Validating a Four-gene Set for H-ARS Severity Prediction in Peripheral Blood Samples of Irradiated Rhesus Macaques.

Increased radiological and nuclear threats require preparedness. Our earlier work identified a set of four genes (DDB2, FDXR, POU2AF1 and WNT3), which predicts severity of the hematological acute radiation syndrome (H-ARS) within the first three days postirradiation In this study of 41 Rhesus macaques (Macaca mulatta, 27 males, 14 females) irradiated with 5.8-7.2 Gy (LD29-50/60), including some treated with gamma-tocotrienol (GT3, a radiation countermeasure) we independently validated these genes as predictors in both sexes and examined them after three days. At the Armed Forces Radiobiology Research Institute/Uniformed Services University of the Health Sciences, peripheral whole blood (1 ml) of Rhesus macaques was collected into PAXgene® Blood RNA tubes pre-irradiation after 1, 2, 3, 35 and 60 days postirradiation, stored at -80°C for internal experimental analyses. Leftover tubes from these already ongoing studies were kindly provided to Bundeswehr Institute of Radiobiology. RNA was isolated (QIAsymphony), converted into cDNA, and for further gene expression (GE) studies quantitative RT-PCR was performed. Differential gene expression (DGE) was measured relative to the pre-irradiation Rhesus macaques samples. Within the first three days postirradiation, we found similar results to human data: 1. FDXR and DDB2 were up-regulated, FDXR up to 3.5-fold, and DDB2 up to 13.5-fold in the median; 2. POU2AF1 appeared down regulated around tenfold in nearly all Rhesus macaques; 3. Contrary to human data, DDB2 was more up-regulated than FDXR, and the difference of the fold change (FC) ranged between 2.4 and 10, while the median fold changes of WNT3, except days 1 and 35, were close to 1. Nevertheless, 46% of the Rhesus macaques showed down-regulated WNT3 on day one postirradiation, which decreased to 12.2% on day 3 postirradiation. Considering the extended phase, there was a trend towards decreased fold changes at day 35, with median-fold changes ranging from 0.7 for DDB2 to 0.1 for POU2AF1, and on day 60 postirradiation, DGE in surviving animals was close to pre-exposure values for all four genes. In conclusion, the diagnostic significance for radiation-induced H-ARS severity prediction of FDXR, DDB2, and POU2AF1 was confirmed in this Rhesus macaques model. However, DDB2 showed higher GE values than FDXR. As shown in previous studies, the diagnostic significance of WNT3 could not be reproduced in Rhesus macaques; this could be due to the choice of animal model and methodological challenges.

PUM1 and PGK1 are Favorable Housekeeping Genes over Established Biodosimetry-related Housekeeping Genes such as HPRT1, ITFG1, DPM1, MRPS5, 18S rRNA and Others after Radiation Exposure.

