) such as superoxide radicals (O₂⁻), hydroxyl radicals (OH·), and hydrogen peroxide (H₂O₂) from water radiolysis. OH· is highly reactive and can produce several DNA modifications^15,16,17 . The most abundant and studied oxidative DNA base modification is 8-hydroxy-2′-deoxyguanosine (8-OHdG), which is produced by the reaction of OH· with guanine^16,18 . 8-OHdG pairs with adenosine and cytosine during DNA replication and causes a G: C to T: A transversion mutation¹⁹ . If such radiation-induced mutations occur in the germ cells of organisms inhabiting areas exposed to high radiation levels, the accumulated unexpected mutations can be inherited by their progeny, and the likelihood of hereditary diseases increases. However, organisms possess mechanisms for eliminating ROS and repairing DNA that counteract mutagenesis. Antioxidant enzymes such as superoxide dismutase (SOD), glutathione peroxidase, and catalase eliminate radiation-induced ROS. When DNA is damaged by radiation, DNA glycosylases that participate in base excision, such as 8-oxoguanine DNA glycosylase (OGG1), repair the DNA damage by hydrolysing oxidised deoxynucleotides (dNTPs).

OGG1 hydrolyses the N-glycosyl bond, releases 8-OHdG as a free base, and cleaves the sugar-phosphate backbone^17,20,21 . In mice, the highest expression level of OGG1 mRNA is observed in the testes, implying that OGG1 plays an important role in minimising germline mutations^21,22 . OGG1 mRNA expression is elevated by oxidative stressors^23,24,25 .

This study determined whether radiation-induced oxidation following the FDNPP accident affected the reproductive organs of large Japanese field mice. We compared the effects of radiation exposure on the testes of field mice captured near the highly contaminated Fukushima area, designated by the Japanese government as a difficult-to-return zone, with that in mice captured in other (uncontaminated) prefectures. Testicular SOD1 expression, 8-OHdG accumulation, and OGG1 transcription were compared among field mice captured from areas with different radiation levels.

Daily dose rate and total lifetime dose in mouse testis

Cs-134 and Cs-137 concentrations were higher in the bodies of mice from Fukushima than in those from Toyama and Aomori throughout the experimental period (Table 1). Radioactive Cs concentrations were transformed into internal doses in the mouse testes using the Electron Gamma Shower (EGS5) Monte Carlo simulation program. In 2012, at Fukushima, the average daily internal Cs-134 and Cs-137 dose rates and the external dose rate from radioactive substances in the environment in testis were 0.061 ± 0.019, 0.086 ± 0.026, and 0.338 ± 0.027 mGy/d, respectively (Fig. 1a). The average daily internal dose rates of Cs-134 and Cs-137 rapidly decreased to 0.005 ± 0.000 and 0.010 ± 0.001 mGy/d in 2013 and 0.014 ± 0.003 and 0.034 ± 0.008 mGy/d in 2014, respectively. A gradual decrease in the average external dose rate (0.226 ± 0.018 mGy/d in 2013 and 0.177 ± 0.020 mGy/d in 2014) was observed. The ratios of the internal dose rate to the total dose rate in mouse testes were 30.4% in 2012, 6.1% in 2013, and 21.5% in 2014 (Fig. 1a), suggesting that radiation exposure in mouse testes was mainly external rather than internal. In contrast to Fukushima, the total dose rate at Toyama and Aomori was < 0.005 mGy/d throughout the study period. Internal dose rates were < 10% of the total dose rate (Toyama: 1.5% in 2012, 0.1% in 2013, and 0.4% in 2014; Aomori: 7.2% in 2012, 1.6% in 2013, and 9.0% in 2014).

The lifetime dose to which each mouse testis was exposed was calculated using the animal’s age derived from the degree of tooth wear (Fig. 1b). The average total lifetime dose of mouse testis in Fukushima was 184.7 ± 26.4 mGy in 2012, 65.0 ± 6.7 mGy in 2013, and 80.0 ± 9.6 mGy in 2014 (Fig. 1b). In the control groups, the 2012 Aomori sample showed a total lifetime dose of 1.1 ± 0.1 mGy in the testis, but other samples had values below 1 mGy (Fig. 1b). Similar values and tendencies were observed when calculating whole-body exposure (Supplementary Fig. S1).

