Glycophorin A mutant frequency in radiation workers at the nuclear power plants and a hospital

Mutation Research 501 (2002) 45–56

Glycophorin A mutant frequency in radiation workersat the nuclear power plants and a hospital

Mina Haa,c, Keun-Young Yoob, Soo-Hun Chob,c,∗a Department of Preventive Medicine, College of Medicine, Dankook University, Chung-Nam, Chun-An 330-714, South Korea

b Department of Preventive Medicine, Seoul National University, College of Medicine, Seoul 110-799, South Koreac Institute of Environmental Medicine, SNUMRC, Seoul 110-799, South Korea

Received 10 November 2001; received in revised form 19 December 2001; accepted 20 December 2001

Abstract

We studied to assess the validity of the glycophorin A (GPA) mutant assay as a biological marker of the cumulativeeffects of chronic low doses of ionizing radiation. In 144 nuclear power plants workers and 32 hospital workers, informationon confounding factors, such as age and cigarette smoking, was obtained through a self-administered questionnaire. Theinformation on physical exposure doses was obtained from the registries for radiation exposure monitoring and controlat each facility. The range of cumulative exposure doses were 0–12.02 cGy. GPA mutant assay was performed by the BR6method with modification using a FACScanTM flow cytometer. Potential confounders, such as, age and cigarette smoking habitsshowed increasing trends with GPA variants, but were not of statistical significance. The hospital workers showed higherfrequency of the GPA NO variant than nuclear power plant workers. Significant dose–response relationships were foundbetween cumulative exposure to radiation and variants levels by simple and multiple linear regression models. The slope ofregression equation of the dose–response of nuclear power plants workers was much smaller than that of hospital workers.These findings suggest that there may be dose-rate effects. In a population exposed to chronic low-dose radiation, the GPAassay shows potential to be used as an effective biologic marker for assessing the cumulative exposure dose although it couldnot be able to see a dose relation below 10 cGy of cumulative exposure dose. © 2002 Elsevier Science B.V. All rights reserved.

Keywords: Low-dose ionizing radiation; Somatic mutation; Glycophorin A mutation; Dose–response relationship

1. Introduction

A certain amount of controversy remains on thehealth effects of low doses of ionizing radiation.High-dose ionizing radiation studies on atomic bombsurvivors in Hiroshima and Nagasaki provides evi-dence of carcinogenicity. Research using classicalepidemiological studies and biomarkers to detect theearly effects induced by chronic low doses of ionizingradiation has provided some of the answers.

∗ Corresponding author. Fax.:+82-2-7474830.E-mail address: [email protected] (S.-H. Cho).

The glycophorin A (GPA) mutant frequency assayhas recently been reported as a persistent indicator ofpast radiation exposure in populations with high-doseexposure [1]. The exposure markers may provideinformation on biodosimetry, and can indicate theinternal, rather than the physical dose. Thus, expo-sure markers can provide information on the exposuredose if individual physical exposure monitoring is notavailable.

GPA is the major cell surface glycoprotein of humanerythrocytes and is present at about 5×105 copies percell. The measurement of GPA mutant frequencieshas been used as an index for the level of somatic

0027-5107/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved.PII: S0027-5107(02)00009-X

46 M. Ha et al. / Mutation Research 501 (2002) 45–56

cell mutation in bone marrow stem cells, for theassessment of background spontaneous mutationrates, and as biomarkers of exposure and the effectsof occupational carcinogens [2]. GPA assay reflect thetotal life-time exposure because the variants releasefrom the damaged bone marrow stem cells.GPA genehas been mapped to chromosome 4q and occurs intwo co-dominantly expressed allelic forms (M and N)which are identical except for two amino acid residues.For technical reasons the GPA mutant assay can onlybe performed in MN heterozygotes, but this applies toapproximately half the human population irrespectiveof race. Fluorescently labeled monoclonal antibodiesspecific for the M or N alleles enable rare varianterythrocytes which lack the expression of one or theother to be detected. In theory, it is possible to detectfour classes of variants, hemizygous variants lackingthe expression of one allele (NO or MO) and homozy-gous variants with double expressions of one allele(NN or MM). Therefore, we can detect the eventswhich lead to homozygosity, such as chromosomemissegregation, somatic recombination or gene con-version, in addition to point mutations and deletionsleading to hemizygous expression. In application, theassay has tended to be used for the detection of NOand NN variants. The GPA assay has several technicaladvantages as a routine population screening tool (notissue culture is involved, stability of red blood cells),but one disadvantage of the current method is that adetailed mutational spectrum cannot be obtained dueto the absence of nucleic acids in erythrocytes.

