Review
 Future Drugs Ltd. All rights reserved. ISSN 1473-7159 89
CONTENTS
Current biodosimetry
methods
Molecular profiling by
gene expression
Expert opinion
Five-year view
Key issues
Reference
www.future-drugs.com
Biological indicators for the
identification of ionizing radiation
exposure in humans
Sally A Amundson, Michael Bittner, Paul Meltzer, Jeffrey Trent and
Albert J Fornace, Jr.
While the effects of acute high-dose irradiation are well-documented, less is known about
the effects of low level chronic radiation exposure. Physical dosimetry cannot always be
relied upon, so dose estimates and determination of past radiation exposure must often be
based upon biological indicators. Some of the established methods used in the assessment
of nuclear accidents are reviewed here, including cytogenetic analyses, mutation-based
assays and electron spin resonance. As interest in research on low-level radiation
exposures expands, there is an increasing need for new biomarkers that can identify
exposed individuals in human populations. Developments in high-throughput gene
expression profiling may enable future development of a rapid and noninvasive testing
method for application to potentially exposed populations.
Expert Rev. Mol. Diagn. 1(2), (2001)
Author for correspondance:
NIH, National Cancer Institute
37 Convent Dr., Bldg. 37,
Bethesda, MD 20892, USA
Tel.: +1 301 594 7772
Fax: +1 301 480 2514
amundson@mail.nih.gov
KEYWORDS:
biodosimetry, cDNA microarray,
cytogenetics, dose-response, gene
induction, mutation
Since in cases of accidental or suspected radi-
ation exposure physical dosimetry is often
incomplete or absent, the need for biological
indicators of exposure has long been recog-
nized. In the case of industrial accidents,
information from biodosimetry can assist in
determining dose distribution as well as over-
all exposure, important factors for triage of
affected individuals. The inclusion of
biomarker information in epidemiological
studies may also contribute to our under-
standing of the long-term risks of both acute
and chronic exposures.
While the primary biological indicators of
radiation exposure have been developed and
applied to populations exposed to acute and
often relatively large doses, such as the atom
bomb survivors, the Chernobyl `liquidators'
and even radiotherapy patients, there is
increasing interest in developing biomarkers
for lower doses and more chronic exposures.
Such exposures are generally more common,
both among radiation workers and the popula-
tion at large. Several cohorts of chronically
irradiated people are under study, such as resi-
dents of the Techa River area in Russia and to
a lesser extent the populations surrounding the
Hanford facility in the US and the Sellafield
nuclear plant in England. Attempts are also
being made to define bioassays specific for
more densely ionizing radiations, such as the
-particles produced by radon gas or the heavy
ions in cosmic rays. Such exposures can be of
concern everywhere from homes in high radon
areas, to airline personnel, to extended space-
flight, such as on the new international space
station or the planned mission to Mars.
A number of characteristics determine the
practicality and usefulness of a biological indi-
cator of radiation exposure. The time required
to perform the assay can be critical, especially
in accident or potential military situations
where dosimetry is needed as soon as possible
and may need to be applied to large popula-
tion. A short turnaround time is needed both
to assign victims to appropriate treatment and
to determine which workers have exceeded
allowable dose limits and should be removed
from high-exposure areas. Assays with a poten-
tial for automation are highly desirable in this
context, while those requiring high levels of
expertise are less useful. The time after expo-
sure when an assay is informative requires con-
sideration. Some assays are appropriate in a
Amundson, Bittner, Meltzer, Trent & Fornace
90 Expert Rev. Mol. Diagn. 1(2), (2001)
narrow window of time from hours to days after an acute expo-
sure, while a few biomarkers persist for years. For any marker,
minimal variations in the range of normal background levels
are also important, as in real world situations, individual pre-
exposure measurements are unlikely to be available for compar-
ison. The ideal biomarker would be specific for ionizing radia-
tion exposure. Unfortunately, many of the available markers
can be confounded by effects of age, smoking or other environ-
mental toxins. It may, however, be possible to define a signa-
ture of the more densely ionizing high linear energy transfer
(LET) radiations.
Finally, a clearly demonstrable dose-response relationship is
necessary. While some in vivo relationships have been deter-
mined in animal models or radiotherapy patients, many bio-
dosimetry techniques rely on calibrations to dose-response
curves determined in vitro. When all these factors are consid-
ered, it is clear that there is no one perfect biodosimeter for all
applications. Continuing technological improvements and
combined approaches may provide valuable support for medi-
cal triage, epidemiology and mechanistic insight into the late
effects of ionizing radiation exposure.
Current biodosimetry methods
The hematopoietic system contains some of the most radiation-
sensitive and easily sampled cells in the human body. This has
been exploited by many of the biodosimetry methods developed
to date, including those based on somatic mutation, cytogenetics
and gene expression (TABLE 1). One of the earliest and most direct
methods of dose determination following radiation exposure
involves charting daily counts of different cell types circulating in
the peripheral blood. Total leukocyte counts decline rapidly in
the first week following radiation exposures in excess of about 1
Gy and the extent and duration of the decline and subsequent
recovery have been shown to correlate well with dose [1]. Total
body irradiation (TBI) doses of 1 Gy and higher can also be well
estimated from peripheral blood neutrophil counts. The dose
estimates derived from these two methods agree closely with each
other and were widely used and confirmed following the Cher-
nobyl accident as well as other well documented accidents at
research facilities in Russia [2].
Electron spin resonance of dental enamel
Electron spin resonance (ESR) is a physical measurement of
absorbed dose that can be applied to biological material.
Although samples of fingernails and clothing have been used in
ESR determinations of dose shortly after exposure, dental
enamel is the most widely used material. One advantage of the
dental enamel technique is that the absorbed dose can be deter-
mined many years after exposure with no decrease in accuracy.
