Radiation dose-rate effects on gene expression for human biodosimetry
Ghandhi et al. BMC Medical Genomics
Radiation dose-rate effects on gene expression for human biodosimetry
Shanaz A. Ghandhi
Lubomir B. Smilenov
Carl D. Elliston
Sally A. Amundson
Background: The effects of dose-rate and its implications on radiation biodosimetry methods are not well studied in the context of large-scale radiological scenarios. There are significant health risks to individuals exposed to an acute dose, but a realistic scenario would include exposure to both high and low dose-rates, from both external and internal radioactivity. It is important therefore, to understand the biological response to prolonged exposure; and further, discover biomarkers that can be used to estimate damage from low-dose rate exposures and propose appropriate clinical treatment. Methods: We irradiated human whole blood ex vivo to three doses, 0.56 Gy, 2.23 Gy and 4.45 Gy, using two dose rates: acute, 1.03 Gy/min and a low dose-rate, 3.1 mGy/min. After 24 h, we isolated RNA from blood cells and these were hybridized to Agilent Whole Human genome microarrays. We validated the microarray results using qRT-PCR. Results: Microarray results showed that there were 454 significantly differentially expressed genes after prolonged exposure to all doses. After acute exposure, 598 genes were differentially expressed in response to all doses. Gene ontology terms enriched in both sets of genes were related to immune processes and B-cell mediated immunity. Genes responding to acute exposure were also enriched in functions related to natural killer cell activation and cell-to-cell signaling. As expected, the p53 pathway was found to be significantly enriched at all doses and by both dose-rates of radiation. A support vectors machine classifier was able to distinguish between dose-rates with 100 % accuracy using leave-one-out cross-validation. Conclusions: In this study we found that low dose-rate exposure can result in distinctive gene expression patterns compared with acute exposures. We were able to successfully distinguish low dose-rate exposed samples from acute dose exposed samples at 24 h, using a gene expression-based classifier. These genes are candidates for further testing as markers to classify exposure based on dose-rate.
Low dose-rate effects; Radiation biodosimetry; Prediction
To optimize biodosimetry methods for estimating
radiation exposure after a large-scale radiological event, all
likely radiation qualities, modes of exposure and
exposure times should be considered while designing assays
that will be useful for triage . It is important to
determine whether an individual received a dose by a lower
dose-rate, which can be from both internal and external
sources of radiation, and which may pose a moderate
health risk as compared with a single acute dose. Low
dose-rates of exposure may also confound the estimation
of total dose if dosimetry assays are not tailored to
distinguish dose-rate effects .
There have been many studies addressing the
development of a gene expression-based signature for estimation
of dose, in peripheral blood irradiated ex vivo [3–6], in
blood from total body irradiated (TBI) patients [7–9],
isolated human monocytes , CD4+ lymphocytes ,
skin from biopsies [12, 13], and cell lines from humans
[14–16]; and a few that address effects of similar doses
delivered over a period of hours or days in cell lines [15, 16],
but little is known about the gene expression response of
human blood to low dose-rates (LDR). Development of
a gene signature in blood that is able to discriminate
between irradiated samples without a matching
preexposure sample has been shown to be a powerful tool
in biodosimetry assay development , and the goal of
this study was to use a similar approach and identify
genes that would discriminate between both dose and
dose-rates. There are in vivo studies on transcriptomic
changes in radiation workers; and also changes induced by
internal emitters in mice, that have determined dose and
dose-rate effects in organs and blood [17–23]. These
studies have revealed that gene expression differences can be
detected after prolonged exposure times.
In the study presented in this paper, exposure of
human blood ex vivo to LDR and acute irradiation gave a
robust gene expression response as measured by
microarrays and validated by qRT-PCR. We identified genes
that responded uniquely to LDR and not to acute doses.
Class prediction by dose-rate successfully identified
samples as LDR-exposed or acute. This is an important first
step towards developing and further refining
geneexpression based assays that can be used to determine
the contribution of dose-rate to overall dose.
Irradiation and culture of blood
We collected blood from healthy volunteers (5 females
and 3 males) between the ages of 26 and 59 years, with
informed consent in compliance with the Columbia
University Institutional Review Board (protocol approval
number IRB-AAAF2671). 27 mL of blood from each
donor was collected into Sodium Citrate tubes (Becton
Dickinson, New Jersey, catalog# 366415) and mixed well.
Blood was diluted in equal volumes of RPMI solution
(supplemented with 10 % heat-inactivated fetal bovine
serum and 1 % penicillin streptomycin) prior to irradiations
in 50 mL Tube Spin® Bioreactor 50 tubes (TPP,
Switzerland), which are optimized for culture
incubation and gas exchange.
All irradiations were performed in an X-Rad 320
Biological Irradiator (Precision X-Ray, North Branford CT).