In gene expression (GE) studies, housekeeping genes (HKGs) are required for normalization purposes. In large-scale inter-laboratory comparison studies, significant differences in dose estimates are reported and divergent HKGs are employed by the teams. Among them, the 18S rRNA HKG is known for its robustness. However, the high abundance of 18S rRNA copy numbers requires dilution, which is time-consuming and a possible source of errors. This study was conducted to identify the most promising HKGs showing the least radiation-induced GE variance after radiation exposure. In the screening stage of this study, 35 HKGs were analyzed. This included selected HKGs (ITFG1, MRPS5, and DPM1) used in large-scale biodosimetry studies which were not covered on an additionally employed pre-designed 96-well platform comprising another 32 HKGs used for different exposures. Altogether 41 samples were examined, including 27 ex vivo X-ray irradiated blood samples (0, 0.5, 4 Gy), six X-irradiated samples (0, 0.5, 5 Gy) from two cell lines (U118, A549), as well as eight non-irradiated tissue samples to encompass multiple biological entities. In the independent validation stage, the most suitable candidate genes were examined from another 257 blood samples, taking advantage of already stored material originating from three studies. These comprise 100 blood samples from ex vivo X-ray irradiated (0-4 Gy) healthy donors, 68 blood samples from 5.8 Gy irradiated (cobalt-60) Rhesus macaques (RM) (LD29/60) collected 0-60 days postirradiation, and 89 blood samples from chemotherapy-(CTx) treated breast tumor patients. CTx and radiation-induced GE changes in previous studies appeared comparable. RNA was isolated, converted into cDNA, and GE was quantified employing TaqMan assays and quantitative RT-PCR. We calculated the standard deviation (SD) and the interquartile range (IQR) as measures of GE variance using raw cycle threshold (Ct) values and ranked the HKGs accordingly. Dose, time, age, and sex-dependent GE changes were examined employing the parametrical t-test and non-parametrical Kruskal Wallis test, as well as linear regression analysis. Generally, similar ranking results evolved using either SD or IQR GE measures of variance, indicating a tight distribution of GE values. PUM1 and PGK1 showed the lowest variance among the first ten most suitable genes in the screening phase. MRPL19 revealed low variance among the first ten most suitable genes in the screening phase only for blood and cells, but certain comparisons indicated a weak association of MRPL19 with dose (P = 0.02-0.09). In the validation phase, these results could be confirmed. Here, IQR Ct values from, e.g., X-irradiated blood samples were 0.6 raw Ct values for PUM1 and PGK1, which is considered to represent GE differences as expected due to methodological variance. Overall, when compared, the GE variance of both genes was either comparable or lower compared to 18S rRNA. Compared with the IQR GE values of PUM1 and PGKI, twofold-fivefold increased values were calculated for the biodosimetry HKG HPRT1, and comparable values were calculated for biodosimetry HKGs ITFG1, MRPS5, and DPM1. Significant dose-dependent associations were found for ITFG1 and MRPS5 (P = 0.001-0.07) and widely absent or weak (P = 0.02-0.07) for HPRT1 and DPM1. In summary, PUM1 and PGK1 appeared most promising for radiation exposure studies among the 35 HKGs examined, considering GE variance and adverse associations of GE with dose.

Applicability of Gene Expression in Saliva as an Alternative to Blood for Biodosimetry and Prediction of Radiation-induced Health Effects.

As the great majority of gene expression (GE) biodosimetry studies have been performed using blood as the preferred source of tissue, searching for simple and less-invasive sampling methods is important when considering biodosimetry approaches. Knowing that whole saliva contains an ultrafiltrate of blood and white blood cells, it is expected that the findings in blood can also be found in saliva. This human in vivo study aims to examine radiation-induced GE changes in saliva for biodosimetry purposes and to predict radiation-induced disease, which is yet poorly characterized. Furthermore, we examined whether transcriptional biomarkers in blood can also be found equivalently in saliva. Saliva and blood samples were collected in parallel from radiotherapy (RT) treated patients who suffered from head and neck cancer (n = 8) undergoing fractioned partial-body irradiations (1.8 Gy/fraction and 50-70 Gy total dose). Samples were taken 12-24 h before first irradiation and ideally 24 and 48 h, as well as 5 weeks after radiotherapy onset. Due to the low quality and quantity of isolated RNA samples from one patient, they had to be excluded from further analysis, leaving a total of 24 saliva and 24 blood samples from 7 patients eligible for analysis. Using qRT-PCR, 18S rRNA and 16S rRNA (the ratio being a surrogate for the relative human RNA/bacterial burden), four housekeeping genes and nine mRNAs previously identified as radiation responsive in blood-based studies were detected. Significant GE associations with absorbed dose were found for five genes and after the 2nd radiotherapy fraction, shown by, e.g., the increase of CDKN1A (2.0 fold, P = 0.017) and FDXR (1.9 fold increased, P = 0.002). After the 25th radiotherapy fraction, however, all four genes (FDXR, DDB2, POU2AF1, WNT3) predicting ARS (acute radiation syndrome) severity, as well as further genes (including CCNG1 [median-fold change (FC) = 0.3, P = 0.013], and GADD45A (median-FC = 0.3, P = 0.031)) appeared significantly downregulated (FC = 0.3, P = 0.01-0.03). A significant association of CCNG1, POU2AF1, HPRT1, and WNT3 (P = 0.006-0.04) with acute or late radiotoxicity could be shown before the onset of these clinical outcomes. In an established set of four genes predicting acute health effects in blood, the response in saliva samples was similar to the expected up- (FDXR, DDB2) or downregulation (POU2AF1, WNT3) in blood for up to 71% of the measurements. Comparing GE responses (PHPT1, CCNG1, CDKN1A, GADD45A, SESN1) in saliva and blood samples, there was a significant linear association between saliva and blood response of CDKN1A (R2 = 0.60, P = 0.0004). However, the GE pattern of other genes differed between saliva and blood. In summary, the current human in vivo study, (I) reveals significant radiation-induced GE associations of five transcriptional biomarkers in salivary samples, (II) suggests genes predicting diverse clinical outcomes such as acute and late radiotoxicity as well as ARS severity, and (III) supports the view that blood-based GE response can be reflected in saliva samples, indicating that saliva is a "mirror of the body" for certain but not all genes and, thus, studies for each gene of interest in blood are required for saliva.

