[CANCER RESEARCH 50, 5049-5054, August 15. 1990]
Micronucleated Erythrocytes as an Index of Cytogenetic Damage in Humans:
Demographic and Dietary Factors Associated with Micronucleated
Erythrocytes in Splenectomized Subjects1
Daniel F. Smith,2 James T. MacGregor,3 Robert A. Hiatt, N. Kim Hooper, Carol M. Wehr, Beverly Peters,
Lynn R. Goldman, Leslie A. Yuan, Philip A. Smith, and Charles E. Becker
Environmental Epidemiology and Toxicology Branch, California Department of Health Services, Emeryville, California 94608 [D. F. S., L. R. G., L. A. Y.J; Food Safety
Research Unit, Western Regional Research Center, United States Department of Agriculture, Albany, California 94710 [J. T. M., C. M. W., P. A. S.J; The Permanente
Medical Group, Division of Research, Oakland, California 94611-5463 [R. A. H., B. P.J; Reproductive and Cancer Hazard Assessment Section, California Department
of Health Services, Berkeley, California 94704 [N. K. H.J; Northern California Occupational Health Center, University of California, San Francisco, California 94110
[C. E. B.I
ABSTRACT
Erythrocytes containing micronuclei serve as an indicator of genotoxic
exposure in splenectomized
individuals. Micronucleated erythrocytes,
derived from cytogenetically damaged RBC precursors, are not selectively
removed from peripheral blood in individuals who lack splenic function.
The relationship between micronucleated cell frequencies and demo
graphic, environmental, and dietary factors was examined in 44 subjects
with previous splenectomy due to trauma. Their micronucleated cell
counts fit a log-normal distribution, with geometric means of 3.3 micronucleus-containing cells/1000 reticulocytes and 2.7/1000 normochromatic
erythrocytes. A multiple regression analysis showed that drinking five
cups of coffee or tea/day (relative to none) was associated with an
approximately 2-fold higher frequency of micronucleated cells. Weaker
statistical associations were also noted with micronucleus frequency and
the consumption of calcium supplements (associated with a higher fre
quency) and vitamins A, C, or E (lower frequency). An apparent trend of
higher micronucleus counts with age was attenuated when other factors
were considered in the regression. Cigarette smoking and decaffeinated
coffee consumption were among the factors not associated with elevated
micronucleated cell frequencies. Because the occurrence of micronuclei
in reticulocytes reflects cytotoxic exposures within the past 3-8 days, it
may be possible to test directly the relationship of these factors to
micronucleus formation through intervention studies.
INTRODUCTION
With the current concern for the potential genotoxic and
carcinogenic effects of environmental exposures (in air, water,
diet, occupation, etc.), sensitive indicators of DNA damage are
needed to determine the extent to which such exposures actually
contribute to genotoxic damage in humans.
The measurement of micronuclei in human erythrocytes and
lymphocytes is one such marker of the effect of genotoxic
exposures (1). The measurement of micronuclei in bone marrow
erythrocytes has become established as a routine screening test
for in vivo cytogenetic damage in experimental animals (2) and
has been applied to humans with exposures to chemotherapeutic drugs or with occupational exposures (3-7). However, the
invasive nature of bone marrow sampling severely limits the
usefulness of this procedure as a routine assay for cytogenetic
damage in humans. Interest in the peripheral blood lymphocyte
assay (8) has been stimulated recently by the introduction of
the cytokinesis-block method (9), which allows first-division
Received 11/1/89; revised 4/11/90.
The costs of publication of this article were defrayed in part by the payment
of page charges. This article must therefore be hereby marked advertisement in
accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
' Support for this study was provided in part by a grant from the Northern
California Cancer Program.
2To whom requests for reprints should be addressed, at the Environmental
Epidemiology Section, California Department of Health Services, 5900 Hollis
Street, Suite E, Emeryville, CA 94608.
'Present address: SRI International, 333 Ravenswood Avenue, Menlo Park.
CA 94025.
nuclei to be accumulated in culture as binucleate cells which
are easily recognized and scored. The ability to determine the
cell division being scored is a critical requirement because the
micronucleus frequency depends on the proportion of cells
which have divided following the induction of DNA damage
and the fate of micronuclei in cells which have divided more
than once (9). Nevertheless, the assay is still limited by a lack
in sensitivity to agents which do not produce long-lived lesions
in nondividing cells, due to the test cell population not actively
dividing in vivo.
This report suggests using micronucleated peripheral blood
erythrocytes in splenectomized subjects as a marker for chro
mosomal damage. The end point measured is the prevalence of
micronuclei (also known as Howell-Jolly bodies) among periph
eral erythrocytes. Under normal conditions, the human spleen
removes micronucleus-containing erythrocytes from the periph
eral blood. But in asplenic individuals, micronucleated cells
persist and accumulate with repeated genotoxic exposure (10).