Evaluation of oxidative stress in mouse testes

For Fukushima samples, we first measured Sod1 expression. Sod1 encodes an enzyme that removes radiation-generated free radicals. Sod1 transcription levels in the testes of mice captured in Fukushima in 2012 were higher than those in mice from control sites. A significant difference in Sod1 transcription was observed between Fukushima and Aomori (Steel–Dwass; Fukushima (FK) vs. Toyama (TY): W = −3.259, p = 0.051; FK vs. Aomori (AO): W = −4.472, p < 0.001; TY vs. AO: W = −1.039, p = 0.794; Fig. 2a). In 2013 and 2014, testicular Sod1 expression in Fukushima mice decreased, but the values were not significantly different from those in Toyama and Aomori mice (Steel–Dwass; 2013, FK vs. TY: W = 0.000, p = 1.000, FK vs. AO: W = −2.877, p = 0.103, TY vs. AO: W = −2.210, p = 0.265; 2014, FK vs. TY: W = 1.244, p = 0.670, FK vs. AO: W = −0.387, p = 0.979, TY vs. AO: W = −0.956, p = 0.815; Fig. 2a).

We also investigated the accumulation of 8-OHdG in the testes and cauda epididymis to elucidate the extent of free radical DNA oxidation (Fig. 2b). The percentage of 8-OHdG-positive cells in the testicular seminiferous tubules of mice from Fukushima in 2012 was significantly higher than that of mice from Toyama and Aomori (Steel–Dwass; FK vs. TY: W = −4.472, p < 0.001; FK vs. AO: W = −5.020, p = 0.000; TY vs. AO: W = 0.1148, p = 1.000; Fig. 2c).

Overall, the proportion of accumulated 8-OHdG-positive cells decreased from 2012 to 2013; however, the proportion of accumulated 8-OHdG-positive cells in Fukushima mice remained significantly higher than that in mice at other sites (Steel–Dwass; FK vs. TY: W = −4.855, p < 0.001; FK vs. AO: W = −6.121, p = 0.000; TY vs. AO: W = −2.242, p = 0.261; Fig. 2c). In 2014, no maximum proportion of accumulated 8-OHdG-positive cells was found in mice captured in Fukushima, nor was there a significant difference in the proportion of 8-OHdG-positive cells between Fukushima and Aomori mice (Steel–Dwass; FK vs. TY: W = −4.768, p = 0.000; FK vs. AO: W = −0.817, p = 0.861; TY vs. AO: W = 2.191, p = 0.313; Fig. 2c).

When the positive cell rate was tabulated for each spermatogenesis stage using the data from five individuals with a particularly high 8-OHdG-positive cell rate, a high percentage of positive cells was found in stages 9–14 (Supplementary Fig. S2). In contrast, no positive cells were found in the cauda epididymis of mice captured in Fukushima in 2012 or 2013.

In 2012, the transcript level of Ogg1 was five times higher in Fukushima mouse testes than in those of mice from other sites (Steel–Dwass; FK vs. TY: W = −4.116, p = 0.003; FK vs. AO: W = −4.472, p < 0.001; TY vs. AO: W = −2.079, p = 0.343; Fig. 2d). In contrast, there were no significant differences in Ogg1 transcript levels between the sites in 2013 and 2014. In 2014, however, the expression of this gene was slightly higher in the testes of mice from Fukushima than in those from other sites (Steel–Dwass; 2013, FK vs. TY: W = 0.074, p = 1.000, FK vs. AO: W = 0.662, p = 0.897, TY vs. AO: W = 0.375, p = 0.966; 2014, FK vs. TY: W = −0.560, p = 0.936, FK vs. AO: W = −1.291, p = 0.673, TY vs. AO: W = −1.673, p = 0.514; Fig. 2d).

Relationship between oxidative stress markers, radiation dose rate, and lifetime dose in testis

The contribution of radiation to oxidative stress marker induction was examined by constructing linear mixed-effects models that included thresholds (Table 2). Akaike’s information criterion with correction (AICc) scores were determined without Sod1 or the lifetime dose in the mouse testis and body (Supplementary Methods and Supplementary Fig. S3). Subsequently, a model with a threshold was reconstructed. If the threshold value is in a range, the model was constructed using the median value. The coefficient values indicated a significant relationship between the daily dose rate in the testis and Sod1 mRNA expression, the ratio of 8-OHdG-positive cells, and Ogg1 mRNA expression (Table 2; Fig. 3a), and between the lifetime dose in the testis and the ratio of 8-OHdG-positive cells and Ogg1 mRNA expression (Table 2; Fig. 3b). For models of Sod1 and Ogg1, the assumption of normality of residuals was met (Shapiro–Wilk normality test; Table 2). A similar tendency was observed when calculating the whole-body exposure (Supplementary Fig. S4 and Supplementary Table S1).