The atomic bomb survivals were one of the orig-inal populations to be investigated using the GPAmethod, and from the beginning, they showed signifi-cant radiation dose–response [3,4]. By now, well over1000 survivals have been studied, with remarkablyconsistent results [5,6]. Apart from the backgrounddifference, the GPA assay reflects total life-time expo-sure and so this biological dosimeter could be helpfulin situations where the physical or chemical doses areeither unavailable or uncertain.

A small number of studies have addressed the appli-cation of this biomarker to low-dose exposure popula-tions. In Sellafield nuclear facility (UK), GPA NO/NNvariants from 36 workers revealed no correlation withdose [7]. The study suggested that stable chromosomalaberrations are more sensitive indicators of chronicradiation exposure than GPA variants. Some studies

have also been done about the effects of exposure oncleanup workers at the Chernobyl nuclear reactor site,and here the GPA variant showed no significant dif-ference between exposed personnel and the controlsfor either the NO or the NN variant cell frequency inabout 150 cleanup workers [8]. On the other hand, ina similar study at the same site the GPA mutant fre-quency showed a significant dose-associated increasein 1500 cleanup workers [9]. Moreover, a later studysuggested that the dose-dependent effects show a muchshallower linear dose–response in workers with predo-minantly protracted exposures at low-dose rate. Nume-rous radiobiological studies involving a number ofexperimental systems, including specific locus muta-tion studies in mice, have demonstrated attenuated bio-logical effects of equal radiation doses when deliveredas fractionated doses as compared to a single dose orwhen delivered at low-dose rates [10,11]. The purposeof this study was to evaluate the use of the GPA mutantassay as a biomarker for low-dose ionizing radiationexposure. In Korea, the construction of nuclear powerplants began at Kori in 1978 and at present there arefour sites. Nuclear power plant workers and hospitalradiation workers were used as subjects in this study.

2. Materials and methods

2.1. Population studied

The subjects included for analysis in this paper were144 workers in two nuclear power plants and 32 work-ers in one university hospital, with the MN blood type.

The average age of subjects was 38.4-year-old forthe nuclear workers and 35.2 years for the hospitalworkers. With the exception of five hospital workers,all of the study subjects were male (Table 1). Bothgroups of workers had been monitored for exposureto radiation, with film badges or thermoluminiscentdosimeters (TLDs), throughout their working lives attheir worksites. The average individual dose of totalcumulative exposure was 2.09 cGy in nuclear workersand 0.92 cGy in hospital workers (Table 2).

2.2. Glycophorin A mutant assay

Blood samples were MN typed with commer-cial typing sera (Ortho Diagnostics, Raritan, NJ), as

M. Ha et al. / Mutation Research 501 (2002) 45–56 47

Table 1General characteristics of the study subjects

Variables Nuclear workers Hospital workers P-value

Number of subjects – 144 32 –Tenure (months) Mean± S.D. 166.2± 75.7 109.1± 80.2 <0.05∗Age (years) Mean (range) 38.4 (26–56) 35.2 (27–50) <0.05∗∗Sex Male 144 27 <0.05∗∗∗

Female 0 5 –Smoking No 70 13 <0.05∗∗∗

Yes 74 19 –Alcohol drinking No 32 8 –

Yes 112 23 –Past medical history No 136 29 –

Yes 8 3 –Family history of cancer No 120 30

Yes 24 2 –

∗ P < 0.05 based ont-test.∗∗ P < 0.05 based on analysis of variance.∗∗∗ P < 0.05 based onχ2-test.

