Comparison of gene expression response to neutron and x-ray irradiation using mouse blood

The goal of radiation biodosimetry is to accurately predict radiation dose, as a surrogate of radiological injury, in a large-scale radiation emergency. Most approaches focus on the first week after an event, when radiological triage is most needed. A wide range of different methods have been used to identify cellular responses to radiation, ranging from cytogenetic measurements, specifically dicentric assays, to high throughput proteomic, metabolomic, and transcriptomic profiling [1]. Several groups have reported gene expression-based analyses, with the development of predictive gene sets that correlate with dose [2–5].

Radiation quality plays an important role in driving gene expression, and different signaling pathways may be triggered in response to the different types of irradiation. In addition, various genes and biological functions are induced in cells at different time points after radiation exposure [6–8]. To date, most of the gene expression signatures for the purpose of radiation biodosimetry have been obtained in response to photon (γ- and x-rays) exposure. However, there have been a limited number of studies that have compared gene expression profiles in mice or human blood cells after exposures involving heavy ions and α-particles, the latter representing isotopes likely to be used in a radiological dispersal device [9]. For example, using human peripheral blood mononuclear cells, Chauhan et al. [10] found that the gene expression profile induced by α-particle radiation was very similar to the x-ray responses, despite the fact that α-particles are characterized by a higher linear energy transfer coefficient compared with x-rays [11]. In contrast, comparison of α-particle and x-ray irradiation impact on human tumor and endothelial cells [12], as well as human epidermal keratinocytes [13] produced distinct gene expression profiles.

An improvised nuclear device (IND) detonation is capable of producing a mixture of γ-rays and fast neutrons. Furthermore, as neutrons generally have a higher Relative Biological Effectiveness (RBE) for most physiological endpoints, it is important to understand the impact that neutrons would have on the biodosimetry methods that are being developed for medical triage purposes. Radiation affects multiple biological processes, including immune system and inflammatory responses, cell cycle progression, cell death, DNA repair, as well as metabolism. However, the impact of neutron radiation on any of these processes has not been studied in detail. This study presents an initial characterization of gene expression responses following exposure to IND-spectrum neutrons or x-rays.

In the present study, gene expression signatures were analyzed after in vivo irradiation of mice with either neutrons or x-rays. The doses used were selected based on the limited RBE information available for IND-spectrum neutrons. Using the in vitro cytokinesis-block micronucleus assay [14] in peripheral human blood lymphocytes, Xu et al. [15] calculated the RBE of our neutron spectrum to be between 3 and 5 for micronucleus induction with 250 kVp x-rays used as the reference radiation, with similar results for micronucleus induction in mice (Helen Turner, personal communication). We therefore exposed mice to 0.25 Gy or 1 Gy of neutron radiation and, separately, 1 Gy or 4 Gy of x-ray radiation. We describe biological processes significantly over-represented only in the neutron response, or only in the x-ray radiation response, as well as processes shared by both types of radiation. The ultimate goal is to investigate whether we can separately estimate the photon and the neutron component after a mixed photon/neutron exposure, and to develop gene expression signatures capable of discriminating between neutron and photon exposures.

The overall goal of this study was to identify differentially regulated genes in response to neutron or x-ray irradiation and perform a comparative analysis of biological processes between the two types of radiation at time points spanning the range of interest for biodosimetry. Our data suggest that the gene transcriptional response varied widely depending on radiation quality, dose, and time since exposure. It should be noted that these characteristics of gene expression response may contribute to the apparently large number of “unique” genes responding to only one radiation quality. Previous studies have shown differences in the timing of gene expression responses at high and low doses, and following exposure to different radiation qualities. It is likely that many of the genes seen to respond to only one radiation quality in this study would show a response to the other radiation quality in a different time-dose combination. Some of the observed differences may also be attributable to the different nature of x-rays vs. neutrons. Up to 2/3 of damage from low-LET (e.g. x-rays) is due to indirect action, (mediated by free radicals), whereas high-LET neutrons cause direct damage to the DNA (Hall and Giaccia, 2012), which is more complex and difficult to repair, and may result in different signaling responses.