This study serves as an addendum to our previous work on the effects of radiation on wild large Japanese field mice²⁶ . In the present study, we also included data from 2012, a year with higher radiation levels than 2013 and 2014, and calculated specific exposure levels (daily dose rate and total lifetime dose). The previous report only noted the ambient dose rates of trapping sites and caesium accumulation levels within the mouse bodies. The specific daily dose rate and the total lifetime dose were not calculated. To accurately understand the effects of radiation on biological processes, it is essential to calculate both the specific daily dose rate and the total lifetime dose. The previous study on mice captured in difficult-to-return zones in 2013 and 2014 reported that the apoptotic cell frequencies and the frequencies of morphologically abnormal sperm when compared with non-contaminated sites showed no significant statistical difference. In this study, we used oxidative stress markers to evaluate DNA damage caused by radiation.

The testes of large Japanese field mice collected from Fukushima were exposed to higher radiation dose rates than those of large Japanese field mice collected from the control sites. The external dose rate accounted for approximately 70% of the total dose rate. These results suggest that the Japanese field mice in Fukushima were mainly exposed to radiation from their habitats. Our results showed that internal exposure (dose rate and total lifetime dose, Fig. 1a, b) to Cs-134 and Cs-137 was higher in 2014 than 2013. This observed trend may be attributed to a transient elevation in the Cs concentration in the soil resulting from the deposition of falling leaves and rainwater runoff at the capture site. This phenomenon may affect the Cs concentrations in the food of mice, increasing their internal exposure levels.

In 2012, Sod1 transcription, 8-OHdG accumulation, and Ogg1 transcription in the testes of Fukushima mice were significantly greater than those in the testes of mice from control sites; however, the differences were no longer clear in 2013 and 2014 (Fig. 2a, c, d), presumably because of the small sample sizes in each study area and annual decreases in the total radiation dose rate. To clarify the general trends in the relationships between the radiation dose rate and levels of oxidative stress markers, we constructed linear mixed-effects models incorporating year and site as random effects. The same model was also constructed to determine the relationship between the lifetime dose and oxidative stress markers.

The models, including the threshold, were considered the best for all stress markers according to the AICc information criterion (Supplementary Fig. S3, Supplementary Table S2).

For the relationship between markers Sod1 and Ogg1 and the daily dose rate in mouse testes, the Sod1 and Ogg1 models were successfully constructed but showed no significant or negative intercepts and initial slopes (Table 2). Additionally, there was a threshold, but it was close to zero, due to the distribution of the data and lack of biological significance (Fig. 3a). Therefore, care must be taken when interpreting the meaning of the thresholds. When models were analysed by year, no significant Sod1 and Ogg1 models were established (Supplementary Data S1). One reason why a significant model could be constructed with combined data from 3 years but not with data from each year may be the annual environmental fluctuations during the trapping period (i.e., the breeding season). For example, temperature, rainfall, and food quality during the breeding season varied each year, which may have affected gene expression levels at the time of capture. By combining data from three years, the effect of these annual variations on the results can be mitigated. In contrast, when analysing data from each year separately, these variations may have a substantial impact, making it more difficult to obtain significant results. Considering that the maximum reported lifespan for the Japanese field mouse is 26 months¹² , combining the data from 3 years in our analysis averaged out these environmental factors. However, considering the discrepancies between the results of the combined three-year analysis and the yearly analyses, we could not definitively infer that the expression levels of Sod1 and Ogg1 increase in a dose-rate-dependent manner.

For 8-OHdG, a threshold was observed at 0.18 to 0.19 mGy/d for the daily dose rate in mouse testis (Supplementary Fig. S3 and Supplementary Table S2). The ratio of 8-OHdG-positive cells increased around this threshold, with the response being initially strong and slowing after the threshold (Fig. 3a). When 8-OHdG data were analysed for each year, models were constructed in 2012, 2013, and 2014 and a significant relationship between the daily dose rate and the ratio of 8-OHdG-positive cells in 2012 and 2013 was confirmed (Supplementary Data S1). After pooling three years of data, it was confirmed that significant models could also be developed as mentioned above. Therefore, the proportion of 8-OHdG-positive cells increases in a dose-rate-dependent manner. During the study period in Fukushima, the average internal and external radiation dose rates ranged from 0.1 to 1.0 mGy/d. According to the ICRP^11,27 , such a level of radiation induces certain effects in reference rats with a very low probability. The result suggested that one of the effects of radiation in large Japanese field mice in Fukushima could be the accumulation of 8-OHdG.