This device provides a system for precise delivery of
radiation doses to specimens in a self-contained, shielded
cabinet, and features an adjustable shelf, exchangeable
beam hardening filters, and a programmable control panel
that allows tube current ranging from 0.1 to 12.5 mA at its
maximum. To achieve the lowest possible dose-rate using
this device, we designed and built a custom Thoraeus filter
(1.25 mm Sn, 0.25 mm Cu, 1.5 mm Al). This filter provides
a dose rate of ~4 Gy/day at the maximum SSD
(source-tosurface distance), and a dose rate of ~1Gy/min, an
acceptable “acute” dose rate, at 40 cm SSD. This custom beam
filter was designed to enable both acute and low dose rate
irradiations to be performed using the same quality of
x rays, while changing only the mA and SSD.
The protracted irradiations of blood samples also
required the maintenance of a tissue culture environment,
with control of temperature, humidity, and carbon dioxide
content. Commercially available incubators were deemed
unsuitable, due to the large amounts of metal in their
construction. We did not want any metal in the x-ray beam
because the increased scatter would affect dose
homogeneity. We therefore, created an all plastic incubator (Fig. 1)
capable of incubating blood in 50 ml conical tubes, angled
to maintain a higher surface area to volume ratio for
efficient gas exchange, and to keep the blood within a 20 cm
diameter target area in order to minimize planar dose
variation. The samples were rotated at a speed of three
rotations per hour to further minimize any dose
inhomogeneity. Temperature was controlled through solid state
heaters on a feedback loop attached to carbon fiber walls
to distribute the heat evenly. This setting maintained a
temperature of 37 °C (±0.5 °C) for 24 h. The CO2
concentration and humidity were maintained by perfusing the
incubator at a rate of 1.7 l/min with a gas mix (5 % CO2
and 95 % air) that was humidified using a bubble
humidifier and monitored directly by numeric readouts from a
GMP70 Hand-held CO2 meter (Vaisala, Finland). The
temperature was also monitored directly by the readout of
the solid-state heater controller. Temperature and relative
humidity were also recorded using a data-logger
(EL-USB2-LCD, Lascar Electronics, Inc., Erie, PA). For irradiations,
the incubator was placed in the X-ray machine and the
incubation parameters were allowed to stabilize before
blood samples (in 50 ml tubes) were placed in the sample
Acute exposures were performed at a dose rate of
1.03 Gy/min x rays at a machine setting of 320 keV/
12.5 mA. The samples were exposed to doses of 0.56 Gy,
2.23 Gy and 4.45 Gy and then returned to the cell
culture incubator at 37 °C, 5 % CO2 for the rest of the 24 h
incubation. For low dose-rate exposures, blood samples
Fig. 1 Custom incubator for low-dose ex vivo blood irradiations,
consisting of (a) Incubator chamber containing the rotating sample
platform; (b) Heating elements attached to the chamber walls; (c).
Temperature controller; (d) Turntable motor; (e) CO2 and humidity
monitor; and (f) Temperature and humidity logger. The sample platform
completes three rotations per hour to provide dose homogeneity. The
lid that seals this chamber during use is removed in this image
were exposed to 0.56 Gy (3 h at 3.1 mGy/min, followed
by 21 h in a standard incubator), 2.23 Gy (12 h at
3.1 mGy/min, followed by 12 h in a standard incubator),
and 4.45 Gy (24 h at 3.1 mGy/min). RNA was isolated
from all samples at 24 h after the start of exposure. All
irradiated samples were compared with a matching
sham-irradiated control sample from the same donor.
RNA was isolated 24 h after the start of exposure
following the recommended protocol for the PerfectPure RNA
kit from 5Prime (Gaithersburg, MD). Globin transcripts
were reduced using the Ambion GLOBINclear-Human kit
(Life Technologies, Grand Island, NY, catalog# AM1980).
RNA yields were quantified using the NanoDrop ND1000
Spectrophotometer (Thermo Scientific) and RNA quality
was checked using the 2100 Bioanalyzer (Agilent
Technologies). RNA used for microarray hybridization had an
RNA Integrity Number of >8.5.
Cyanine-3 (Cy3) labeled cRNA was prepared with the
One-Color Low Input Quick Amp Labeling Kit (Agilent
Technologies) according to the manufacturer’s
instructions. Dye incorporation and cRNA yield were verified
with the NanoDrop ND1000 Spectrophotometer; 1.6
microgram of cRNA, >9 pmol Cy3 per microgram cRNA
was fragmented and hybridized (17 h with rotation at
65 °C) to Agilent Whole Human Genome Microarrays
4X44K v2 (G4112F), and then washed using the Gene
Expression Hybridization Kit and GE Wash Buffers as
recommended by Agilent. Slides were then scanned with
the Agilent DNA Microarray Scanner (G2505B), and the
images were analyzed (Agilent Feature Extraction Software
ver. 10.7) with default parameters for background
correction and flagging non-uniform features.
Background-corrected hybridization intensities were
imported into BRB-ArrayTools, Version 4.2.1 
log2transformed and median normalized; after normalization
the distribution of signals remained uniform across all
arrays. Non-uniform outliers or features not significantly
above background intensity in 25 % or more of the
hybridizations were filtered out. A further filter requiring a
minimum 1.4 fold change in at least 20 % of the
hybridizations was then applied to remove genes with no
variation across the dataset, with >10,000 genes remaining
which is optimal for data analyses; probes were also
averaged to one probe per gene and duplicate genes were
reduced by selecting the one with maximum signal
intensity, yielding a final set of 12,073 features that were
used for subsequent analyses. The microarray data is
available through the Gene Expression Omnibus with
accession number GSE65292.