A small number of genes showed similar changes after exposure to neutrons and x-rays, and displayed a bidirectional mode of regulation. A similar temporal pattern of expression for some genes has been described previously in mice injected with ¹³⁷CsCl [6]. In that study, genes were upregulated by day 2 or 3, and then downregulated by day 20 or 30 after isotope administration. Many gene ontology categories, including actin and the cytoskeleton and integrin signaling pathways, showed the same temporal pattern. In the present study, we identified a number of differentially expressed genes that were significantly different from controls at all radiation types and doses. A number of these genes were downregulated at day 1, and upregulated at day 7. Gene ontology analysis revealed that one biological function, cell cycle, was significant among down-regulated genes at day 1 and then significant among up-regulated genes at day 7. All other biological functions were either significant at one dose and time of irradiation, or they were uniformly up- or down-regulated irrespective of time.

Most of the genes differentially expressed in both neutron and x-ray exposures were related to immune response, and B and T cell physiology. These genes were downregulated starting at day 1 and reduced expression persisted until the end of the experiments at day 7. Widespread decreased expression of immune function genes has been shown previously in both human blood irradiated ex vivo, as well as in vivo mouse peripheral blood following ionizing radiation exposure or ⁹⁰Sr as an internal emitter [22–24]. High-dose radiation (> 1 Gy) has been shown to disrupt immune cell functions, leading to increased cell death of blood cells in mice [25, 26]. Moreover, in patients with acute radiation syndrome, hematopoietic cell proliferation is inhibited by radiation exposure [27]. A preponderance of down-regulated genes has previously been associated with higher doses or later times after irradiation (e.g. [6]), perhaps reflecting greater damage or a failure of repair. In our study, even 0.25 Gy of neutron radiation resulted in the downregulation of genes involved in immune cell function.

A significant number of genes that were downregulated by both neutron and x-ray exposure at day 1 were related to DNA and mRNA metabolism, gene transcription, RNA processing and splicing. However, at day 7, these processes were no longer overrepresented following x-ray exposure, but they were present after neutron exposure. Moreover, in response to neutron exposure, these processes were enriched with additional related processes, such as tRNA metabolism/processing, RNA transport/localization, and noncoding RNA functions.

The contribution of DNA and RNA related functions in response to DNA damage has only recently been appreciated. A genome-wide siRNA screen looking for modulators of DNA damage signaling revealed that the largest number of hits were those targeting gene products responsible for nucleic acid metabolism, particularly those involved in mRNA binding and processing [28]. Furthermore, a phosphoproteomic analysis showed a close link between genome stability and RNA synthesis metabolism [29]. Likewise, it has been shown that the largest subset of ATM/ATR/DNA-PK substrates identified in a phosphoproteomic screen were proteins linked to RNA and DNA metabolism, particularly those proteins involved in posttranscriptional mRNA regulation [30]. These observations, employing different experimental approaches, highlight the importance of regulatory circuits controlling RNA metabolism and stability in DNA repair and checkpoint function. In addition to these findings, our study suggests that neutron (but not photon) irradiation affects biological processes enriched in tRNA regulation and RNA transportation and localization, as well as non-coding RNA metabolism and processing. tRNAs have been viewed as passive players involved in protein synthesis. However, recent evidence suggests that they have more active roles and tRNA modulation represents a mechanism by which cells achieve altered expression of specific transcripts and proteins. tRNA pools in cells can be divided into those that favor proliferation and those that support differentiation [31]. As a result, modifications in tRNA and their corresponding enzymes are implicated in diseases, including diabetes and cancer. For example, upregulation of certain tRNAs increases metastasis in breast cancer patients [32]. Furthermore, control of RNA transport and localization would be expected to impact on the rate of protein translation [33].

A difference observed between neutron and x-ray response was the enrichment in biological processes involved in lipid biosynthesis and metabolism that was seen only in response to x-ray exposure. It has long been known that the cellular targets of ionizing radiation, such as x-rays, are not limited to nuclear DNA, but that proteins and lipids in other cellular compartments, such as the plasma membrane [34], are also affected. The action of x-rays has been attributed to the generation of reactive oxygen species that oxidize DNA, lipids and proteins [35]. We can speculate that in response to x-rays, cells upregulate lipid, coenzyme, and vitamin biosynthetic and metabolic processes as a means of repairing the damage caused by x-ray irradiation to the cell membrane. The latter two processes could also serve as anti-oxidant responses [36]. In addition, fatty acid oxidation processes, which are also overrepresented in the x-ray irradiation, would be required by the cells to meet the energy demands of various metabolic processes.

In response to DNA damage, cells activate the DNA damage and repair signaling pathway. DNA damage that cannot be repaired efficiently leads to cell death or senescence. Although protein abundance and activity do not always follow gene expression changes, our Gene Ontology analysis suggests an apparent down regulation of several DNA repair pathways after neutron but not after x-ray irradiation, especially at day 7, perhaps reflecting a failure to repair the more complex damage resulting from high LET radiation. At day 1 after 1 Gy neutron exposure, GO analysis suggested suppression of MMR, whereas at day 7, in addition to MMR, NER, BER, and NHEJ genes were all significantly over-represented among downregulated genes. In contrast, exposure to 4 Gy of x-rays transiently downregulated expression of MMR genes at day 1 and BER genes at day 7. Homologous recombination, which along with NHEJ constitutes the major DNA double-strand break repair mechanism, was not significantly over-represented among differentially expressed genes. It has been shown previously that high-LET radiation induces complex DNA damage that is not easily repaired and NHEJ is not involved [37–40]. More recently, it has been shown that high-LET irradiation with protons or carbon ions causes a shift away from NHEJ toward HRR in the repair of double-strand breaks [41]. Consistent with these data, our observed downregulation of genes in the NHEJ and other DNA repair pathways in response to neutron exposure may reflect the fact that these lesions are not repaired by these processes. Regulation at the gene expression level suggests a potential mechanism for favoring the homologous recombination pathway in the attempted repair of neutron damage, and is worthy of further investigation.

A major biological function that is affected by radiation is the cell cycle. Cell cycle regulating genes are important determinants of radiosensitivity and cell fate in response to DNA damage. We studied the effect of neutron radiation on mouse cell cycle-regulated genes and compared it with that of x-rays. Unlike genes in other biological processes, which were either up or downregulated after neutron or x-ray irradiation, many cell cycle genes showed a bidirectional expression based on time. A group of 5 genes, namely Ube2c, Fzr1, Ccna2, Cdc25b, and Nusap1, were downregulated 1 day after irradiation, whereas the same genes were overexpressed 7 days post-irradiation. These temporal changes were further confirmed by qRT-PCR. A literature search revealed that these genes play important roles in the control of mitosis. Additionally, their protein products are related to the anaphase promoting complex/cyclosome (APC/C) either as subunits of APC/C (Ube2c, Fzr1) or as substrates (Ccna2, Nusap1).

The APC/C is an E3 ubiquitin ligase, which is composed of at least 14 core subunits. The APC/C is active during mitosis and G1 phase of the cell cycle. Because of its role in cell cycle regulation, APC/C is important for maintaining genomic integrity [42]. Furthermore, APC/C has been implicated in an array of diverse functions ranging from cell differentiation to apoptosis and senescence, as well as cellular metabolism, cell motility, and gene transcription through the degradation of specific substrates [42].

APC/C targets a large repertoire of substrates and recruits them for ubiquitylation via one of two co-activators, Cdc20 and Cdh1 (Fzr1, the mouse homolog) [42]. The physiological role of Cdh1 has been extensively studied in the context of human cancer, since downregulation of Cdh1 has been reported in many cancers, including those of prostate, ovary, liver, brain, and during the malignant progression of a B-lymphoma cell line. In mice, Fzr1 heterozygosity results in the development of epithelial tumors, suggesting that Fzr1 may be a haploinsufficient tumor suppressor [43]. Downregulation of Cdh1 in post-mitotic neurons has been implicated in neurodegenerative diseases, such as Alzheimer’s disease [44].

In addition to its role in mitosis, Cdh1 has important functions in mediating DNA damage response to genotoxic stress [45] that ensure genomic integrity [46]. Cdh1-null cells fail to maintain DNA damage-induced G2 arrest and APC/C^cdh1 is activated by x-irradiation-induced DNA damage (but not UV irradiation). Interestingly, the levels of mitotic cyclins in Cdh1^-/- cells after DNA damage were similar to those of wild-type cells. These data imply that cyclin A and cyclin B cannot be substrates for APC/C^cdh1 when it is activated irregularly by DNA damage at G2 [47].

