Chromosomal aberration and sister-chromatid exchange frequencies in peripheral blood lymphocytes of a large human population sample

Mutation Research, 204 (1988) 421-433

421

Elsevier MTR 04474

Chromosomal aberration and sister-chromatid exchange frequencies in peripheral blood lymphocytes of a large human population sample * Michael A Bender 1, R. Julian Preston 2, Robin C. Leonard 1, Beatrice E. Pyatt 1, P. Carolyn Gooch 2 and Michael D. Shelby 3 1 Medical Department, Brookhaven National Laboratory * *, Upton, N Y 11973, 2 Biology Division, Oak Ridge National Laboratory * * *, P.O. Box Y, Oak Ridge, TN 37830, and 3 National Toxicology Program, National Institute of Environmental Health Sciences, P.O. Box 12233, Research Triangle Park, NC 27709 (U.S.A.)

(Received 7 April 1987) (Accepted 11 June 1987)

Keywords: Peripheral blood; Lymphocytes; Chromosomal aberrations; SCE frequencies; Population sample.

Summary In order to assess the potential of cytogenetic determinations on peripheral blood lymphocytes as a means of monitoring human populations subject to low level occupational and environmental exposures to chemical mutagens and carcinogens, accurate baseline data are required. Accordingly, we have determined mean frequencies of chromosomal aberrations and of sister-chromatid exchanges, their variances, and the sources of this variance in a cohort of 353 healthy employees of the Brookhaven National Laboratory. A detailed protocol was adopted for blood sampling, lymphocyte culture, cytogenetic preparation and scoring in order to minimize variation from these potential sources. Scoring was divided between the Oak Ridge and the Brookhaven groups with duplicate scoring sufficient to evaluate and minimize the effect of any differences between laboratories or between individual scorers. In all, the data include 71 950 cells scored for chromosomal aberrations and 16 898 cells scored for sister-chromatid exchanges. The mean unadjusted frequency of sister-chromatid exchanges was 8.29 + 0.08/cell. As reported in other studies, cigarette smoking very significantly influenced sister-chromatid exchange frequencies; in our study the mean for smokers was 9.0 + 0.2, while that for non-smokers was 8.1 + 0.1/cell. The mean frequency was statistically higher in females than in males, regardless of smoking status. On the other hand, age of the subject did not significantly influence sister-chromatid exchange frequencies. Curiously, the subject's total white cell count did influence sister-chromatid exchange frequency. N o other source of variation was found. The frequencies of chromosomal aberrations of all types were determined. The frequency of the most c o m m o n unequivocal chromatid type, the chromatid deletion, was 0.81 + 0.05%, that of the most common unequivocal chromosome type, the dicentric, was 0.16 + 0.02%. N o statistically significant influence was

Correspondence: Michael A Bender, Medical Department, Brookhaven National Laboratory, Upton, NY 11973 (U.S.A.). * Research supported by Interagency Agreements Y01-ES20099 and Y01-ES-20101 between the National Toxicology Program and U.S. Department of Energy. * * Operated by Associated Universities, Inc. for the U.S. Department of Energy under Contract No. DE-ACO 2-

* *

76CH000160; accordingly, by acceptance of this article, the publisher and/or recipient acknowledges the U.S. Government's fight to retain a nonexclusive, royalty-free license in and in any copyright covering this paper. * Operated by Martin Marietta Energy Systems, Inc. for the U.S. Department of Energy under Contract No. DEAC05-840R-21400.

0165-1218/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)

422

found of age or sex, nor of any other parameter tested, on the frequency of any chromosomal aberration type, with the single exception of long acentric fragments, often "supernumerary", believed to represent X chromosomes precociously separated at the centromere. Such fragments were significantly more frequent in samples from females than those from males, and showed a significant positive regression on age.

Determination of chromosome aberration frequencies in human peripheral blood lymphocytes (PBL) allowed to divide in short-term, phytohemagglutinin-stimulated tissue culture is a wellestablished technique for monitoring exposures to ionizing radiation, useful as a biological dosimeter in individuals (Bender and Gooch, 1966) and as a population monitor for occupational exposure (Evans et al., 1979; Lloyd et al., 1980). The parallel use of cytogenetic endpoints in PBL to monitor occupational or environmental human population exposures to chemical mutagens and carcinogens has often been advocated, and actually attempted in some cases. In order to evaluate and properly design cytogenetic population monitoring studies it is of course necessary to know as much as possible about normal spontaneous background frequencies and their variances and about the factors (confounding variables) influencing them. Extensive information of this sort has been accumulated for the chromosome-type chromosomal aberrations useful for dosimetry and monitoring in the ionizing radiation cases. Unfortunately, ionizing radiation and most chemical mutagens and carcinogens produce initially aberrations of the chromatid types, not chromosome types. Also, sister-chromatid exchange frequencies, a sensitive measure of chemical damage, are influenced little by ionizing radiation. Feeling that what information there was in the literature on the chromatid aberration types and on SCE frequencies was unsatisfactorily sparse, we undertook the present study. Because of possible (and in some cases, reported) influence of such things as culture conditions, inoculum size and scoring criteria on the cytogenetic result of such studies, we adopted a rigid study protocol designed to minimize such possible sources of variation. Because of the size of the study contemplated, it was undertaken as a collaboration between our groups at the Brookhaven and the Oak Ridge National Laboratories

with the scoring burden shared equally by the two. Because cytogenetic scoring clearly involves elements of subjective judgment, material from approximately 10% of the samples was scored independently at both laboratories so that interlaboratory and interscorer variability could be assessed. To provide another measure of variability, particularly with time, some subjects were sampled more than once over the 3-year course of the study. Materials and methods

