Somatic Mutations at the Glycophorin A (GPA) locus (2) has been found to be a sen- sitive indicator of exposure to in (GPA) Locus Measured in Red Cells of atom bomb survivors (3), individuals exposed in the Goiania accident (4), and in Liquidators Who Immigrated Chernobyl-exposed groups (5-7). to Israel In the present work, our objective was to apply the GPA somatic mutation assay to a Vered Y. Wishkerman,112 Michael R. Quastel,l group of Chernobyl clean-up workers who immigrated to Israel and to compare them Amos Douvdevani,2 and John R. Goldsmiths' with unexposed control subjects in Israel Departments of 1Nuclear , 2Nephrology, 3Epidemiology, Soroka and immigrants who came from various Medical Center and Faculty of Health Sciences, Ben-Gurion University of areas ofthe FSU. the Negev, Beer Sheva, Israel Materials and Methods

Glycophorin A (GPA) assays for human erythrocytes with gene expression loss and duplication Subjects phenotypes (N0, NN) were carried out on 15 Chernobyl clean-up workers (liquidators) who Of 41 volunteer clean-up workers who had immigrated to Israel within the preceding 5 years, 19 local Israeli controls, and 14 Russian (nonliquidator) immigrants. GPA phenotype variants in red blood cells of the 15 liquidators official papers identifying themselves as showed values ranging from 1 to 101 events/106 cells, with a mean ±SD of 25.6 ±7.0. In liquidators or clean-up workers at the comparison, the 19 Israeli controls had values ranging from 0 to 13 GPA events per 106 cells, Chernobyl , only 15 were identified as having the MN blood with a mean ± SD of 3.9 ± 0.8. The difference was highly significant (p< 0.001). Another group of 14 volunteer control subjects (nonliquidators) who had emigrated from the former to type. There were two groups of control Israel during the past 5 years showed values ranging from 0.0 to 35.0 events per 106 cells, with a subjects. The first group consisted of 40 mean ±SD of 6.1 ±2.7. The difference between this group and the liquidator group was Israeli residents who had not come from significant at p<0.01. The results are compatible with past exposure to radiation in the group FSU countries, of whom 19 had the MN identified as liquidators. - Environ Health Perspect 105(Suppl 6):1451-1454 (1997) blood type. The second group of 26 con- trols were emigrants from FSU countries to Keywords: radiation, Chernobyl, clean-up workers, liquidators, glycophorin A, biological Israel who were selected randomly; of these dosimetry, biological indicators of exposure 14 had the MN blood type. The age range of the liquidators with blood type MN was 30 to 63 years of age Introduction (73% male), Israeli MN controls were 20 On 26 April 1986 there was an explosion areas were exposed to variable amounts of to 69 years of age (47% male), and the and fire in reactor number 4 of the P- and y- irradiations from the emitted FSU MN controls were 27 to 64 years of Chernobyl power plant. More than 1018 radioisotopes. A subgroup of this popula- age (29% male). Bq of radioactive substances were released tion consisted of salvage or clean-up work- into the atmosphere, including 1.4 x 108 ers, called liquidators, sent to work during Methodology Ci of 1-131 and other short-lived radio- and after the accident in different jobs near Peripheral blood was obtained by venipunc- iodines that were dispersed by winds and and within the damaged reactor. Their ture from each volunteer. Each sample con- precipitated with rain. People over vast total number has been estimated to be as sisted of 5 to 10 ml of heparinized blood. high as 600,000. Although their exposure The samples were coded and each sample was supposed to have been limited to 25 was tested for determination of the blood cSv, the actual radiation doses that these type-MM, MN, or NN. This was done This paper is based on a presentation at the people received is not clear. using commercial anti-M and anti-N anti- International Conference on Radiation and Health held 3-7 November 1996 in Beer Sheva, Israel. During 1990, mass migration to Israel bodies obtained from Ortho Diagnostic Abstracts of these papers were previously published from the former Soviet Union (FSU) Systems Inc. (Raritan, NJ). Identification of in Public