To determine the relationship between markers and the total lifetime dose in mouse testes, a linear mixed-effect model for Sod1, Ogg1, and the proportion of 8-OHdG-positive cells was attempted. The Sod1 model was not constructed (Table 2). On the other hand, significant Ogg1 models were successfully constructed (Table 2). When models were analysed by year, no significant models relevant to Sod1 and Ogg1 were established (Supplementary Data S1). These results suggested that Sod1 expression does not increase with lifetime dose. The transcriptional regulation of Sod1 is highly controlled, and mRNA levels of the gene increase in response to various stressors such as chemicals, heat shock, and irradiation²⁸ . Increased Sod1 transcription in response to elevated ROS levels has also been reported^29,30 . As oxidative stress markers respond quickly to stimuli, the dose rate per unit time may be a better fit than the total integrated dose. With regard to Ogg1, annual environmental fluctuations affected the construction of models, similar to their effect on the construction of Sod1 and Ogg1 models in relation to daily dose rate. When analyzing data from each year separately, these fluctuations may have a substantial impact, making it more difficult to obtain significant results. In contrast, analyzing combined data from three years averaged out these environmental factors. However, considering the discrepancies between the results of the combined three-year analysis and the yearly analyses, we believe that these results do not clearly indicate that the expression levels of Ogg1 increase in a total lifetime dose-dependent manner. Although it is unclear whether Ogg1 increases in a dose-dependent manner, as will be shown later, it is likely that DNA repair occurs at the observed Ogg1 expression levels.

Analysis of 8-OHdG also showed a threshold of 94–105 mGy for the lifetime dose in mouse testes. The ratio of 8-OHdG-positive cells increased around this threshold, with the response being initially strong and slowing after the threshold (Fig. 3b). When 8-OHdG data were analysed for each year, models were constructed in 2012, 2013, and 2014, and a significant relationship between the lifetime dose and the ration of 8-OHdG-positive cells in 2012 was confirmed (Supplementary Data S1). After pooling three years of data, it was confirmed that significant models could also be developed as mentioned above. Therefore, we concluded that the ration of 8-OHdG-positive cells is considered to increase in a lifetime dose-dependent manner. In addition, the results of X-ray irradiation experiments at a dose of 100 mGy using captive-bred Japanese field mice showed 8-OHdG accumulation in the testes of mice (Supplementary Note and Supplementary Data S1, S2 and S3). This result suggested that the phenomenon of a high level of 8-OHdG accumulation observed in the testes of mice from Fukushima was due to the radiation effect of radioactive materials in the environment. A high percentage of 8-OHdG-positive cells was found in spermatogenesis stages 9–14 in the testis, which suggests that stage 9–14 spermatocytes are highly sensitive to 8-OHdG. The accumulation of 8-OHdG leads to an increase in DNA mutations and carcinogenesis^31,32,33,34 . Previous studies have shown that Ogg1-deficient mice exhibit elevated hepatic 8-OHdG accumulation and DNA mutation frequencies^31,32 . Mice lacking mutT homologue 1 (Mth1), Ogg1, and mutY homologue (Mutyh), which encode base excision repair enzymes, exhibit an increased number of base substitution mutations from G to T in the germline³³ and a higher lung tumor incidence³⁴ . Additional supportive information has shown that this may increase mutation frequency in wild rodents distributed in areas contaminated with radioactive materials. Chromosomal aberrations have also been observed in wild rodents in contaminated areas. For example, small Japanese field mice (Apodemus argenteus) and house mice (Mus musculus) collected from Fukushima Prefecture were exposed to higher radiation levels (approximately 1–3 mGy/d) than those observed in the present study (maximum 0.91 mGy/d). The frequency of chromosomal aberrations in splenic lymphocytes tended to increase with increasing dose rate⁴ . Based on the above-reported data, large Japanese field mice inhabiting areas with higher radiation exposure than those in the present study may exhibit DNA mutations.

In contrast with the testicular results, 8-OHdG accumulation was not detected in the cauda epididymis of the mice captured in Fukushima. The dose rate and lifetime dose dependency of Ogg1 expression in mouse testes were not clear; however, the base excision repair system in large Japanese field mice functioned effectively even in highly contaminated areas of Fukushima, with most 8-OHdG likely repaired at the observed Ogg1 expression levels. Therefore, DNA mutations in sperm cells were effectively minimised by the DNA repair system in large Japanese field mice in the study area.

To confirm this, it is necessary to compare the genomic information of sperm cells in the cauda epididymis with that of somatic cells to confirm that DNA repair worked effectively. Moreover, extending our research focus, members of our team are currently investigating mutations in proteins related to DNA damage repair pathways. This research aims to understand how repair mechanisms for DNA and double-strand breaks operate under conditions of chronic radiation exposure. By exploring mutations in DNA repair proteins, we hope to elucidate how these proteins adjust under radiation stress and whether they are altered in a way that affects their function.