stated earlier the GPA assay can only be performedon samples from heterozygous MN individuals, andthese comprise approximately half the population. Inthis study the BR6 method was used for the GPAassay using erythrocyte fixation, immunolabeling,and flow cytometric analysis [12]. Blood samplingwas performed for about 10 months. Within at least3 days from the day of sampling, every blood sam-ple was fixed with formalin. The fixed samples werecryopreserved [13], the method involved freezingmedia and placing on dry ice for 1 h and storing in

Table 2Radiation exposure doses of the study subjects

Variables Dose (cGy) Nuclear workers(n = 144)

Hospital workers(n = 32)

Recent 1 year Mean∗ 0.07 ± 0.17 0.13± 0.24Median 0 0.02Range 0–1.16 0–0.87

Recent 2 years Mean∗ 0.18 ± 0.36 0.18± 0.26Median 0.01 0.05Range 0–2.75 0–0.87

Recent 5 years Mean∗ 0.36 ± 0.53 0.45± 0.43Median 0.08 0.37Range 0–2.75 0.002–2.21

Total cumulative Mean∗ 2.09 ± 2.77 0.92± 1.29Median 0.93 0.43Range 0–12.02 0.002–6.79

∗ P < 0.05 based on Wilcoxon rank sum test.

a deep freezer at−80◦C until at most 1 week be-fore analysis. The samples were then immunostainedwith two fluorescently labeled monoclonal antibodiesspecific for the M- and N-forms of GPA. We usedBRIC-157-F as the anti-N antibody and 6A7 as theanti-M antibody, these were purchased from the Inter-national blood group reference laboratory (IBGRL).In case of 6A7, we conjugated biotin and followedby streptavidin-R-phycoerythrin (Caltag Laborato-ries, South San Francisco, CA). Flow cytometry wasused to determine the frequency of erythrocytes that


lacked fluorescence from the M-specific antibody.The labeled red blood cells population was thenexamined at high speed (3000–4000 cells/s) to quan-tify cellular fluorescence intensities using a BectonDickson FACScanTM (Becton Dickinson Immuno-cytometry Systems, San Jose, CA) flow cytometer.A total of 4 × 106 cells were analyzed from eachsample. Results were reported as the number of NOor NN variants per million cells.

2.3. Statistical analysis

Statistical analyses were performed using the SASsystem for Windows [14]. For each nuclear powerplant worker and hospital worker, the mean frequen-cies were calculated and compared using the Wilcoxonrank sum test. To show the dose–response relation-ships between radiation exposure dose and variantfrequency, regression analyses were undertaken usinggeneralized linear models including and excludingoutliers. The definition of outliers was≥30 × 10−6

of NO or NN variant frequency as described by Big-bee et al. [9]. Among the study subjects, two nuclearworkers showed high variant frequencies that fittedthe definition of outliers. Those two workers wereabove 40-year-old (42 and 44) and had no specificpast medical histories or radiation exposure histories,besides those at the nuclear power plants. The reg-ression slopes of nuclear workers were compared tothat of hospital workers. Regression analysis was alsoperformed on the group whose exposure doses werebelow the reported detection limit of 10 cGy [15].Regression analyses were done to assess the effect ofsmoking and age, and the non-parametric Wilcoxonrank sum test was used to compare smokers andnon-smokers.

In addition, multivariate analyses using a multipleregression model were performed to show dose–response adjusted for age, including and excludingoutliers.

To assess the stability of samples over the periodof sample collection and fixation, seven sampleswere assayed twice at 2–3 days intervals and corre-lation analysis was performed between these assays.The correlation coefficients of the repeated assayswere 0.8818, with statistical significance in the NOvariant and 0.5585, without significance in the NNvariant.

3. Results

The means of the work durations were significantlydifferent between the two groups. For nuclear work-ers it was 13.85 years (166.2 months) and for hospitalworkers 9.09 years (109.1 months). The proportionof current smokers was 59% among hospital workersand 51% among nuclear workers. In terms of alcoholintake, past medical history and family history, nosignificant differences were found between the groups(Table 1).