Study population Subjects were normal healthy volunteers from the Brookhaven National Laboratory workforce. During the initial phase of the study 103 volunteers (primarily Medical Department employees) were sampled with no particular sampling scheme in order to provide material with which to evaluate the other protocols adopted. For the bulk of the study, a stratified random sampling scheme was set up. In April 1983, Brookhaven National Laboratory employed 3173 regular full-time workers. About 21% of these were doctoral-level scientific staff; the other 79% were support personnel working in professional, administrative, technical, clerical, and manual occupations. White males constituted 68.7% of this total; white females, 17%. With the cooperation of the Director's office and the Personnel Department, 4 lists were generated - - white males, white females, non-white males, and non-white females. Using the last digit of the BNL identification number (the "life number") and a table of random digits, 250 names were selected from each of the lists of white males and white females, and 50 were selected from the lists of non-white males and non-white females. Letters were sent to these 600 employees' mail-drops explaining the design and purpose of the study and asking for their cooperation. 282 positive responses were obtained, a response rate of 47%; 20

423 of these had also participated in the earlier volunteer stage of the study. Thirty of these respondents were dropped subsequently prior to sampling due to their own withdrawal or because they met one or more of the exclusion criteria - - previous chemotherapy, previous radiotherapy, or viral infection within 1 week of scheduled sampling. The distribution of departments of employment in the sample was not significantly different from that of the laboratory as a whole. All blood sampling was done at Brookhaven, as were all of the tissue cultures and fixations, as well as most of the slide preparation.

Cytogenetic protocol To minimize this possible source of variation, a rigid cytogenetic protocol was adopted for this study. Following the suggestion of Bloom et al. (Bloom, 1981) we avoided media known to favor the expression of fragile sites; the culture medium adopted was RPMI 1640 (Gibco). 10 ml of medium containing 15% fetal calf serum (Gibco, virus screened, and heat inactivated; all was of the same lot number and kept frozen for use in this study), antibiotics and supplemental glutamine, were placed in sterile, capped 15-ml conical plastic centrifuge tubes. An aliquot of each blood sample was used to obtain a total white cell count with a Coulter Model S cell counter. To each culture we added the amount of whole blood required, based upon this count, to give 5 × 106 leukocytes per culture. 0.25 ml of reconstituted phytohemagglutinin (Gibco) was added to each culture prior to incubation at 37 o C. The caps were tightened and the tubes slanted at an angle of 60-75 ° from vertical. 2 h prior to harvesting, colcemid (Gibco) was added to each culture to give a final concentration of 0.1 g g / m l . SCEs are commonly scored in lymphocytes which have replicated their D N A twice in the presence of the thymidine analogue 5-bromodeoxyuridine (BrdUrd). Since the baseline frequencies are sensitive to the concentration of the analogue, increasing from the lower limit imposed for obtaining satisfactory differentiation up to a saturation level, it has been advocated that levels above that necessary to produce saturation should be used in order to preclude variation in SCE frequency due to unequal numbers of lymphocytes

competing for the analogue available (Carrano and Moore, 1982). However, as we desired the greatest sensitivity possible for this cytogenetic endpoint (i.e., the best signal-to-noise ratio), we adopted the alternate strategy of using a low BrdUrd concentration (25 /~M) and seeding each culture with the same number of leukocytes. While it would have been more desirable to seed with a known number of lymphocytes, rather than of total leukocytes, waiting for the differential counts required would have unduly delayed preparing the cultures. Differential counts were made, however, and were available for use later in the data analysis. As has often been pointed out, ideally only first in vitro mitoses should be analyzed for chromosomal aberrations because aberrations are lost or appear in altered form in subsequent cell divisions. Fixation after a particular (early) incubation time is often advocated to preclude this problem, but in our experience, there is no single fixation time at which one can be certain of obtaining only first in vitro mitoses, and at which one can also be sure of obtaining an acceptably high mitotic index to ensure that a culture will afford enough well-spread metaphases for analysis. One solution, often applied in the analysis of ionizing-radiation-induced aberrations (Bender, 1979; Scott and Lyons, 1979) is to score only unequivocal first in vitro mitoses in differentially stained preparations from BrdUrd-containing cultures. In the in vivo ionizing radiation case as well as for G O in vitro treatments with a few chemical clastogens, the BrdUrd incorporation during culture was shown not to influence aberration frequencies (Bender, 1979); however, because the origins of spontaneous aberrations are not known, we preferred not to do this. Instead, we elected to fix our samples for chromosomal aberration analysis after 48 h in culture as often suggested (Bloom, 1981), but also to make a parallel culture with BrdUrd, also fixed after 48 h, from which we could assess the frequency of second or later in vitro mitoses likely to have been scored in the culture made without BrdUrd used for chromosomal aberration frequency assessment. Thus, for chromosome aberration and SCE analysis 3 different kinds of culture were made: with and without BrdUrd (25/~M) and fixed after

424

48 h in culture for assessment of the frequencies of first, second, and third and later in vitro mitoses and for measurement of chromosomal aberration frequencies, respectively, and with BrdUrd and fixed after 56 h in culture to ensure enough second in vitro mitoses for SCE analysis. Duplicate cultures were made where possible, as dictated by the leukocyte count and sample volume. All cultures were handled and incubated in yellow light in order to preclude the possibility of BrdUrd photolysis by blue wavelengths. The cultures were fixed according to standard methods, employing a 15-min hypotonic treatment in 75 mM KC1 and fixation in 3 : 1 methanol : glacial acetic acid. Air-dried slides were stained either by a conventional Giemsa method (for chromosomal aberrations) or by an F P G method utilizing Hoechst 33258 dye, exposure to fluorescent "black light" and Giemsa staining (for assessment of SCE and of frequencies of first and later in vitro mitoses). Staining was done by the laboratory which was to score the material. Slides from half of the total of 374 sets of subjects' cultures were sent to Oak Ridge for scoring there; those from the other half were retained for scoring at Brookhaven. In addition, in order to assess any difference in scoring that might have existed between the two laboratories, material (different slides) from 28 cases selected randomly was scored at both laboratories independently. For chromosomal aberrations, a total of 200 cells were analyzed per sample, 100 by each of two different scorers, each scoring different slides, and no more than 50 cells from any one slide. To insure uniformity of aberration scoring, one scorer in each laboratory scored one-half the cells scored at that laboratory; in addition, each cell scored as containing any aberration other than an achromatic lesion by another scorer was also reviewed by this same scorer so all these aberrations were scored or confirmed by one cytogeneticist in each laboratory. For sister-chromatid exchanges, a total of 50 cells were analyzed per sample, one-half by each of two different scorers, each scoring no more than 25 cells per shde. First or later mitoses were assessed on 100 cells at O R N L and on 200 cells, 100 by each of two scorers, at BNL. Slides were scanned at low magnification (10 X objective) and cells judged suitable for analysis