According to its ecological life history, the Japanese field mouse has two reproductive seasons: early spring and late summer^35,36 . The reported maximum lifespans in the wild are approximately 15 months in Hokkaido and 26 in Honshu, Japan. Considering the information and the estimated age in days, the field mice captured in 2012 represented the first generation exposed to the nuclear accident, the second generation (born in the summer of 2011), and the third generation (born in the spring of 2012). Similarly, field mice captured in 2013 were categorised as individuals from the third, fourth (born in the summer of 2012), and fifth generations (born in the spring of 2013). Finally, the field mice captured in 2014 were classified as individuals of the fifth, sixth (born in the summer of 2013), and seventh generations (born in the spring of 2014). In this study, oxidative stress markers in the mice captured in 2012 showed a clear reaction to radiation. As the dose rate declined, from March 2011 to August 2012, at least three generations of male mice showed signs of oxidative stress in their reproductive systems. Especially in the first generation, immediately after the accident, there may have been severe reproductive suppression, according to the radiation dose estimated from soil samples by Garnier-Laplace et al.⁷ . Our results indicate that the effects of oxidative stress have decreased since 2013.

However, a previous study reported enhanced spermatogenesis in mice captured in 2014³⁷ . The population of field mice inhabiting Akogi (external dose rate from radiocesium: 407–447 μGy/d), close to the study site of the present study, experienced an increase in the number of spermatogenic and proliferating cell-nuclear-antigen-positive cells per seminiferous tubule when compared with that observed in areas with a lower external dose rate—Tanashio (21 μGy/d) and Murohara (304~365 μGy/d). Thus, although oxidative stress levels tend to decrease, other markers, such as spermatogenesis and DNA mutations, can be used to assess the effects of radiation. Furthermore, the potential for radiation to cause genetic instability and its implications for future generations are of considerable concern. Genetic instability may indirectly cause transgenerational effects³⁸ . Additionally, oxidative stress is implicated in the development of genomic instability³⁹ . Therefore, although there are various limitations to research on wild animals, further research should be conducted on wild animals in combination with laboratory experiments to understand the impact of the Fukushima accident on wildlife, including how successive generations of field mice may carry genetic alterations initiated by their ancestors’ exposure to radiation forward.

In conclusion, this study showed that the testes of mice captured in a radiation-contaminated area of Fukushima were exposed to approximately 0.1–1.0 mGy/d, which, according to the ICRP, has a very low probability of affecting reproduction. Regarding the expression levels of Sod1 and Ogg1 in mouse testes, the study findings do not clearly conclude that they increase with increasing dose rate and lifetime dose. Significant accumulation of 8-OHdG in the testes of field mice inhabiting difficult-to-return zones in Fukushima was observed in 2012 and 2013; however, no 8-OHdG-positive cells were observed in the cauda epididymis. Thus, it is likely that DNA repair was effectively performed by enzymes in the testes, such as Ogg1. Although the accumulation of 8-OHdG increases DNA mutations, DNA mutations in sperm cells can be effectively minimised by the DNA repair system. To confirm this, it is necessary to compare the genomic information of sperm with that of somatic cells from mice captured in a radiation-contaminated area of Fukushima, which will confirm whether the mutation rate in the germ cells of the field mice distributed in the difficult-to-return zone was changed by the released radioactive material and whether a trans-generational effect exists.

Study area

Target organs of large Japanese field mice were sampled during each breeding season (July to September) from 2012 to 2014 (Supplementary Data S4) at three forest sites in Japan. If the number of individuals captured during the study period was small, animals captured in the spring breeding season (March–April) were used to satisfy the sample size for statistical analysis. One of the sites was a radiation-contaminated area in Fukushima Prefecture (37°35ʹN, 140°44ʹE). The Japanese government designated this region as a difficult-to-return zone since it was affected by the nuclear power plant accident. Two other sites, one in Aomori Prefecture (40°35ʹN, 140°57ʹE) and the other in Toyama Prefecture (36°35ʹN, 137°24ʹE), were used as controls. The gamma radiation air dose rates at ground level were measured simultaneously with mouse sampling using a portable environmental gamma survey meter (NHE20CY3-131By-S; Fuji Electric Co., Ltd., Tokyo, Japan). Air dose rate data were used to estimate the external daily radiation dose rate, as described below.

Animals

The mice were captured using Sherman-type live traps baited with sunflower seeds. At each study site, 50–80 traps were set at 10 m intervals within a 2 km radius from each coordinate. Captures were conducted once or five times per year. Each sampling period lasted 2–3 d. The traps were checked every morning, and captured mice were immediately transported to the laboratory. Sexually mature males weighing > 30 g⁴⁰ , having paired testes weighing > 0.95 g, and being reproductively active³⁶ were used in the experiments described below. Mice were euthanised by CO₂ asphyxiation and weighed. The testes and cauda epididymis were removed from each mouse and weighed. Half of one testis from each animal was fixed in Bouin solution at room temperature, and the other half was stored in QIAzol Lysis Reagent (QIAGEN, Hilden, Germany) at −20 °C. After removing the head and internal organs, the body was stored at − 20 °C.

