Exposure to ionizing radiation, mechanical trauma, toxic chemicals or infections, or combinations thereof (i.e., combined injury) can induce organic injury to brain tissues, the structural disarrangement of interactive networks of neurovascular and glial cells, as well as on arrays of the paracrine and systemic destruction. This leads to subsequent decline in cognitive capacity and decompensation of mental health. There is an ongoing need for improvement in mitigating and treating radiation- or combined injury-induced brain injury. Cranial irradiation per se can cause a multifactorial encephalopathy that occurs in a radiation dose- and time-dependent manner due to differences in radiosensitivity among the various constituents of brain parenchyma and vasculature. Of particular concern are the radiosensitivity and inflammation susceptibility of: 1. the neurogenic and oligodendrogenic niches in the subependymal and hippocampal domains; and 2. the microvascular endothelium. Thus, cranial or total-body irradiation can cause a plethora of biochemical and cellular disorders in brain tissues, including: 1. decline in neurogenesis and oligodendrogenesis; 2. impairment of the blood-brain barrier; and 3. ablation of vascular capillary. These changes, along with cerebrovascular inflammation, underlie different stages of encephalopathy, from the early protracted stage to the late delayed stage. It is evident that ionizing radiation combined with other traumatic insults such as penetrating wound, burn, blast, systemic infection and chemotherapy, among others, can exacerbate the radiation sequelae (and vice versa) with increasing severity of neurogenic and microvascular patterns of radiation brain damage.
There is growing concern regarding the long-term adverse effects of ionizing radiation on nervous tissues are associated with crucial aspects of brain constituents related to mental activity, when the radiation-induced organic brain injury severely affects mental health (1–3 ). The consistent sequelae of high-dose radiation to the brain include alteration of neuronal architecture, suppression of adult neurogenesis, and induction of neuroinflammation, vascular impairment, autoimmune response, radiation myelopathy and neurological disorders, which can ultimately lead to declining in cognitive capacity (1–16 ). Extensive research has been done to explore the response of brain tissues to X rays and gamma rays, resulting in a dramatic improvement in radiation treatment modalities (4–26 ). Nevertheless, even Gamma Knife and CyberKnife® procedures, with the stereotactic precision, produce scattered radiation to the normal parenchymal and vascular tissues outside the target areas, presenting an ongoing challenge to radiation oncologists (27–28 ).
Notably, the knowledge and perceptions of radiobiological responses/effects has been constantly evolving over the last century, undergoing revisions and encompassing new data from modern molecular and cell biology, molecular histopathology, micro-irradiation techniques for single-cell target, human and animal genetics, functional genomics, and systems biology (3 Omics), along with implementation of computational predictive models, big data research including clinical data on radiation, occupational and military medicine (1–48 ). Ultimately, this effort has led to a paradigm shift towards precision radiation oncology as well as radiation protection and radiation countermeasures based on integrative radiation systems biology (49–51 ).
Referring to radiation oncology, implementation of the systems biology methodology for optimization of ionizing radiation combined with chemotherapy or biological response modifier (BRM) therapy could revolutionize the future therapeutic modalities for cancer treatment. Indeed, while the main objective of combined-modality therapy is to maximize the ratio of normal tissue preservation against tumor cell reduction, clinical observations have often shown worsening long-term adverse radiation effects on the brain during the period after the “double-impacts” produced due to radiation treatment combined with chemo- or BRM therapy (12, 51 ).
Importantly, a crucial feature in the induction of adverse effects of radiation combined injury is the synergy of different etiological factors, i.e., radiation combined with concurrent or subsequent secondary trauma (such as penetrating wound, blunt, blast, thermal, hemorrhage, etc.) or insults from chemical toxins and infections. For example, this has been described in detail in a variety of models employing total-body irradiation combined with a secondary insult occurring from radiation-induced aseptic inflammation, immune suppression, coagulopathy and a high susceptibility to bacterial translocation from wound and Enterobacteriaceae bacteremia (45 ). A decrease in the LD50/30 dose for gamma rays has been reported, in which animals received radiation combined with either skin wound, burn or laceration (45 ).