RNA samples from five donors (3 male and 2 female
to mitigate against the possibility of sex-specific bias in
results [25, 26] were hybridized. A total of 35 RNA
samples were hybridized in this study. Class comparison
analyses were conducted using BRB-ArrayTools to
identify genes that were differentially expressed between
classes using a random-variance t-test . Genes with
p-values less than 0.001 were considered statistically
significant. The false discovery rate (FDR) was also
estimated for each gene using the method of Benjamini and
Hochberg , to control for false positives.
The High-Capacity cDNA Archive Kit (Life
Technologies, Foster City, CA) was used to prepare cDNA from
total RNA from three of the female donors not used in
the microarray hybridization experiments. Quantitative
real-time RT-PCR (qRT-PCR) was performed for
selected genes using Taqman assays (Life Technologies) on
a Low Density array (384-well microfluidic card), to
confirm microarray experiment findings for selected genes.
The 48 genes and corresponding assays are listed in
Additional file 1. In gene expression validation studies,
400 ng cDNA was used as input for PCR. Quantitative
real time PCR reactions were performed with the ABI
7900 Real Time PCR System using Universal PCR Master
Mix (Life Technologies), with initial activation at 50 °C for
120 s and 94.5 °C for 10 min, followed by 40 cycles of
97 °C for 30 s and 59.7 °C for 60 s. Relative fold-change
was calculated by the ΔΔCT method, using SDS version
2.3 (Thermofisher). Data were normalized to RPLPO
gene expression levels. We used Genorm  to assess
the stability of the housekeeping genes included on the
Low Density array panels, and RPLPO gene expression
was found to be most stable in our data. RPLPO was
therefore used to normalize the qRT-PCR data.
Gene Ontology and pathway analyses
Lists of genes significantly over- or under-expressed
relative to controls were imported separately into the
PANTHER database (version 9.0, release 2014-01-24) to
identify enriched biological themes and gene ontology
(GO) terms using the statistical overrepresentation test,
GO-Slim Biological Processes, Molecular Functions and
Pathways . Benjamini corrected p- values <0.05 were
Class prediction analysis
Gene sets for class prediction were determined using
BRB-ArrayTools Class Predictions, which provide
various options for classifier prediction and cross-validation.
Predictions used a cut off significance p-value of 0.0001
(for dose-rate and irradiation classification) and 0.001
(for dose classification) between classes to determine the
classifier gene set. Support Vector Machines  was
used for classification of samples between two
categories, and Diagonal Linear Discriminant Analysis, which
avoids complex models with excessive parameters in
order to avoid over fitting data without loss of
performance  was used for classifications with more than
two categories. The algorithms tested the classifier gene
set for accuracy and sensitivity and specificity  and
we used the Leave-one-out cross-validation method to
compute mis-classification rates.
We analyzed gene expression changes using the
BRBArrayTools Class comparison tool for paired analyses
between classes and the results of significantly
differentially expressed genes (p <0.001) are summarized in
Table 1. We found the broadest changes and highest
number of genes affected at 4.45 Gy by both dose-rates
(354 genes changed after LDR 4.45 Gy exposure and 565
genes changed after acute 4.45 Gy exposures at 24 h);
Additional file 2 contains details of gene expression
changes summarized in Table 1. An intersection of the
differentially expressed genes at each dose (Additional
file 3) indicated that the set of differentially expressed
genes after 4.45 Gy included most of the genes changed
at lower doses.
We performed gene ontology analyses on the sets of
differentially expressed genes. Among the sets of genes
responding to 0.56 or 2.23 Gy at low dose rate,
PANTHER GO-slim analysis reveal only one significantly
affected function after LDR 0.56 Gy exposure, DNA repair
(with a p-value of 3.7 X10-2). Genes in this category were
XPC, DDB2, POLH, GADD45A and PCNA, all of which
are sensitive radiation response genes.
The genes responding to the 4.45 Gy dose at both dose
rates showed additional significantly enriched biological
processes. A comparison of significantly enriched
biological terms and the genes belonging to each category
are shown in Table 2. Cellular processes, immune
processes, B cell mediated immunity and cell
communication were common biological functions affected by both
low-dose rate and acute exposure. Unique to the low
dose-rate response genes, was the pyrimidine nucleobase
metabolic process (p-value 2.7 × 10-2) with genes involved
in DNA editing. After acute 4.45 Gy exposures, biological
processes affected included natural killer cell activation
(p-value 6.7 × 10-5) and other cell signaling processes
(p-value 1.3 × 10-2), not observed after LDR. GO-slim
molecular functions revealed that receptor activity and
binding were significantly affected by both LDR 4.45 Gy
(p-value 2.2 × 10-3) and acute 4.45 Gy (p-value 1.8 × 10-11)
doses, with an additional molecular function category of
chemokines (p-value 1.3 × 10-2) significant only after acute
exposure, not observed after LDR.