Health Reviews 24(3-4):205-431 (1996). started, and by 1996 about 700,000 people the blood types was carried out according Manuscript received at EHP 12 August 1997; accepted 2 October 1997. had come to Israel. Roughly 20% of this to the commercial instructions. We thank L. Merkin and S. Poijak for finding and number is estimated to have emigrated from The flow cytometric assay for N0 and interviewing the liquidators and making blood sam- radiocontaminated areas of Belarus, , NN phenotypes was performed according ples available. The authors express their appreciation and . Within this to the BR6 method of et al. to 1. Yevelson and J. Cwikel for locating appropriate immigrant group Langlois (8). control subjects from the former Soviet Union. were a limited number of dean-up workers. The assay enumerates the frequency of N0 Address correspondence to Prof. M.R. Quastel, The presence of these people in Israel has and NN phenotypes in a cell population of Institute of Nuclear Medicine, Soroka Medical Center, led to the possibility ofestimating the extent MN blood type. This procedure can be P.O. Box 151, Beer Sheva, Israel 84101. Telephone: 972-7-6400754 or 940. Fax: 972-7-6400765. E-mail: ofpast radiation exposure using various bio- performed only on the 50% of the popula- maayl [email protected] logical indicators (1). tion expressing the heterozygous MN phe- Abbreviations used: FACS, fluorescence-activated The somatic mutation test ofhuman ery- notype. The assay uses immunofluorescence cell sorter; FITC, fluorescein isothiocyanate; FSU, for- mer Soviet Union; GPA, glycophorin A; N0, NN, MN, throcytes with gene expression loss and labeling and flow cytometry to enumerate phenotypes of GPA alleles; SD, standard deviation. duplication phenotypes at the glycophorin A the frequency of red blood cells lacking

Environmental Health Perspectives * Vol 105, Supplement 6 * December 1997 1451 WISHKERMAN ETAL.

expression or showing duplication of one Discussion of the other (NN, MM, with the muta- allele. The fluorescent anti-M antibody Various biological methods have been tion being called homozygous). The nat- (6A7) labeled with rodamine, and the fluo- developed to estimate human exposure to ural frequency of variants was estimated rescent anti-N labeled with fluorescein ionizing radiation. These are based mainly to be 7.1 ± 5.6 per 106 cells (16), but nat- isothiocyanate (FITC) were obtained from on chromosomal damage and mutagenic uraly increases with age (17) and with the International Blood Group Reference changes at the cellular level (1). One sensi- smoking (18) have been observed. Laboratory (Bristol, U.K.). tive technique, originally described by The GPA assay was used to show The blood was fixed and labeled with Langlois et al. (2,9), measures the fre- elevated levels of red cell variants due to fluorescent monoclonal antibodies and quency of red cells carrying a mutation at somatic mutation in human groups exposed analyzed by flow cytometry using a Becton the GPA locus in red cells, presumably to radiation such as that from Hiroshima -Dickinson FACSTAR (San Jose, CA) flu- because of mutation of erythroid precur- and Nagasaki (2,3,19-22) and from the orescence-activated cell sorter (FACS). The sor cells. In this technique, fluorescent Goiania radiation exposure accident (4). time taken from drawing the blood to moncolonal antibodies specific for group Increases in the frequency of variants have FACS analysis was 5 to 14 days, during M or N alleles are used to identify the cell been found after exposure to mutagenic which time the whole blood samples were that expresses only one allele. It was found drugs (23) as well as in -susceptible kept at 40C. that the normal frequency of variant cells patients with ataxia telangiectasia (24) and The results are the sum of both N0 missing the M or N allele is about 1 in Bloom's syndrome (25,26). It has been and NN phenotypes. 