The available clinical data on radiation combined injury are limited and mostly originate from the cohorts of atomic bomb survivors and victims of the Chernobyl nuclear accident. Even more limited are clinical observations from these groups on radiation combined brain injury; thus, clarification of its pathogenesis based on these cohorts may be challenging. Nonetheless, presumably, the effects of high-dose radiation can include a decreased ability to recover from mechanical (e.g., wound, blunt or blast) trauma due to suppression of normal neurogenesis, oligodendrogenesis and vasculogenesis, and induction of parenchymal and neurovascular inflammation. Moreover, these conditions can increase the risk of a tertiary effect, i.e., intracranial infections. Overall, importantly, brain trauma, infection or chemotherapy/chemotoxins can exacerbate the radiation encephalopathy sequelae (and vice versa). Predictive models of these outcomes should be included in the integrative radiation systems biology methodology and employ data from translation research and combined radiation therapy and clinical data from nuclear and radiological accidents and incidents (12, 49, 51 ).
This review addresses the molecular and cellular mechanisms mediating differential effects induced by high-dose (>2 Gy) exposure of X rays and gamma-photons in the cerebral parenchymal and brain stroma that are associated with brain dysfunction. Specifically, this review focuses on the countermeasures against cerebrovascular impairment caused by ionizing radiation and/or radiation combined injury.
X-ray and Gamma-Photon Tissue Irradiation: Molecular Effects Induced by Ionization and the Mechanisms of Cell Injury
The generated “primary reactive products” of radiolysis are subjected to different types of recombination yielding either the “secondary cytotoxic” products or adducts with intact “bystander” constituents (54–56 ). Thus, the impact of a pulse radiation to a single cell can per se produce an instant effect on vital pathways that control transcriptional, translational and post-translational events, cell metabolic homeostasis, redox balance, cell-to-cell communication, growth, differentiation or aberrations (56–63 ). Moreover, it appears that living cells are equipped with numerous molecular mechanisms [e.g., redox sensing transcriptional factors, nuclear factor (erythroid-derived 2)-like 2 (Nrf2), NF-κB, activator protein 1 (AP1) and mitogen-activated protein kinase (MAPK)], to respond to different types of cytotoxic stress and/or impacts of the damage-associated molecular patterns (DAMPs), reactive oxygen species (ROS) or the reactive electrophilic species, as a part of the intrinsic defense control. Noticeably, radiation-induced damage to the cell constituents occurs in a dose-dependent manner. Thus, the extent of the subsequent cell remodeling ultimately leads to either cell restoration or a complete cell loss (e.g., cell senescence or death) (64–67 ). In this respect, considering the high energy of the penetrating radiation, the molecular damage to the cell and tissue structures from the direct radiation impact is not specific, although the consequent reactive responses to the radiation-produced injury are. In general, this specificity defines levels of radiosensitivity of cells in different tissues and systems.
As the ionization energy is delivered to the targeted molecules, the electrons are subsequently ejected and transferred to the oxidized molecules, yielding in the biological environment, including the water ion-radicals (H2O•+), an array of carbon-, oxygen-, sulfur-, and nitrogen-centered ion-radicals and the excited water molecule (H2O*) and then, the discharged water-caged electrons (e–aq), hydroxyl radical (HO•), nucleophilic hydril (H•) and other free radical species. There are numerous transient products yielded in irradiated cells which disappear within a millisecond via numerous redox reactions; namely the water ion-radical (H2O+), the hydrated ejected electrons (e–aq), hydroxyl radical (HO•), nucleophilic hydril (H), and an array of carbon-, oxygen-, sulfur-, and nitrogen-centered ion-radicals in different biomolecules (52–54, 56 ). Moreover, these transient oxidized species disappear within a millisecond via numerous redox reactions with antioxidants or other redox-susceptible biomolecules, thus propagating oxidative stress (52, 53, 68, 69 ). In particular, this is referred to as the strong electrophilic hydroxyl radical (HO•), which has a capacity to oxidize organic molecules with diffusion-dependent constant rates (69, 70 ). Biomolecule-centered radicals and the caged e– can either remain due to intramolecular recombination or react with the solvated oxygen to form numerous peroxides and superoxide anion radicals (O2•–) (52, 53, 56 ). These secondary ROS can be either promptly utilized in the Haber-Weiss/Fenton-type reactions and by these means, activate the free-radical and radical-free metabolic pathways, or proceed to quenching by antiradical/antioxidant redox mechanisms (54–57, 70–73 ). As an end point of these redox reactions, a rise in non-radical electrophiles and DAMPs in the irradiated tissues and biofluids can exacerbate cytotoxicity due to the primary radiation impact (74–79 ).