Comparison of low-dose rate and acute exposure
We also directly compared the gene expression response
to 24-hour continuous LDR exposure with that of the
acute exposure at 4.45 Gy (Table 1). There were 243 genes
differentially expressed when comparing the two different
exposure rates, with a moderate range of fold change
between 0.3 and 3.9. Gene ontology analysis of these 243
genes revealed enrichment of two processes: glycolysis
(p-value 3.74 × 10-4) and monosaccharide metabolic
process (p-value 2.5 × 10-3). Genes included in these two
categories were members of the glycolysis pathway (lactate
dehyrogenase A, LDHA; glyceraldehyde 3-phosphate
dehydrogenase, GAPDH; 6-phospho-fructokinase type C,
PKFP; enolase 1, ENO1; and hexokinase 2, HK2), all of
which were expressed at lower levels in cells exposed to
the protracted dose.
Validation of gene expression by quantitative PCR
We validated gene expression changes from microarrays
using real time qRT-PCR in independent biological
replicates. These samples were true independent biological
replicates representing different donors from those used
in the microarray hybridizations. We chose genes that
were in common between all doses and dose-rates and
are also known radiation response genes . Fold changes
by qRT-PCR agreed well with our microarray
measurements and data are shown for both 4.45 Gy exposed
groups (Fig. 2). We also validated gene expression levels
for the lower doses (2.23 Gy and 0.56 Gy) and these data
along with the 4.45 Gy results, mean and SEM, are
included in Additional file 4.
Table 1 Summary of genes differentially changed (p <0.001) in various class comparisons
Number of genes
False discovery rate
Number of up
Number of down
LDR 4 Gy vs 0 Gy
Acute 4 Gy vs 0 Gy
LDR 2 Gy vs 0 Gy
Acute 2 Gy vs 0 Gy
LDR 0.5 Gy vs 0 Gy
Acute 0.5 Gy vs 0 Gy
LDR 4 Gy vs Acute 4 Gy
Table 2 Biological process enrichment analysis using PANTHER
PANTHER GO-Slim Biological Process LDR 4 Gy
Cellular process 110
Immune system process
Pyrimidine nucleobase metabolic process
B cell mediated immunity
Response to stimulus
Natural killer cell activation
aNS not significant
Gene expression patterns
We searched for gene expression patterns that were in
common between LDR and acute dose-rates and also
those that showed differences that could distinguish
samples that received a dose by a lower dose-rate. We
identified more than 20 genes that showed a similar
pattern of response where the dose rate appeared not to
affect the changes. There were both up regulated and
down regulated genes that belonged to this group and
genes showing this characteristic behavior, AEN and
CDKN1A (up in LDR and acute) and MYC and E2F5
(down in both LDR and acute), are shown in Fig. 3a and
b, respectively. The other pattern of interest was genes
that only appeared to respond to LDR and not to acute
Fig. 2 Validation of microarray results by qRT-PCR. Shown here are log2 (fold changes) of genes that were determined to be differentially regulated by
the 4.45 Gy dose by both dose-rates. The graph on the left shows the mean log2 (fold change) after LDR 4.45 Gy; and the graph on the right is the
mean of log2 (fold change) after Acute 4.45 Gy exposure. All microarray (blue bars, five biological replicates) and qRT-PCR (red-bars, three biological
replicates) results are average fold-change from paired analyses; SEM values for all data are included in Additional file 4
Fig. 3 Patterns of gene expression response, shown as microarray results of representative genes, in all panels (a-d), open symbols with solid
lines represent Low dose-rate (ldr) responses and closed symbols with dotted lines represent acute dose rate (acute) gene expression responses.
a AEN (circles) and CDKN1A (squares) are representative genes in the group that showed similar up regulation of mRNA levels by both dose-rates.
b MYC (circles) and E2F5 (squares) are representative genes in the group that showed similar down regulation of mRNA levels by both dose-rates.
c RBM3 (circles) and GRM2 (squares) are representative genes for the group that showed up regulation only by LDR. d DUSP3 (circles) and ID1
(squares) are representative genes for the group that showed down regulation only by LDR. All points are mean of 5 biological replicates (from
paired analyses) and error bars are SEM
dose. There were two types of genes in this group, one
in which the genes were up regulated by LDR only, not
acute doses (RBM3 and GRM2, Fig. 3c) and the other in
which genes were down regulated by LDR only, not
acute doses (DUSP3 and ID2, Fig. 3d). This preliminary
assessment of different gene expression response
patterns by LDR suggested that there are genes that could
distinguish between the dose-rate for the same dose
delivered. For some genes, such as APOBEC3H, FDXR and
PHLDA3, induced at all doses by LDR and acute, the
change in gene expression after the 4.45 Gy dose was
higher in the acute dose group suggesting protection of
response by low dose-rate.
We used the Class Prediction tool in BRB-ArrayTools to
determine a classifier that would correctly discriminate
between dose-rates. Initially, we identified a classifier
that distinguished between un-irradiated and acute or
LDR-irradiated samples; in which a set of 121 genes
(p-value <0.0001) correctly identified samples as 0 Gy
control (CTL), Acute, or LDR with 94 % accuracy
(Table 3) using Diagonal Linear Discriminant Analysis
(DLDA). Next we tested and found that a 62-gene
classifier could correctly classify exposed samples as Acute or
LDR with 100 % accuracy using Support Vector
Machines (Table 4). Lastly, we determined a classifier gene
set comprised of 140 genes that correctly classified
samples by dose without regard to the rate of exposure, with
90 % accuracy using Diagonal Linear Discriminant
Analysis (Table 5). Details of classifier gene sets and
performance of classifiers in cross-validations are included
in Additional file 5; all classification analyses were
performed on normalized gene expression signal intensities.