As a consequence, evidence of cytogenetic damage is uniquely
visible in splenectomized persons as an increased frequency of
micronucleus-containing RBCs. Although limited to splenec
tomized individuals, the advantages of the assay include sim
plicity and greater ease of scoring than other currently applied
assays.
The assay can be conducted on two erythrocyte populations:
the immature polychromatic erythrocytes which stain positively
for residual RNA (hereafter referred to as reticulocytes) and the
mature normochromatic erythrocytes which do not stain posi
tively for RNA (hereafter simply erythrocytes). Newly formed
reticulocytes mature into circulating erythrocytes after only 12 days (11); hence, micronucleated reticulocytes demonstrate
only very recent genotoxic exposures. Erythrocytes have an
average life span of 120 days (11); hence, micronucleated eryth
rocytes reflect genotoxic exposures within approximately 4
months prior to sampling (10).
The studies of micronuclei in circulating erythrocytes which
have been reported to date have examined relatively small
numbers of subjects splenectomized due to illness (idiopathic
thrombocytopenia, Hodgkin's disease, leukemia, etc.) or trau
matic damage to the spleen. Chemotherapeutic agents known
to damage DNA cause > 10-fold increases in micronucleated
cells, and the kinetics of the appearance and disappearance of
micronucleated cells have been shown to be consistent with the
kinetics of RBC formation and turnover (10). A marked effect
of mild folate deficiency on the frequency of micronucleated
cells has also been reported (12).
The present investigation sought to examine the frequency
of micronucleated cells in a group of splenectomized (but oth
erwise healthy) subjects and to identify common demographic,
5049
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1990 American Association for Cancer Research.
MICRONUCLEATED
ERYTHROCYTES
environmental, and dietary factors that may be correlated with
this frequency.
MATERIALS
that was approximately 0.05 times the area of every fourth field and is
based upon at least 100 cells.
The percentage of RNA-positive cells in each sample was determined
by counting the number of RNA-positive cells in each fourth field and
counting the number of RNA-negative cells in a small sector of that
field. For a typical proportion of RNA-positives to RNA-negatives, this
percentage was based upon a total count of approximately 30 RNApositive cells and approximately 100 RNA-negative cells.
Absence of splenic function was verified by determining the frequency
of "pitted" erythrocytes (14) in the glutaraldehyde-fixed blood samples.
AND METHODS
Enrollment of Study Subjects. Potential study subjects were identified
from the records of the Northern California Kaiser Permanente Medical
Care Program. Computer-stored discharge records for the years 19711984 were searched to obtain a list of individuals who were hospitalized
with trauma/injury (International Classification of Diseases, Rev. 9,
800-999 and E codes) and whose medical records had splenectomy
(surgical procedure code 45.1) as the primary discharge diagnosis. This
produced a list of 303 individuals with a trauma-related splenectomy
for whom computer-stored information, including date of birth, date
of surgical procedure, and address, was available.
To each person was mailed a recruitment packet that contained a
letter describing the study and requesting their participation, a human
subjects consent form that followed Kaiser Permanente guidelines, and
a self-administered questionnaire. The questionnaire included basic
demographic information, several dietary questions, and questions
about exposures to potential genotoxic agents, including X-rays,
chemotherapy, and potential occupational exposures. Because of the
limited life span of erythrocytes and the previous demonstration that
micronucleated cell frequencies return to their baseline values within 4
months after cessation of genotoxic exposures (10), the questionnaire
asked only about exposures within the last 4 months. Subjects who
agreed to participate were directed to their local health plan facility for
blood drawing.
Laboratory Procedures. A blood sample was drawn from each subject
into a heparinized Vacutainer tube by Kaiser Permanente laboratory
staff. Two drops of blood were transferred to a vial containing 0.5 ml
of 3% glutaraldehyde in 0.1 M sodium phosphate buffer, pH 7.4. The
sample was gently mixed and held at room temperature at least 24 h.
Six or more standard blood smears were prepared using the freshly
drawn heparinized blood. Smears were air dried, immersed in absolute
methanol for 2-5 min, and air dried again. All samples were delivered
to the United States Department of Agriculture laboratory for analysis.
Upon receipt at the United States Department of Agriculture labora
tory, the glutaraldehyde-fixed preparations were refrigerated until
scored.
The methanol-fixed slides were immersed in a solution of acridine
orange (0.02 mg/ml in 0.01 M sodium phosphate buffer, pH 7.4) for 4
min and were then rinsed in phosphate buffer for 8 min. Prior to
scoring, a coverslip was placed over the wet smear area, and the slide
was blotted to remove excess buffer. The stained slides were scored as
wet mounts.
Blood samples were scored for the frequency of micronucleus-containing cells in newly formed RNA-positive reticulocytes and mature
RNA-negative erythrocytes. Slides were scored as wet mounts at x630
by epifluorescence microscopy using an oil immersion objective and a
Zeiss fluorescein ¡sothiocyanate filter. The excitation wavelength was
450-490 nm, the dichromatic beam splitter was set at 510 nm, and the
barrier filter was 520 nm. Under these conditions of staining and
excitation, RNA-containing bodies fluoresce red-orange, and DNAcontaining bodies (nuclei and micronuclei) fluoresce bright yellow.