It would appear that the “secondary” ROS could potentially amplify the “primary” oxidative stress; however, theoretical calculations indicate that the yields of products, generated as a consequence of a primary ionization event and essential for “secondary” oxidative hit, are lower by orders of magnitude than those produced by normal cellular metabolism (80 ). Therefore, the concept of amplification of the radiation-induced “primary” oxidative stress has been further refocused on the shift of metabolic and pro-inflammatory redox pathways in irradiated cells (62, 63, 66, 77 , 81–85 ). Indeed, numerous observations indicate that the radiation-induced radiolysis can affect crucial cellular constituents and tilt calcium homeostasis, and thus can activate a cascade of metabolic responses leading to a prolonged oxidative stress which propagates systemically (64, 80–87 ). Moreover, there are several redox mechanisms, which have been proposed to drive the metabolic oxidative stress. Among these are activation of: 1. Constitutive and inducible nitric oxide synthases (62, 81 ); 2. NADPH oxidase (76, 83 ); 3. Monoamine oxidase (88 ); 4. Radiation-inducible microRNA miR-193a-3p (89 ); 5. Transient receptor potential (TRP) proteins (90 ); and 6. ER-mitochondrial axis (83, 91–93 ).
In this particular role, the radiation-induced imbalance of mitochondrial redox machinery is the major consumer of oxygen in aerobes as well as the major source of metabolically produced ROS in most cells (83, 91–94 ). This is especially important, considering that the mitochondrial volume represents a substantial radiation target, i.e., 4–30% cell volume depending on the cell type (95 ). Recent extensive investigations of the radiation effects on mitochondria have shown that these organelles have a superior role in response to radiation hit by triggering: 1. Short- and long-term metabolic responses (e.g., a decrease in oxidative phosphorylation); 2. Metabolic amplification of the “primary” ROS yield from radiolysis; and 3. The cytochrome C-dependent mechanism of programmed cell death (84, 91–94 ). Numerous in vitro and in vivo observations indicate that radiation effects on mitochondria have multifactorial characteristics. First, mitochondrial DNA is very susceptible to radiation damage. In response to the formation of lesions in irradiated mitochondria, expression of specific mitochondrial genes that are related to cell survival can be upregulated. They can also induce a compensatory increase in the mitochondrial DNA copy number, i.e., “mitochondrial polyploidization” (84 ). The reactive products of radiolysis such as ROS, RNS, RCS, RLS can produce post-translational modifications of mitochondrial proteins followed by functional alterations, such as the following: 1. A prolonged dysregulation of the respiratory electron transport chain (affecting complex I, complex II, complex III and succinate/pyruvate-mediated respiratory capacity) with subsequent increases in ROS production; 2. Alterations in oxidative phosphorylation; 3. Remodeling of the mitochondrial-ER network, activation of mitophagy, and changes in the mitochondrial mass; 4. Increases in intraorganelle communications and [Ca2+]-mediated propagation of oxidative stress; and 5. Increases in permeability and swelling (83, 84, 91–96 ). Altogether, these reactions may constitute the “mitochondrial ROS-induced ROS” mechanism (97 ), and thus sustain the effect of prolonged oxidative stress in irradiated tissues (83, 84, 87, 93, 94 ). Thus, the radiation-associated “metabolic” activation of ROS can further sustain and propagate the systemic electrophilic stress to form a variety of non-targeted detrimental effects on tissues and organ systems.
Oxidation of biomolecules due to radiolysis and the Fenton mechanism has been extensively investigated over the past decades. For a long time, the radiobiological effects were associated with radiation-induced clustered DNA lesions, i.e., two or more individual lesions within one or two helical turns of the DNA that occur after passage of a single radiation track through a nucleus. Mitochondrial DNAs (mtDNAs) are equally susceptible to radiation injury (98 ). Radiation can induce clustered DNA cleavage through the direct impact of the ionizing photons on the DNA as well as through the indirect action of reactive chemical species formed near the DNA due to radiolysis and oxygenation. Indirect effects are attributed to oxidative damage by ROS, primarily by hydroxyl radicals generated in radiolysis and the Fenton reactions (70 , 85 , 98, 99 ). However, the “direct” damage of DNA induced by photons occurs randomly in sugar and base moieties leading to strand breakage, release of free (unaltered) DNA bases, phosphates, and the formation of intermediate DNA free radicals and the TBA, i.e., 2-thiobarbituric acid, reactive products. The hydroxyl radicals react with the bases of DNA rather than the sugars (70 ). In these events, the main reaction is the addition of the hydroxyl radical to the πbonds of the bases. In the presence of oxygen, the resulting pyrimidine carbon-based radicals can be converted to the corresponding peroxyl radicals, as are sugar-based radicals. Hydroxyl radical adducts of purines can be further subjected to cleavage or undergo recombination. The ultimate intramolecular recombination of base-centered radicals can lead to oxidation of the base and sugar moieties and the oxidative DNA cleavage (55, 70, 71, 95–100 ).