A variety of exposure types and combinations of
exposures could result from an improvised nuclear device
Table 3 Performance of the Diagonal Linear Discriminant
Analysis classifier (121 genes) on irradiation and dose rate
(IND) or radioactive dispersal device (RDD). In order to
develop appropriate radiological triage approaches,
biodosimetry assays must be tested and optimized for their
ability to detect the contribution of various factors such
as dose and dose-rate. A tiered approach to triage in
large-scale scenarios [33, 34] would ideally include a
low-dose rate detection assay to identify individuals who
have received exposure over a prolonged time, so that
their treatment can be adjusted accordingly. In a recent
NATO study involving different laboratories that
crossvalidated results for different radiation biodosimetry
assays, the dicentric chromosome assay, micronucleus
assay, γ-H2AX and gene expression were assessed for
their sensitivity and it was concluded that a combination
of assays would be optimal for the estimation of dose
and “never versus ever” exposure [35, 36]. Therefore, it
is possible that a gene expression signature that can
discriminate between low dose-rates and acute exposures,
in combination with other assays that estimate dose, will
enhance the ability to identify individuals with an
immediate need for clinical treatment in a large-scale event.
Low dose-rate studies have also been done for very
low cumulative doses, to assess the gene expression
response. One such study on a prostate cancer cell line
measured gene expression changes after a 24-hour
chronic exposure to dose-rates as low as 7-17 μGy/min,
and unexpectedly, found that the gene expression
response was more similar to that of a 2 Gy acute dose
than a 10 cGy acute dose . In another study in mice,
which were given a 5-week continuous dose of radiation
at 2 μGy/min (cumulative dose 10.5 cGy, which was
previously shown to be effective on gene expression as a single
acute dose ) there was no significant change in gene
expression. The effects observed at these very low doses may
be the result of various factors inherent to the study
design, the extended time in the second study or the origin
and type of cells in the first, but it emphasizes the need
for biodosimetry experiments to be designed to establish
Table 5 Performance of the Diagonal Linear Discriminant
Analysis classifier (140 genes) on dose
the range of dose rates and exposure times likely to
impact on biodosimetric estimates and triage decisions.
Using an ex vivo blood irradiation model that has
previously shown changes in radiation response genes that
are sensitive to dose and time [3, 37], we exposed blood
samples to relatively high total doses (up to 4 Gy), in
the range of Acute Radiation Syndrome (ARS), either
acutely, or over a period of 24 h. The prolonged
radiation exposure time was to approximate exposure to
fallout from an improvised nuclear device (IND) or
radioactive dispersal device (RDD) that might occur
before first responders arrive on the scene and are able to
begin taking samples for biodosimetry.
In the current study, we detected gene expression
changes at all doses, with increasing numbers of genes
responding with increasing dose, as expected (Table 1).
The number of genes differentially expressed at lower
doses (0.56 Gy and 2.25 Gy) by LDR (p-value <0.001)
was slightly higher than those after acute doses, however,
the period of time from end of exposure to harvest was
shorter after low dose-rate exposures, which may
contribute to this difference. In the case of LDR 4.45 Gy
however, the number of genes differentially expressed
was less than acute 4.45 Gy (Additional file 2).
We focused our analyses on the 4.45 Gy dose
responses, because they allowed us to determine and
compare maximum differences in gene expression and
biological function. These comparisons revealed that the
acute exposure elicits many responses similar to LDR,
but may also affect additional processes related to
natural killer cells and cell-cell signaling via chemokines. In
contrast, the gene expression response to LDR initiated
some of the same functions as acute exposure, but
additional processes related to nucleotide metabolism
(Table 2) and DNA repair were also affected.
Further GO analyses of the LDR gene expression
response against PANTHER pathways revealed that the
p53 pathway was found to be significantly affected at all
doses (LDR 0.56Gy, p-value 5.0 × 10-4; LDR 2.23 Gy,
pvalue 4.8 × 10-5; and LDR 4.45 Gy, 1.5 × 10-5). This
suggests that at 24 h after the start of exposure at all doses,
p53-regulated functions in cell stress, cell cycle and
DNA damage repair are important; even after the
shortest LDR exposure. We focused on p53 target genes as
this pathway was the top scoring biological pathway
controlling gene regulation across all doses and dose-rates.