RNA-negative erythrocytes are a dull green color and are easily distin
guished from the red-orange RNA-positive cells. Micronuclei were
defined as intracellular bodies with the characteristic yellow fluores
cence of DNA in addition to the other characteristics of micronuclei
described in the literature (13). The frequencies of micronucleus-containing erythrocytes in each sample were determined using a semiautomated procedure in which the count of total cells scored was based
upon specified subfields of known area.
The numbers of micronucleus-containing cells among 2,000 RNApositive erythrocytes and among at least 10,000 RNA-negative eryth
rocytes were determined for each sample. The total number of RNApositive cells scored was estimated by counting the number of RNApositive cells in every fifth field scored and was based upon a count of
at least 400 cells. The total number of RNA-negative cells scored was
estimated by counting the number of RNA-negative cells in an area
IN HUMANS
Pitted cells accumulate in the erythrocyte population of subjects who
lack a functioning spleen but occur at a very low frequency (<2%) in
individuals with normal spleen function. A sample of fixed blood in 12 drops of fixative solution was placed on a glass slide with coverslip.
The specimen was examined under Normarski optics at xlOOO mag
nification. Pitted cells were defined by the presence of an apparent cell
surface depression with a continuous, smooth, and defined edge and
evident shadowing. The size of this apparent depression may range
from 0.05-0.2 times the cell diameter. Cells with one or more surface
depressions meeting these criteria were classified as pitted. At least 500
cells were scored/specimen. Subjects with a frequency of pitted eryth
rocytes of >12% were considered to lack splenic function (14).
Statistical Analysis. Laboratory and questionnaire data were merged
and analyzed using the statistical software packages BMDP (15) and
SAS (16, 17). Questionnaire responses were used to define various
demographic, dietary, or life-style factors that might be associated with
the frequency of micronucleated cells. All of the codings and groupings
of these factors were made without knowledge of any individual's
micronucleus frequency.
Coffee and tea consumption were combined to create a variable
representing daily total caffeine intake from these two sources. The
daily consumption of each were weighted proportionately to their
approximate caffeine content (18) and summed as follows: 1 x (cups
of caffeinated coffee/day) + 0.4 x (cups of tea/day) = cups of coffee
equivalents/day.
We also asked subjects whether they took regular doses of vitamin
or mineral supplements. Because relatively few subjects reported taking
each specific vitamin or mineral, the responses were combined for those
taking B vitamins and those taking vitamins C, E, or A. Grouping of
vitamins C, E, and A was based on evidence that their antioxidant
properties may defend against genetic and carcinogenic damage (19).
Occupational title, job duties, and reported exposures to solvents,
metal fumes, pesticides, or other chemicals were used to group subjects
into those with a high potential for genotoxic exposure, those with
relatively less potential, and those with little or no potential.
To examine possible effects of X-ray exposures, we considered only
X-rays to major portions of the body (such as trunk, pelvis, limbs) and
excluded dental X-rays or X-rays of distal extremities.
The frequencies of micronucleus-containing cells in the two cell
populations (reticulocytes and erythrocytes) were analyzed separately.
To investigate the association between various demographic or life
style factors and the frequency of micronucleus-containing cells, we
compared the frequencies between subgroups of categorical variables
(such as gender) or examined scatter plots of micronucleus-containing
cell counts against continuous variables, such as age and coffee con
sumption.
The frequency distribution of micronucleated cells/1000 cells was
skewed to the right and appeared to follow a log-normal distribution.
Therefore, we derived a log transformation of the frequency of micronucleus-containing cells as In (number of micronucleus-containing
cells/number of cells scored). Hence, the means we report for micronucleated cell counts/1000 cells are geometric means.
To assess the independent statistical relationships of several variables
simultaneously, we used a multiple linear regression analysis (20).
Model goodness of fit was assessed by R2, which is interpreted as the
proportion of the variance in the log micronuclei frequency accounted
for by the independent variables in the model (20). The dependent
variable in each model was the natural log of the number of micronu
cleus-containing cells/1000 cells, as described above. The regressions
used as independent variables the demographic and life-style factors
5050
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1990 American Association for Cancer Research.
MICRONUCLEATED
ERYTHROCYTES
that were associated with micronucleus counts in the unadjusted analy
sis or that we thought may be associated a priori. For ease of presen
tation and interpretation, the independent variables were kept untransformed. More complicated representations (e.g., logarithmic, quadratic)
did not significantly improve the fit. Hence, the regression coefficient
for each independent variable indicates the change in log frequency of
micronucleated cells associated with a one unit change in the independ
ent variable. Alternatively, the antilog of the regression coefficient can
be interpreted as the multiplier of the micronucleated cell frequency
associated with a unit change in the independent variable.