No subject had more than 2 cGy during the yearprior to the study. The distribution of total cumulativeexposure was extremely skewed to the left (Fig. 1).The arithmetic means of total cumulative exposurewere 2.089 cGy (S.D. = 2.766) for nuclear workersand 0.915 cGy (S.D. = 1.288) for hospital workers.While hospital workers showed significantly higherdoses than nuclear workers in terms of cumulativeexposure doses during the 1 year, recent 2 yearsand recent 5 years prior to the study, the hospitalworkers showed significantly lower doses than thenuclear workers in terms of the total cumulative dose(Table 2).

In terms of NO variant, hospital workers showedhigher variant frequency than nuclear workers. Themeans were 1.38 variant cells per 106 red blood cellsin nuclear workers and 3.35 in hospital workers. Therewas no difference between the two groups of workersin terms of the NN variant (Table 3).

The linear regression without outliers indicated thatthe number of NO variant cells had a significant pos-itive association with the total cumulative exposuredose in all the subjects. The parameter estimate was0.181× 10−6/cGy (Fig. 2). However, in the modelwith outliers, the significance of the positive associa-tion disappeared. In terms of NN variant, there werepositive association but it did not carry statistical sig-nificance with or without outliers.

Each group of nuclear and hospital workers showedsignificant dose–response relationships. Nuclearworkers were statistically significant for both NO andNN variant; the values of the slopes being 0.175 and0.034, respectively. Hospital workers also showedsignificant dose–response relationships for both typeof variants (1.881 for the NO variant and 2.332 forthe NN variant) (Fig. 3). The slopes of the hospitaland nuclear workers were statistically significantly


Fig. 1. Distribution of the reported physical doses (total cumulative dose). (a) Nuclear power plants workers; (b) hospital workers.

different for both types of variants (NO;P = 0.001,NN; P = 0.0631).

Although regression analyses of age showed in-creasing trends of variant frequency with age, the pa-rameter estimates were very small, 0.061× 10−6 peryear for the NO variant and 0.131× 10−6 per year

for the NN variant, but both were without statisticalsignificance (Fig. 4).

Smokers showed higher mean variant frequenciesthan non-smokers (Fig. 5). In smokers, the geometricmeans were 2.21×10−6 for the NO variant and 4.73×10−6 for the NN variant. In nonsmokers, these were


Table 3GPA variant frequencies among 106 red cells in exposed workers

Nuclear workers (n = 144) Hospital workers (n = 32)

NO variantsArithmetic meana,∗ 1.38 ± 4.58 3.35± 4.70Geometric meanb 0.98 (0.06–15.56) 13.18 (0.72–239.88)Median 0.25 1Range 0–42 0–18.75

NN variantsArithmetic meana,∗ 2.43 ± 7.55 2.72± 4.33Geometric meanb 2.99 (0.20–40.93) 11.22 (0.93–134.90)Median 0.75 1Range 0–72 0–21.5

a Arithmetic means and S.D.b Exponentiated values of the means and the asymmetric dispersion interval calculated from the logarithmic mean± S.D. of GPA

variant frequency.∗ P < 0.05 based on Wilcoxon rank sum test.

Table 4Multivariate dose–response analysis of GPA variant frequencies using a generalized linear model among 106 red cells

Populations Variables NO variants P-value NN variants P-value

All exposed subjects Dose 0.1663 0.2301 0.0230 0.9127Age 0.0479 0.3889 0.1291 0.1274

All exposed subjects (without the outlier) Dose 0.1919 0.0398 0.0955 0.3518Age 0.0066 0.8599 0.0272 0.5024

1.18×10−6 and 2.72×10−6, respectively. There wereno statistically significant differences between smok-ers and non-smokers.

In multivariate analyses adjusted by age, NO vari-ant showed a significant dose–response relationshipafter excluding outliers, but was not significant withoutliers included; the slopes were 0.192× 10−6/cGy(P = 0.0398) without outliers and 0.166× 10−6/cGy(P = 0.2301) with outliers (Table 4). NN variantswere not significant with or without outliers.

In a subgroup of subjects, whose total cumulativeexposure doses were less than 10 cGy, there was nosignificant dose–response relationship.

4. Discussion

As has been found in several previous studies,this study showed a wider range of frequency ofNN variant than NO variant, these findings areprobably because the NN variant has a greater

potential for spontaneous mutation than the NOvariant [16].