The free electrons produced in the dental enamel by radiation
exposure are measured by ESR and most commonly compared
to a calibration curve generated in the laboratory by irradiation
of dental materials with known doses of radiation [3]. ESR is
now considered sensitive enough to detect low doses of about
Table 1. Comparison of some of the major methods of biodosimetry for radiation exposure.
Assay Ease of assay Baseline
variability
Detection
limit (Gy)
Factors confounding
specificity
Post-exposure
duration
Types of
exposures
AP
ESR Large equipment
requirement
To 3-fold 0.1 Ingested  emitters Indefinite (years) Acute or chronic
TBI
No
Blood counts Simple, daily counts
over several weeks
Low 0.5­1 Physical stress, injury
other toxic exposures
Weeks­months Acute TBI only No
Somatic mutation
gpa by flow cytometry Rapid, simple Moderate 1­2 Age, smoking Indefinite (years) Acute or chronic Yes
hprt ­ T-cell cloning Simple, 2+ weeks Moderate 1­2 Other toxic exposures Transient (<1 year) No
Autoradiography Rapid, cheap, simple Transient (<1 year) Yes
Chromosomal
Unstable ­ dicentrics Technical expertise Moderate 0.5 Relatively rad. specific Several years? Acute No
Micronuclei Rapid, simple Moderate 0.1­0.3 Exposure to clastogens 1 year half-life Partial vs. TBI Yes
PCC Rapid, cheap, simple Moderate 0.1­0.5 Exposure to clastogens Days­weeks Partial vs. TBI Yes
Stable ­ FISH Technical expertise Moderate 0.1­0.25 Low F-ratio specific for
high LET (?)
Years High vs. low
LET(?)
No
Gene expression Technical expertise Low 0.2 Specificity At least 72 h Acute or chronic Yes
Profiles Potential to simplify Undetermined TBI
AP: Automation potential.
Radiation biomarkers
www.future-drugs.com 91
0.1 Gy and for higher doses retrospective ESR measurements
have agreed well with dose estimates made by blood counts and
chromosome aberration methods [2].
Cytogenetic methods
Dicentric assay
Ionizing radiation is a strong clastogen, causing chromosome
breakage and resulting in cytogenetic aberrations in exposed
cells. A number of cytogenetic methods have therefore been
developed as measures of radiation exposure and when applied
in accident situations ­ such as that at Chernobyl ­ have gen-
erally produced dose estimates agreeing well with physical
dosimetry or ESR [1,2].
The dicentric assay in particular has been one of the most
widely applied techniques for radiation biodosimetry and among
all available assays is still considered the most specific for ionizing
radiation damage [4]. In this assay, peripheral lymphocytes are
separated from the blood and stimulated to divide in culture.
Then, using standard staining or hybridization with centromere-
specific FISH probes, metaphase chromosome spreads are scored
for the observed frequency of chromosomes that have two cen-
tromeres, the so-called dicentric chromosomes. Radiation dose is
then estimated from comparison to a standardized curve
obtained from human lymphocytes irradiated in vitro. While sig-
nificant increases in dicentric frequencies have been documented
following in vitro doses above 0.02 Gy [5], practical detection
limits for in vivo exposures appear to be closer to 0.5 Gy [2].
Studies of radiotherapy patients have also suggested that in vivo
yields of dicentrics may indeed be considerably lower than
those predicted by in vitro calibration curves [6].
As centromeres are the site of chromosome attachment to the
mitotic spindle, chromosomes with two centromeres will be
unable to segregate properly into daughter cells at mitosis. This
means that dicentrics are unstable aberrations and the lym-
phocytes bearing these informative chromosomes in the periph-
eral blood decline over time with kinetics that are not yet fully
understood. A study following 15 people exposed in a radiolog-
ical accident in Goiania, Brazil, suggests that the rate of decline
of the dicentric frequency may depend on the initial dose, with
higher doses declining most rapidly and lower doses producing
more stable dicentric frequencies [7]. Despite uncertainties in
interpretation of dicentric frequencies obtained at long times
after radiation exposure, this assay remains one of the most
practical shortly after exposure.
Micronucleus assay
Another cytogenetic assay used for biodosimetry is the scoring
of micronucleus formation. This method has several advan-
tages over the dicentric assay in that it requires less specialized
expertise, is more rapid and hence can more readily be applied
to monitoring large populations. In this assay, lymphocytes
are mitogenically stimulated in culture, then cytokinesis is
blocked by cytochalasin-B. This results in mitosis and nuclear
division without cell division. Extranuclear chromatin parti-
cles (micronuclei) are then counted in binucleated cells.
Unlike the fairly radiation-specific dicentric assay, micronu-
clei can be induced by a range of other clastogens, including
cigarette smoking and exposure to clastogenic chemicals.
While most spontaneously arising micronuclei contain cen-
tromeres, it has been found that the majority of radiation-
induced micronuclei represent acentric fragments formed by
chromosome breakage. The use of centromeric FISH probes
to allow rapid scoring of only acentric micronuclei has
enhanced specificity and lowered the dose detection limit of
this assay to between 0.1­0.2 Gy [8,9].
Premature chromosome condensation
Radiation damage can also be detected in interphase cells by the
premature chromosome condensation (PCC) assay. This method
classically uses fusion of the test cells with mitotic cells, which
transmit a signal for dissolution of the nuclear membrane and con-
densation of the interphase chromosomes as if in preparation for
mitosis. Fragments in excess of the expected 46 chromosomes
occur when breaks were present within the interphase chromo-
somes. More recent refinements using chemical induction of PCC
and FISH probes for chromosome painting have increased the
speed and accuracy of the assay [10­12]. Excess PCC fragments have
been shown to increase with increasing radiation exposure [13].