The heads were removed for age determination. The ages of the mice in months were estimated based on molar wear and presented as median age in days (Supplementary Table S3).

Mouse bodies without internal organs or heads were used to measure the mouse body Cs concentration. Tissue was minced using a food processor, packed in a 100 mL polystyrene container, and sample weight and volume were measured. The Cs-134 and Cs-137 activities of the samples were measured using high-purity germanium (HpGe) detectors (GMX45P4-76, ORTEC, TN or GCW7023, Canberra Industries Inc., Marion, CT, USA). Gamma Studio software (SEIKO EG&G Co., Ltd., Tokyo, Japan) and Spectrum Explorer software (Canberra) were used to analyse the gamma-ray spectra. A standard radiation source (MX033U8PP, Japan Radioisotope Association, Tokyo, Japan) was used for the calibration. The radioactivity levels of Cs-134 and Cs-137 were corrected for radioactive decay to obtain expected values at the time of sampling. Values below the detection limit were considered 0 Bq/kg. The Cs concentrations in the mouse bodies were used to estimate the internal daily radiation dose rate.

All rodent capture experiments were approved by the National Institute for Environmental Studies for the analysis and experimentation of environmental samples contaminated with radioactive materials (Permission No. 平成24年10月10日) and performed according to the guidelines for the use of wild mammals of the Mammal Society of Japan⁴¹ and the Guidelines for the Care and Use of Laboratory Animals of the National Institute for Environmental Studies. The study was carried out in compliance with the ARRIVE guidelines.

Calculation of daily radiation dose rates in testes

The daily radiation dose rate in the testicles (mGy/d) was expressed as the sum of the daily external and internal radiation dose rates. The external daily radiation dose rate for the whole body was calculated based on the radiation air dose measured using a portable environmental gamma survey meter (NHE20CY3-131By-S; Fuji Electric Co., Ltd., Tokyo, Japan). The ambient dose equivalent rate was used to calculate the daily external dose rate using the following formula with the radiation converting factor of γ-ray 1.0⁴² :

Daily external dose rate (mGy/d) = (Ambient dose equivalent rate of trapping site [μSv/h]/1.2) ×ばつ (24 h/1) ×ばつ (radiation weighting factor of γ-ray/1,000).

The air dose rates were assessed at arbitrarily chosen locations at each site in 2012. Air dose rates were measured at five random locations at intervals of at least 10 m in 2013 and 2014; average values were used. Due to missing data for Toyama in 2014, the average values measured in 2012 and 2013 were substituted. We confirmed that the obtained results did not differ, and the data for Toyama in 2014 were hypothesised to be both the maximum value from 2013 and the minimum value from 2012. The dose rate of beta rays from the ground was not included because the maximum range of beta rays from Cs-134 and Cs-137 in water is < 2 mm⁴³ and the testes were covered with fur (2.8 mm)⁴⁴ and skin (0.8 mm)⁴⁵ and protected by the peritoneum and tunica albuginea; therefore, the contribution of beta rays was considered negligible.

Internal daily radiation dose rates in the testicles were estimated using the EGS5 program⁴⁶ with mouse body Cs concentration data. The energy distribution of the beta particles generated from Cs-134 and Cs-137 was calculated according to the Fermi theory⁴⁷ using experimental data⁴⁸ . The shape of the field mouse testes was regarded as ellipsoid.

The minor and major axes were calculated from the relational formula with testis weight (Ishiniwa et al., unpublished data). The composition of the testes was recorded as TISSUE and SOFT (ICRP; soft tissue), while the tissues surrounding the testes were recorded as ADIPOSE TISSUE (ICRP; adipose tissue). The composition data were obtained from the NIST website (http://physics.nist.gov/cgi-bin/Star/compos.pl?matno=261 and http://physics.nist.gov/cgi-bin/Star/compos.pl?refer=ap&matno=103). Cs was assumed to be uniformly distributed throughout the body of the mice. The specific radioactivities of Cs-134 and Cs-137 were determined based on Ge semiconductor detector measurements. In the model, it was assumed that beta rays were emitted in all directions from the Cs-134 and Cs-137 present. The probability of an interaction between the beta particles and tissue atoms was determined using the EGS5 program. The radiation dose rate of the testes was calculated from the simulated beta-particle energy deposition within a fixed period. MeV (cm² g^{− 1} s^{− 1}) was converted to Gy/h (J kg^{− 1} h^{− 1}).

Calculation of lifetime radiation dose in testes

The lifetime radiation dose (mGy) in the testicles of individual mice was defined as the sum of the lifetime external and internal radiation doses. The lifetime external radiation dose for each mouse was calculated as the daily external dose rate multiplied by the estimated age in days.