were scored at high magnification (100 X objective). Any cell observed at high magnification was included in the analysis unless a specific reason justified its rejection (i.e., uninterpretable because of scratches or debris obscuring one or more chromosomes). Only spreads with from 44 to 48 centromeres were included in the analysis ( O R N L included only cells with 45-48 centromeres, and the results have been adjusted to that basis). Cells with centromere counts outside this range, including tetraploids, and those with "shattered" or "pulverized" chromosomes were excluded from the aberration totals but recorded separately. All aberration types were included in the scoring; these were achromatic lesions, chromatid deletions, isochromatid deletions and chromatid exchanges among the chromatid types, and centric rings, dicentrics and symmetrical exchanges (mainly balanced translocations) among chromosome types. Long, apparently acentric fragments of the size of a C group chromosome) were recorded separately as either "acentric fragments" or "supernumerary acentric fragments"; since the preparations had not been banded they could not be assigned as X chromosomes with precocious centromere separation, their most probable origin (see Discussion). Because chromosome-type deletions cannot be distinguished from non-sister-union isochromatid deletions, all acentric fragments shorter than an average C group chromosomal and without sister union were recorded (arbitrarily) as isochromatid deletions.

Data management and analysis Primary cytogenetic data acquisition was done manually on printed forms. Periodically this was entered into FORTRAN-generated, keyed, indexed files. All entries were carefully checked for accuracy. Four files were created; one for aberration data at each laboratory, and one for SCE, white count and percentages of first, second and subsequent in vitro cell cycle data for each laboratory. The files were linked by a unique accession number given each subject, plus a "sequence" digit to distinguish different samples from the same subject in cases of multiple samples obtained on different dates. Though 200 cells were usually scored for aberrations by the scoring laboratory for aberrations

425

and 50 for SCE, this was not always possible. For the purpose of the present analysis it was convenient to arbitrarily include aberration data only on blocks of 50 cells (i.e., if a scorer ran out of material to score after scoring 67 cells, only the first 50 scored were utilized for analysis; the remaining 17 were excluded). For SCE analysis, we arbitrarily excluded from analysis cases where fewer than 10 cells could be scored. A 4-page questionnaire was given each participant. This questionnaire covered standard demographic questions, as well as a brief occupational, medical, and family history. The cytogenetic and demographic data were linked in a SIR (Scientific Information Retrieval) database. The keyed F O R T R A N files of cytogenetic data have been maintained as well, and the SIR database is periodically updated from these files. Transfer of data to the database is checked by page summaries, which are compared directly with the score sheets. All statistical analyses have been done using BMDP programs on files generated by the SIR database. With the exception of analyses done specifically to investigate time trends and effect of laboratory of scoring, the data from multiple samples from the same persons have been averaged and treated as one sample.

we have data. Totals of 71950 cells (40150 scored at Brookhaven and 31 800 scored at Oak Ridge) and 16898 cells (10138 from Brookhaven and 6760 from Oak Ridge) are presently available for aberration analysis and SCE analysis, respectively. The sample of subjects included 167 females and 186 males. Despite stratification of the sample in order to include more non-white volunteers, it includes only 38 (20 female and 18 male). All were black, except for 2 oriental females. The ages of the subjects ranged from 18.5 to 67.5 years. Their distribution is shown in Fig. 1. Among the many pieces of information we gathered about the subjects, 4 - - cigarette smoking status, exposure to organic solvents, the use of prescription drugs and classification as a radiation worker (i.e., badged to record occupational exposure) - - seem of particular possible relevance to cytogenetic endpoints. Table 1 shows the distribution of the study population in these respects. Using the cytogenetic data from the 28 subjects for whom material was scored at both laboratories, we tested to see whether there were any statistically significant differences between the frequencies of the various types of chromosomal aberrations or of SCE ascertained at Brookhaven and those ascertained at Oak Ridge. No significant differences were found. Therefore, for the

Results A total of 358 subjects were sampled. Of these, one was discovered later to have had radiation therapy, an exclusion criterion, and the cultures for another four were lost through a technical accident, and were not replaced by resampling, so the actual sample from which data could reasonably be expected was 353. Of these, 12 did not yield enough scorable material for cytogenetic analysis, a failure rate of 3.4%. In many instances, the problem with the 12 failures was clearly that their cells grew too slowly for the protocol; the 48-h cultures had blast cells but few mitoses, while the 56-h cultures had plentiful mitoses but few if any in their second in vitro division. Doubtless, a protocol with later fixation times would eliminate this problem. More than one sample which yielded cytogenetic data was obtained from 25 cases (2 from 19, 3 from 5, and 5 from 1), for a total of 374 samples from the 341 cases in the study for whom

z2t 29

51

~JJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJ~'~l -

-

I

8

2

AGE (yr) Fig. 1. Histograms showing the age distribution of the sample at the time their blood samples were obtained.