Heterogeneous exposures resulting in overdispersion of
cytogenetic markers
The distribution of cytogenetic damage among cells can also
be used to estimate partial body exposures, an important
aspect of biological dose reconstruction. Following TBI the
observed aberrations would be expected to follow a Poisson
distribution among the scored cells, while an overdispersion
of aberrations would indicate that only a subset of the cells
were in the radiation field. Two mathematical approaches for
this type of analysis were initially developed with the dicentric
assay and continue to provide improved dosimetry in cases of
heterogeneous exposure [14­16]. The extent of overdispersion
of PCC fragments has also been shown to correlate with the
irradiated fraction in human lymphocyte cultures in vitro, as
well as for rhesus monkeys irradiated in vivo [17,18].
Determining LET of exposure from relative aberration frequencies
It may also be possible to distinguish between high and low
LET radiation exposures by comparing different types of
chromosome aberrations. High LET radiation, such as neu-
trons or -particles are more densely ionizing than X- or -
rays. When such radiations interact with a cell, they produce
multiple lesions in close proximity to each other, as ioniza-
tions occur along the particle track. The creation of more
sites of damage closer together increases the probability that
multiple double-strand breaks (dsb) will result within a sin-
gle chromosome, whereas the dsb produced by -rays and
most chemical clastogenic agents are more or less randomly
distributed (FIGURE 1). Thus, high LET irradiation has been
predicted to result in a greater number of aberrations involv-
ing a single chromosome, such as pericentric inversions and
Amundson, Bittner, Meltzer, Trent & Fornace
92 Expert Rev. Mol. Diagn. 1(2), (2001)
ring chromosomes, while -rays should favor interchromo-
somal events, such as translocations and dicentrics [19].
Consistent with this prediction, the ratio of interchromo-
somal (between different chromosomes) to intrachromosomal
(exchanges within a single chromosome) aberrations, termed
the F value, has been found to be significantly reduced in two
cell lines irradiated with -particles as well as in human periph-
eral blood lymphocytes irradiated by neutrons [20,21]. Analysis
of F values for the atomic bomb survivors has also indicated
biological evidence supporting a major neutron component in
the Hiroshima dose [22]. The usefulness of low F values as a spe-
cific marker for high LET radiation exposure remains contro-
versial, however, not all studies have found significant differ-
ences in this ratio for high and low LET radiations [23].
Comparisons of yields of intra-arm interchanges (paracentric
inversions or acentric rings) to interarm intrachanges (pericen-
tric inversions or centric rings) may eventually provide a supe-
rior marker for high LET exposures [24]. Finally, as FISH and
chromosome painting techniques have allowed scoring of more
complex chromosome aberrations, especially in samples
exposed to high LET. Recent experimental evidence suggests
that insertions may also provide a specific marker for past
exposure to high LET radiation [25].
Somatic mutation
Repair of DNA damage caused by radiation in hematopoietic
stem cells can result in somatic mutations in marker loci that
can be monitored as biological indicators of dose. Mutations in
several different loci have been exploited for detection of radia-
tion exposure, including expression of hemoglobin (Hb) and
glycophorin A (gpa) variants in erythrocytes and mutations at
the HLA or hypoxanthine-guanine phosphoribosyltransferase
(hprt) loci in T-lymphocytes. A drawback common to these
somatic mutation end-points is their relative lack of specifi-
city for radiation exposure as other environmental exposures
or physiological states can also increase the observed mutant
frequencies in vivo.
GPA variants
The gpa assay in erythrocytes has been widely used for bio-
dosimetry. Although mature human red blood cells do not have
nuclei, mutations occurring in progenitor cells in the bone
marrow can be monitored by the measurement of phenotypic
variants among circulating erythrocytes. Two alleles of gpa
encode the cell surface proteins that determine the M and N
blood group antigens. In M/N constitutional heterozygotes,
variant red blood cells expressing only one allele can be quanti-
fied rapidly by flow cytometric techniques [26]. The obvious
drawback to this method is that it can only be applied to heter-
ozygotes, approximately 50% of the population. The mutant
progenitor cells appear to persist for years in the bone marrow,
however a significant dose-response relationship has been
found for gpa variants even years after high-dose acute expo-
sures in atomic bomb survivors and victims of the Chernobyl
accident and in lower dose-rate exposures in patients treated
with Iodine-131 [27­29]. No correlation with dose was found
among workers at the Sellafield nuclear plant, perhaps due to a
relatively high apparent threshold dose (about 1­2 Gy) for
detection of significant gpa variants [30].
Hprt and mutant spectra
Functional inactivation of the hprt gene has probably been the
most extensively used of the T-cell biodosimetry assays. In con-
trast to the erythrocyte assays, the T-cell assays monitor muta-
tions occurring directly in the circulating peripheral cells. The
hprt gene codes for a salvage pathway enzyme that allows the
phosphoribosylation of hypoxanthine and guanine as precur-
sors for DNA synthesis. It can also utilize purine analogs, such
as 6-thioguanine, which can then incorporate into DNA and
kill the cells. Mutant cells that have lost this enzyme can grow
in concentrations of 6-thioguanine that are toxic to wild type
Cell nucleus
Centromere
Site of
chromosome
breaks
Stable aberration
(pericentric inversion)
or
+
Unstable aberration
(centric ring & acentric
fragment)
High LET
Low LET
Stable aberration
(translocation)
+
+
or
Unstable aberration
(dicentric & acentric
fragment)
Figure 1. Representation of radiations of different quality interactingFigure 1. Representation of radiations of different quality interactingFigure 1. Representation of radiations of different quality interactingFigure 1. Representation of radiations of different quality interacting
with chromosomes in a cell nucleus.with chromosomes in a cell nucleus.with chromosomes in a cell nucleus.with chromosomes in a cell nucleus. The dashed arrows represent the track
of a particle or photon passing through the nucleus. In the case of high LET
radiation, the ionizations occur close together. An ionization within a
chromosome may result in a break in the DNA. If such breaks are improperly
repaired, they may be resolved as either stable or unstable chromosome
aberrations as illustrated to the right of this figure. The same dose of low LET
radiation will create more diffuse damage from a larger number of photon
tracks with ionization events occurring further apart, resulting in fewer
multiple lesions within the same chromosome. When two or more different
damaged chromosomes interact to resolve breaks, different types of stable
and unstable aberrations result.