While the targeted DNA damage and epigenetic alterations are considered to be crucial mechanisms of the radiation-induced cell death, mutagenesis and genomic instability, the emerging role of other types of biomolecules modified due to radiation is of growing interest. Thus, like nucleic acids and nucleotides, deleterious effects of radiation on proteins (or peptides and amino acids), carbohydrates and lipid polyunsaturated fatty acids (PUFA) can occur either due to their radiolysis followed by formation of RC•, ROO•, RO•, RS• and RN•, or due to oxidation by ROS, primarily by hydroxyl radicals, resulting from radiation exposure and the Fenton reactions. This free radical oxidation results in an array of reactive electrophiles, which can cause further post-translational modification proteins, activation of cell defense mechanisms, or detrimental effect on homeostatic responses (76, 85, 100, 101 ), Recently defined important electrophiles produced due to radiation exposure and peroxidation are the protein-, amino acid- and lipid-derived carbonyls (55, 67, 73, 104, 105 ). Direct protein carbonylation, i.e., post-translational modification resulting in the addition of reactive carbonyl groups to proteins, occurs as a consequence of oxidation of lysine, arginine, histidine, proline, glutamate, and threonine residues, and fragmentation products of peptide bond cleavage reactions. Protein carbonyls are reported to be detectable for a while after radiation exposure (55, 85, 102, 103 ).
The radiation-induced oxidation of lipids generates a large number of reactive intermediates, i.e., reactive lipid species (RLS). Thus, radiation-induced cleavage of PUFA or abstraction of a proton from PUFA by hydroxyl radical leads to addition of molecular oxygen to form PUFA-peroxides followed by their chemical degradation (Hock cleavage). These peroxidation reactions yield a plethora of PUFA-derived electrophilic (“soft”) carbonyls, which includes α,β-unsaturated aldehydes and ketones, such as 2-alkenals, 4-hydroxy-2-alkenals (4-HNE), 4-oxo-2-alke-nals (4-ONE), acrolein, and malondialdehyde (MDA) (85, 102 ). The molecular construction containing α,β-unsaturated carbonyl conjugated with the diene displays efficient electron-withdrawing properties when reacting with nucleophiles, such as cysteine thiol, lysine, or histidine residues, via the Michael mechanism, producing a variety of intra- and intermolecular covalent adducts. These covalent modifications result in a free carbonyl attached to the protein that appears as “secondary” carbonylation (75, 85, 102 ). In addition to carbonylation, the PUFA-derived aldehydes can react with the amine moiety of lysine residues to form the Schiff base adducts, which can further undergo intra-protein recombinations (102 ). Importantly, RLS such as MDA or 4-HNE can react with DNA and RNA (nuclear and mitochondrial) as well, in particular, with the guanine and adenine bases yielding etheno adducts (i.e., ethenobases). The produced aberrations in the nucleic acids can lead to mis-transcriptions and thus to altered transcriptional products. Moreover, by building Schiff bases with histones and other transcription regulatory proteins in the nucleus, these RLS can provoke clastogenic effects and promote epigenetic alterations. Notably, PUFA derivatives with conjugated dienes are particularly susceptible to in vivo nitration with endogenously produced RNS yielding nitroalkenes, another group of electrophilic RLS (75 ). Biological effects of α,β-unsaturated aldehydes and ketones have been well addressed in the literature recently and the multidiversified interference of these products has been associated with the cell protein machineries situated in cytosol and crucial organelles (e.g., mitochondria, ER and nuclei), and formation of immunogens, inflammagens, and DAMPs in the biofluids recognizable by PRRs on non-targeted cells (76, 77, 98, 100, 102, 106, 107 ). Based on these observations, post-translational modification of proteins with PUFA-derived carbonyls, i.e., RLS, has also been proposed as a novel signaling mechanism that modulates the cell redox-stress responses, including those mediated by Nrf2, NF-κB, heat-shock proteins 70 and 90, heat-shock factor 1 (HSF-1) and APE/EpRE (75, 100, 102, 105, 106 ).