426 TABLE 1 DISTRIBUTION OF THE 353 SUBJECTS WITH RESPECT TO CIGARETTE S M O K I N G STATUS, O R G A N I C SOLVEN'I EXPOSURE, USE OF PRESCRIPTION DRUGS, A N D CLASSIFICATION AS " R A D I A T I O N WORKERS", BY SEX Sex

Smoking

Solvents No

Drugs

Radiation

Current

Former

Never

Unknown

Yes

Unknown

Yes

No

Unknown

Yes

No

Male

30

69

70

17

87

80

19

33

136

17

101

67

18

Female

43

50

61

13

42

114

11

52

104

11

43

114

10

rest of the analyses all of the data from both laboratories were simply pooled. We have also compared the data from the various chromosomal aberration types and for SCE from the 103 subjects sampled at the beginning of the study who were not later included in the stratified random sample with that from the subjects in the stratified random sample. Only one was found. This was for aneuploid cells; the percentage for the volunteers was 4.91 + 0.27, while that for the random sample was 6.63 + 0.26 ( p < 0.0001). We have therefore pooled the data from the two classes of subjects for subsequent analyses of aberration types and of SCE. Finally, because the sampling of subjects for this study spanned a relatively long period (almost 3 years), we tested for significant regression of either aberration or SCE frequencies on time of sampling. Again, none was found. Because, inadvertently, one laboratory included cells with only 44 chromosomes (centromeres) in its aberration analysis, while the other did not, we have applied a correction to the aberration data from that laboratory, assuming simply that the odds on finding an aberration involving a single chromosome were reduced in such cells by the fraction 1 / 4 6 (i.e., assuming the aberrations randomly distributed among chromosomes, and that an average length chromosome was lost). In fact, this correction is so small that it does not affect the rounded frequencies presented below at all, but its use in the significance tests precluded any possible bias. Chromosomal aberrations The mean frequencies of the various types of chromosomal aberrations, tetraploids and aneuploids, are given in Table 2 together with their

Unknown

standard errors and the range of values seen in individuals in the study. As already noted, concern over possible loss of aberrations if cells in mitoses later than the first in vitro are included in the sample scored has led to the recommendation that cultures intended for aberration frequency determination be fixed at 48 h, as we did in the present study. However, our analyses of differentially stained preparations from parallel cultures containing BrdUrd and fixed at 48 h demonstrate that fixation at 48 h does not guarantee that only first in vitro mitoses will be present: the mean frequency of second and later mitoses in our 48-h cultures was 92.65 _+ 0.35, with a range from 0 (this occurred in 0.3% of the cultures) to 2 cases with 45% second or later mitosis at 48 h. To see whether we could find any evidence that the presence of second or later mitoses in the sample actually reduces spontaneous aberration frequencies significantly, we tested for regression of aberrations on the frequency of second or later mitoses observed, both by individual aberration type and by chromatid- and chromosome-type totals, but none that even approached statistical significance was found. As already mentioned, 25 subjects had 2 or more samples drawn so that we could assess the possible contribution of temporal variation to total variance. However, the inter-sample variance was no greater between samples from the same subject than between samples from different subjects, so it does not appear that the time of sampling is an important contribution to overall variance. As discussed more fully in the Discussion, age, sex, cigarette smoking, exposure to organic solvents or to ionizing radiation and use of prescription drugs are all factors either already reported to influence chromosomal aberration levels or at least

427 TABLE 2 MEANS, STANDARD ERRORS AND RANGES, AS PERCENTS, OF ANEUPLOIDS AND OF THE VARIOUS CLASSES OF CHROMOSOMAL ABERRATIONS Class

Type

Mean

S.E.M.

Range

Aneuploid

2n = 45 2n = 47 Tetraploid

6.061 0.173 0.025

0.196 0.019 0.006

0-23 0- 2.75 0- 1

Chromatid

achromatic lesions chromatid deletions isochromatid deletions a chromatid exchanges

4.249 0.809 0.034 0.048

0.117 0.046 0.007 0.008

0-12.5 0- 6 0- 1 0- 1

Chromosome

deletion b ring dicentric translocation

0.421 0.019 0.160 0.050

0.032 0.006 0.016 0.009

0000-

Other

"acentric fragment" "supernumerary acentric fragment"

0.314 0.125

0.033 0.018

0- 3 0- 3.5

4.5 1 1,5 1.5

a Includes only sister-union types; i.e., only the unequivocally isochromatid-type deletions. b Includes all breaks of both chromatids (isolocus) without sister union; i.e., both true chromosome-type deletions and the non-sister-union isochromatid deletions.

reasonable candidates. We accordingly tested to see whether any of these factors contributed significantly to the variation of aberration frequencies, both by individual aberration type and in aggregate. However, only 1 of the 6 could be seen to influence any aberration frequencies significantly. This influence was that of age on the frequency of dicentric chromosomes, for which the regression analysis yielded a relationship which just reached significance (r = 0.148, p = 0.006). Because asymmetrical c h r o m o s o m e - t y p e exchanges (rings and dicentrics) m a y arise either directly or from chromatid-type exchanges following replication in a subsequent cell cycle, and because they m a y persist for m a n y years after their induction, it is of interest to k n o w how m a n y of those seen in our study were of recent origin, and how m a n y were "old", having arisen at some earlier time in the subject's lifetime. Because the acentric fragments which must arise at the time any asymmetrical exchange aberration is formed tend to be rapidly lost as a result of cell division, the presence or absence of the expected a c c o m p a nying acentric fragment is a measure of the " o l d ness" of such aberrations. In our samples, 11 out of the 117 dicentrics and 2 out of 17 of the

c h r o m o s o m e rings, or 9.7% for the two classes combined, lacked a c c o m p a n y i n g acentric fragments, indicating that about 90% were " n e w " .