In the analyses of dose–response relationship whenexcluding outliers, the NO variant showed a signifi-cant positive relationship with total cumulative dose,but the NN variant did not. In a separate analysisof hospital and nuclear workers, however, the NNvariant frequencies showed significant positive asso-ciations with exposure doses in simple regressionmodels at each worker group. This seems to indicatethat similar environmental conditions in each workergroup may decrease the range of variant frequencies,and the findings obtained from this study correspondwith several other reports that NO variant is a moresensitive indicator than the NN variant of ionizingradiation exposure [7–9,15].

Hospital workers showed higher mean frequencyof the NO variant than nuclear workers but not of theNN variant. The range of variant for nuclear workerswas more than twice that of hospital workers, andin addition the regression slopes of nuclear workers


Fig. 2. Dose–response relationships after exclusion of outliers with values less than 30. (a) Dose–response curve of NO variant; (b)dose–response curve of NN variant.


Fig. 3. Dose–response relationships of GPA variant in each group. The line a and the closed circles are those of hospital workers and theline b and open circles of nuclear power plant workers. (a) Dose–response relationships of NO variant in each group; (b) dose–responserelationships of NN variant in each group.


Fig. 4. Glycophorin A variants distribution by age. (a) NO variant distribution by age; (b) NN variant distribution by age.


Fig. 5. Frequency of GPA variant distribution by cigarette smoking. (a) NO variant frequency by smoking; (b) NN variant frequency bysmoking.

were less than those of hospital workers. These areprobably because while the total cumulative exposuredoses and the mean tenure of hospital workers wasless than nuclear workers, the more recent exposuredoses of hospital workers, such as over the previ-ous 1–2 and 5 years, were significantly higher thanthose of nuclear workers. In other words, these resultsprovide some insights into the effect of dose-rate

on the induction of GPA mutations. Previously anumber of studies have shown that mutation frequen-cies are much lower at comparable total doses formice or human that are exposed at low-dose ratesthan at high-dose rates [9,15,17]. On the other hand,the slope of the regression line for NO variant was0.18× 10−6/cGy (0.192× 10−6/cGy in the multiplemodel) in the present study. Acute and highly exposed


individuals, such as atomic bomb survivors and Cher-nobyl victims, showed a dose–response in regressionlines of ca. 0.25 × 10−6/cGy, and those who hadbeen exposed to a relatively low level, such as Cher-nobyl cleanup workers, showed ca. 0.04× 10−6/cGy[9]. The dose–response values found in the presentstudy lay between the two. Among all of the exposedpeople, including previous and present studies, thenuclear workers involved in this study are the longestand lowest exposed group. The slope for the hospitalworkers was some 10 times higher value than that ofthe nuclear workers. As shown in Fig. 3, however, thesteep slopes for hospital workers may simply reflectthe single high point with a dose of 70 Gy, the workerinvolved had no specific medical history and was anon-smoker. He had worked since 1976, and was thelongest serving member in the department of nuclearmedicine, where the workers’ mean radiation expo-sure dose of total cumulative in the hospital was thehighest. His recent 5 years exposure dose was belowthe mean of recent 5 years exposure doses in the hos-pital and majority of his exposure dose came from theearly period of working in 1970s, when the systemof radiation protection was not strictly operated inthe hospital. In that early period (1976–1983), he hadrelatively high cumulative exposure dose (53.85 cGyper year). It seems that the single high point showeda high-dose rate effect during that early working pe-riod compared to the majority of hospital workers.However, additional comprehensive studies need tobe made in a larger population with other exposedgroups to confirm the dose-rate effect.

In this study, GPA variant frequencies showed in-creasing trends according to the age of the participants.The degrees of increase were 0.0610×10−6 per year inthe NO variant and 0.1309× 10−6 per year in the NNvariant, but these values were not significant. Thesefindings correspond with the results of previous stud-ies, in which the values were 0.13×10−6 in NN [20],0.14× 10−6in NO [19] and 0.17× 10−6 in MO [16].