Radiation biomarkers
www.future-drugs.com 93
cells, thus allowing mutant selection. Furthermore, the location
of the hprt gene on the human X-chromosome means it is func-
tionally hemizygous, allowing detection of the loss of a single
allele. An assay using T-cell cloning and hprt mutant fraction
determination has been used to show a strong relationship
between dose and induced mutations in atomic bomb survivors
and patients receiving high doses of radiation therapy [31,32]. An
increase in hprt mutant fractions may also be detectable follow-
ing lower dose exposures but these results seem more variable
depending on the time of sampling [33].
An additional benefit of the T-cell cloning assay is the poten-
tial for molecular analysis of mutants arising in vivo and the
subsequent determination of induced and spontaneous muta-
tional spectra to enhance the specificity of the assay. While
spontaneous hprt mutants show a wide variety of point muta-
tions and deletions [34,35], an increase in gross structural
changes has been seen in individuals undergoing radiation ther-
apy [36]. The observed mutations also show an increase in size
and frequency of deletions correlating with dose. In vitro expo-
sure of human T-cells to radon resulted in an increase in small
partial deletions with a low frequency of total gene deletions
[36], suggesting a possible LET-dependent component in the
hprt mutant spectrum. Although molecular mutant spectra
have been derived largely from in vitro results, there does
appear to be general agreement between these studies and the
existing data on in vivo exposures [36].
One of the drawbacks of the cloning assays is the length of
time required before results are obtained. A short-term autora-
diography or immunofluorescence assay can also be used to
more rapidly quantitate in vivo hprt mutations [37]. This assay is
based on the fact that only cells that have lost expression of hprt
will incorporate tritiated thymidine or bromodeoxyuridine in
the presence of 6-thioguanine in short-term phytohemaggluti-
nin-stimulated cultures. The labeled cells can then be quanti-
fied using either standard or automated image analysis. Results
obtained from individuals suspected of radiation exposure in
two different accidents suggest this assay may be useful as an
indicator of exposure [38]. As with the gpa assay, a relatively high
threshold of 1­2 Gy for detection limits the usefulness of the
hprt assay in low-dose exposure situations.
Molecular profiling by gene expression
Recent technological advances may allow an additional exploita-
tion of the molecular responses of cells to ionizing radiation.
Exposure of cells to DNA-damaging agents elicits a highly com-
plex molecular response, much of which is mediated through
changes in gene expression. A transcriptional response to genoto-
xic stress, estimated to involve 1% or more of the genome, was
initially identified in yeast and similar complex transcriptional
responses were soon confirmed in mammalian cells [39­42]. The
stress response pathways responding to different environmental
and physiological stresses have many overlapping components,
including growth factors, cytokines, oncogenes and genes
involved in cell cycle, apoptosis, signaling pathways and DNA
repair. The recent development of functional genomic
approaches to simultaneously quantify expression of thousands
of genes in a single experiment may allow the determination of
expression signatures indicative of exposure to ionizing radiation
or other environmental toxins. Although presently still in the
speculative realm, this approach would be highly attractive as it
would be amenable to rapid, even automated, noninvasive analy-
sis and may additionally have the potential to discern competing
effects from incidents involving different quality radiations or
mixed chemical and physical exposure components.
A number of high-throughput gene expression measurement
methods are currently available, including serial analysis of gene
expression (SAGE), oligonucleotide arrays and cDNA arrays
[43,44]. In our own laboratory, we have used cDNA microarrays
constructed by printing PCR amplified cDNA sequences onto
glass followed by hybridization to 2-color fluorescent labeled
probes [101]. The measurements made by this technique agreed
with those obtained by more conventional single probe quanti-
tative hybridization analysis of genes induced by -irradiation
in a human tumor cell line [45].
As gene expression profiles of radiation exposure are devel-
oped, it will remain crucial to validate the findings by inde-
pendent means, such as single probe quantitative hybridiza-
tion or real-time PCR. While the majority of published
studies of gene induction by ionizing radiation have used
large, sometimes supra-lethal doses to ensure easily measura-
ble effects, extrapolation of the results of such studies to doses
relevant to human exposures is not entirely satisfactory.
Changes in mRNA levels have been documented at doses of
less than 1 Gy and we have recently shown five genes with a
linear relationship between induction and -ray dose in the
range of 0.02­0.5 Gy in vitro, indicating that such molecular
responses may provide potential indicators of exposure in
physiologically relevant ranges [46].
As an initial step towards a gene expression assay with biomon-
itoring applicability, we have recently used cDNA microarray
analysis to identify a set of genes with linear dose-dependent ele-
vation in human peripheral blood lymphocytes (PBL) as long as
72 h after ex vivo irradiation with between 0.2 and 2 Gy -rays
(FIGURE 2) [47]. Of the genes examined, there was only slight varia-
tion in expression levels in untreated PBL from different individ-
uals, supporting the possibility of establishing ranges of expres-
sion correlating with normal and exposed populations. Elevated
gene expression appears to be maintained for at least a day in sev-
eral organs following TBI of mice with doses from 0.2­2 Gy and
in the PBL of human patients undergoing TBI prior to bone
marrow transplantation. Although these experiments are still in
preliminary stages and require more extensive studies for valida-
tion, the possibility of using gene expression changes to monitor
for radiation exposure is extremely enticing.