Development of a Point-of-Care Microfluidic RNA Extraction Slide for Gene Expression Diagnosis after Irradiation.

In times of war, radiological/nuclear emergency scenarios have become a reemphasized threat. However, there are challenges in transferring whole-blood samples to laboratories for specialized diagnostics using RNA. This project aims to miniaturize the process of unwieldy conventional RNA extraction with its stationed technical equipment using a microfluidic-based slide (MBS) for point-of-care diagnostics. The MBS is thought to be a preliminary step toward the development of a so-called lab-on-a-chip microfluidic device. A MBS would enable early and fast field care combined with gene expression (GE) analysis for the prediction of hematologic acute radiation syndrome (HARS) severity or identification of RNA microbes. Whole blood samples from ten healthy donors were irradiated with 0, 0.5 and 4 Gy, simulating different ARS severity degrees. RNA quality and quantity of a preliminary MBS was compared with a conventional column-based (CB) RNA extraction method. GE of four HARS severity-predicting radiation-induced genes (FDXR, DDB2, POU2AF1 and WNT3) was examined employing qRT-PCR. Compared to the CB method, twice as much total RNA from whole blood could be extracted using the MBS (6.6 ± 3.2 µg vs. 12.0 ± 5.8 µg) in half of the extraction time, and all MBS RNA extracts appeared DNA-free in contrast to the CB method (30% were contaminated with DNA). Using MBS, RNA quality [RNA integrity number equivalent (RINe)] values decreased about threefold (3.3 ± 0.8 vs. 9.0 ± 0.4), indicating severe RNA degradation, while expected high-quality RINe ≥ 8 were found using column-based method. However, normalized cycle threshold (Ct) values, as well as radiation-induced GE fold-changes appeared comparable for all genes utilizing both methods, indicating that no RNA degradation took place. In summary, the preliminary MBS showed promising features such as: 1. halving the RNA extraction time without the burden of heavy technical equipment (e.g., a centrifuge); 2. absence of DNA contamination in contrast to CB RNA extraction; 3. reduction in blood required, because of twice the biological output of RNA; and 4. equal GE performance compared to CB, thus, increasing its appeal for later semi-automatic parallel field applications.

Validating Radiosensitivity with Pre-Exposure Differential Gene Expression in Peripheral Blood Predicting Survival and Non-Survival in a Second Irradiated Rhesus Macaque Cohort.