Cigarette smoke is a well-known carcinogen.Smokers showed higher variant frequencies thannon-smokers in terms of both types of variant. Inhealthy people, not exposed to ionizing radiation,smoking level has been correlated with NO variantfrequency [18]. In the atomic bomb survival groupwho were classified as unexposed to ionizing radia-tion, NO variant frequency showed a slight increasing

trend according to smoking level, but with statisticalsignificance [19]. In the present study, total smokingpack-year did showed a small positive correlationwith variant frequency but this was not statisticalsignificant for either variant type.

The two outliers in the regression analyses basedstudy of dose–response relationships were more than40-year-old and in a previous study, most of the iden-tified outliers were also more than 40-year-old [18].

The three measurements of high variant frequen-cies, which seriously affected the regression curves,were due to two workers (Figs. 2 and 3). They hadvery long tenure of service (upper 95, 90% of tenure)and were 40- and 48-year-old. They were current andrecently-stopped smokers, of 20 and 4 pack-years,but had no specific past medical histories. In termsof their past histories of occupational exposures, theywere exposed to unusually high radiation doses forthe most recent 3–10 years compared to the otherpeople. Therefore, it is possible that their high variantfrequencies were due to the combined effects of oldage, smoking and high cumulative exposure doses,including specific short periods of very high doses.

The detection limit of the GPA assay for bio-dosimetry proposed by previous authors is 10 cGy ofcumulative exposure [15]. In this study, 56% of thesubjects had received less than a total dose of 10 cGy.There were no significant dose–response relation-ships in these low exposure groups in terms of eithervariant type, as was noted in other studies [15].

The correlation coefficients of repeated measuresin this study correspond with the nature of the accu-racy of the BR6 method [12], which has a theoreticallimitation due to Poisson counting errors caused byrun-to-run variability. The smaller correlation coeffi-cient (and no significance) of NN variants than NOvariants are similar to the findings of a previous study,which found a larger coefficient of variation of NNthan NO [2].

Acknowledgements

We wish to thank Professor Kee-Hyun Chang atthe Department of Diagnostic Radiology, ProfessorMyung-Chul Lee in the Department of NuclearMedicine, Professor Sung Whan Ha at the Depart-ment of Therapeutic Radiology, SNU Hospital and


Dr. Chong Soon Kim in Radiation Health ResearchCenter for enrolling subjects in this study and theirreview of this paper. This work was supported byGrant no. 981-0715-115-2 from the Basic Researchprogram of the KOSEF.

References

[1] M.L. Mendelsohn, The current applicability of large scalebiomarker programs to monitor cleanup workers, in: M.L.Mendelsohn, J.P. Peers, M.J. Normandy (Eds.), Biomarkersand Occupational Health, Joseph Henry Press, Washington,DC, 1995, pp. 9–19.

[2] R.H. Jensen, W.L. Bigbee, Direct immunofluorescencelabeling provides and improved method for the glycophorin Asomatic cell mutation assay, Cytometry 23 (1996) 337–343.

[3] R.G. Langlois, W.L. Bigbee, S. Kyoizumi, N. Nakamura,M.A. Bran, M. Akiyama, R.H. Jensen, Evidence for increasedsomatic cell mutations at the glycophorin A locus in atomicbomb survivors, Science 236 (1987) 445–448.

[4] S. Kyoizumi, S. Umeki, M. Hakoda, et al., Detection ofsomatic mutations at the glycophorin A locus in erythrocytesof atomic bomb survivors using a single beam flow sorter,Cancer Res. 49 (1989) 581–588.

[5] M. Akiyama, S. Kyoizumi, M. Hakoda, et al., Studies onchromosome aberrations and HPRT mutations in lymphocytesand GPA mutation in erythrocytes of atomic bomb survivors,in: M.L. Mendelsohn, R.J. Albertini (Eds.), Mutation andthe Environment. Part C. Somatic and Heritable Mutation,Adduction and Epidemiology, Progress in Clinical andBiological Research, Vol. 340C, Wiley, New York, 1990,pp. 69–80.

[6] M. Akiyama, Y. Kusunoki, S. Umeki, Y. Hirai, N. Nakamura,S. Kyoizumi, Evaluation of four somatic mutations assays asbiological dosimeters in humans, in: W. Dewey, et al. (Eds.),Radiation Research. Part A. Twentieth-Century Perspective,Academic Press, San Diego, 1992, pp. 172–176.