Sister-chromatid exchanges The mean frequency of SCE observed was 8.29 + 0.08 SCE/cell, with a range of individual values from 4.0 to 12.8 SCE/cell. Values observed in single cells ranged from 0 (in 13 cells) to a high of 31 in 2 cells. The distribution of individual means is clearly binomial, though that of the values for cells within the samples for individuals clearly failed to fit a binomial distribution, with high values skewing the distributions to the right, as is illustrated in Fig. 2. Cigarette smoking has widely been reported to elevate mean SCE levels. Dividing the data according to whether subjects reported never smoking cigarettes, former but not current cigarette smoking, or current cigarette smoking, we found a large statistically significant elevation when current smokers were c o m p a r e d to those who never smoked; for 65 smokers the mean SCE value was 9.02 + 0.18, while for 220 who never smoked the mean was 8.08 + 0 . 0 8 (difference = 0.94 + 0.19 SCE/cell, p < 0.001 by simple t test), but no significant elevation was seen in

428 I

(a)

24 A 22 ~' 2O o

~ m Z b.l

J

O W b-

I

I

o

4

i

8

6

I0

L_

I

12

14

M E A N SCE FREQUENCY

p

13

I

i

i

~

i

I

i

i

i

i

I

i

i

4

6

8

I0 12 14 16 18 2 0 22 2 4 2 6 28 NUMBER OF SCE

(b)

12

ic -6

c

7

i

0

2

i

i

i

s'o-

0.0001. On the other hand, when the SCE data, adjusted for smoking and sex, were regressed on subject age, no evidence for an age effect was found (r = 0.06; p --- 0.28). We also tested whether reported exposure to organic solvents, use of prescription drugs, or radiation worker status influenced the mean SCE frequencies for our subject population, but no evidence of such influences was found. As noted in Materials and methods, concern has been expressed as to whether the reported effect of the level of BrdUrd in the culture medium might influence SCE frequencies if different numbers of lymphocytes were competing for available BrdUrd in cultures from different subjects. Our protocol of adjusting inoculum size according to white blood cell count was an attempt to minimize this possible effect. When we regressed individual mean SCE values on the white blood cell count upon which inocuhim size was calculated, however, we found a statistically significant relationship ( r = 0.321; p < 0.001. Even when adjusted for the smoking and sex effects, this relationship remained highly significant (r = 0.364; p < 0.001) as is shown in Fig. 3. The slope of this relationship (increasing SCE frequency with increasing white blood count) is in the opposite direction from that expected if the larger inoculum sizes were produc-

Fig. 2, Histograms of (a) distribution of mean SCE values for individuals, and (b) distribution of actual individual cell values in the whole sample of 18786 cells.

I

I

I

I

I

I

I

I

I

>_ 9 . 0

former smokers ( p = 0.18). 40 of the males in the sample reported current cigar or pipe smoking, of whom many were former cigarette smokers. When the mean SCE values for these 40 were compared to either the non-smokers or the former smokers, no evidence for any influence on SCE frequencies was seen, however. Both the sex and the age of the subject have also been reported to influence SCE frequencies. When corrected for the smoking effect, we did indeed find a small but statistically highly significant difference between male and female SCE means. The adjusted female mean was 8.38 + 0.04, while that for males was 8 . 1 1 _ 0.03. This difference, amounting to 5%, yielded by t test p <

(...) Z I.Lt

8.7 b-

w 8.4

""

s." "....".*

"

--.- • • • ".

Z

7.8

•

.

.... ,• •

i 2.0

i 4.0

., .

I 6.0

*.

.

.

.

o

i 8.0

WHITE BLOOD

i i I i I I0.0 12.0 14.0 16.0 18.0

CELL COUNT

(rn)

Fig. 3. Regression of mean SCE values, adjusted for smoking status and sex, on the subjects' white cell count. The regression is significant at p < 0.001.

429

ing more competition for available BrdUrd, which could reduce lymphocyte incorporation levels and hence, presumably, SCE frequency. Discussion

Determination of chromosomal aberration frequencies in peripheral blood lymphocytes as a means of biological dosimetry for human exposures to ionizing radiation is a well-accepted technique (Bender, 1969; Sasaki, 1983; Bauchinger, 1984; Lloyd e t a l . , 1980) and there are consequently a large number of published studies of the spontaneous frequencies of the chromosome-type aberrations which are useful for this purpose. It is of interest to compare the present results. Lloyd (1984) has summarized published data on dicentric and acentric fragment (chromosome deletions and non-sister-union isochromatid deletions) frequencies from 65 different studies involving 2000 subjects and a total of 211 661 cells analyzed. The range of dicentric frequencies for those studies large enough for one or more dicentrics to be observed, was 0.1 x 10 - 3 to 2.0 x 1 0 - 3 ; that for acentric fragments was 2 x 10 -3 to 4.5 x 10 -3 (for only 187105 cells, as some studies did not report on these aberrations). The means were 0.78 x l0 -3 and 3 . 7 x 1 0 -3, respectively. For the largest 6 studies, each involving over 100 subjects, these values were 1.1 x 10 -3 for dicentrices and 3.4 x 10 -3 for acentric fragments. Among more recent studies, that of Evans et al. (1979) reported 1.1 × 10 -3 for dicentrics and 2.3 × 10 .3 for acentric fragments in 7900 cells from 79 subjects, that of Obe and Beek (1982) reported a dicentric frequency of 0.3 x 10 -3 in almost 15000 cells from 83 subjects, that of Ltonard et al. (1984) reported a dicentric frequency of 1.1 × 10-3 and an acentric fragment frequency of 2.4 x 10 -3 for acentric fragments in 11 500 cells from 23 subjects, while that of Galloway et al. (1986) reported a dicentric frequency of 2.1 × 10 -3 and an acentric fragment frequency of 3.2 × 10 -3 in over 41 000 cells from 304 individuals (see below). These values all lie within the range of the earlier studies reviewed by Lloyd et al. Our present values of 1 . 6 + 0 . 2 × 1 0 3 and 4 . 6 + 0 . 3 x 1 0 -3 for dicentric and acentric fragments, respectively, agree quite well. The much rarer centric chromosome