Radiosensitivity differs in humans and possibly in closely related nonhuman primates. The reasons for variation in radiosensitivity are not well known. In an earlier study, we examined gene expression (GE) pre-radiation in peripheral blood among male (n = 62) and female (n = 60) rhesus macaques (n = 122), which did or did not survive (up to 60 days) after whole-body exposure of 7.0 Gy (LD66/60). Eight genes (CHD5, CHI3L1, DYSF, EPX, IGF2BP1, LCN2, MBOAT4, SLC22A4) revealed significant associations with survival. Access to a second rhesus macaque cohort (males = 40, females = 23, total n = 63) irradiated with 5.8-7.2 Gy (LD29-50/60) and some treated with gamma-tocotrienol (GT3, a radiation countermeasure) allowed us to validate these gene expression changes independently. Total RNA was isolated from whole blood samples and examined by quantitative RT-PCR on a 96-well format. cycle threshold (Ct)-values normalized to 18S rRNA were analyzed for their association with survival. Regardless of the species-specific TaqMan assay, similar results were obtained. Two genes (CHD5 and CHI3L1) out of eight revealed a significant association with survival in the second cohort, while only CHD5 (involved in DNA damage response and proliferation control) showed mean gene expression changes in the same direction for both cohorts. No expected association of CHD5 GE with dose, treatment, or sex could be established. Instead, we observed significant associations for those comparisons comprising pre-exposure samples with CHD5 Ct values ≤ 11 (total n = 17). CHD5 Ct values ≤ 11 in these comparisons were mainly associated with increased frequencies (61-100%) of non-survivors, a trend which depending on the sample numbers, reached significance (P = 0.03) in males and, accordingly, in females. This was also reflected by a logistic regression model including all available samples from both cohorts comprising CHD5 measurements (n = 104, odds ratio 1.38, 95% CI 1.07-1.79, P = 0.01). However, this association was driven by males (odds ratio 1.62, 95% CI 1.10-2.38, P = 0.01) and CHD5 Ct values ≤ 11 since removing low CHD5 Ct values from this model, converted to insignificance (P = 0.19). A second male subcohort comprising high CHD5 Ct values ≥ 14.4 in both cohorts (n = 5) appeared associated with survival. Removing these high CHD5 Ct values converted the model borderline significant (P = 0.051). Based on the probability function of the receiver operating characteristics (ROC) curves, 8 (12.3%) and 5 (7.7%) from 65 pre-exposure RNA measurements in males, death and survival could be predicted with a negative and positive predictive value ranging between 85-100%. An associated odds ratio reflected a 62% elevated risk for dying or surviving per unit change (Ct-value) in gene expression, considering the before-mentioned CHD5 thresholds in RNA copy numbers. In conclusion, we identified two subsets of male animals characterized by increased (Ct values ≤ 11) and decreased (Ct values ≥ 14.4) CHD5 GE copy numbers before radiation exposure, which independently of the cohort, radiation exposure or treatment appeared to predict the death or survival in males.

RENEB Inter-Laboratory Comparison 2021: The Gene Expression Assay.