The total lifetime internal radiation dose for each mouse was calculated by multiplying the daily internal dose rate with the estimated age in days.

Immunohistochemistry staining for testicular 8-OHdG

Testes were fixed in Bouin’s solution and processed for paraffin embedding. Sections were cut at a thickness of 5 μm, dewaxed, washed with phosphate-buffered saline (PBS) for 5 min, and boiled in 10 mM citric acid (pH 6.0) for 5 min in an autoclave. The sections were then rinsed three times with PBS for 2 min. After 30 min of incubation at room temperature with goat serum albumin (3%), the sections were covered with the primary antibody, mouse anti-8-OHdG monoclonal antibody (1:20, Japan Institute for the Control of Aging, Shizuoka, Japan), and incubated overnight at 4 °C. The sections were then washed thrice for 2 min with PBS and covered with the secondary antibody, Histofine Simple Stain Rat MAX PO (MULTI; NICHIREI BIOSCIENCES Inc., Tokyo, Japan), for 20 min at room temperature. The sections were stained with diaminobenzidine (DAB; Histofine DAB Kit; NICHIREI BIOSCIENCES, Inc.). Haematoxylin was used as a counterstain, and the sections were observed using a NIKON Bioresearch ECLIPSE 80i Microscope. A negative control section was prepared using the same procedure described above, without the primary mouse anti-8-OHdG monoclonal antibody (Fig. 2b).

To calculate the ratio of 8-OHdG-positive cells in the testes, 100 seminiferous tubules were selected from tissue sections of each mouse. The seminiferous tubes were selected such that the three stages reported by Oakberg⁴⁹ and Perey et al.⁵⁰ , were represented by one-third each; between 31 and 34 seminiferous tubes were selected from each of the three stages. We then counted all the cells in the selected seminiferous tubules, classifying them as either 8-OHdG-positive or -negative. The percentage of 8-OHdG-positive cells was calculated by dividing the number of 8-OHdG-positive cells in selected seminiferous tubules by the total number of cells in the selected seminiferous tubules. For mice with a high percentage of 8-OHdG-positive cells, data showing the spermatogenesis stages³⁶ with a high percentage of 8-OHdG-positive cells are presented in Supplementary Figure S2. In addition, immunohistochemistry staining for 8-OHdG was also performed on the cauda epididymis of mice captured in Fukushima in 2012 (n = 11) and 2013 (n = 20).

Transcript levels of Sod1, Ogg1, and β-actin determined using quantitative real-time PCR

Total testicular RNA was isolated using the EZ1 RNA Universal Tissue Kit (QIAGEN). Single-stranded cDNA was synthesised from total RNA (1 μg) using PrimeScriptTM Master Mix (TaKaRa Bio Inc., Shiga, Japan). The transcript levels of Sod1, Ogg1, and β-actin mRNA were measured by quantitative real-time PCR performed using KOD SYBR^® qPCR Mix (TOYOBO CO., LTD., Okasa, Japan) and a Light Cycler^® 480 (Roche Diagnostics, Basel, Schweiz). The Sod1 and Ogg1 primers were as follows: As-SOD1_forward, 5′-TCCAT TGAAG ATCGA GTGAT CTC-3′; As-SOD1_reverse, 5′-TCACA CCACA AGCCA AGC-3′; As-OGG1_forward, 5′-GATTG CCCAT CGTGA CTAC-3′; and As-OGG1_reverse, 5′-CACTG AACAG CACTG CTTG-3′. The β-actin primers⁵¹ were as follows: As-ACTB_forward, 5′-CAGGA TGCAG AAGGA GATTA C-3′ and As-ACTB_reverse, 5′-ACCGA TCCAC ACAGA GT-3′. Real-time PCR amplifications of Sod1, Ogg1, and β-actin were performed with forty cycles of denaturing (95 °C, 10 s), annealing (58 °C, 10 s), and extension (68 °C, 30 s) using a single fluorescence measure following initial denaturation (98 °C, 2 min). After the amplification and quantification cycles, a melting curve analysis (65–95 °C with a heating rate of 0.1 °C/s and continuous measurement of fluorescence) was conducted to validate the specific PCR products. A standard curve was constructed by plotting the cross point (Cp) against the continuous dilution of a particular sample. The PCR efficiency (E) of one cycle was calculated as E = 10^−1/S (where S is the slope of the straight line between Cp and dilution). The relative concentration was calculated using the following equation: Concentration _{(SOD1 or OGG1)} = E_{(SOD1 or OGG1)}^{Cp(SOD1 or OGG1)}/E_(actb)^Cp(actb).