105. Langlois et al. (8) also described an suggested, therefore, that the GPA somatic improved version of this method using a mutation assay has value for the assessment Statistical Analysis single-beam cytometer (BR6), which led to of cancer risk (20,27). Comparison was made between the results improved separation of normal and mutant Jensen et al. (5) evaluated clean-up obtained from the liquidators and the two phenotypic cells. workers from Russia and Ukraine and control groups using the Student's t-test The GPA gene is present in single copy found in 1 15 individuals a significant and the Mann-Whitney test. in the genome of chromosome 4 (10,11). increase of N0 red cells corresponding to The protein coding region of the GPA the estimated radiation exposure. The Results gene is 453 nucleotides in length and con- increased frequency of N0 variants was The results of the study are summarized in sists of 7 exons and 6 introns extending stable in 10 donors over a 7-year period. Figure 1. The 19 Israeli controls showed over 40 kb. The GPA gene occurs as two Somatic mutation analyses using the tight distribution within the range of 0 to 13 common alleles, GPAM and GPAN, that GPA assay were also carried out by per 106 cells. The mean ± SD was 3.9 ± 0.8. are responsible for MN blood groups. GPA Bigbee et al. (6,7) on Estonian clean-up In contrast, the 15 liquidators showed values can carry either blood group M or N anti- workers exposed to ionizing as ranging from 1 to 101 per 106 cells, with a gen, as determined by the amino acid a result of the. Chernobyl accident. mean ± SD of 25.5 ± 7.0. The difference was sequence at residue 1 and residue 5 of the However, in the latter group, the fre- highly significant atp< 0.001. mature glycoprotein that is. Ser1/Gly5 for quency of variant erythrocytes was low, To determine whether the difference M and Leul/Glu5 for N (12,13). These suggesting that the dose to the workers was due to radiation exposure at Chernobyl antigens differ by only 2 amino acids out did not exceed 10 cGy, probably too low or to differences between the people living of 131. The two alleles are co-dominantly a dose to result in detectable damage to in areas of the FSU and to those in Israel expressed as MN heterozygotes and erythroid stem cells. who did not come from the FSU, a second approximately equally represented in the The natural frequency of N0 red cells group of controls was obtained. The 14 human population, so that heterozygotes was reported by Jensen et al. to be 7.1 ± 5.6 MN control individuals in the latter group represent about 50% of all individuals. per 106 cells in U.S. residents not exposed had values ranging from 0.0 to 35.0, with a Hence, only 50% of the population can be to radiation (16) and 6.6 ±4.7 per 106 mean ±SD of 6.1 ±2.7 per 106 cells. analyzed because they are heterologous at cells (5) for unexposed Ukrainian and Compared with the liquidator group, the this locus (14). Russian subjects. These values compare difference was also significant at p < 0.01. Exposure to mutagens such as ionizing well with those for our Russian control radiation can cause damage to precursor group (6.1 ± 2.7 per 106 cells), whereas the 125-, erythroid cells of the bone marrow. Grant frequency we observed in Israeli residents and Bigbee (15) reported that the GPA was somewhat lower at 3.0± 0.8 per 106 assay distinguishes between two indepen- cells though the difference was not statisti- co 100- -a dent classes of variant cells. In one class, cally significant. co the expression of one allelic form of GPA In this study we have used the GPA ° 75- coe surface protein is lost, and in the second, assay to measure the frequency of muta- in addition to allelic loss the cells tions in red cells of workers , 50- express salvage (liq- a) the remaining homolog at twice the level uidators) of the Chernobyl accident who CD of the heterozygote. Thus, in individuals had immigrated to Israel. Our measure- 25- 0 having MN blood type, the GPA mutants ments, which were carried out 9 to 10 * *.

*. ^ . . . * . are recognized on rare red blood cell vari- years after the accident, showed signifi- Israeli controls Liquidators FSU controls ants as either the simple loss of M or N cant elevation in the number of variant expression (N0, M0), termed hemizy- red cells of liquidators compared to the Figure 1. Frequency of variant red cells having N0 or gous mutations, or loss of expression of number observed among Israeli and FSU NN pheonotypes. one of the alleles, with double expression immigrant control subjects.