rings are less often reported (at least separately) in the literature. The reports of Obe and Beek, Ltonard et al., and Galloway et al. report frequencies of zero, zero and 0.35 × 10 -3, respectively, from which one might naively calculate a mean for the 3 studies of 0.19 x 10 -3. Our value of 0.19 + 0.06 x 10 -3 is clearly not in disagreement with the published data! Recently, Awa and Neel (1986) have reported on a phenomenon they call "rogue" cells occurring in a sample of over 100 000 cells from peripheral blood leukocytes from 9818 individuals from Hiroshima, Japan. In that sample 24 cells were seen which had multiple dicentric or tricentric chromosomes, acentric fragments and double minutes. No correlation was seen with radiation exposure, or with age or sex, and no explanation could be offered as to their origin. In our present study we have not seen this phenomenon; only 4 cells were found out of the 71950 ceils scored for aberrations which had more than 1 ring a n d / o r dicentric, and each had but 2, with no accompanying aberrations of other types. Because chromatid-type aberrations are not induced in the circulating G O peripheral blood lymphocytes used in radiation dosimetry, and have only relatively recently been widely recognized to in fact constitute the class of chromosomal aberrations directly induced by most chemical clastogens, it is not surprising that chromatid aberration frequencies in populations have been rarely reported. Often, when they are reported, the different types are lumped together, sometimes as "cells with aberrations". This situation was in fact a large part of our reason for the present study. However, Galloway et al. (1986) have recently published data on chromatid aberration frequencies in 41282 cells from 304 control subjects studied as part of a study of cytogenetic effects of occupational exposure to ethylene oxide. Among the chromatid types, they reported frequencies of achromatic lesions of 8.11% (range 0-51), of chromatid deletions of 0.64% (range 0-6) and of chromatid exchanges of 0.11% (range 0-4). These may be compared with our results shown in Table 2. Also, as summarized in Galloway et al., frequencies of chromatid exchanges have also been presented in several other recent studies. Tonomura et al. (1983) reported a frequency of 0.8 x 10 -3 in

430

92467 cells from 96 subjects. Gundy and Varga (1983) reported finding none in a study of 17 500 cells from 175 persons. Obe et al. (1982) studied smokers and non-smokers separately. Their frequencies for non-smokers and for smokers were 0.4 and 0.7, based upon samples of almost 24 000 cells from 20 subjects and about 32 000 cells from 24 subjects, respectively. Clearly, the frequency reported by Galloway et al. of achromatic lesions is higher than what we found, and their frequency of chromatid exchanges is nigher than both ours and of other published studies, though the reasons for such differences remain unclear. The class of "aberration", sometimes called "supernumerary acentric fragment", where an apparently acentric chromosomal object about as long as a large C group chromosome appears in addition to a normal complement of 46 chromosomes, and the parallel case where the long apparent fragment occurs in a spread with only 45 centric chromosomes, requires some discussion. Because in many cases the object replaces a C group chromosome, it was concluded by Jacobs et al. (1964) that it might be a precociously divided X chromosome. In a careful study using autoradiography, and C- and G-banded preparations, Fitzgerald (1975) concluded that they were, in fact, X chromosomes in which the centromere had prematurely divided. They observed a range of possible karyotypes, including 46, - C + f (C = Cgroup chromosome, f = fragment), 46, - 2 C + 2f, and 47, - C + 2f, where all the "fragments" were X chromosomes. The frequency of cells with "fragments" was considerably higher in females than in males, and the frequency also increased with age for the female subjects. Though our own preparations were not banded, we assume the same origin, particularly as their frequency was also significantly nigher in females than in males (0.62 + 0.07 vs. 0.28 + 0.04 per 100 cells; p < 0.0001), and there was a significant regression on age for the female subjects (r = 0.36; p < 0.001), as shown in Fig. 4. Galloway and Buckton (1978) and Galloway et al. (1986) have reported the same phenomena. The latter found mean frequencies of 0.23% in 9516 cells from male subjects, and 0.84% in 8788 cells from female subjects, values not very different from our own. Many published studies report frequencies of

I

I

I

1" 20

]

"1" w

25

50

I

I

I

I

I

I

I

I

I

4.0

I

5.5 W

o 5.0 rtla-

2.5 Z ILl (.9

2.0

rr

u_ 1.5

~

1.0

Z

~ o.5 =

0.0

'r"1

55

4 0 45 50 AGE (yr)

I

!

I

55

60

65

~

Fig. 4. Regression of "acentric fragments", both "supernumerary" and not, on age for female subjects. The regression is significant at p < 0.001.

sister-chromatid exchanges in human peripheral blood lymphocytes. These are difficult to compare meaningfully with our present SCE results, however, as the frequencies reported by various laboratories differ widely, at least in part because of their known sensitivity to technical factors, most notably the concentration of BrdUrd used and the serum used for culture (Wolff and Perry, 1974; Kato and Sandberg, 1977). For example, Carrano and Moore (1982) have published SCE data on 42 normal individuals. 80 cells were scored per subject, and the mean SCE frequency per cell was 8.99 (calculated from their data presented on a per chromosome basis), not very different from our own value of 8.24+ 0.08, notwithstanding Carrano and Moore's use of a much higher BrdUrd concentration. More recently, workers from the same group published data on 40 non-exposed control subjects as part of a study of possible cytogenetic effects of occupational exposure to ethylene dibromide (Steenland et al., 1986). Again using a high BrdUrd level, and scoring 80 cells per subject, they found a mean of about 8.44 SCE/cell (again calculated from their data). On the other hand, in what is probably the largest normal population study published to date, Soper et al. (1984) found a mean of 9.9 SCE/cell in cultures from 479 subjects which contained the same

431

BrdUrd concentration as used by Carrano and Moore and by Steenland et al. A number of factors have been reported to influence the frequencies of either chromosomal aberrations or SCE, or both, in human peripheral blood lymphocytes. In addition to the obvious case of ionizing radiation exposure and chromosomal aberrations, age, sex, and cigarette smoking have been reported to influence chromosomal aberration frequencies (Obe and Herha, 1978; Evans, 1979; Vijayalaxmi and Evans, 1982; Galloway et al., 1986), while age, sex, cigarette smoking, and even race have been reported to influence SCE frequencies (Lambert et al., 1978; KrishnaMurthy, 1979; Hopkin and Evans, 1980; Husgafvel-Pursiainen et al., 1980; Butler, 1981; Obe et al., 1982; Soper et al., 1984; Margolin and Shelby, 1985; Steenland et al., 1986; Wulf et al., 1986). The only strong correlation on which most (though not all) authors seem to agree, however, is that between cigarette smoking and SCE frequency, perhaps because this is very substantial effect. In the present study it amounts to an elevation of 12% in current cigarette smokers over the nonsmokers' levels. Sex also had a highly significant effect on SCE frequency in our study, with females showing about a 5% excess over males after adjustment for the cigarette smoking effect, in reasonable agreement with the excesses reported in Margolin and Shelby's (1985) analysis of published data and by Steenland et al. (1986). Interestingly, Wulf and Niebuhr (1985) reported a similar increase in frequency of spontaneous SCE in the XX ceils compared to XY cells from a pair of chimeric human twins. On the other hand, we have not seen the sort of age relationship for SCE frequency reported by Soper et al. (1984) and by Wulf et al. (1986). However, that reported by Soper et al. was rather weak (p < 0.042), and that seen by Wulf et al. was in a study of children between 1 and 18 years of age, an age group not sampled in the present study. Neither do we see any suggestion of an effect of the race of the subject on SCE frequencies, as suggested by the analysis of Margolin and Shelby (1985). Our sample of non-white subjects is clearly not large enough to be definitive (36 blacks and 2 orientals), but is, however, larger than the non-white sample assessed by Margolin and Shelby (1985).