Early and high-throughput individual dose estimates are essential following large-scale radiation exposure events. In the context of the Running the European Network for Biodosimetry and Physical Dosimetry (RENEB) 2021 exercise, gene expression assays were conducted and their corresponding performance for dose-assessment is presented in this publication. Three blinded, coded whole blood samples from healthy donors were exposed to 0, 1.2 and 3.5 Gy X-ray doses (240 kVp, 1 Gy/min) using the X-ray source Yxlon. These exposures correspond to clinically relevant groups of unexposed, low dose (no severe acute health effects expected) and high dose exposed individuals (requiring early intensive medical health care). Samples were sent to eight teams for dose estimation and identification of clinically relevant groups. For quantitative reverse transcription polymerase chain reaction (qRT-PCR) and microarray analyses, samples were lysed, stored at 20°C and shipped on wet ice. RNA isolations and assays were run in each laboratory according to locally established protocols. The time-to-result for both rough early and more precise later reports has been documented where possible. Accuracy of dose estimates was calculated as the difference between estimated and reference doses for all doses (summed absolute difference, SAD) and by determining the number of correctly reported dose estimates that were defined as ±0.5 Gy for reference doses <2.5 Gy and ±1.0 Gy for reference doses >3 Gy, as recommended for triage dosimetry. We also examined the allocation of dose estimates to clinically/diagnostically relevant exposure groups. Altogether, 105 dose estimates were reported by the eight teams, and the earliest report times on dose categories and estimates were 5 h and 9 h, respectively. The coefficient of variation for 85% of all 436 qRT-PCR measurements did not exceed 10%. One team reported dose estimates that systematically deviated several-fold from reported dose estimates, and these outliers were excluded from further analysis. Teams employing a combination of several genes generated about two-times lower median SADs (0.8 Gy) compared to dose estimates based on single genes only (1.7 Gy). When considering the uncertainty intervals for triage dosimetry, dose estimates of all teams together were correctly reported in 100% of the 0 Gy, 50% of the 1.2 Gy and 50% of the 3.5 Gy exposed samples. The order of dose estimates (from lowest to highest) corresponding to three dose categories (unexposed, low dose and highest exposure) were correctly reported by all teams and all chosen genes or gene combinations. Furthermore, if teams reported no exposure or an exposure >3.5 Gy, it was always correctly allocated to the unexposed and the highly exposed group, while low exposed (1.2 Gy) samples sometimes could not be discriminated from highly (3.5 Gy) exposed samples. All teams used FDXR and 78.1% of correct dose estimates used FDXR as one of the predictors. Still, the accuracy of reported dose estimates based on FDXR differed considerably among teams with one team's SAD (0.5 Gy) being comparable to the dose accuracy employing a combination of genes. Using the workflow of this reference team, we performed additional experiments after the exercise on residual RNA and cDNA sent by six teams to the reference team. All samples were processed similarly with the intention to improve the accuracy of dose estimates when employing the same workflow. Re-evaluated dose estimates improved for half of the samples and worsened for the others. In conclusion, this inter-laboratory comparison exercise enabled (1) identification of technical problems and corrections in preparations for future events, (2) confirmed the early and high-throughput capabilities of gene expression, (3) emphasized different biodosimetry approaches using either only FDXR or a gene combination, (4) indicated some improvements in dose estimation with FDXR when employing a similar methodology, which requires further research for the final conclusion and (5) underlined the applicability of gene expression for identification of unexposed and highly exposed samples, supporting medical management in radiological or nuclear scenarios.

Gene Expression Changes in a Prefinal Health Stage of Lethally Irradiated Male and Female Rhesus Macaques.

Radiation-induced gene expression (GE) changes can be used for early and high-throughput biodosimetry within the first three days postirradiation. However, is the method applicable in situations such as the Alexander Litvinenko case or the Goiania accident, where diagnosis occurred in a prefinal health stage? We aimed to characterize gene expression changes in a prefinal health stage of lethally irradiated male and female rhesus macaques. Peripheral blood was drawn pre-exposure and at the prefinal stage of male and female animals, which did not survive whole-body exposure with 700 cGy (LD66/60). RNA samples originated from a blinded randomized Good Laboratory Practice study comprising altogether 142 irradiated rhesus macaques of whom 60 animals and blood samples (15 samples for both time points and sexes) were used for this analysis. We evaluated GE on 34 genes widely used in biodosimetry and prediction of the hematological acute radiation syndrome severity (H-ARS) employing quantitative real-time polymerase chain reaction (qRT-PCR). These genes were run in duplicate and triplicate and altogether 96 measurements per time point and sex could be performed. In addition, 18S ribosomal RNA (rRNA) was measured to depict the ribosome/transcriptome status as well as for normalization purposes and 16S rRNA was evaluated as a surrogate for bacteremia. Mean differential gene expression (DGE) was calculated for each gene and sex including all replicate measurements and using pre-exposure samples as the reference. From 34 genes, altogether 27 genes appeared expressed. Pre-exposure samples revealed no signs of bacteremia and 18S rRNA GE was in the normal range in all 30 samples. Regarding prefinal samples, 46.7% and 40% of animals appeared infected in females and males, respectively, and for almost all males this was associated with out of normal range 18S rRNA values. The total number of detectable GE measurements was sixfold (females) and 15-fold (males) reduced in prefinal relative to pre-exposure samples and about tenfold lower in 80% of prefinal compared to pre-exposure samples (P < 0.0001). An overall 11-fold (median) downregulation in prefinal compared to pre-exposure samples was identified for most of the 27 genes and even FDXR appeared 4-14-fold downregulated in contrast to a pronounced up-regulation according to cited work. This pattern of overall downregulation of almost all genes and the rapid reduction of detectable genes at a prefinal stage was found in uninfected animals with normal range 18S rRNA as well. In conclusion, in a prefinal stage after lethal radiation exposure, the ribosome/transcriptome status remains present (based on normal range 18S rRNA values) in 60-67% of animals, but the whole transcriptome activity in general appears silenced and cannot be used for biodosimetry purposes, but probably as an indicator for an emerging prefinal health stage.