[7] J.D. Tucker, E.J. Tawn, D. Holdsworth, S. Morris, R.Langlois, M.J. Ramsey, P. Kato, J.D. Boice Jr., R.E. Tarone,R.H. Jensen, Biologic dosimetry of radiation workers at theSellafield nuclear facility, Rad. Res. 148 (1997) 216–226.

[8] D.H. Moore II, J.D. Tucker, I.M. Jones, R.G. Langlois, P.Pleshanov, I. Vorobtsova, R. Jensen, A study of the effectsof exposure on cleanup workers at the Chernobyl nuclearreactor accident using multiple end points, Rad. Res. 148(1997) 463–475.

[9] W.L. Bigbee, R.H. Jensen, T. Veidebaum, M. Tekkel,M. Rahu , A. Stengrevics, A. Auvinen, T. Hakulinen, K.

Servomaa, T. Rytomaa, G.I. Obrams, J.D. Boice Jr.,Biodosimetry of Chernobyl cleanup workers from Estoniaand Latvia using glycophorin A in vivo somatic cell mutationassay, Rad. Res. 147 (1997) 215–224.

[10] United Nations Scientific Committee on the Effects of AtomicRadiation, Sources, Effects and Risk of Ionizing Radiation,1993 Report to the General Assembly with Annexes, UN,New York, 1994.

[11] National Research Council, Committee on the BiologicalEffects of Ionizing Radiation, Health Effects of Exposureto Low Levels of Ionizing Radiation (BEIR V), NationalAcademy Press, Washington, DC, 1990.

[12] R.G. Langlois, B.A. Nisbet, W.L. Bigbee, D.N. Ridinger,R.H. Jensen, An improved flow cytometric assay for somaticmutations at the glycophorin A locus in humans, Cytometry11 (1990) 513–521.

[13] R.H. Jensen, W.L. Bigbee, Direct immunofluorescencelabeling provides an improved method for the glycophorinA somatic cell mutation assay, Cytometry 23 (1996) 337–343.

[14] SAS Institute, SAS/STAT Guide for Personal Computers,Version 6.02, SAS Institute, Cary, NC„ 1996.

[15] R.H. Jensen, R.G. Langlois, W.L. Bigbee, S.G. Grant, D.Moore II, M. Pilinskaya, I. Vorobtsova, P. Pleshanov, Elevatedfrequency of glycophorin A mutations in erythrocytes fromChernobyl accident victims, Rad. Res. 141 (1995) 129–135.

[16] J. Cole, T.R. Skopek, Somatic mutant frequency, mutationrates and mutational spectra in the human population in vivo,Mutat. Res. 304 (1994) 33–105.

[17] W.L. Russell, E.M. Kelly, Mutation frequencies in male miceand the estimation of genetic hazards of radiation in men,Proc. Natl. Acad. Sci. U.S.A 79 (1982) 542–544.

[18] W.L. Bigbee, J.C. Fuscoe, S.G. Grant, I.M. Jones, A.E.Gorvad, K. Harrington-Brock, C.L. Strout, C.B. Thomas,M.M. Moore, Human in vivo somatic mutation measuredat two loci: individuals with stably elevated backgrounderythrocyte glycophorin A (GPA) variant frequencies exhibitnormal T-lymphocytehprt mutant frequencies, Mutat. Res.397 (1998) 119–136.

[19] S. Kyoizumi, M. Akiyama, J.B. Cologne, K. Tanabe, N.Nakamura, A.A. Awa, Y. Hirai, Y. Kusunoki, S. Umeki,Somatic cell mutations at the glycophorin A locus inerythrocytes of atomic bomb survivors: implications forradiation carcinogenesis, Rad. Res. 146 (1996) 43–52.

[20] W.L. Bigbee, S.G. Grant, R.G. Langlois, R.H. Jensen, A.Anttila, P. Pfaffli, K. Pekari, H. Norppa, Glycophorin Asomatic mutation frequencies in Finnish reinforced plasticsworkers exposed to styrene, Cancer Epidemiol. BiomarkersPrev. 5 (1996) 801–810.

Documents

Glycophorin A mutant frequency in radiation workers at the nuclear power plants and a hospital