Our finding of a statistically significant influence of age on the frequency of dicentric chromosomes does not really parallel the reports of age effects on aberration frequencies of Bochkov (1972), Obe and Herha (1978), Evans (1979), or Galloway et al. (1986). We found such a relation for only chromosome-type dicentrics, and not for other kinds of chromosome or chromatid exchanges, nor for overall aberration frequency. Finally, it is of interest to compare our present data with that from an earlier study we conducted using precisely the same procedures, protocols and materials, and which was in fact conducted concurrently with the early samplings for the present study. This was our study of persons living near the Love Canal chemical dump site in Niagara Falls, NY (Heath et al., 1984). In that study we studied 2 possibly exposed groups (a total of 45 subjects with 11000 cells analyzed for aberrations) and 2 control groups (a total of 44 subjects with 11 000 cells analyzed for aberrations). There were no statistically significant differences in aberration or SCE frequencies for the different groups, and

TABLE 3 COMPARISON OF SOME RESULTS FROM T H E PRESENT STUDY W I T H THOSE FROM O U R EARLIER STUDY, U S I N G PRECISELY THE SAME PROTOCOL, OF PERSONS R E S I D I N G N E A R LOVE CANAL (Heath et al., 1984) The Love Canal values have been recalculated from the raw data using the same methods as in the present study

Chromatid

Chromosome

Mean SCE a

achromatic lesions chromatid deletions isochromatid deletions chromatid exchanges deletion ring dicentric translocation

Present study

Love Canal

4.25+0.12

8.21+0.37 *

0.81 + 0.05

1.14 + 0.07 *

0.03 + 0.01

0.06 + 0.02

0.05 + 0.01

0.08 + 0.02

0.42+0.03 0.02 + 0.01 0.16 + 0.02 0.05 + 0.01

0.93+0.10 * 0.06 + 0.02 0.33 + 0.05 0.13 + 0.03 *

8.29+0.08

8.62+0.12

a Unadjusted for age or smoking. * Statistically significantly different ( p < 0.01 by simple t test).

432 so for purposes of c o m p a r i s o n here we have simply pooled results for the 2 studies a n d the 2 control groups, giving a " N i a g a r a Falls, N Y study group" consisting of 89 subjects. A total of 22000 cells were a n a l y z e d for aberrations, a n d 5000 cells were analyzed for SCE. U s i n g the raw data from the N i a g a r a Falls study group, we have calculated in precisely the same way as for the present study the values shown i n T a b l e 3. It will be seen that in every case the N i a g a r a Falls values are n u m e r i cally greater than those from the present study. F o r 4 (achromatic lesions, c h r o m a t i d deletions, c h r o m o s o m e deletions a n d dicentrics), the differences are statistically highly significant. As the two studies were identical except for the subject populations, it is t e m p t i n g to speculate that the differences are real, perhaps reflecting a biological impact of some e n v i r o n m e n t a l factor i n the N i a g a r a Falls area. Acknowledgments We t h a n k the following for their expert technical assistance, a n d in particular for their persistence in the tedious scoring operation: A.M. Boccio, T.Ho, R. Kali, H.E. Luippold, M.S. Makar, a n d M.H. T h o m p s o n .

References Awa, A.A., and J.V. Neel (1986) Cytogenetic "rogue" cells: What is their frequency, origin, and evolutionary significance? Proc. Natl. Acad. Sci. (U.S.A.)., 83, 1021-1025. Bauchinger, M. (1984) Cytogenetic effects in human lymphocytes as a dosimetry system, in: W.G. Eisert and M.L. Mendelsohn (Eds.), Biological Dosimetry, Springer, Berlin, pp. 15-24. Bender, M.A (1969) Human radiation cytogenetics, in: L.G. Augenstein, R. Mason and M. Zelle (Eds.), Advances in Radiation Biology, Vol. 3, Academic Press, New York, pp. 215-275. Bender, M.A (1979) New trends and techniques in chromosome aberration analysis, in: Routine Use of Chromosome Analysis in Radiation Protection, TECDOC-224, International Atomic Energy Agency, Vienna, pp. 109-117. Bender, M.A, and P.C. Gooch (1966) Somatic chromosome aberrations induced by human whole-body irradiation: the "Recuplex" criticality accident, Radiat. Res., 29, 568-582. Bloom, A,D. (Ed.) (1981) Guidelines for Studies of Human Populations Exposed to Mutagenic and Reproductive Hazards, March of Dimes Birth Defects Foundation, New York.