Examining potential confounding factors in gene expression analysis of human saliva and identifying potential housekeeping genes.

Isolation of RNA from whole saliva, a non-invasive and easily accessible biofluid that is an attractive alternative to blood for high-throughput biodosimetry of radiological/nuclear victims might be of clinical significance for prediction and diagnosis of disease. In a previous analysis of 12 human samples we identified two challenges to measuring gene expression from total RNA: (1) the fraction of human RNA in whole saliva was low and (2) the bacterial contamination was overwhelming. To overcome these challenges, we performed selective cDNA synthesis for human RNA species only by employing poly(A)+-tail primers followed by qRT-PCR. In the current study, this approach was independently validated on 91 samples from 61 healthy donors. Additionally, we used the ratio of human to bacterial RNA to adjust the input RNA to include equal amounts of human RNA across all samples before cDNA synthesis, which then ensured comparable analysis using the same base human input material. Furthermore, we examined relative levels of ten known housekeeping genes, and assessed inter- and intra-individual differences in 61 salivary RNA isolates, while considering effects of demographical factors (e.g. sex, age), epidemiological factors comprising social habits (e.g. alcohol, cigarette consumption), oral hygiene (e.g. flossing, mouthwash), previous radiological diagnostic procedures (e.g. number of CT-scans) and saliva collection time (circadian periodic). Total human RNA amounts appeared significantly associated with age only (P ≤ 0.02). None of the chosen housekeeping genes showed significant circadian periodicity and either did not associate or were weakly associated with the 24 confounders examined, with one exception, 60% of genes were altered by mouthwash. ATP6, ACTB and B2M represented genes with the highest mean baseline expression (Ct-values ≤ 30) and were detected in all samples. Combining these housekeeping genes for normalization purposes did not decrease inter-individual variance, but increased the robustness. In summary, our work addresses critical confounders and provides important information for the successful examination of gene expression in human whole saliva.

Early molecular markers for retrospective biodosimetry and prediction of acute health effects.

Radiation-induced biological changes occurring within hours and days after irradiation can be potentially used for either exposure reconstruction (retrospective dosimetry) or the prediction of consecutively occurring acute or chronic health effects. The advantage of molecular protein or gene expression (GE) (mRNA) marker lies in their capability for early (1-3 days after irradiation), high-throughput and point-of-care diagnosis, required for the prediction of the acute radiation syndrome (ARS) in radiological or nuclear scenarios. These molecular marker in most cases respond differently regarding exposure characteristics such as e.g. radiation quality, dose, dose rate and most importantly over time. Changes over time are in particular challenging and demand certain strategies to deal with. With this review, we provide an overview and will focus on already identified and used mRNA GE and protein markers of the peripheral blood related to the ARS. These molecules are examined in light of 'ideal' characteristics of a biomarkers (e.g. easy accessible, early response, signal persistency) and the validation degree. Finally, we present strategies on the use of these markers considering challenges as their variation over time and future developments regarding e.g. origin of samples, point of care and high-throughput diagnosis.

Inter-laboratory comparison of gene expression biodosimetry for protracted radiation exposures as part of the RENEB and EURADOS WG10 2019 exercise.