In order to compare the strengths of the associations of various study
variables, we used the fitted models to calculate the micronucleus counts
predicted by varying each independent variable, holding the others
constant (21).
IN HUMANS
n
r>
CO
6
Micronucleated
10
12
14
Cells per 1000 Reticulocytes
RESULTS
Enrollment of Study Subjects. Of the 303 patients identified
with a discharge diagnosis of splenectomy, 174 had either
terminated their membership in the health plan, had moved
without providing the health plan with a forwarding address,
or were deceased. Of the remaining 129 persons, 48 (37%)
declined to participate, or failed to provide both a blood speci
men and a completed questionnaire. This left 81 subjects (63%
of those who could be contacted) for whom we received both
laboratory and questionnaire data.
0
4
8
10121416
In order to eliminate subjects with evidence of splenosis or
Micronucleated Cells per 1000 Erythrocytes
regenerated splenic function, we excluded 37 subjects who had
Fig. 1. Frequency distribution of micronucleated cells/1000 reticulocytes (top)
pitting in <12% of the erythrocytes examined. The statistical
and erythrocytes (bottom) from 44 asplenic individuals.
analysis below is based on the remaining 44 confirmed asplenic
subjects.
mean of 2.7/1000. These values are similar to values reported
The asplenic subjects contained more males than females (26 for bone marrow erythrocytes (7).
and 18, respectively), and the predominant ethnic group was
In the unadjusted analysis, micronucleated cell frequencies
white, non-Hispanic. The median age of the group was 42 years,
appeared directly associated with age and consumption of coffee
with the youngest member being 5 years old and the eldest 80 equivalents, while there was no apparent trend with cigarette
years old. Nearly half (46%) of the group reported a total
consumption (Fig. 2).
household income of >$30,000/year. None of these subjects
For the multivariate analysis, Table 1 lists the independent
reported receiving chemotherapy within the previous 4 months,
variables included in the regression, the units for each variable,
although eight persons reported receiving major X-ray expo
its regression coefficient and P value, and the predicted microsure.
nucleus frequency multiplier (the exponentiation of the regres
The questions about dietary and personal habits asked re
sion coefficient) with its 95% confidence interval. These coef
spondents to report their average experience during the pre
ficients are "adjusted" for all other factors included in the
vious 4 months. Sixteen (36%) reported smoking cigarettes,
model, in the sense that they are calculated in the presence of
with a median consumption in the category of 0.5-1 pack/day.
the
other factors.
Twenty-nine (66%) drank coffee (median consumption, 1-2
For reticulocytes, the regression model has an R- of 0.41,
cups/day), 10 (23%) drank decaffeinated coffee (median, <1
indicating that 41% of the variance in log micronucleus fre
cup/day), and 12 (27%) drank tea (median, <1 cup/day).
quency was accounted for by the 11 variables in the regression.
Ten (23%) subjects reported taking daily doses of at least one
of the vitamins A, C, or E. Four (9%) took regular doses of B The variables in this model that were statistically significantly
vitamins, 10 (23%) took multiple vitamin preparations, and six associated with higher micronucleus frequencies were the con
sumption of coffee equivalents and taking calcium supplements.
subjects (14%), all of whom were females, reported taking
Taking vitamin A, C, or E supplements was significantly asso
calcium supplements.
with lower frequencies.
The four subjects classified as having a job with a high ciated
The equivalent analysis for erythrocytes had an R2 of 0.43
potential for genotoxic exposure included three workers (a
(Table 1). Male gender and coffee equivalents consumption
radiation worker, a pipefitter, and a shipfitter) whose reported
were
significantly associated with increased micronuclei. The
exposures included halogenated organic solvents or welding
materials and a hairstylist who worked with hair dyes and coefficients for taking vitamins A, C, or E and calcium supple
ments, while approximately similar in magnitude to those in
permanent wave solutions. Subjects in the medium potential
group included a wallpaper hanger and a rancher who used the reticulocyte analysis, did not reach statistical significance
(0.1 < />< 0.2).
herbicides.
Results of the Micronucleus Assay. The numbers of microThe regression models given in Table 1 were used to show
the changes in the mean micronucleus-containing cell counts
nucleus-containing cells/1000 reticulocytes in the 44 asplenic
subjects ranged from 0.5-15.9/1000, with a geometric mean of (and 95% confidence intervals) that would be predicted from
3.3/1000 (Fig. 1). The number of micronucleus-containing
varying each independent variable (Fig. 3). These predictions
cells/1000 erythrocytes was 0.3-8.6/1000, with a geometric
were calculated against the predicted micronucleated cell counts
5051
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1990 American Association for Cancer Research.