Expert opinion
There is general consensus in the field that there is as yet no
perfect biological indicator of radiation exposure. An
approach of using combined dosimetry from the most
appropriate methods in a given situation has been advocated
Amundson, Bittner, Meltzer, Trent & Fornace
94 Expert Rev. Mol. Diagn. 1(2), (2001)
and is likely to be the best current option [48]. Overall, bio-
logical indicators of exposure are currently most informative
in situations of acute uniform exposure to the entire body in
relatively high doses. In such situations, measurements of
chromosome aberrations can be made within the first few
days after exposure and can assist in determining the best
course of treatment for radiation casualties. Among these
methods, scoring of dicentric chromosomes remains the
most broadly applied, i.e., while micronuclei and excess
PCC fragments may also be manageable in large populations
with suspected exposure [49]. Improvements in automation
of these techniques will add to their attractiveness for triage
and other field applications.
An additional benefit of these analyses is their usefulness in
reconstructing partial exposures, a critical concern in many
accident situations where physical dosimetry is generally inade-
quate. As the use of FISH probes in conjunction with the clas-
sical chromosome aberration techniques is further refined,
cytogenetic methods of dose determination are likely to
become more precise and increasingly informative in the time
immediately following exposure. It should be noted, however,
that some of these more complex analyses may require a high
degree of expertise and so may be more useful for research
purposes than in the field.
Close monitoring of blood counts will also remain impor-
tant in tracking the course of disease. It is only through con-
tinued comparisons of cytogenetic measurements with dose
estimates derived from direct clinical observation and any
available physical dosimetry that calibration of dose-response
curves obtained from in vitro irradiations can be refined to
more accurately reflect the in vivo response. Measurements
from ESR or mutation in lymphocytes may also assist in
refining the estimates of absorbed dose that can be made by
more immediately applicable techniques.
It is important to continue these refinements of dose esti-
mates made by biological indicators because these measure-
ments are not only of use to prevent exceeding occupational
dose limits, but also for assignment to acute medical care. Ret-
rospective dosimetry can be important in epidemiological stud-
ies and ultimate prediction of long-term risks of radiation expo-
sure. Without accurate dosimetry information, estimates of
cancer and other risks cannot be made and exposure limits for
radiation protection cannot be set effectively.
Accurate means to distinguish between high and low LET
components of exposure are also sorely needed. The relative
biological effectiveness (RBE) of the same amount of energy
absorbed as different qualities of radiation is still a matter of
much debate and the assignment of different RBE values can
greatly alter the prediction of long-term risks for a specific
exposure. Techniques comparing the yields of various spe-
cific complex types of chromosome aberrations are currently
the best available hope in this regard. While these tech-
niques require complex FISH probes and expert analysis, if
they can be proven in vivo they will provide a powerful tool
for biological monitoring.
There are as yet no truly satisfactory biological indicators
for chronic low-dose and low-dose-rate radiation exposures.
This is of concern as the majority of exposures ­ both occu-
pational and among the population at large ­ are of this low-
dose chronic type. Since epidemiological doses are less cer-
tain and even in vitro effects can be slight and difficult to
study, the risks of these exposures are not clearly understood.
Recent scientific initiatives have stimulated basic research in
this area and as mechanisms of chronic radiation action and
biological response become better understood, it is likely that
biomarkers will be refined and developed to assist in the
monitoring of potentially exposed populations. Aside from
the slightly increased risks of cancer which may be associated
with low-dose chronic radiation exposure, the psychological
impact on populations fearing potential exposures also needs
to be considered. A screening method able to detect signifi-
cant exposure or the lack of it is likely to have a positive
effect on public health in this regard.
Figure 2.Figure 2.Figure 2.Figure 2. Overview of cDNA microarray experiment detecting geneOverview of cDNA microarray experiment detecting geneOverview of cDNA microarray experiment detecting geneOverview of cDNA microarray experiment detecting gene
expression changes in human PBL.expression changes in human PBL.expression changes in human PBL.expression changes in human PBL. The black wavy lines represent mRNA
from untreated control PBL and the open wavy lines represent mRNA from
irradiated PBL. In this case, RNA was harvested 24 h after radiation treatment.
Labeled cDNA was synthesized from the mRNA by a single round of reverse
transcription to produce the complex probe for hybridization. Array targets
were prepared by the PCR amplification of inserts from EST clones. These
were printed onto glass slides using a robotic print head and the control and
irradiated cDNA targets were mixed and hybridized to the microarray. The
slide was then scanned and analyzed with a fluorescent reader. The results of
a typical dose-response experiment are shown in the graph for a single
induced gene, DDB2. Relative induction has been normalized to one for
expression in untreated control PBL and the line was fit by linear regression.
RelativeDDB2induction
-ray dose (cGy)
0 50 100 150 200
8
6
4
2
0
Printed
array
Scan
Separate lymphocytes
from whole blood
Irradiate, culture,
then extract RNA
Control
-ray treated
Incorporate fluorochrome
in cDNA by RT reaction
Cy5
Cy3
Hybridize
to array
together
DDB2
Monitor
dose-response
Radiation biomarkers
www.future-drugs.com 95
One of the major areas offering hope for potential new bio-
logical indicators of radiation exposure is expression profiling.