Large-scale radiation emergency scenarios involving protracted low dose rate radiation exposure (e.g. a hidden radioactive source in a train) necessitate the development of high throughput methods for providing rapid individual dose estimates. During the RENEB (Running the European Network of Biodosimetry) 2019 exercise, four EDTA-blood samples were exposed to an Iridium-192 source (1.36 TBq, Tech-Ops 880 Sentinal) at varying distances and geometries. This resulted in protracted doses ranging between 0.2 and 2.4 Gy using dose rates of 1.5-40 mGy/min and exposure times of 1 or 2.5 h. Blood samples were exposed in thermo bottles that maintained temperatures between 39 and 27.7 °C. After exposure, EDTA-blood samples were transferred into PAXGene tubes to preserve RNA. RNA was isolated in one laboratory and aliquots of four blinded RNA were sent to another five teams for dose estimation based on gene expression changes. Using an X-ray machine, samples for two calibration curves (first: constant dose rate of 8.3 mGy/min and 0.5-8 h varying exposure times; second: varying dose rates of 0.5-8.3 mGy/min and 4 h exposure time) were generated for distribution. Assays were run in each laboratory according to locally established protocols using either a microarray platform (one team) or quantitative real-time PCR (qRT-PCR, five teams). The qRT-PCR measurements were highly reproducible with coefficient of variation below 15% in ≥ 75% of measurements resulting in reported dose estimates ranging between 0 and 0.5 Gy in all samples and in all laboratories. Up to twofold reductions in RNA copy numbers per degree Celsius relative to 37 °C were observed. However, when irradiating independent samples equivalent to the blinded samples but increasing the combined exposure and incubation time to 4 h at 37 °C, expected gene expression changes corresponding to the absorbed doses were observed. Clearly, time and an optimal temperature of 37 °C must be allowed for the biological response to manifest as gene expression changes prior to running the gene expression assay. In conclusion, dose reconstructions based on gene expression measurements are highly reproducible across different techniques, protocols and laboratories. Even a radiation dose of 0.25 Gy protracted over 4 h (1 mGy/min) can be identified. These results demonstrate the importance of the incubation conditions and time span between radiation exposure and measurements of gene expression changes when using this method in a field exercise or real emergency situation.

Predicting the Radiation Sensitivity of Male and Female Rhesus Macaques Using Gene Expression.

Radiosensitivity differs in humans and likely among closely-related primates. Reasons for variation in radiosensitivity are not well known. We examined preirradiation gene expression in peripheral blood among male and female rhesus macaques which did or did not survive (up to 60 days) after whole-body irradiation with 700 cGy (LD66/60). RNA samples originated from a blinded randomized Good Laboratory Practice study in 142 irradiated rhesus macaques. Animals were untreated (placebo), or treated using recombinant human IL-12, G-CSF or combination of the two. We evaluated gene expression in a two-phase study design where phase I was a whole genome screen [next generation sequencing (NGS)] for mRNAs (RNA-seq) using five RNA samples from untreated male and female animals per group of survivor and non-survivor (total n = 20). Differential gene expression (DGE) was defined as a statistically significant and ≥2-fold up- or downregulation of mRNA species and was calculated between groups of survivors and non-survivors (reference) and by gender. Altogether 659 genes were identified, but the overlapping number of differentially expressed genes (DGE) observed in both genders was small (n = 36). Fifty-eight candidate mRNAs were chosen for independent validation in phase II using the remaining samples (n = 122) evaluated with qRT-PCR. Among the 58 candidates, 16 were of significance or borderline significance (t test) by DGE. Univariate and multivariate logistic regression analysis and receiver operating characteristic (ROC) curve analysis further refined and identified the most outstanding validated genes and gene combinations. For untreated male macaques, we identified EPX (P = 0.005, ROC=1.0), IGF2BP1 (P = 0.05, ROC=0.74) and the combination of EPX with SLC22A4 (P = 0.03, ROC=0.85) which appeared most predictive for the clinical outcome for treated and combined (untreated and treated) male macaque groups, respectively. For untreated, treated and both combined female macaque groups the same gene (MBOAT4, P = 0.0004, ROC = 0.81) was most predictive. Based on the probability function of the ROC curves, up to 74% of preirradiation RNA measurements predicted survival with a positive and negative predictive value ranging between 85-100% and associated odds ratios reflecting a 2-3-fold elevated risk for surviving per unit change (cycle threshold value) in gene expression. In conclusion, we identified gender-dependent genes and gene combinations in preirradiation blood samples for survival prediction after irradiation in rhesus macaques.