MICRONUCLEATED
ERYTHROCYTES
S
Q
fio
InY . -6.384 + 001IX
(p.O 09)
6 399 t 0016X
(p.001)
20
40
60
20
40
60
Age in Years
Age m Years
£100-
I
10-
IN HUMANS
average of 5 cups/day and taking regular doses of calcium
supplements are approximately comparable (an increase of
about 2/1000). An increase of about 1/1000 reticulocytes is
predicted/30-year increase in age, but this association was not
observed for erythrocytes. Regular daily doses of vitamin A, C,
or E would be predicted to lower micronucleus-containing cell
counts to approximately one-half the baseline value.
The results in Fig. 3 are intended to illustrate the relative
strengths of the statistical associations predicted by the factors
investigated in this cross-sectional sample. These are, of course,
hypothetical predictions derived from a statistical model. In
practice, changing the value of one variable (e.g., reducing a
person's coffee consumption) might well result in the changing
of another as well (e.g., increasing tea or decaffeinated coffee
consumption).
InY . -5936 . 0 IdSX
InY . -6119 . 0133X
(p.O 03)
(P.Oon
0246
Collée
Equn/aleniCups per Day
5 663
O
05
10
InY . -5 912 (P-091Õ
•¿
002BX
O 191X
(p-0 38)
0
15
Packs ol Cigarettes per Day
DISCUSSION
0246
Coffee EquivalentCups per Day
05
10
'5
acKSof Cigarettes per Day
Fig. 2. Micronucleated cells/1000 reticulocytes (A) and erythrocytes (•)from
44 asplcnic individuals by ago of the individual (lop), caffeine intake measured as
cups of coffee equivalents/day (middle), and packs of cigarettes smoked/day
(bottom). Note that the vertical axis is a log scale. The least squares fit line and
the regression equation are given for each plot, as well as the P value for the test
that the slope is zero.
for a hypothetical splenectomized subject with "standard" val
ues for each independent variable: a woman aged 42 years (the
mean) who did not drink coffee, decaffeinated coffee, or tea,
did not take vitamin or calcium supplements, did not work in
a potentially genotoxic job. had not had an X-ray in the last 4
months, and did not smoke. The regression models predict that
the mean micronucleus frequencies for such a person would be
2.4 micronucleus-containing
reticulocytes/1000 and 1.4 micronucleus-containing erythrocytes/1000. Fig. 3 indicates that
the predicted effects of increasing coffee consumption by an
This study has demonstrated the feasibility of recruiting
splenectomized subjects for an epidemiological study and has
shown statistically significant associations between the micronucleus frequency in circulating reticulocytes or erythrocytes
and three dietary factors: consumption of caffeinated coffee
and/or tea, vitamin A, C, or E, and calcium supplements.
Caffeine has long been a suspected carcinogen, primarily for
cancers of the pancreas, bladder, kidney, and ovary (22). Caf
feine is known to have adverse effects on DNA synthesis and
mitosis as well as DNA repair processes (23), but, in general,
human epidemiological studies have not revealed any strong or
consistent relationship to any cancer (22). Pancreatic cancer
has been most studied because of an initially positive association
with coffee consumption (24), but subsequent studies have not
borne out this relationship (25). Nevertheless, on the basis of
the long-term suspicion of caffeine as a carcinogen, the findings
of the current study are of special interest and deserve further
study. We also note that reported intake of decaffeinated coffee
was not associated with a similar increase in micronuclei.
Consumption of coffee has recently been reported to be asso
ciated with increased chromosomal fragility (26) and sister
chromatid exchange frequency (27) in peripheral blood lym
phocytes.
The possible protective role of vitamins A, C, and E in the
etiology of cancer has received an increasing amount of support
from animal as well as human observational and intervention
studies (28). Vitamin A is important for cell differentiation and
proliferation, a deficiency of which is an important factor in
Table 1 Multiple regression analysis of micronucleated cell frequency" among reticulocytes and erythrocytes in 44 splenectomized subjects
ReticulocytesVariableGenderAgeCoffee
of
of
coefficient (95%
coefficient (95%
confidence
limits)1.34(0.85-21.11 coefficient0.6040.0250.1540.085-0.3590.587-0.6410.020-0.1960.5290.0
confidence
limits)1.83(1.13-2.97)1
coefficient0.2920.1090.135-0.018-0.6480.7370.3780.018-0.1030.021-0.159P"0.220.100.050.870.020.050.360.940.780.940.47Ex
equivalentsDecaffeinated
coffeeVitamin
ECalciumA. C. or
supplementsB
vitaminsMultiple
vitaminsGenotoxic
exposuresX-rayCigarette
smokingUnitMale
female10
vi.
y1
cup/day1
cup/dayDaily
vs.Daily
dose
vi.Daily
dose
vi.Daily
dose
vi.nonononoHigh
dose
potentialYes
i'i. low
no1 vi.
pack/dayRegression
(0.98-11.14(1.01-10.98
(0.90-1.17(1.02-1.09(0.87-0.70(0.41-1.80(
.03
(0.79-112)26)30)22)0.52
(0.32-0.86)2.09(1.04-41.46(0.66-31.02(0.64-10.90(0.44-11.02(0.58-10.85(0.56-1A2
(0.23-1.02(0.62-0.82
(0.38-.17).34).37).19)(.76).22).68).76)1.70(0.92
= 0.4119)22)63)85)81)31)Regression
=Exponential
0 Dependent variable expressed as )n(micronucleated cells/1000 cells).
* P value for two-tailed test of hypothesis that regression coefficient is 0.
5052
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1990 American Association for Cancer Research.
MICRONUCLEATED
ERYTHROCYTES
Fig. 3. Micronucleated cells/1000 reticuloeytes (top) and erythrocytes (bottom) pre
dicted by the multiple regression model for
changes in the independent variables. The
point estimates (*) and 95% confidence inter
vals (bars) represent the frequency of micronucleated cells predicted by varying each inde
pendent variable separately, as described. The
shifts in micronucleus counts are compared to
a baseline predicted value (
) calculated
when all independent variables are at their
reference levels.
IN HUMANS
8
o
ü
<"
O
in
E
co
•¿..4.4.4.
carcinogenesis. Also, carotene is known to have antioxidant
effects which protect against damage to DNA (29). In epide
miológica!studies, an association with reduced lung cancer risk
has been most strongly supported (28, 30) and enough evidence
exists to justify randomized double-blind controlled trials of
the effect of ß-carotenesupplements on cancer outcomes (31).
Vitamins C and E are also antioxidants which reduce DNA
damage and mutations and could thereby protect against car
cinogenesis (28). Little evidence currently exists from human
studies for a protective effect of vitamin C, but a recent study
of serum vitamin E in a large cohort of Finnish men suggests a
reduced risk of all cancers among those with high levels com
pared to those with low levels (32).
Calcium has, until recently, not been considered of etiological
importance in cancer. However, recent studies from several
sources have suggested that dietary calcium may be associated
with a reduced risk of colon cancer (33, 34). Calcium may bind
with fats in the intestine which make these fats less accessible
to metabolism by bile salts and the subsequent formation of
carcinogenic breakdown products (35). Our findings are not
consistent with the epidemiological literature in that persons
with high calcium intake in our study had higher levels of
micronuclei. It may be that taking calcium supplements was a
marker for something else not measured in our study, such as
low estrogen levels in women or contaminants in over-thecounter calcium supplements.
Several studies have reported a strong association of micronucleus frequency with age in human peripheral blood lympho
cytes (9, 36-38). Simple scatter plots of our erythrocyte data
against age (Fig. 2) showed a trend similar to that reported for
lymphocytes. However, the multiple regression analysis sug
gested that the age effect apparent in our unadjusted analysis
was partly due to confounding by other factors (such as coffee
consumption) that were correlated with age. An analysis of our
data relating micronucleated cell counts to age only, as, for
example, was done by Fenech and Morley (9), would have
predicted a 17% increase in micronuclei among reticulocytes,
and a 12% increase among erythrocytes, per 10-year increase
in age. In contrast, the multivariate analysis predicted an in
crease in micronuclei per 10-year increase in age of only 11%
for reticulocytes and 3% for erythrocytes.
Although smoking has been reported to be associated with
an increased micronucleus frequency in both peripheral blood
lymphocytes (37, 39) and in erythropoietic bone marrow cells
(39), we observed no evidence of an association between smok
ing and increased micronucleus frequency in the present study.
This may have been due in part to the relatively low exposures
in this study group. No subject reported smoking as many as 2
packs of cigarettes/day, and only three smoked >1 pack/day.
Serum folate deficiency has recently been shown to increase
the frequency of micronuclei in erythrocytes (12) but could not
be evaluated in the present study. Recent observations in labo
ratory animals demonstrated a strong enhancement of caffeineinduced micronucleus formation by folate deficiency (40). These
findings suggest an interaction between folate status and caffeinated beverage consumption. It will be important to monitor
folate status in future studies and to assess its possible interac
tion with other factors.
The main limitation of this study was its small sample size.
Our group of 44 confirmed asplenic subjects was too small to
adequately assess the effects of factors with low prevalence,
such as specific occupational exposures. Even though 81 sub
jects had completed questionnaire and laboratory work, we
were required to eliminate 37 subjects (46%) because they
showed evidence of residual or regenerated splenic function.
The proportion of individuals with such regenerated splenic
5053
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1990 American Association for Cancer Research.
MICRONUCLEATED ERYTHROCYTES IN HUMANS
function is believed to be highest among those splenectomized
because of trauma (14). In spite of this, we chose to restrict the
study to those with traumatic splenectomies in order to assem
ble a group that had no underlying illness.
Although a multiple regression analysis was used in this
cross-sectional study to examine the associations of several
factors simultaneously, confirmation of some of these associa
tions must await controlled intervention studies. Because the
cell population studied is rapidly dividing in vivo, short-term
experimental studies in which suspected factors are added or
withheld for several days should permit this confirmation.