Statistical analysis

To compare values of oxidative stress markers among study sites each year, we conducted a Steel–Dwass test in R using the ‘NSM3’ software package⁵² . To assess the relationship between the radiation exposure daily dose rate and the degree of oxidative stress response, we used a linear mixed-effect model with the ‘lmerTest’⁵³ , ‘lme4’⁵⁴ , and ‘glmmTMB’ software packages⁵⁵ . The same model was used to assess the relationship between the lifetime dose of radiation and oxidative stress response. The explanatory variables were the total (external + internal) mouse testis radiation, body radiation dose rate, or lifetime dose, and year and site were included as random effects to account for confounding effects because the environment at each study site could have varied greatly among the three sites or from year to year. We constructed models for the response variables of Sod1 expression, 8-OHdG-positive cell proportion, and Ogg1 expression. The linear regression model assumes the normality of error terms⁵⁶ . The sample residuals are useful for assessing normality instead of error terms⁵⁷ . The normality of the residuals was confirmed using the Shapiro–Wilk test, which is the most powerful normality test⁵⁷ . For the 8-OHdG-positive cell data, models were constructed using ‘glmmTMB’, which can include the offset. The existence of ‘thresholds’ in the linear regression models was verified using AICc (Supplementary Methods). An α value of 0.05 was used to indicate statistical significance for all tests. All analyses were conducted using R ver. 3.6.2⁵⁸ . The above statistical analyses were conducted not only on the combined data from 2012, 2013, and 2014 but also on the individual data from each year.

All data used in this study are included in this manuscript and the Supplementary Materials.

We thank Dr. Yoshiki Tate for technical support with 8-OHdG immunostaining and Dr. Takehiko Hiwatari for assistance with the measurement of radioactive Cs in the samples. This study was financially supported by the project "Study on the dynamics of radioactive materials in a multimedia environment" commissioned by the Ministry of the Environment, Japan.

Authors and Affiliations

1. Ecological Risk Assessment and Control Section, Center for Environmental Biology and Ecosystem Studies, National Institute for Environmental Studies, 16-2, Onogawa, Tsukuba, Ibaraki, 305-8506, Japan

   Hiroko Ishiniwa & Manabu Onuma

2. Radioecologcial Transfer and Effects Division, Institute of Environmental Radioactivity, Fukushima University, 1 Kanayagawa, Fukushima, Fukushima, 960-1296, Japan

   Hiroko Ishiniwa

3. Ecological Genetics Analysis Section, Center for Environmental Biology and Ecosystem Studies, National Institute for Environmental Studies, 16-2, Onogawa, Tsukuba, Ibaraki, 305-8506, Japan

   Tsukasa Okano

4. Department of Radiation Biology, School of Veterinary Medicine, Rakuno Gakuen University, 582, Bunkyodai-Midorimachi, Ebetsu, Hokkaido, 069-8501, Japan

   Daiji Endoh

5. High Energy Accelerator Research Organization, 1-1, Oho, Tsukuba, Ibaraki, 305-0801, Japan

   Hideo Hirayama

6. Environmental Impact Assessment Section, Fukushima Regional Collaborative Research Center, National Institute for Environmental Studies, 10-2, Fukasaku, Miharu, Fukushima, 963-7700, Japan

   Akira Yoshioka

7. Faculty of Science, Academic Assembly, University of Toyama, Gofuku 3190, Toyama, Toyama, 930-8555, Japan

   Yasushi Yokohata

8. Laboratory of Wildlife Science, School of Veterinary Medicine, Kitasato University, 23-35-1, Higashi, Towada, Aomori, 034-8628, Japan

   Junji Shindo

9. Division of Bio-Resources, Department of Biotechnology, Frontier Science Research Center, University of Miyazaki, 5200, Kihara, Kiyotake, Miyazaki, 889-1692, Japan

   Chihiro Koshimoto, Akio Shinohara & Shinsuke H. Sakamoto

10. Department of Animal and Grassland Sciences, Faculty of Agriculture, University of Miyazaki, Gakuen-kibanadai-nishi-1-1, Miyazaki, 889-2192, Japan

    Shinsuke H. Sakamoto

11. Biodiversity Division, National Institute for Environmental Studies, 16-2, Onogawa, Tsukuba, Ibaraki, 305-8506, Japan

    Masanori Tamaoki & Manabu Onuma

Contributions

H.I., T.O., M.O., Y.Y., and J.S. collected the specimens; H.I., T.O., M.O., K.C., A.S., and S.H.S. performed the experiments; D.E. and H.H. performed the EGS5 simulations; A.Y. performed the statistical analysis; M.O. and M.T. designed the study; and H.I. and M.O. wrote the paper. All authors contributed extensively to the discussion regarding this work and reviewed the manuscript.

Corresponding author

Correspondence to Manabu Onuma.

Competing interests

The authors declare no competing interests.

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions