Gamma-tocotrienol, a radiation countermeasure, reverses proteomic changes in serum following total-body gamma irradiation in mice

Radiological incidents or terrorist attacks would likely expose civilians and military personnel to high doses of ionizing radiation, leading to the development of acute radiation syndrome. We examined the effectiveness of prophylactic administration of a developmental radiation countermeasure, γ-tocotrienol (GT3), in a total-body irradiation (TBI) mouse model. CD2F1 mice received GT3 24 h prior to 11 Gy cobalt-60 gamma-irradiation. This dose of radiation induces severe hematopoietic acute radiation syndrome and moderate gastrointestinal injury. GT3 provided 100% protection, while the vehicle control group had 100% mortality. Two-dimensional differential in-gel electrophoresis was followed by mass spectrometry and Ingenuity Pathway Analysis (IPA). Analysis revealed a change in expression of 18 proteins in response to TBI, and these changes were reversed with prophylactic treatment of GT3. IPA revealed a network of associated proteins involved in cellular movement, immune cell trafficking, and inflammatory response. Of particular interest, significant expression changes in beta-2-glycoprotein 1, alpha-1-acid glycoprotein 1, alpha-2-macroglobulin, complement C3, mannose-binding protein C, and major urinary protein 6 were noted after TBI and reversed with GT3 treatment. This study reports the untargeted approach, the network, and specific serum proteins which could be translated as biomarkers of both radiation injury and protection by countermeasures.


Results
The purpose of this study was to report the comprehensive changes in mouse serum protein expression by GT3, a promising radiation countermeasure, and correlate these with survival against 11 Gy irradiation, which induces severe H-ARS and moderate GI injury. One group of mice (n = 16) was administered GT3 and the other group received vehicle 24 h prior to 11 Gy TBI, and all were observed for 30 days post-irradiation. Vehicle-treated animals started dying on day 7 and by day 17 post-irradiation, all animals in this group expired. On the other hand, all GT3-treated mice survived for 30 days until the experiment was terminated (Fig. 1, p < 0.001). www.nature.com/scientificreports/ The molecular mechanism of this radioprotective agent was also investigated by considering the pathways and networks involved. The mice were prophylactically treated with GT3 or vehicle 24 h prior to supralethal TBI of 11 Gy. This TBI dose was selected because it is known to lead to severe H-ARS and moderate GI injury in CD2F1 mice. After 24 h post-irradiation, serum samples were collected. The 2D-DIGE method was employed to investigate changes in protein expression. Spot detection software was used to select 100-200 spots with isoelectric points ranging between pI 4 and 9, and molecular weights ranging from 10 to 150 kDa for trypsin digestion from the approximately 2500 proteins of interest from each gel (Fig. 2). The analyses identified proteins using MALDI-TOF/TOF and the MASCOT search engine (Fig. 3).
The gel analysis software identified 132 unique spots, of which 78 protein identities were determined. Of this cohort, 53 spots of these proteins exhibited a 1.5-fold or greater change in expression when comparing the serum samples of unirradiated animals to those from the TBI groups. The 53 spots represented 23 unique proteins because there were multiple instances of the same protein, likely the result of potential post-translational modifications. From these 23 proteins, 19 of the appearances reflected either a consistent increase or decrease in expression in response to TBI with respect to the unirradiated control group. Finally, 18 of these 19 proteins reflected an expression reversal when compared to the prophylactic GT3-treated group.
There were 18 proteins detected in the serum samples of the mice that received TBI and prophylactic treatment with GT3 that showed a marked change in expression which was reversed compared to non-prophylactic treatment with GT3, indicating a potentially protective effect of GT3 treatment (Fig. 4). Protein expression ratio changes are presented with a color scale to indicate increase (red) or decrease (green). One protein, beta-2-glycoprotein 1, showed a decrease in expression while the remaining 17 proteins exhibited an increase in expression in response to TBI with respect to the unirradiated group. With GT3 prophylactic treatment, beta-2-glycoprotein 1 showed an increase in expression while the remaining 17 proteins exhibited a decrease in expression, reversing the initial observation and indicating a treatment specific effect.
The profile of serum proteins showed a change in expression in response to TBI with respect to the unirradiated group and then a corresponding reversal with GT3 prophylactic treatment were offered to the Ingenuity Pathway Analysis (IPA) software program to determine potential pathways and to chart a network based upon previously reported protein relationships (Fig. 5). The network suggested by IPA identified a number of potential relationships particularly associated with immune response involving interleukins, transcription factors, receptors, and apoptosis-related proteins.