REFERENCES
1. Vainio, H. Current trends in the biological monitoring of exposure to
carcinogens. Scand. J. Work Environ. Health. //: 1-6, 1985.
2. Heddle, J. A., Hile, M.. Kirkhart, B., Mavournin, K., MacGregor, J. T.,
Newell, G. W., and Salamone, M. F. The induction of micronuclei as a
measure of genotoxicity: a report of the U. S. Environmental Protection
Agency Gene-Tox Program. Mutât.Res.. 123: 61-118, 1983.
3. Goetz, P., Sram, R. J., and Dohnalova, J. Relationship between experimental
results in mammals and man: cytogenetic analysis of bone marrow injury
induced by a single dose of cyclophosphamide. Mutât.Res., 31: 247-254,
1975.
4. Högstedt, B., Gullberg. B., Mark-Vendei, E., Mitleman, F., and Skerfving,
S. Micronuclei and chromosome aberrations in bone marrow cells and
lymphocytes of humans exposed mainly to petroleum vapors. Hereditas.. 94:
179-187, 1981.
5. Krogh Jensen, M., and Huttel, M. S. Assessment of the effect of azathioprine
on human bone marrow cells in vivo, combining chromosome studies and the
micronucleus test. Dan. Med. Bull., 23:152-154, 1976.
6. Krogh Jensen, M., and Nyfors, A. Cytogenetic effect of methotrexate on
human cells in vivo. Mutât.Res., 64: 339-343, 1979.
7. Abe, T., Isemua. T., and Kikuchi, Y. Micronuclei in human bone marrow
cells: evaluation of the micronucleus test using human leukemia patients
treated with antileukemic agents. Mutât.Res., 130: 113-120. 1984.
8. Heddle, J. A.. Benz, R. D., and Countryman, P. I. Measurement of chro
mosomal breakage in cultured cells by the micronucleus technique. In: H. J.
Evans, and D. C. Lloyd (ed.), Mutagen-induced Chromosomal Damage in
Man, pp. 191-200. Edinburgh, Scotland: Edinburgh University Press, 1978.
9. Fenech, M., and Morley, A. A. Cytokinesis-block micronucleus method in
human lymphocytes: effect of in vivo ageing and low dose X-irradiation.
Mutât.Res., 161: 193-198, 1986.
10. Schlegel, R., MacGregor, J. T., and Everson, R. B. Assessment of cytogenetic
damage by quantitation of micronuclei in human peripheral blood erythrocytes. Cancer Res., 46: 3717-3721, 1986.
11. Harris, J. W., and Kellermeyer, R. W. The Red Cell. Cambridge, MA:
Harvard University Press, 1970.
12. Everson, R. B., Wehr, C. M., Erexson, G. L., and MacGregor, J. T. Marginal
folate depletion increases human chromosomal damage in vivo: demonstra
tion by analysis of micronucleated erythrocytes. J. Nati. Cancer Inst., 80:
525-529, 1988.
13. Schmid, W. The micronucleus test for cytogenetic analysis. In: W. Hollaender
(ed.), Chemical Mutagens: Principles and Methods for Their Detection, pp.
31-53. New York: Plenum Press, 1976.
14. Pearson, H. A., Johnston, D., Smith, K. A., and Touloukian, R. J. The bornagain spleen: return of splenic function after splenectomy for trauma. N.
Engl. J. Med., 29«:1389-1392, 1978.
15. Dixon, W. J., Brown, M. B., Engelman, L., Frane, J. W., Hill, M. A.,
Jennrich, R. I., and Toporek, J. D. BMDP Statistical Software. Berkeley,
CA: University of California Press, 1983.
16. SAS Institute. SAS User's Guide: Basics, Version 5. Cary, NC: SAS Institute,
Inc., 1985.
17. SAS Institute. SAS User's Guide: Statistics, Version 5. Cary, NC: SAS
Institute, Ine, 1985.
18. Roberts, H. R., and Barone, J. J. Biological effects of caffeine: history and
use. Food Toxico!., 37: 32-39, 1983.
19. Ames, B. N. Food constituents as a source of mutagens, carcinogens, and
anticarcinogens. In: I. Knudsen (ed.). Genetic Toxicology of the Diet, pp. 332. New York: Alan R. Liss, 1986.
20. Kleinbaum, D. G., Kupper, L. L., and Muller, K. E. Applied Regression
Analysis and Other Multivariable Methods, Ed. 2. Boston: PWS-Kent, 1988.
21. Greenland, S.. Schlesselman, J. J., and Criqui, M. H. The fallacy of employ
ing standardized regression coefficients and correlations as measures of effect.
Am. J. Epidemiol., 123: 203-208, 1986.
22. Ernster, V. L. Epidemiologie studies of caffeine and human health. Prog.
Clin. Biol. Res., 158: 377-400, 1984.