Many changes in gene expression have been documented fol-
lowing radiation exposure both in vivo and in vitro. With the
advent of high-throughput screening and information-intensive
analysis, expression profiles correlating with specific types of
exposure may be defined. Although it remains to be seen what
the limits of in vivo detection would be with such methods,
they could provide a very powerful tool for biomonitoring.
Many approaches are developing to sift large data sets for
informative patterns. Clustering programs developed for the
analysis of expression array data look for correlations
between the most similar patterns, allowing molecular defi-
nition and classifying expression profiles [50,102]. Similar
mathematical analysis of cancer cell expression profiles has
identified previously unclassified subtypes of lymphoma and
melanoma [51,52]. Several organizations including Phase I and
the Health and Environmental Sciences Institute of ILSI
(International Life Sciences Institutes) are actively involved
in defining the gene response signatures resulting from
chemical exposures as informative for general toxicology and
environmental monitoring. The application of similar tech-
niques to radiation exposure may eventually enable exposure
profiling based on expression signatures unique to acute or
chronic, high or low LET radiation or absorbed dose. Once
defined, such profiles could be assessed accurately using
real-time PCR or other rapid and cost-effective methods
amenable to high-throughput and automated analysis.
Five-year view
The next 5 years are likely to see rapid advances in the arena of
expression profiling. Mathematical models and computer algo-
rithms are being developed and refined for exploitation of
expression databases tailored to specific questions and the ques-
tion of in vivo response to ionizing radiation is sure to be
included. Developments in the nascent field of proteomics ­
the quantitation of protein expression profiles ­ may also
impact on biodosimetry. As our ability to monitor protein
changes in a cell catches up with our current ability for tran-
scription profiling, proteomics may ultimately prove more
informative. While changes in transcription are common medi-
ators of cellular response to radiation, protein changes may be
more relevant, longer lasting or more indicative of long-term
cellular damage correlating with biological impact and risk.
As molecular biological tools are now available to address
mechanistic issues of cellular response to low doses and
chronic irradiation states, an explosion of basic research is
likely to inform our application of all available biological
markers. However, the continued refinement of extrapolation
between in vitro experiments and in vivo effects will perhaps
remain the most crucial factor in the success or failure of any
biological indicator of radiation exposure.
Acknowledgements
This work was supported in part by United States Department
of Energy grant ER62683. The authors would like to thank Ms
Khanh T Do, Mr Bill Giasi and Mr Sohrab Shahab for techni-
cal assistance.
Key issues
ˇ A number of factors determine the usefulness of potential
radiation biomarkers:
­ Time after exposure in which they are informative
­ Simplicity and rapidity of assay, automation potential
­ Applicability to large populations (considerations of
invasiveness)
­ Good dose-dependence in vivo as well as in vitro
­ Minimal variation in the normal range of unirradiated
individuals
­ Radiation specificity or few confounding factors
ˇ While no single method meets all these criteria in every
situation, cytogenetic assays, such as PCC, dicentrics and
chromosome aberrations scored with the aid of FISH probes
are currently among the most informative and widely used
methods for acute exposure situations. Further refinements
of the assays and validation of results in vivo are still needed.
ˇ None of the currently available biological indicators of
exposure is very satisfactory for low-dose, low-dose-rate
chronic radiation exposure. As the effects of such chronic
exposures become better understood, it will be increasingly
important to find means of monitoring populations for these
exposures.
ˇ A specific radiation signature remains the `Holy Grail' of
molecular radiation biology. While no absolute specificity for
radiation has yet been found, some types of mutations and
chromosome aberrations (i.e., dicentrics) are strongly
associated with radiation exposure, especially high LET. More
research is still needed in this area.
ˇ Expression profiling by such methods as microarray analysis
is one newly developing technology that offers great promise
for biomonitoring, both in terms of exposure assessment and
for predicting the specific risks of exposure. More research is
required, however, to prove the validity of profiling in vivo
and its predictive power.
Reference
Papers of special note have been highlighted as:
ˇ of interest
ˇˇ of considerable interest
1 Vorobiev AI. Acute radiation disease and
biological dosimetry in 1993. Stem Cells
15(Suppl. 2), 269­274 (1997).
2 Baranov AE, Guskova AK, Nadejina NM,
Nugis VY. Chernobyl experience: biological
indicators of exposure to ionizing radiation.
Stem Cells 13(Suppl.1), 69­77 (1995).
3 Pass B, Aldrich JE. Dental enamel as an in
vivo radiation dosimeter. Med. Phys 12,
305­307 (1985).
4 Hoffmann W, Schmitz-Feuerhake I. How
radiation-specific is the dicentric assay? J.
Amundson, Bittner, Meltzer, Trent & Fornace
96 Expert Rev. Mol. Diagn. 1(2), (2001)
Expo. Anal. Environ. Epidemiol. 9, 113­133
(1999).
5 Lloyd DC, Edwards AA, Leonard A et al.
Chromosomal aberrations in human
lymphocytes induced in vitro by very low
doses of X-rays. Int. J. Radiat. Biol. 61,
335­343 (1992).
6 Steinstrasser A, Kemmer W. Biophysical
investigations of the dose-effect
relationship in chromosome aberrations
of human lymphocytes caused by
thorotrast deposits. II. Biological and
medical aspects. Radiat. Environ. Biophys.
19, 17­28 (1981).
7 Ramalho AT, Costa ML, Oliveira MS.
Conventional radiation-biological
dosimetry using frequencies of unstable
chromosome aberrations. Mutat. Res. 404,
97­100 (1998).
8 Vral A, Thierens H, De Ridder L. In vitro
micronucleus-centromere assay to detect
radiation-damage induced by low doses in
human lymphocytes. Int. J. Radiat. Biol.