The heat map in Fig. 4 depicts genes that are regulated
by p53 across all three doses in the LDR treatment
group. More p53-regulated genes were involved in the
response after the highest dose (4.45 Gy) where blood
cells were exposed continuously to the LDR radiation
with no recovery period. In the case of the lowest dose
(0.56 Gy) and intermediate dose (2.23 Gy) a subset of
these genes responded and changes in mRNA were still
Fig. 4 p53 pathway gene expression in LDR exposed blood cells. The genes contributing to the significance of the PANTHER p53-pathway are
listed and their fold change depicted in the heat map, the three columns represent the mRNA changes at 24 h after 0.56 Gy, 2.23 Gy and 4.45 Gy
dose; irradiation at low dose rate. Missing values that were not significant at a particular dose are shown in grey; the scale shows shades of yellow
to red (upregulated genes) and shades of green (down-regulated genes)
detectable at 24 h after the start of exposure. This
suggests that there is an accumulation of stress and damage
during the protraction of dose, possibly mediated by p53
and also other transcriptional regulators [38, 39], that
persists and is not completely resolved at 24 h after the
start of exposure.
We were also able to identify groups of genes that
were more responsive to LDR than acute dose rate
(Fig. 3c and d). Genes such as RBM3 and GRM2 that
were up regulated after LDR appeared to respond to
doses ≥2.23 Gy. In the case of down regulated genes,
DUSP3 and ID1, the response was significant (p-value
<0.001) at even the lowest dose, LDR 0.56 Gy. The down
regulation of ID1 mRNA after LDR was interesting
because it is a known radiation response gene to acute
γ-irradiation in cell lines . In our study it was not
induced significantly above background by acute dose
but persistently down-regulated by LDR. Other genes
shown here, RBM3 (RNA binding motif protein 3) and
GRM2 (Glutamate receptor, metabotropic 2) and DUSP3
(Dual specificity phosphatase 3) have not been
previously shown to be affected by radiation in blood and
may represent candidate genes for further studies on
dose-rate effects in radiation biodosimetry.
A comparison of gene expression changes by dose rate
between LDR and acute only, revealed the lowered
expression of genes encoding glycolytic enzymes, by the
protraction of exposure at 24 h. These genes are
involved in glucose metabolism and energy production in
cells and the continuous delivery of radiation at LDR
may have a dampening effect on the activity of these
metabolic functions. In another study using the same
LDR, 3.1 mGy/min for 24 h in mice, measurement of
metabolites in urine at 48 h after the beginning of
exposure showed that citrate in the TCA cycle was
decreased by both LDR and acute exposure . In the
same study, hexanoylcarnitine and tiglylcarnitine from
fatty-acid oxidation pathways were decreased by LDR
exposure compared to controls or acute exposure to the
same dose, consistent with the perturbations we found
reflected in the gene expression data.
The changes observed in blood gene expression after
24 h in the current study support the development of
dosimetry signatures to distinguish between dose-rates of
exposure, as well as between doses. We performed class
predictions by irradiation status and dose-rate (Tables 3
and 4) and dose (Table 5). The classifier gene set that
performed best and distinguished between LDR and
acute exposure included the LDR-only response genes
DUSP3 and ID1 (Fig. 3d). This suggests that it may be
possible to develop a gene-based signature that can detect
protracted exposures without the need for a pre-exposure
sample. Further independent validation studies will of
course be needed, but such a test could also be used in
conjunction with other strictly dosimetric assays, including
gene expression or cytogenetic endpoints, to provide a
better and more practical biodosimetry assay.
This study investigated the effects of dose-rate on human
blood cell gene expression, over a 24-hour period. Although
there were many similarities in immune function and stress
response genes, we found that low dose-rate exposure can
result in distinctive gene expression patterns compared
with acute exposures. Typical p53 gene responses were also
seen at all doses delivered by the lower dose rate. We were
able to successfully distinguish low dose-rate exposed
samples from acute dose exposures, using classification
algorithms on our gene expression data. These genes are
candidates for further validation studies to develop a
genebased signature that can detect low dose-rate exposures for
Availability of supporting data
The data set supporting the results of this article is
available in the NCBI GEO repository
Below is the link to the electronic supplementary material.
The authors declare that they have no competing interests.
SAG was involved in sample collection and assisted with irradiations;
performed all microarray and qPCR experiments and analyses and drafted
the manuscript. LS was involved in design of the incubator and performed
all irradiation treatments, CDE was involved in design of the low dose-rate
incubator, performed the dosimetry, and helped prepare the manuscript. MC
helped with data analysis. SAA designed the study and the low dose-rate
incubator and was involved with preparation of the manuscript. All authors
read and approved the final manuscript.
We thank the Center for Radiological Research members Gary Johnson and
Dr. Gerhard Randers-Pehrson for designing and constructing the custom-built
incubator and Thoraeus filter for low dose-rate exposures. We thank Dr. Helen
Turner and Dr. Antonella Bertucci for sample collection. We also thank Dr.
Congju Chen for assisting with irradiations. Analyses were performed using
BRB-ArrayTools developed by Dr. Richard Simon and BRB-ArrayTools Development
Team. This work was supported by the Center for High-Throughput
Minimally-Invasive Radiation Biodosimetry, National Institute of Allergy and
Infectious Diseases grant number U19AI067773.
1. Ainsbury EA , Bakhanova E , Barquinero JF , Brai M , Chumak V , Correcher V , et al. Review of retrospective dosimetry techniques for external ionising radiation exposures . Radiat Prot Dosimetry . 2011 ; 147 ( 4 ): 573 - 92 .