23. Bishun, N., Williams, D., and Mills, J. The cytogenetic effects of caffeine on
two tumour cell lines. Mutât.Res., 26: 151-155, 1974.
24. MacMahon, B., Yen, S., Trichopoulos, D., Warren, K., and Nardi, G. Coffee
and cancer of the pancreas. N. Engl. J. Med.. 304: 630-633, 1981.
25. Hiatt, R. A., Klatsky, A. L.. and Armstrong, M. A. Pancreatic cancer, blood
glucose, and beverage consumption. Int. J. Cancer.. 41: 794-797, 1988.
26. Chen, A. T. L., Reidy, J. A., Annest, J. L., Welty, T. K., and Zhou, H. G.
Increased chromosome fragility as a consequence of blood folate levels,
smoking status, and coffee consumption. Environ. Mol. Mutagenesis, 13:
319-324, 1989.
27. Reidy, J. A., Annest, J. L., Chen, A. T. L., and Welty, T. K. Increased sister
chromatid exchange associated with smoking and coffee consumption. En
viron. Mol. Mutagenesis, 12: 311-318, 1988.
28. Willet, W. C., and MacMahon, B. Diet and cancer—an overview. N. Engl.
J. Med., 310: 633-638. 1984.
29. Krinsky. N. I., and Deneke, S. M. The interaction of oxygen and oxy-radicals
with carotenoids. J. Nat. Cancer Inst., 69: 205-210, 1982.
30. Menkes, M. S., Comstock. G. W., Vuilleumier, J. P., Helsing, K. J., Rider.
A. A., and Brookmeyer, R. Serum beta-carotene, vitamins A and E, selenium,
and the risk of lung cancer. N. Engl. J. Med., 315: 1250-1254, 1986.
31. Hennekens, C. Strategies for a primary prevention trial of cancer and cardi
ovascular disease among U. S. physicians. Am. J. Epidemiol., 118:453-454,
1983.
32. Knekt, P., Aromaa, A., Maatela, J., Aaran, R. K.. Nikkari, T., Hakama, M.,
Hakulinen. T., Peto, R., Saxen, E., and Teppo, L. Serum vitamin E and risk
of cancer among Finnish men during a 10-year follow-up. Am. J. Epidemiol.,
12 7: 28-41, 1988.
33. Garland, C, Barrett-Connor. E., Rossof, A. H., Shekelle. R. B., Criqui, M.
H., and Paul, O. Dietary vitamin D and calcium and risk of colorectal cancer:
a 19-year prospective study in men. Lancet, /.•
307-309, 1985.
34. Slattery, M. L., Sorenson, A. W., and Ford, M. H. Dietary calcium intake
as a mitigating factor in colon cancer. Am. J. Epidemiol., 128: 504-514,
1988.
35. Newmark, H. L., Wargovich, M. J., and Bruce, W. R. Colon cancer and
dietary fat, phosphate, and calcium: a hypothesis. J. Nati. Cancer Inst., 72:
1323-1325, 1984.
36. Fenech, M., and Morley, A. A. The effect of donor age on spontaneous and
induced micronuclei. Mutât.Res., 148: 99-105, 1985.
37. Högstedt,B. Micronuclei in lymphocytes with preserved cytoplasm: a method
for assessment of cytogenetic damage in man. Mutât.Res., 130:63-72, 1984.
38. Yager, J. W., Sorsa, M., and Selvin, S. Micronuclei in cytokinesis-blocked
lymphocytes as an index of occupational exposure to alkylating cytostatic
drugs. In: H. Bartsch, K. Hemminki, and I. K. O'Neill (eds.). Methods for
Detecting DNA Damaging Agents in Humans: Applications in Cancer Epi
demiology and Prevention, pp. 213-216. Lyon, France: International Agency
for Research in Cancer, 1988.
39. Högstedt,B., Gullberg, B., Hedner. K.. Kolnig, A.-M., Mitelman, F., Skerf
ving, S., and Widegren, B. Chromosome aberrations and micronuclei in bone
marrow cells and peripheral blood lymphocytes in humans exposed to ethylene oxide. Hereditas., 98: 105-113, 1983.
40. MacGregor, J. T. Dietary factors affecting spontaneous chromosomal dam
age in man. In: M. Pariza, H. U. Aeschebacher, J. Felton, and S. Sato (ed.).
Carcinogens and Mutagens in the Diet. New York: Alan R. Liss, 1989.
5054
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1990 American Association for Cancer Research.
Micronucleated Erythrocytes as an Index of Cytogenetic
Damage in Humans: Demographic and Dietary Factors
Associated with Micronucleated Erythrocytes in
Splenectomized Subjects
Daniel F. Smith, James T. MacGregor, Robert A. Hiatt, et al.
Cancer Res 1990;50:5049-5054.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/50/16/5049
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1990 American Association for Cancer Research.