Impact of Environmental Radiation on the Health and Reproductive Status of Fish from Chernobyl.

Aquatic organisms at Chernobyl have now been chronically exposed to environmental radiation for three decades. The biological effects of acute exposure to radiation are relatively well documented, but much less is known about the long-term effects of chronic exposure of organisms in their natural environment. Highly exposed fish in freshwater systems at Chernobyl showed morphological changes in their reproductive system in the years after the accident. However, the relatively limited scope of past studies did not allow robust conclusions to be drawn. Moreover, the level of the radiation dose at which significant effects on wildlife occur is still under debate. In the most comprehensive evaluation of the effects of chronic radiation on wild fish populations to date, the present study measures specific activities of 137Cs, 90Sr, and transuranium elements (238Pu, 239,240Pu, and 241Am), index conditions, distribution and size of oocytes, as well as environmental and biological confounding factors in two fish species perch ( Perca fluviatilis) and roach ( Rutilus rutilus) from seven lakes. In addition, relative species abundance was examined. The results showed that both fish species are, perhaps surprisingly, in good general physiological and reproductive health. Perch, however, appeared to be more sensitive to radiation than roach: in the most contaminated lakes, a delay of the maturation of the gonads and the presence of several undeveloped phenotypes were evident only for perch and not for roach.


Introduction 27
Wildlife has been chronically exposed to environmental radiation from the Chernobyl accident 28 for the past 30 years. The biological effects of acute exposure to radiation in laboratory settings 29 have been relatively well studied (Frederica radiation database: www.frederica-online.org 1 ), but 30 relatively little is known about the effect of long-term chronic exposure of organisms in the 31 natural environment. The fate of wildlife remaining in the Chernobyl Exclusion Zone (CEZ) is 32 under debate and controversy continues on the dose rate at which significant environmental 33 impacts occur. Previous studies found no evidence of effects of radiation on aquatic 34 macroinvertebrate or mammal populations 2,3 whereas others found reduced abundance of insect, 35 spider, bird and mammal populations 456 at Chernobyl and Fukushima. Environmental studies on 36 the long-term effects of radionuclide contamination at Chernobyl are of crucial importance for 37 refining the environmental protection regulations, underpinning the public and political debate 38 on risks of exposure to ionizing radiation and predicting the long-term impact on the 39 environment of the more recent nuclear accident at Fukushima. 40 Fish are considered to be the most radiosensitive aquatic organisms 7 and have been highly 41 exposed in freshwater systems at Chernobyl since the accident on the 26 th of April 1986. At 42 Fukushima, both freshwater and marine fish have been exposed since the March 2011 accident. 43 At Chernobyl, the highest dose rate to fish after the accident was estimated to be 400 μGy/h 8 . 44 Doses rates rapidly declined after the accident due to decay of short-lived isotopes, decreased 45 bioavailability of 137 Cs and its accumulation to bottom sediment 9 . In the first month after the 46 accident, the 137 Cs activity concentration were the highest in prey fish whereas a few years later, 47 10 A standardised cross section of liver and gonad were fixed and processed according to standard 152 protocols described in SI. The liver sections were examined for microscopic pre-tumour and 153 tumour lesions according to BEQUALM and ICES criteria 18 , and lesions associated to nuclear 154 and cellular polymorphism, cell death, inflammation and regeneration. For the female gonad 155 sections, the distribution of immature or mature oocytes was determined by counting the number 156 of perinuclear and cortical alveolar oocytes in a defined surface area at magnification 10 using a 157 microscope (Zeiss axiozoom), and the relative frequency of a germ cell stage was calculated as 158 follows: (number of oocytes at a given stage/total number of oocytes) x 100. Oocyte surface was 159 measured using Zen Pro software. 160

Statistical analyses 161
Statistical analyses were performed using R version 3.1.2. After satisfying the assumptions of 162 the normal distribution of the residuals, generalised linear models were used. If the normality of 163 the residuals was not respected, a Kruskal-Wallis rank test was applied. When significant, a 164 Wilcoxon rank test and a Bonferroni correction of the α error were performed. Pearson 165 correlation tests were performed. Further details are provided in SI. 166 167