2. Hall EJ , Brenner DJ . The dose-rate effect revisited: radiobiological considerations of importance in radiotherapy . Int J Radiat Oncol Biol Phys . 1991 ; 21 ( 6 ): 1403 - 14 .
3. Paul S , Amundson SA. Development of gene expression signatures for practical radiation biodosimetry . Int J Radiat Oncol Biol Phys . 2008 ; 71 ( 4 ): 1236 - 44 .
4. Boldt S , Knops K , Kriehuber R , Wolkenhauer O. A frequency-based gene selection method to identify robust biomarkers for radiation dose prediction . Int J Radiat Biol . 2012 ; 88 ( 3 ): 267 - 76 .
5. Dressman HK , Muramoto GG , Chao NJ , Meadows S , Marshall D , Ginsburg GS , et al. Gene expression signatures that predict radiation exposure in mice and humans . PLoS Med . 2007 ; 4 ( 4 ): e106 .
6. Amundson SA , Do KT , Shahab S , Bittner M , Meltzer P , Trent J , et al. Identification of potential mRNA biomarkers in peripheral blood lymphocytes for human exposure to ionizing radiation . Radiat Res . 2000 ; 154 ( 3 ): 342 - 6 .
7. Paul S , Barker CA , Turner HC , McLane A , Wolden SL , Amundson SA . Prediction of in vivo radiation dose status in radiotherapy patients using ex vivo and in vivo gene expression signatures . Radiat Res . 2011 ; 175 ( 3 ): 257 - 65 .
8. Amundson SA , Grace MB , McLeland CB , Epperly MW , Yeager A , Zhan Q , et al. Human in vivo radiation-induced biomarkers: gene expression changes in radiotherapy patients . Cancer Res . 2004 ; 64 ( 18 ): 6368 - 71 .
9. Templin T , Paul S , Amundson SA , Young EF , Barker CA , Wolden SL , et al. Radiation-induced micro-RNA expression changes in peripheral blood cells of radiotherapy patients . Int J Radiat Oncol Biol Phys . 2011 ; 80 ( 2 ): 549 - 57 .
10. El-Saghire H , Michaux A , Thierens H , Baatout S. Low doses of ionizing radiation induce immune-stimulatory responses in isolated human primary monocytes . Int J Mol Med . 2013 ; 32 ( 6 ): 1407 - 14 .
11. Nosel I , Vaurijoux A , Barquinero JF , Gruel G. Characterization of gene expression profiles at low and very low doses of ionizing radiation . DNA Repair . 2013 ; 12 ( 7 ): 508 - 17 .
12. Goldberg Z , Schwietert CW , Lehnert B , Stern R , Nami I. Effects of low-dose ionizing radiation on gene expression in human skin biopsies . Int J Radiat Oncol*Biol*Phys . 2004 ; 58 ( 2 ): 567 - 74 .
13. Goldberg Z , Rocke DM , Schwietert C , Berglund SR , Santana A , Jones A , et al. Human in vivo dose-response to controlled, low-dose low linear energy transfer ionizing radiation exposure . Clin Cancer Res . 2006 ; 12 ( 12 ): 3723 - 9 .
14. Khodarev NN , Park JO , Yu J , Gupta N , Nodzenski E , Roizman B , et al. Dosedependent and independent temporal patterns of gene responses to ionizing radiation in normal and tumor cells and tumor xenografts . Proc Natl Acad Sci U S A . 2001 ; 98 ( 22 ): 12665 - 70 .
15. McDonald JT , Briggs C , Szelag H , Peluso M , Schneider D , Perepletchikov A , et al. Chronic low dose-rate radiation down-regulates transcription related to mitosis and chromosomal movement similar to acute high dose in prostate cells . Int J Radiat Biol . 2014 ; 90 ( 3 ): 231 - 40 .
16. Amundson SA , Lee RA , Koch-Paiz CA , Bittner ML , Meltzer P , Trent JM , et al. Differential responses of stress genes to low dose-rate gamma irradiation . Mol Cancer Res . 2003 ; 1 ( 6 ): 445 - 52 .
17. Paul S , Ghandhi SA , Weber W , Doyle-Eisele M , Melo D , Guilmette R , et al. Gene expression response of mice after a single dose of 137CS as an internal emitter . Radiat Res . 2014 ; 182 ( 4 ): 380 - 9 .
18. Schuler E , Parris TZ , Rudqvist N , Helou K , Forssell-Aronsson E. Effects of internal low-dose irradiation from 131I on gene expression in normal tissues in Balb/c mice . EJNMMI Res . 2011 ; 1 ( 1 ): 29 .
19. Templin T , Amundson SA , Brenner DJ , Smilenov LB . Whole mouse blood microRNA as biomarkers for exposure to gamma-rays and (56)Fe ion . Int J Radiat Biol . 2011 ; 87 ( 7 ): 653 - 62 .
20. Schuler E , Rudqvist N , Parris TZ , Langen B , Helou K , Forssell-Aronsson E. Transcriptional response of kidney tissue after 177Lu-octreotate administration in mice . Nucl Med Biol . 2014 ; 41 ( 3 ): 238 - 47 .