l1452 Environmental Health Perspectives * Vol 105, Supplement 6 - December 1997 GLYCOPHORIN-A SOMATIC MUTATIONS IN RED CELLS OF LIQUIDATORS

These results suggest the presence of H = [(N0W -N0b)/slope ofstandard 3.9 ± 0.8 per 106 cells, whereas 15 liquidators clones of phenotypically variant red cells curve] x Q, showed values averaging 25.5 ± 7.0 per 106 in which the N antigen is missing or where cells. Using the slope based on the Goiania duplicated, presumably an indication of H = dose equivalent in Sv, data, a rough estimate of 1.7 Sv mean expo- mutated red cell erythroid precursor cells = frequency ofN0 variants in the sure dose is made. Using the slope from in the bone marrow. N0W Hiroshima, the estimate is 0.5 Sv. If the Kyoizumi et al. (20) reported a slightly radiation exposed person, value of 10 events per 106 cells is used for greater frequency of GPA N0 cells in N0b = frequency ofN0 variants in background frequency of variant cells, then males than in females (by -10%) in unex- unexposed controls the mimimal dose is 15.5/48=0.32 Sv. posed survivors in Hiroshima and Nagasaki, Q= quality factor Values in the range 0.3 to 1.7 Sv are consis- but no effect of sex on dose response was tent with average doses of 0.51 (acute) and observed. Although control groups con- The slope N0 per Gy was estimated by 0.96 Gy (chronic) estimated by Lucas (30) tained a lower proportion of males, com- Straume et al. (28) on the basis of results using fluorescence in situ hybridization for pared to the liquidator group, the marked obtained from Hiroshima atomic bomb sur- 12 exposed Chernobyl liquidators without difference reported between liquidators vivors and from Goiania radiation accident medical symptoms. Thus, these rough and and controls is not explainable on the basis victims. Values for Hiroshima survivors led rather tentative estimates suggest that of the natural male-to-female difference in to slope values of 48x 10-6 per Gy from radiation exposures occurred above 0.25 Sv, N0 frequency. data of Langlois et al. (3), and independent which was the provisional upper limit The effects noted in this study may be GPA measurements by Kyoizumi et al. (19) specified at the time ofthe accident. attributable to radiation exposure on the provided N0 values for Hiroshima sur- Nevertheless, it must be pointed out the part of the liquidators during the course vivors of 40 x 10-6 per Gy. These values nature of the exposures at Hiroshima dif- of their activities at the Chernobyl site were of course based on radiation exposures fered considerably from those of Goiania, 9 to 10 years earlier. Therefore, it was that occurred about 40 years before the and both differed in terms of intensity, important to compare these effects with GPA measurements were made. duration, source of radiation, and time of those of unexposed persons who had More recent exposures from the 1987 GPA analysis from the exposure incurred by emigrated from the FSU as well as with Goiania radiation accident were tested by the Chernobyl liquidators. Therefore, calcu- those of residents of Israel. The fact that Straume et al. (4). In this accident, Cs-137 lations of dose to liquidators based on GPA the emigrant FSU control subjects did exposure was received over a period ofabout analyses of atom bomb survivors and not show the effect on N0 frequency 2 weeks. Using dose esimates of Ramalho et Goiainia findings are not necessarily reliable. as seen in liquidators does not support al. (29), the slope was estimated to be The difference between these results, the possibility that mutagic agents 12.8 x 104 per Gy. which support those of Jensen et al. (5), (apart from radiation) in their previous The Chernobyl liquidators were exposed and the low values found in the studies by residential or working environments primarily to external 'shine' radiation and Bigbee et al. (6,7) with groups of liquida- caused these changes. inhalation of emitted radionuclides, primar- tors from Baltic countries is of interest, and A rough estimation of radiation expo- ily 3 and y, although some at emitters were may indicate that radiation exposures to sure dose based on the results of the GPA also present. Therefore, the quality factor can different liquidator groups were not uni- somatic mutation assay may be obtained by be assumed to be equal to one. form. Correlating the results of biological the method of Straume et al. (28) who Control variant cell frequencies (from indicators ofexposure with dose reconstruc- used the following equation: unirradiated Israeli residents) averaged tion estimates probably would be useful.