Differentiating Total- or Partial-Body Irradiation in Baboons Using mRNA Expression Patterns: A Proof of Concept.

To better predict clinical outcome after radiation exposure, it is very important to know the absorbed dose and body areas exposed. Previously we found that 22 miRNAs appeared to predict total- and partial-body irradiation (TBI and PBI, respectively) patterns and were suggestive of the percentage of the body exposed in a baboon model. Motivated by these results, we performed a similar analysis on the transcriptional level (mRNAs) using whole genome microarrays. From 17 irradiated baboons, blood samples were taken before, and at 1, 2, 7, 28 and 75-106 days postirradiation to an equivalent TBI dose of 2.5 or 5 Gy applied either to the total body or to different parts of the body such as the upper body (UBE) or left hemibody (LHB). We compared quantile normalized log2-transformed gene expression values with three exposure pattern comparisons, namely TBI vs. PBI, TBI vs. LHB and UBE vs. LHB using Kruskal-Wallis and logistic regression analysis for receiver-operator characteristic (ROC) calculation. We found several hundred significantly (P < 0.05) and ≥2-fold deregulated mRNAs per exposure pattern comparison with a peak of 163-860 mRNAs at day 28. Lower numbers on day 2 (60 mRNAs) and day 7 (91-162 mRNAs) were observed, with the lowest number of deregulated mRNAs at day 75-106 (22-58 mRNAs). The 14 most promising mRNAs (e.g., LTF, DEFA3, OLFM4) appeared 10.1-46.2-fold upregulated and the exposure groups were completely or almost completely discriminated (ROC between 0.8-1.0). Several of the mRNA gene expression changes were significantly associated with the percentage of the body exposed. The numbers of overlapping genes used for diagnosis on consecutive days postirradiation were mostly 0 or less than 10. Bioinformatic analysis confirmed that at each time point different biological processes predominated. Our results suggest mRNA changes over time may be used to retrospectively determine radiation exposure patterns as partial or total body. mRNA gene expression changes likely could be applied over a longer time frame (2-75 days postirradiation) than miRNA, but due to the transient gene expression changes a different set of candidate mRNAs appears to be required at each day after irradiation.

miRNA Expression Patterns Differ by Total- or Partial-Body Radiation Exposure in Baboons.

In a radiation exposure event, a likely scenario may include either total-body irradiation (TBI) or different partial-body irradiation (PBI) patterns. Knowledge of the exposure pattern is expected to improve prediction of clinical outcome. We examined miRNA species in 17 irradiated baboons receiving an upper-body, left hemibody or total-body irradiation of 2.5 or 5 Gy. Blood samples were taken before irradiation and at 1, 2, 7, 28 and 75-106 days after irradiation. Using a qRT-PCR platform for simultaneous detection of 667 miRNAs, we identified 55 miRNAs over all time points. Candidate miRNAs, such as miR-17, miR-128 or miR-15b, significantly discriminated TBI from different PBI exposure patterns, and 5-to-10-fold changes in gene expression were observed among the groups. A total of 22 miRNAs (including miR-17) revealed significant linear associations of gene expression changes with the percentage of the exposed body area (P < 0.0001). All these changes were primarily observed at day 7 postirradiation and almost no miRNAs were detected either before or after 7 days. A significant association in the reduction of lymphocyte counts in TBI compared to PBI animals corresponded with the number of miRNA candidates. This finding suggests that our target miRNAs predominantly originated from irradiated lymphocytes. In summary, gene expression changes in the peripheral blood provided indications of the exposure pattern and a suggestion of the percentage of the exposed body area.