Bochkov, N.P. (1972) Spontaneous chromosome aberrations in human somatic cells, Humangenetik, 26, 159-164. Butler, M.G. (1981) Sister chromatid exchange in 4 human races, Mutation Res., 91,377-379. Carrano, A.V., and D.H. Moore (1982) The rationale and methodology for quantifying sister chromatid exchanges in humans, in: J.A. Heddle (Ed.), Mutagenicity: New Horizons in Genetic Toxicology, Academic Press, New York, pp. 267-304. Evans, H.J. (1979) The induction of aberrations in human chromosomes following exposure to mutagens/carcinogens in: P. Emmelot and E. Kriek (Eds.), Environmental Carcinogens, Elsevier/North-Holland, Amsterdam, pp. 57-74. Evans, H.J., K.E. Buckton, G.E. Hamilton and A. Carothers (1979) Radiation-induced chromosome aberrations in nuclear-dockyard workers, Nature (London), 277, 531-534. Fitzgerald, P.H. (1975) A mechanism of X chromosome aneuploidy in lymphocytes of aging women, Humangenetik, 28, 153-158. Galloway, S.M., and K. Buckton (1978) Aneuploidy and aging: Chromosome studies on a random sample of the population using G-banding, Cytogenet. Cell Genet., 28, 78-95. Galloway, S.M., P.K. Berry, W.W. Nichols, S.R. Wolman, K.A. Soper, P.D. Stolley and P. Archer (1986) Chromosome aberrations in individuals occupationally exposed to ethylene oxide, and in a large control population, Mutation Res., 170, 55-74. Gundy, S., and L.P. Varga (1983) Chromosomal aberrations in healthy persons, Mutation Res., 120, 187-191. Heath, C.W., M.R. Nadel, M.M. Zack, A.T. Chen, M.A Bender and R.J. Preston (1984) Cytogenetic findings in persons living near the Love Canal, J. Am. Med. Assoc., 251, 1437-1440. Hopkin, J.M., and H.J. Evans (1980) Cigarette smoke-induced DNA damage and lung cancer risks, Nature (London), 283, 388-390. Husgafvel-Pursiainen,K., J. Maki-Paakkanen, H. Norppa and M. Sorsa (1980) Smoking and sister chromatid exchange, Hereditas, 92, 247-250. Jacobs, P.A., M. Brunton and W.M. Court Brown (1964) Cytogenetic studies in leukocytes on the general population: Subjects of ages 65 and more, Ann. Hum. Genet. 27, 353-365. Kato, H., and A.A. Sandberg (1977) The effect of sera on sister chromatid exchanges in vitro, Exp. Cell Res., 109, 445-448. Krishna-Murthy, P.B. (1979) Frequency of sister chromatid exchanges in cigarette smokers, Hum. Genet., 52, 343-345. Lambert, B., A. Lindblad, M. Nordenskjold and B. Werelius (1978) Increased frequency of sister-chromatid exchanges in cigarette smokers, Hereditas, 88, 147-149. Lronard, A., G.H. Deknudt, E.D. L~onard and G. Decat (1984) Chromosome aberrations in employees from fossilfueled and nuclear-power plants, Mutation Res., 138, 205-212. Lloyd, D.C. (1984) An overview of radiation dosimetry by conventional methods, in: W.G. Eisert and M.L. Mendelsohn (Eds.), Biological Dosimetry, Springer, Berlin, pp. 3-14.

433 Lloyd, D.C., R.J. Purrott and E.J. Reeder (1980) The incidence of unstable chromosome aberrations in peripheral blood lymphocytes from unirradiated and occupationally exposed people, Mutation Res., 72, 523-532. Margolin, B.H., and M.D. Shelby (1985) Sister chromatid exchanges: A reexamination of the evidence for sex and race differences in humans, Environ. Mutagen., 7, Suppl. 4, 63-72. Obe, G., and B. Beck (1982) The human leucocyte system, in: F.J. de Serres and A. Hollaender (Eds.), Chemical Mutagens and Methods for Their Detection, Vol. 7, Plenum, New York, pp. 337-400. Obe, G., and J. Herha (1978) Chromosomal aberrations in heavy smokers, Hum. Genet., 41,259-263. Obe, G., H.-J. Vogt, S. Madle, A. Fahning and W.D. Heller (1982) Double-blind study on the effect of cigarette smoking on the chromosomes of human peripheral blood lymphocytes in vivo, Mutation Res., 92, 309-319. Sasaki, M.S. (1983) Use of lymphocyte chromosome aberrations in biological dosimetry: possibilities and limitations, in: T. Ishihara and M.S. Sasaki (Eds.), Radiation-Induced Damage in Man, Progress and Topics in Cytogenetics, Vol. 4, Liss, New York, pp. 585-604. Scott, D., and C.Y. Lyons (1979) Homogeneous sensitivity of human peripheral blood lymphocytes to radiation-induced chromosome damage, Nature (London), 278, 756-758. Soper, K.A., P.D. Stolley, S.M. Galloway, J.G. Smith, W.W.

Nichols and S.R. Wolman (1984) Sister chromatid exchange (SCE) report on control subjects in a study of occupationally exposed workers, Mutation Res., 129, 77-88. Steenland, K., A. Carrano, J. Ratcliffe, D. Clapp, L. Ashworth and T. Meinhardt (1986) A cytogenetic study of papaya workers exposed to ethylene dibromide, Mutation Res., 170, 151-160. Tonomura, A., K. Kishi and F. Saito (1983) Types and frequencies of chromosome aberrations in peripheral lymphocytes of general populations, in: T. Ishihara and M.S. Sasaki (Eds.), Radiation-Induced Chromosome Damage in Man, Progress in Mutation Res., Vol. 4, Liss, New York, pp. 605-616. Vijayalaxmi, and H.J. Evans (1982) In vivo and in vitro effect of cigarette smoke on chromosomal damage and sisterchromatid exchange in human peripheral blood lymphocytes, Mutation Res., 92, 321-332. Wolff, S., and P. Perry (1974) Differential Giemsa staining of sister chromatids and the study of sister chromatid exchanges without autoradiography, Chromosoma, 48, 341-353. Wulf, H.C., and E. Niebuhr (1985) Different sister chromatid exchange rates in XX and XY cells of a pair of human chimeric twins, Cytogenet. Cell Genet., 39, 105-108. Wulf, H.C., N. Kousgaard and E. Niebuhr (1986) Sister chromatid exchange in childhood in relation to age and sex, Mutation Res., 174, 309-312.