REFERENCES 1. Albertini RJ, Nicklas JA, Fuscoe JC, Skopek TR, Branda RF, 6. Bigbee WL, Jensen RH, Veidebaum T, Tekkel M, Rahu M, O'Neill JP. In vivo mutations in human blood cells: biomarkers Stengrevics A, Kesminiene A, Kurtinaitis J, Auvinen A, for molecular epidemiology. Environ Health Perspect 99: Hakulinen T et al. Glycophorin A biodosimetry in Chernobyl 135-141 (1993) cleanup workers from the Baltic countries. Br Med J 2. Langlois RG, Bigbee WL, Jensen RH. Measurements of the 312:1078-1079 (1996). frequency ofhuman erythrocytes with gene expression loss phe- 7. Bigbee WL, Jensen RH, Veidebaum T, Tekkel M, Rahu M, notypes at the glycophorin A locus. Hum Genet 74:353-362 Stengrevics A, Auvinen A, Hakulinen T, Servomaa K, Rytomaa (1986). T et al. Biodosimetry of Chernobyl cleanup workers from 3. Langlois RG, Bigbee WL, Kyoizumi S, Nakamura N, Bean and using the glycophorin A in vivo somatic cell MA, Akiyama M, Jensen RH. Evidence for increased somatic mutation assay. Radiat Res 147:215-224 (1997). cell mutations at the glycophorin A locus in atomic bomb sur- 8. Langlois RG, Nisbet BA, Bigbee WL, Ridinger DN, Jensen vivors. Science 236:445-448 (1987). RH. An improved flow cytometric assay for somatic mutations 4. Straume T, Langlois RG, Lucas J, Jensen RH, Bigbee WL, at the glycophorin A locus in humans. Cytometry 11:513-521 Ramalho AT, Brandao-Mello CE. Novel biodosimetry meth- (1990). ods applied to victims of the Goiania accident. Health Phys 60: 9. Langlois RG, Bigbee W'L, Jensen RH. Flow cytometric charac- 71-76 (1991). terization of nomal and variant cells with monoclonal antibod- 5. Jensen RH, Langlois RG, Bigbee WL, Grant SG, Moore D ies specific for glycophorin A. J Immunol 134:4009-4017 2nd, Pilinskaya M, Vorobtsova I, Pleshanov P. Elevated fre- (1985). quency of glycophorin A mutations in erythrocytes from 10. Siebert PD, Fukuda PJ. Human glycophorin A and B are Chernobyl accident victims. Radiat Res 141:129-135 (1995) encoded by separate, single copy genes coordinately regulated

Environmental Health Perspectives * Vol 105, Supplement 6 * December 1997 1453 WISHKERMAN ETAL.