71, 61­68 (1997).
9 Thierens H, Vral A, Barbe M, Aousalah B,
De Ridder L. A cytogenetic study of
nuclear power plant workers using the
micronucleus-centromere assay. Mutat. Res.
445, 105­111 (1999).
ˇ Study of a population of radiation workers
exposed to low does using the
micronucleus assay and illustrating the
current state-of-the-art biomonitoring.
10 Coco-Martin JM, Begg AC. Detection of
radiation-induced chromosome aberrations
using fluorescence in situ hybridization in
drug-induced premature chromosome
condensations of tumour cell lines with
different radiosensitivities. Int. J. Radiat.
Biol. 71, 265­273 (1997).
11 Kanda R, Eguchi-Kasai K, Hayata I.
Phosphatase inhibitors and premature
chromosome condensation in human
peripheral lymphocytes at different cell-
cycle phases. Somat. Cell Mol. Genet. 25, 1­
8 (1999).
12 Prasanna PG, Escalada ND, Blakely WF.
Induction of premature chromosome
condensation by a phosphatase inhibitor
and a protein kinase in unstimulated
human peripheral blood lymphocytes: a
simple and rapid technique to study
chromosome aberrations using specific
whole-chromosome DNA hybridization
probes for biological dosimetry. Mutat. Res.
466, 131­141 (2000).
ˇ Demonstration of a rapid method for
PCC analysis without the need for mitotic
cells combining chemical induction of
PCC in unstimulated lymphocytes with
FISH painting.
13 Prasanna PG, Kolanko CJ, Gerstenberg
HM, Blakely WF. Premature chromosome
condensation assay for biodosimetry:
studies with fission-neutrons. Health Phys.
72, 594­600 (1997).
14 Saskai MS, Miyata H. Biological dosimetry
in atomic bomb survivors. Nature 220,
1189­1193 (1968).
15 Dolphin GW. Biological dosimetry with
particular reference to chromosome
aberration analysis. A review of methods.
Handling of Radiation Accidents (STI/PUB/
22.9) 215­224 (1969).
16 Voisin P, Assaei RG, Heidary A et al.
Mathematical methods in biological
dosimetry: the 1996 Iranian accident. Int. J.
Radiat. Biol. 76, 1545­1554 (2000).
ˇ Comparison of different approaches to
reconstruction of dosimetry in partial
body exposures using data from the field,
not the laboratory.
17 Blakely WF, Prasanna PG, Kolanko CJ et al.
Application of the premature chromosome
condensation assay in simulated partial-body
radiation exposures: evaluation of the use of
an automated metaphase-finder. Stem Cells
13(Suppl. 1), 223­230 (1995).
18 Darroudi F, Natarajan AT, Bentvelzen PA et
al. Detection of total- and partial-body
irradiation ina monkey model: a comparative
study of chromosomal aberration,
micronucleus and premature chromosome
condensation assays. Int. J. Radiat. Biol. 74,
207­215 (1998).
ˇ Comparison of not only of several
different biodosimetry methods, but also
of in vitro and in vivo differences.
19 Brenner DJ, Sachs RK. Chromosomal
`fingerprints' of prior exposure to densely
ionizing radiation. Radiat. Res. 140, 134­
142 (1994).
20 Schwartz JL, Hsie AW. Genetic and
cytogenetic markers of exposure to high-
linear energy transfer radiation. Radiat. Res.
148, S87­92 (1997).
21 Sasaki MS. F-value as a chromosomal
fingerprint of the quality of radiation. J.
Radiat. Res. (Tokyo) (1999) 40(Suppl.),
165­71.
22 Brenner DJ. Direct biological evidence for a
significant neutron dose to survivors of the
Hiroshima atomic bomb. Radiat. Res. 145,
501­507 (1996).
23 Schmid E, Braselmann H, Bauchinger M.
Does the limiting F value at very low doses
depend systematically on linear energy
transfer? Radiat. Res. 152, 563­566 (1999).
24 Sachs RK, Brenner DJ, ChenAM, Hahnfeldt
P, Hlatky LR. Intra-arm and interarm
chromosome intrachanges: tools for probing
the geometry and dynamics of chromatin.
Radiat. Res. 148, 330­340 (1997).
ˇ Theoretical basis of the use of
chromosome aberration ratios to
characterize past radiation exposures and
radiation quality.
25 Anderson RM, Marsden SJ, Wright EG,
Kadhim MA, Goodhead DT, Griffin CS.
Complex chromosome aberrations in
peripheral blood lymphocytes as a potential
biomarker of exposure to high-LET -
particles. Int. J. Radiat. Biol. 76, 31­42
(2000).
ˇ Presentation of experimental evidence
obtained from human lymphocytes
supporting complex cytogenetic
aberrations and insertions as markers of
high LET radiation exposures.
26 Langlois RG, Nisbet BA, Bigbee WL,
Ridinger DN, Jensen RH. An improved
flowcytometric assay for somatic mutations
at the glycophorin A locus in humans.
Cytometry 11, 513­521 (1990).
27 Langlois RG, Akiyama M, Kusunoki Y et al.
Analysis of somatic cell mutations at the
glycophorin A locus in atomic bomb
survivors: a comparative study of assay
methods. Radiat. Res. 136, 111­117 (1993).
28 Jensen RH, Langlois RG, Bigbee WL et al.
Elevated frequency of glycophorin A
mutations in erythrocytes from Chernobyl
accident victims. Radiat. Res. 141, 129­135
(1995).
29 Jensen RH, Reynolds JC, Robbins J et al.
Glycophorin A as a biological dosimeter for
radiation dose to the bone marrow from
iodine-131. Radiat. Res. 147, 747­752
(1997).