Discussion
GT3 has been found to be optimally effective in mice and NHPs when administered 24 h prior to irradiation 29,30,34,40 . A dose of 11 Gy total-body cobalt-60 gamma radiation exposure induces severe H-ARS and moderate GI injury. A GT3 dose of 200 mg/kg has consistently demonstrated radioprotective efficacy in the murine model 29 . We conducted GT3 survival as well as proteomics studies using the above treatment schedule at a dose of 200 mg/kg, radiation dose of 11 Gy, and CD2F1 mice. We observed 100% survival in the GT3-treated  www.nature.com/scientificreports/ group by the end of the study, while all mice in the control group died by day 17 post-irradiation ( Fig. 1). Suboptimal doses of GT3 (25 and 50 mg/kg) have also demonstrated moderate radioprotective efficacy in irradiated mice 41 . Proteomic analysis has proven to be an essential tool for dissecting the mechanisms contributing to survival following exposure to ionizing radiation 42 . While tocols have shown the potential for radioprotection against ARS in studies, the underlying mechanisms and pathways of protection are not well understood 30,40 . The present study was designed to probe the serum proteins circulating following TBI and contribute new evidence of the specific proteins and networks involved in the radioprotection of GT3 treatment. We collected blood samples from animals at 24 h post-irradiation for proteomic analysis based on earlier studies where changes had been observed in serum cytokines and jejunum proteins at this time point 29  www.nature.com/scientificreports/ A total of 18 proteins found in the serum were determined to show a change in expression following TBI which was reversed with GT3 prophylactic treatment. Only one protein, beta-2-glycoprotein 1, showed a decrease in expression with TBI compared to the unirradiated group. Beta-2-glycoprotein 1, also known as ApoH, is a phospholipid binding protein which has anti-oxidant properties 43 . While the present study found beta-2-glycoprotein 1 declined with 11 Gy exposure, interestingly, Rithidech et al. 44 found an increase in expression with a sub-lethal exposure of 3 Gy at 48 h. Of course, it is not surprising that the proteomic profile is temporally dynamic and dose dependent, making the case for further investigation into time course and irradiation dose to enhance our understanding of the proteomic dynamics which can inform biomarker identification.
The remaining 17 proteins identified showed an increase in expression following 11 Gy radiation. Nine of these proteins were also found to be elevated in expression by Rithidech et al. 44 at 48 h following 3 Gy irradiation including alpha-1-acid glycoprotein 1, alpha-2-macroglobulin, clusterin, complement C3, hemopexin, haptoglobin, kininogen-1, mannose-binding protein C, and major urinary protein 6. While we found alpha-1-antitrypsin 1-1 to be elevated at 24 h, Rithidech et al. 44 found expression decreased after 1 week with respect to the unirradiated control. This investigation used C57BL/6 mice and a 9.0 Gy (0.76 Gy/min) dose of totalbody cobalt-60 gamma irradiation, which was approximately LD 40/30 , while the radiation dose in our study was supralethal leading to 100% mortality in untreated mice by day 17 post-irradiation.
Fu et al. 45 found serum amyloid A, interleukin (IL)-22, urokinase, resistin, and IL-6 to be good predictors of mortality in a mouse model. This study used CBA/CaJ mice and a total-body sublethal dose of 3 Gy (0.72/Gy) Cesium-137 gamma-radiation. This study did not distinguish between the specific forms of serum amyloid A, but our study also revealed changes in serum amyloid A-2. While our study did not uncover the other proteins identified by Fu et al., it is likely that low abundant interleukins would be less likely to be detected by the 2-D gel method. Another recent study observing long term exposure to low-dose radiation found interleukin IL-5, IL-12p40, P-selectin, and serum amyloid A1 to be good predictors of low-dose radiation exposure with biodosimetric potential 46 .