Results 168
Water chemistry 169 Electrical conductivity varied from 120 to 318 μS/cm and the pH from 6.3 to 8.6 at the water 170 surface of the lakes. Dissolved oxygen varied from 48 to 125% and the temperature from 15.6 to 171 11 20.1°C at the water surface of the lakes (Table S2). Surface water concentrations of NO3and 172 NO2varied from 49 to 259 μg/L and from 1.5 to 11.9 μg/L respectively across the lakes (Table  173 S2). Surface water concentrations of PO4 3varied from 1.4 to 15 μg/L across the lakes (Table  174 S2). Concentrations of toxic trace elements (As, Cd, Pb and U) were low (< 1 μg/L) in surface 175 waters of all lakes (Tables S3).  176 Full results are presented in SI. 177 Significant contamination of fish from the CEZ with 137 Cs,90 Sr,241 Am,239,240 Pu and 238 Pu 178 Thirty years after the Chernobyl accident, fish from the lakes located in the CEZ and in 179 Belarus (Svyatoye (M)) are still significantly contaminated with 137 Cs (p < 0.001) (Figure 2A). Fish from the CEZ are significantly contaminated with 241 Am 239,240 Pu and 238 Pu ( Figure S1, 195 Table S4). Concentration levels in liver were significantly higher than in muscle for both species 196 (p < 0.001). For instance, 241 Am concentration levels in liver were 7 to 11 times higher than in 197 muscle. Concentration levels of 241 Am 239,240 Pu and 238 Pu in liver and muscle of perch from 198 Glubokoye (H) were significantly higher than in perch from Yanovsky lake and Cooling Pond (p 199 < 0.05) ( Figure S1, Table S4). Concentration levels of 241 Am 239,240 Pu and 238 Pu in liver and 200 muscle of roach did not vary significantly across the lakes (p > 0.05) ( Figure S1). Further details 201 are provided in SI. 202

Dose rate to fish 203
The 90 Sr internal dose rates (whole body average) varied from 0.1 (Cooling Pond (H)) to 7.7 204 (Glubokoye (H)) μGy/h in roach and from 0 (Cooling Pond (H)) to 8.4 (Glubokoye (H)) μGy/h 205 in perch (Table 1). The 137 Cs internal dose rates ranged from 0 (Gorova (L), Dvoriche (L) and 206 Stoyacheye (L)) to 0.5 and 1.4 (Glubokoye (H)) μGy/h in roach and perch respectively. The 207 highest 137 Cs external dose rates to fish were calculated in lakes from the CEZ and varied from 208 5.9 μGy/h in Glubokoye (H) to 7.3 μGy/h in Cooling Pond (H) ( Table 1). The external dose to 209 fish from Svyatoye (M) was 10 times lower (0.7 μGy/h). The external doses were very low for 210 the three other lakes (L). The total  and  dose rate ranged from 7.6 μGy/h in roach from the 211 Cooling Pond (H) to 15.7 μGy/h in perch from Glubokoye (H) lake. The total  dose rate in 212 perch and roach ranged from 0.11 to 0.24 and from 0.01 to 0.02 μGy/h in the liver and muscle 213 13 respectively ( Table 2). The 241 Am internal dose rate contributes to 63% and 87% of the total  214 dose rate in liver and muscle of fish respectively. The 239,240 Pu internal dose rate contributes to 215 27% and 10% of the total  dose rate in liver and muscle of fish respectively. The 238 Pu internal 216 dose rate contributes to 10% and 4% of the total  dose rate in liver and muscle of fish 217 respectively. The total  dose rate contributes to 1.2-1.9% of the total (,  and ) dose rate in 218 fish from the three highly contaminated lakes. 219 Fish species abundance 220 The relative abundance of fish species does not differ between lakes (p = 0.59) therefore there 221 is no evidence of negative effects of radiation exposure on fish biodiversity ( Figure S2). 222