30 Tucker JD, Tawn EJ, Holdsworth D et al.
Biological dosimetry of radiation workers at
the Sellafield nuclear facility. Radiat. Res.
148, 216­226 (1997).
31 Hakoda M, Akiyama M, Kyoizumi S, Awa
AA, Yamakido M, Otake M. Increased
somatic cell mutant frequency in atomic
bomb survivors. Mutat. Res. 201, 39­48
(1988).
Radiation biomarkers
www.future-drugs.com 97
32 Messing K, Bradley WE. In vivo mutant
frequency rises among breast cancer
patients after exposure to high doses of -
radiation. Mutat. Res. 152, 107­112
(1985).
33 Messing K, Ferraris J, Bradley WE, Swartz J,
Seifert AM. Mutant frequency of
radiotherapy technicians appears to be
associated with recent dose of ionizing
radiation. Health Phys. 57, 537­544 (1989).
34 Albertini RJ, Nicklas JA, Fuscoe JC, Skopek
TR, Branda RF, O'Neill JP. In vivo mutations
in human blood cells: biomarkers for
molecular epidemiology. Environ. Health
Perspect. 99, 135­141 (1993).
35 Cariello NF, Skopek TR. In vivo mutation
at the human HPRT locus. Trends Genet. 9,
322­326 (1993).
36 Albertini RJ, Clark LS, Nicklas JA, O'Neill
JP, Hui TE, Jostes R. Radiation quality
affects the efficiency of induction and the
molecular spectrum of HPRT mutations in
human T cells. Radiat. Res. 148, S76­86
(1997).
37 Strauss GH, Albertini RJ. Enumeration of
6-thioguanine-resistant peripheral blood
lymphocytes in man as a potential test for
somatic cell mutations arising in vivo.
Mutat. Res. 61, 353­379 (1979).
38 Ostrosky-Wegman P, Montero R, Palao A,
Cortinas de Nava C, Hurtado F, Albertini
RJ. 6-Thioguanine-resistant T-lymphocyte
autoradiographic assay. Determination of
variation frequencies in individuals
suspected of radiation exposure. Mutat. Res.
232, 49­56 (1990).
39 Ruby SW, Szostak JW. Specific
Saccharomyces cerevisiae genes are expressed
in response to DNA-damaging agents. Mol.
Cell Biol. 5, 75­84 (1985).
40 Fornace AJ, Jr. Mammalian genes induced
by radiation: activation of genes associated
with growth control. Ann. Rev. Genet. 26,
507­526 (1992).
41 Herrlich P, Ponta H, Rahmsdorf HJ. DNA
damage induced gene expression: signal
transduction and relation to growth factor
signaling. Rev. Physiol. Biochem. Pharm.
119, 187­223 (1992).
42 Holbrook NJ, Liu Y, Fornace AJ, Jr.
Signaling events controlling the molecular
response to genotoxic stress. Stress-inducible
Cellular Responses 273­288 (1996).
43 Miller G, Fuchs R, Lai E. Image cDNA
clones, UniGene clustering and ACeDB: an
integrated resource for expressed sequence
information. Genome Res. 7, 1027­1032
(1997).
44 Schena M, Shalon D, Heller R, Chai A,
Brown PO, Davis RW. Parallel human
genome analysis: microarray-based expression
monitoring of 1000 genes. Proc. Natl Acad.
Sci. USA 93, 10614­10619 (1996).
ˇˇ Ground-breaking introduction of
microarrays for gene expression analysis.
45 Amundson SA, Bittner M, Chen YD, Trent
J, Meltzer P, Fornace AJ, Jr. cDNA
microarrayhybridization reveals complexity
and heterogeneity of cellular genotoxic
stress responses. Oncogene 18, 3666­3672
(1999).
46 Amundson SA, Do KT, Fornace AJ, Jr.
Induction of stress genes by low doses of
gamma rays. Radiat. Res. 152, 225­231
(1999).
47 Amundson SA, Do KT, Shahab S et al.
Identification of potential mRNA
biomarkers of human ionizing radiation
exposure in peripheral blood lymphocytes.
Radiat. Res. 154(3), 342­346 (2000).
ˇ Introduction of the use of microarray
analysis to identify new radiation
biomarkers, including ex vivo human
lymphocyte dose-response data.
48 Pass B, Baranov AE, Kleshchenko ED,
Aldrich JE, Scallion PL, Gale RP. Collective
biodosimetry as a dosimetric `gold
standard': a study of three radiation
accidents. Health Phys. 72, 390­396
(1997).
49 Lloyd DC, Edwards AA, Moquet JE,
Guerrero-Carbajal YC. The role of
cytogenetics in early triage of radiation
casualties. Appl. Radiat. Isot. 52, 1107­
1112 (2000).
50 Eisen MB, Spellman PT, Brown PO,
Botstein D. Cluster analysis and display of
genome-wide expression patterns. Proc.
Natl Acad. Sci. USA 95, 14863­14868
(1998).
ˇˇ Descriptions of the mathematical and
statistical underpinnings of agglomerative
hierarchical cluster analysis and details of
its application to expression data.
51 Alizadeh AA, Eisen MB, Davis RE et al.
Distinct types of diffuse large B-cell
lymphoma identified by gene expression
profiling. Nature 403, 503­511 (2000).
52 Bittner M, Meltzer P, Chen Y et al.
Molecular classification of cutaneous
malignant melanoma by gene expression
profiling. Nature 406, 536­540 (2000).
Websites
101 www.nhgri.nih.gov/DIR/Microarray/
main.html
ˇˇ Details specific cDNA microarray
methodology and analysis.
102 www.discover.nci.nih.gov