The proteins identified in our study can be grouped into binding proteins, inflammatory response proteins, coagulation proteins/factors, protease inhibitors, and proteins involved in metabolism and growth. The binding proteins include beta-2-glycoprotein 1, hemopexin, haptoglobin, mannose-binding protein C, and major urinary protein 6. It should be noted that these binding proteins also play a role in inflammatory response 44 .
The inflammatory response proteins identified include alpha-1-acid glycoprotein 1, alpha-1-antitrypsin 1-1, inter alpha-trypsin inhibitor, heavy chain 4, serum amyloid A-2 protein, and haptoglobin. Interestingly, both alpha-1-acid glycoprotein 1 and serum amyloid A-2 protein have recently been reported as potential biomarkers of radiation injury, and these protein expression levels were reversed in our study with prophylactic GT3 treatment 47 . Seven of the proteins identified in our study can be classified as coagulation-related proteins. These members include alpha-2-macroglobulin, serum albumin, beta-2-glycoprotein 1, clusterin, hemopexin, haptoglobin, and kininogen-1. Of particular note is alpha-2-macroglobulin, which was recently reported as a potential radioprotectant in the esophagus in patients receiving thoracic radiotherapy 48 . Further evidence of radioprotection was previously reported in a rat model where the animals received 6.7 Gy X-ray TBI [49][50][51] . In these three studies, rats were exposed to total-body X-rays at a dose of 6.7 Gy (LD 50/30 ), while the radiation dose used for our study (11 Gy) was supralethal.
Other proteins detected fall under the metabolism and growth category. These include epidermal growth factor receptor, leucine-rich alpha-2-glycoprotein precursor, and keratin, type II cytoskeletal 73. Finally, we   www.nature.com/scientificreports/ Additionally, we decided not to remove abundant serum proteins for this study, but this added step may be helpful in resolving proteomic changes of less prominent proteins. Many of the proteins detected in the present study are known to be glycosylated. These proteins include alpha-1-acid glycoprotein 1, alpha-1-antitrypsin 1-1, alpha-2-macroglobulin, beta-2-glycoprotein 1, kininogen-1, hemopexin, and haptoglobin. While this study does not report on the glycosylation status of the serum proteins identified, this may be an interesting area of further inquiry. Changes in sialylation of glycoproteins following TBI and serum N-glycome following local irradiation exposure in mouse serum have been reported 54,55 . Sitespecific glycosylation of seven major plasma proteins in response to partial-body irradiation therapy in cancer patients was recently reported 56 . This group included beta-2-glycoprotein 1, a protein found in the present study to show a reversible change in expression with GT3 treatment. However, the extent of the expression and time course exhibited variability between subjects. Clerc and collaborators have highlighted how N-glycosylation of human plasma proteins can significantly alter protein structure and function 57 . Furthermore, these changes can be considered consequences of disease state and can potentially serve as more precise biomarkers of disease. Five of the proteins identified in the current study found to be increased in expression with radiation exposure are major plasma glycoproteins including alpha-1-acid glycoprotein 1, alpha-1-antitrypsin 1-1, alpha-2-macroglobulin, beta-2-glycoprotein 1, kininogen-1, hemopexin, and haptoglobin. Therefore, we postulate that post-translational modifications, such as glycosylation may have utility in identifying more precise biomarkers of radiation exposure.