General health condition 223
The lengths of the fish were not significantly different across lakes (p = 0.85) ( Table S4) Glubokoye (H) were smaller than for perch from Stoyacheye (L) (p < 0.01) but were similar to 230 the FC of the perch from Svyatoye (L), Dvoriche (L) and Gorova (L) ( Table S5). The 231 hepatosomatic index (HSI) of perch did not significantly vary across sites (p = 0.5). The HSI of 232 roach from Glubokoye (H), Yanovsky (H) and Cooling Pond (H) were significantly higher than 233 for roach from Svyatoye (M) (p < 0.05) but were similar to the HSI of the roach from Dvoriche 234 14 (L) (Table S5). No disease nor gross tumours or malformations were recorded in any of the fish 235 collected. Parasites were observed in liver of the perch from Yanovsky (H), Gorova (L), 236 Stoyacheye (L) and Dvoriche (L) and the prevalence was 55%, 6%, 31% and 14% respectively. 237 The histological analyses of the liver did not reveal any pre-tumour (Foci of cellular alterations) 238 and tumour (Hepatocellular adenoma and carcinoma) lesions nor more lesions associated with 239 nuclear and cellular polymorphism, cell death, inflammation, regeneration and melano-240 macrophage centers in exposed fish. 241

Reproductive status 242
The gonadosomatic index (GSI) of perch and roach were significantly lower at Yanovsky ( The female perch and roach gonads and oocytes did not display any structural damage. 288 No chromosomal damage was evidenced in blood cells of exposed fish 289 The number of micronuclei did not significantly vary across the sites for both species (p = 0.14 290 > 0.05) (Table S6). 291

Discussion 292
Physico-chemical values correspond to good quality waters according to the European surface 293 water quality standards (OECD, Annex1) and nutrient concentrations are typical of oligotrophic 294 waters (nitrate <1-3 mg/L; phosphate < 0.04 mg/L, OECD, Annex1). 295 137 Cs, 90 Sr,241 Am,239,240 Pu and 238 Pu specific activities 296 Thirty years after the accident, the activity concentration of 137 Cs and 90 Sr are still higher than 297 the EU (1250 Bq/kg for 137 Cs; 750 Bq/kg for 90 Sr), Ukrainian (150 Bq/kg for 137 Cs; 35 Bq/kg for 298 90 Sr) and Japanese (100 Bq/kg) maximum permissible level for human consumption in some of 299 the lakes affected by the Chernobyl nuclear accident (see SI for details). 300 Activity concentrations of 137 Cs measured in perch from Glubokoye (H) and Svyatoye (M) 301 lakes were the highest and reach 7844 and 6090 Bq/kg respectively. For Svyatoye (M), which is 302 the most distant of the 7 lakes, the high values are due to the high initial amount of 137 Cs 303 deposited in this area, as well as hydrology and hydrochemistry of the water body. This is a 304 closed lake with a very low water exchange rate and low natural potassium concentration, which 305 explains the slow decontamination 10, 22 . 306 The most important pathway of 137 Cs accumulation in fish is through the diet route 23,24 . The 307 results evidenced a 137 Cs biomagnification phenomenon. As a carnivorous fish, the perch 308 significantly accumulates 2-3 times more 137 Cs than its omnivorous prey, the roach. This is 309 consistent with previous studies led on perch and non-predatory fish from Chernobyl lakes where 310 the 137 Cs contamination levels exceeded that of non-predatory ones by 2 in smaller fish 22

Doses span the lowest protection level for an ecosystem 328
The estimation of the radiation dose is discussed in Supplementary material. In the present 329 study, the total dose rate to roach and perch from the CEZ lakes was estimated to range from 7.6 330 to 14.1 μGy/h and from 7.9 to 15.7 μGy/h respectively. The dose rate in fish from Glubokoye 331 (H) lake was the highest especially due to the contribution of the high 90 Sr internal dose rate (7.7 332 and 8.4 μGy/h for roach and perch respectively). However, as a β emitter and mainly 333 accumulated in calcium-rich tissues, a higher dose is expected in near bone tissues than in other 334 tissues, more than a few millimetres distant. The total internal  dose rate was estimated to range 335 from 0.01 to 0.2 μGy/h in perch and roach which is approximately two orders of magnitude 336 below the total β and  total dose rate although the relative biological effectiveness of the  dose 337 is considered an order of magnitude higher. The doses observed in the study lakes span the