Protein ubiquitination-mediated degradation involves two distinct steps: the covalent attachment of ubiquitin to proteins catalyzed by the sequential actions of the activating (E1), conjugating (E2), and ligating (E3) enzymes, followed by the degradation of the poly-ubiquitylated protein by the 26S proteasome complex. For APC/C the E2 enzymes are Ube2c, which is one of the temporally controlled cell cycle genes in this study, and Ube2s [48], which is not differentially regulated in response to either neutrons or x-rays. Abundant experimental evidence has shown a role for Ube2c in human tumor initiation and progression. On the other hand, there are very few reports that implicate Ube2c in DNA damage response to radiation [49, 50]. Moreover, the mechanistic details of Ube2c response to radiation, as well as their pathophysiological significance remain unexplored.

APC/C regulates spindle formation by promoting the degradation of a number of spindle-binding proteins, including Nusap1 [51]. The nucleolar spindle-associated protein 1 (Nusap1) is a protein highly expressed in proliferating cells and interacts with microtubules [52]. Depletion of Nusap1 caused faulty mitotic spindles, aberrant chromosome segregation, and defective cytokinesis. Overexpression of Nusap1 caused microtubule bundling and cell cycle arrest at the G2/M checkpoint [53].

Cyclin A2 interacts with Cyclin-dependent kinase 2 and controls essential functions in DNA replication and cellular proliferation [54]. Cyclin A2 expression is associated with a poor prognosis in several types of cancer [55]. Cyclin A2 mRNA as well as protein are cell cycle regulated [56] with mRNA and protein abundance increasing 4-fold and 20-fold, respectively, as cells progress from G1 to G2 phase. APC/C degrades Cyclin A2 at the end of mitosis, while mRNA persists longer than the protein in cells.

To fully appreciate the significance of the temporal differential expression of APC-related genes in the radiation response, we will require knowledge about the status of protein levels and their posttranslational modifications (e.g. phosphorylation). However, we can speculate that these changes may be relevant to cell cycle progression, and especially mitosis, thus ensuring genomic stability after irradiation.

Another cell cycle regulated gene that appears in the list of temporally bi-directionally expressed genes in the present study is E2f2. While this gene was consistently downregulated at day 1 post-irradiation, it always appeared upregulated at day 7. This was in sharp contrast with other E2f genes that showed no change at day 1 and downregulation (E2f1, E2f3, E2f4) at day 7 post neutron irradiation, or significant upregulation (E2f1, E2f8) at day 7 post x-ray irradiation. The E2f family of transcription factors has well known functions in the control of cell cycle, and E2f1-3, especially, in promoting G1/S cell cycle transition and thus cell proliferation [57]. Target genes of E2f include several hundred genes that are involved not only in DNA replication and cell cycle progression, but also in DNA damage repair, apoptosis, differentiation and development [58]. The role of E2fs in mitosis has been shown in cancer cells [59–63]. Although the specific function of E2f2 in response to radiation has not been studied, E2f2 transcript and protein levels increase in response to genotoxic stress and maintain genomic stability in neuronal cells [64]. Why E2f2 shows bi-directional expression changes, whereas other E2f members (i.e., E2f1, E2f3, E2f4, and E2f8) do not, is currently not known.

In the current work we identify genes that are differentially expressed following exposure to neutrons or x-rays, as well as genes that showed similar responses to the two radiation modalities. In summary, our results show that genes involved in cell cycle regulation are differentially regulated by neutron and x-ray radiation. However, a few cell cycle genes show a consistent temporal regulation that is common across the two radiation modalities. These genes are functionally interconnected and play important roles during mitosis.

We have found differing patterns of gene expression response to x-rays and neutrons that vary with both dose and time since exposure. These findings support the possibility of using gene expression to detect the neutron component of exposures resulting from an IND detonation, thus providing triage information more relevant to the actual radiation injury than an estimate of total dose alone. Further work will be needed to develop gene expression patterns specific to neutron exposure that would be useful for triage following an IND event. The identification of biomarkers predictive of both dose and type of radiation would be an important advancement in biodosimetry to determine an individual’s exposure and allow more accurate triage for further medical treatment.

The ultimate goal of our neutron studies is to investigate whether we can separately estimate the photon and the neutron component after a mixed photon/neutron exposure, and to develop gene expression signatures capable of discriminating between photon/neutron and pure photon exposures. Following our initial investigation of the gene expression response to neutrons, reported here, our future studies will focus on mixed exposures, including the possibility of synergistic responses.