Material and methods
Animals and animal care. We procured 6 to 8-week-old male CD2F1 mice from Harlan Laboratories, Inc. Experimental design. There were three groups: untreated and sham irradiated, vehicle-treated and irradiated, and GT3-treated and irradiated. There were eight mice in each treatment group for serum sample collection and proteomic analysis. In addition, there were two groups with 16 mice each for the survival study continuing for 30 d post-irradiation. The drug was administered 24 h prior to 11 Gy irradiation, which induces severe H-ARS in CD2F1 mice. This agent has been found to be optimally effective when administered 24 h before radiation exposure.

Irradiation of mice.
Irradiation boxes consisted of compartmentalized Plexiglas boxes designed to hold eight mice. Mice were exposed to a bilateral, midline dose of 11 Gy 60 Co γ-radiation at a dose rate of 0.6 Gy/min, as described earlier 59 . Radiation dosimetry was established primarily on the alanine/EPR (electron paramagnetic resonance) system, currently accepted as one of the most accurate methods for relatively high radiation doses and was used for intercomparisons between national metrology institutions. The calibration curves (spectrometer e-Scan, Bruker Biospin, Inc., Madison, WI, USA) used in dose measurements are based on standard alanine calibration sets procured from the U.S. National Institute of Standards and Technology (NIST, Gaithersburg, MD, USA) 60,61 . The alanine dosimeters obtained from NIST had been calibrated in terms of absorbed dose to water using the US national standard radiation sources. Identical alanine dosimeters were placed midline within mouse phantoms (Plexiglas 1″ diameter, 3″ length) and irradiated for predefined periods of time. Measurement of their EPR signals were taken using the calibration curve constructed with alanine dosimeters from NISTprovided dose rates to water in the core bodies of mice. A small correction was subsequently applied for the difference in mass energy absorption coefficients between water and soft tissue.
Drug preparation and administration. Yasoo Health, Inc. (Johnson City, TN, USA) supplied GT3 formulation in 5% Tween-80 in saline. GT3 formulation was adjusted to a final concentration for injection of 200 mg/kg in 0.1 ml total volume. Equal volumes of olive oil were administered as the vehicle (olive oil in 5% Tween-80 in saline) 33 . A Luer-Lock syringe with a 23 G needle was used for all drug administrations. Subcutaneous injections were administered at the nape of the neck 24 h before irradiation.

Blood collection and serum separation. Blood was collected from anesthetized (Isoflurane, Abbott
Laboratories, Chicago, IL, USA) mice via the inferior vena cava with a 23-gauge needle. After collection, blood was transferred to CapiJect serum separator tubes (3T-MG; Terumo Medical Corp., Elkton, MD, USA), allowed to clot for 30 min, and centrifuged at 400g for 10 min. The serum was collected and stored at − 70 °C until analysis as described earlier 37 . Based on earlier studies, the 24 h post-irradiation time point was selected for sample collection. At this time point, several cytokines and protein biomarkers have been reported to be induced in response to irradiation and GT3 prophylaxis in murine and NHP models [32][33][34] . www.nature.com/scientificreports/ Protein identification. To quantify changes in serum protein expression, Cy dyes (Cy3 or Cy5) were added to serum samples for labeling. Both eight vehicle control and GT3-treated serum samples were combined and labeled with Cy2, as previously described 33 . The 2D-DIGE separation method was employed in triplicate 62 . The Amersham Biosciences 2D-gel system (Amersham Biosciences, Piscataway, NJ, USA) was used to analyze the 2D-gels, and images were captured with the Typhoon TRIO and analyzed by ImageQuantTL software (GE Healthcare, Chicago, IL, USA). Decyder software version 6.5 (GE Healthcare) in combination with Ettan Spot picker (Amersham Biosciences) was used to select between 150 and 200 protein spots per gel for identification. Digestion of the gel spots was performed in sequencing-grade modified trypsin protease at 37 °C before desalting and incorporation into the α-cyano-4-hydroxy-cinnamic acid matrix, which was then spotted onto a MALDI plate. The Applied Biosystems Proteomic Analyzer Spectrometer was used for matrix assisted laser desorption/ionization time of flight tandem mass spectrometry (MALDI-TOF/TOF) combined with peptide fingerprinting mass mapping, and peptide fragmentation mapping yielded specific protein identification. A GPS Explorer workstation equipped with MASCOT search engine (Matrix Science, Boston, MA, USA) was used to resolve peptide mass and associated fragmentation spectra from a primary sequence database.
Pathway and network analyses were prepared using IPA software (Qiagen, Redwood City, CA, USA) and were based on statistical significance as well as functional and biological relevance of the detected proteins. The IPA software was directed to focus on proteins that have the potential to appear in mouse serum for the purpose of more relevant pathway and network analyses.
Statistical analysis. For survival data, a log-rank test was used to compare survival curves. Fisher's exact test was used to compare survival rates at the end of 30 days. A Student's t test was employed for all other statistical analyses and a threshold of 1.5-fold difference was set to determine a significant change in protein expression. Each data point is the result of three biological replicates with a p value of less than 0.05. A protein score confidence interval percentage or ion confidence interval percentage of greater than 95% was considered significant for the MALDI-TOF/TOF analysis.

Data availability
All data generated or analyzed during this study are included in this article.