21. Bong JJ , Kang YM , Shin SC , Choi SJ , Lee KM , Kim HS . Differential expression of thymic DNA repair genes in low-dose-rate irradiated AKR/J mice . J Vet Sci . 2013 ; 14 ( 3 ): 271 .
22. Gruel G , Lucchesi C , Pawlik A , Frouin V , Alibert O , Kortulewski T , et al. Novel microarray-based method for estimating exposure to ionizing radiation . Radiat Res . 2006 ; 166 ( 5 ): 746 - 56 .
23. Fachin AL , Mello SS , Sandrin-Garcia P , Junta CM , Ghilardi-Netto T , Donadi EA , et al. Gene expression profiles in radiation workers occupationally exposed to ionizing radiation . J Radiat Res . 2009 ; 50 ( 1 ): 61 - 71 .
24. Simon R. Analysis of DNA microarray expression data . Best Pract Res Clin Haematol . 2009 ; 22 ( 2 ): 271 - 82 .
25. Paul S , Amundson SA. Gene expression signatures of radiation exposure in peripheral white blood cells of smokers and non-smokers . Int J Radiat Biol . 2011 ; 87 ( 8 ): 791 - 801 .
26. Meadows SK , Dressman HK , Muramoto GG , Himburg H , Salter A , Wei Z , et al. Gene expression signatures of radiation response are specific, durable and accurate in mice and humans . PLoS One . 2008 ; 3 ( 4 ): e1912 .
27. Wright GW , Simon RM. A random variance model for detection of differential gene expression in small microarray experiments . Bioinformatics . 2003 ; 19 ( 18 ): 2448 - 55 .
28. Hochberg Y , Benjamini Y. More powerful procedures for multiple significance testing . Stat Med . 1990 ; 9 ( 7 ): 811 - 8 .
29. Vandesompele J , De Preter K , Pattyn F , Poppe B , Van Roy N , De Paepe A , et al. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes . Genome Biol . 2002 ; 3 ( 7 ): RESEARCH0034 .
30. Mi H , Muruganujan A , Thomas PD . PANTHER in 2013: modeling the evolution of gene function, and other gene attributes, in the context of phylogenetic trees . Nucleic Acids Res . 2013 ; 41 (Database issue): D377 - 86 .
31. Brown MP , Grundy WN , Lin D , Cristianini N , Sugnet CW , Furey TS , et al. Knowledge-based analysis of microarray gene expression data by using support vector machines . Proc Natl Acad Sci U S A . 2000 ; 97 ( 1 ): 262 - 7 .
32. Dudoit S , Fridlyand J , Speed TP . Comparison of discrimination methods for the classification of tumors using gene expression data . J Am Stat Assoc . 2002 ; 97 ( 457 ): 77 - 87 .
33. Flood AB , Boyle HK , Du G , Demidenko E , Nicolalde RJ , Williams BB , et al. Advances in a framework to compare bio-dosimetry methods for triage in large-scale radiation events . Radiat Prot Dosimetry . 2014 ; 159 ( 1-4 ): 77 - 86 .
34. Waselenko JK , MacVittie TJ , Blakely WF , Pesik N , Wiley AL , Dickerson WE , et al. Medical management of the acute radiation syndrome: recommendations of the Strategic National Stockpile Radiation Working Group . Ann Intern Med . 2004 ; 140 ( 12 ): 1037 - 51 .
35. Rothkamm K , Beinke C , Romm H , Badie C , Balagurunathan Y , Barnard S , et al. Comparison of established and emerging biodosimetry assays. Radiat Res . 2013 ; 180 ( 2 ): 111 - 9 .
36. Badie C , Kabacik S , Balagurunathan Y , Bernard N , Brengues M , Faggioni G , et al. Laboratory intercomparison of gene expression assays . Radiat Res . 2013 ; 180 ( 2 ): 138 - 48 .
37. Paul S , Smilenov LB , Amundson SA . Widespread decreased expression of immune function genes in human peripheral blood following radiation exposure . Radiat Res . 2013 ; 180 ( 6 ): 575 - 83 .
38. Daino K , Ichimura S , Nenoi M. Early induction of CDKN1A (p21) and GADD45 mRNA by a low dose of ionizing radiation is due to their dose-dependent post-transcriptional regulation . Radiat Res . 2002 ; 157 ( 4 ): 478 - 82 .
39. Amundson SA , Lee RA , Koch-Paiz CA , Bittner ML , Meltzer P , Trent JM , et al. Differential responses of stress genes to low dose-rate gamma irradiation . Mol Cancer Res . 2003 ; 1 ( 6 ): 445 - 52 .
40. Komarova EA , Diatchenko L , Rokhlin OW , Hill JE , Wang ZJ , Krivokrysenko VI , et al. Stress-induced secretion of growth inhibitors: a novel tumor suppressor function of p53 . Oncogene. 1998 ; 17 ( 9 ): 1089 - 96 .
41. Goudarzi M , Mak TD , Chen C , Smilenov LB , Brenner DJ , Fornace AJ . The effect of low dose rate on metabolomic response to radiation in mice . Radiat Environ Biophys . 2014 ; 53 ( 4 ): 645 - 57 .