by a tumor-promoting phorbol ester. J Biol Chem 261: 21. Langlois RG, Akiyama M, Kusunoki Y, DuPont BR, Moore 12433-12436 (1986). DH II, Bigbee WL, Grant SG, Jensen RH. Analysis of somatic 11. Shows TB, McAlpine PJ. The 1981 catalogue of assigned cell mutations at the glycophorin A locus in atomic bomb sur- human genetic markers and report of the nomclature commit- vivors: a comparative study of assay methods. Radiat Res tee. Cytogenet Cell Genet 32:221-245 (1982). 136:111-117 (1993). 12. Furthmayer H. Structural analysis of a membrane glycoprotein: 22. Akiyama M, Kyoizumi S, Kusunoki Y, Hirai Y, Tanabe K, glycophorin A. J Supramol Struct 7:121-134 (1977). Cologne JB. Monitoring exposure to atomic bomb radiation by 13. Siebert PD, Fukuda PJ. Molecular cloning of a human gly- somatic mutation. Environ Health Perspect 104(Suppl cophorin B cDNA: nucleotide sequence anr genomic relation- 3):493-496 (1996) ship to glycophorin A. Proc Natl Acad Sci USA 84:6735-6739 23. Bigbee WL, Wyrobek AJ, Langlois RG, Jensen RH, Everson (1987). RB. The effect of chemotherapy on the in vivo frequency of 14. Race RR, Sanger R. Blood Groups in Man. 6th ed. glycophorin A 'null' variant erythrocytes. Mutat Res 240: Oxford:Blackwell, 1975;92-188. 165-176 (1990). 15. Grant SG, Bigbee WL. In vivo somatic mutations and segrega- 24. Bigbee WL, Langlois RG, Swift M, Jensen RH. Evidence for tion at the human glycophorin A (GPA) locus:phenotypic vari- an elevated frequency of in vivo somatic cell muations in ataxia ation encompassing both gene-specific an chromosomal telangiectasia. Am J Hum Genet 44:402-408 (1989). mechanisms. Mutat Res 288:163-172 (1993). 25. Langlois RG, Bigbee WL, Jensen RH, German J. Evidence for 16. Jensen RH, Grant SG, Langlois RG, Bigbee WL. Somatic cell increased in vivo mutation and somatic recombination in Bloom's genotoxicity at the glycophorin A locus in humans. Prog Clin syndrome. Proc Natl Acad Sci USA 86:670-674 (1989). Biol Res 372:329-339 (1991). 26. Kyoizumi S, Nakamura N, Takebe H, Tatsumi K, German J, 17. Akiyama M, Kyoizumi S, Hirai Y, Kusunoki Y, Iwamoto KS, Akiyama M. Frequency of variant erythrocytes at the gly- Nakamura N. Mutation frequency in human blood cells cophorin-A locus in two Bloom's syndrome patients. Mutat increases with age. Mutat Res 338:141-149 (1995). Res 214:215-222 (1989). 18. Bigbee WL, Grant SG, Langlois RG, Jensen RH, Auttila A, 27. Akiyama M, Umeki S, Kusunoki Y, Kyoizumi S, Nakamura N, Pfaffli P, Pekari K, Norppa H. Glycophorin A somatic cell muta- Mori T, Ishikawa Y, Yamakido M, Ohama K, Kodama T et al. tion frequencies in Finnish reinforced plastics workers exposed to Somatic-cell mutations as a possible predictor of cancer risk. styrene. Cancer Epidemiol Biomarkers Prev 5:801-810 (1996). Health Phys 68:643-649 (1995). 19. Kyoizumi S, Nakamura N, Hakoda M, Awa AA, Bean MA, 28. Straume T, Lucas JN, Tucker JD, Bigbee WL, Langlois RG. Jensen RH, Akiyama M. Detection of somatic mutations at the Biodosimetry for a radiation worker using multiple assays. glycophorin A locus in erythrocytes of atomic bomb survivors Health Phys 62:122-130 (1992). using a singlebeam flow sorter. Cancer Res 49:581-588 (1989). 29. Ramalho AT, Nascimento ACH, Natarajan AT. Dose assess- 20. Kyoizumi S, Akiyama M, Cologne JB, Tanabe K, Nakamura ments by cytogenetic analysis in the Goiania (Brazil) radiation N, Awa AA, Hirai Y, Kusunoki Y, Umeki S. Somatic cell muta- accident. Radiat Prot Dosim 23:97-100 (1988). tions at the glycophorin A locus in erythrocytes of atomic 30. Lucas JN. Chromosome translocations: a biomarker for retro- bomb survivors: implications for radiation carcinogenesis. spective biodosimetry. Environ Health Perspect 105(Suppl 6) Radiat Res 146:43-52 (1996). 1433-1436 (1997).

1454 Environmental Health Perspectives * Vol 105, Supplement 6 * December 1997