Notably, excessive generation of RLS implicated in the Michael addition/reaction can spread carbonyl (electrophilic) stress from the targeted cells to bystander cells via the gap junction network or the extracellular vesicle mechanisms, thus producing non-targeted effects including mitochondrial dysfunction, ER stress, disturbance of calcium homeostasis, and epigenetic and clastogenic dysregulations (86, 100, 102 ). Overall, while the “radiolysis phase” is short lived, it causes devastating systemic effects, i.e., radiation-induced damage to proteins and lipids, post-translational modification of proteins, impairment of DNA, epigenetic alterations and formation of aberrant organelles generates an array of mediators of stress, danger and inflammation that interfere with cell communication systems and homeostatic control (55, 64, 87, 108, 109 ). In conjunction with these events, activation of free radical reactions, formation of ROS, RNS, RCS, RLS, the Schiff and etheno adducts, the products of protein sulfhydryl oxidation as well as depletion of antioxidants are major features of radiolysis in tissues and fluids, and therefore, are considered to be biomarkers of oxidative stress after radiation exposure.
Differential Susceptibility of Brain Structures and Systems to Gamma-Photon Irradiation and Combined Injury
Impairment of the brain neuroregulatory activity and functions as well as cognitive capacity due to radiation or radiation combined injury is the result of induced organic injury to the brain tissues, the structural disarrangement of interactive networks of neurovascular and glial cells, as well as on arrays of the paracrine and systemic destructive cues originated from reactive responses to the impacts. The response and structural and functional changes in brain cells to radiation vary among different animal species (and humans), cell phenotype, age and brain morphological area and zones (e.g., the cerebral cortex vs. the hippocampus; the lateral ventricles vs. the third ventricle) and are largely suggested to fall in a linear dose-response pattern (4, 7, 18, 29, 30, 46–48, 110–118 ). However, the scope of injury and the consequences of the impacts are determined by cell susceptibility and their spatial-temporal status in the cell interactive networks of the tissues, whether they are neuronal, glial or neurovascular (7, 23, 24, 26, 117, 119–123 ). Notably, the differentiation stage and radiation dose/dose rate can predispose the fate of affected cells. Indeed, there are some particular phenotypes in the brain that are able to proliferate, e.g., the cells of the subependymal zone, glial and vascular endothelial progenitors, which are more sensitive to radiation damage and undergo apoptosis at lower doses compared to non-proliferating (e.g., terminally differentiated and G0 stage) parenchymal and vascular cells. Proliferating cells are also prone to senescence (117, 119–122, 124–135 ).
Evidently, the pathogenesis of the radiation-induced brain disease has a multifactorial character and depends on a latency of cell responses, the dynamic of the radiation-induced structural alterations and severity of the organic injury (117, 122, 127, 128, 136, 137 ).
Moreover, a growing body of data from clinical and translational research indicate that radiation effects on the parenchyma cells and brain physiology are aggravated by the associated damage to the microvascular endothelium leading to cerebrovascular inflammation, breach of the brain blood barrier, and impairment of metabolic homeostasis (6, 14, 26, 41, 117, 118, 136 ). Therefore, structural and functional integrity, and the homeostatic and mental functions may decline gradually within weeks through years postirradiation (117, 118, 121, 138 ).
Overall, gamma photons per se, with their high ability to penetrate the cranial cavity, can, to a certain degree, directly affect all brain cell phenotypes (parenchymal and stromal), as well as interstitial fluids in all parts and areas of the brain regardless of radiation dose(s) given. However, it is widely accepted that radiation exposure below some certain single dose (as threshold) may not produce either radiation sickness or long-term adverse health effects in humans and animals, despite a presence of transient lesions in photon-targeted cells (3, 4, 29, 48, 117, 118, 125 ). Thus, the risk of brain cancer is considerable at radiation doses >0.2 Gy (110, 111, 118, 119, 139–141 ), while a single dose to human <0.02 Sv (or <0.02 Gy gamma photons) or cumulative lifetime exposures to <0.1 Sv count as “no evidence for irradiation-related diseases” (142–145 ).
The estimation of limiting doses for direct radiation effects in the CNS is based on morphological and functional deficits. Thus, in humans, radiation doses between 0.35–0.50 Gy cause transient radiation sickness, with nausea and vomiting, that can develop a few days postirradiation (4, 46, 47, 145 ). In addition, delayed neurological disorder (>30 days) induced by 0.1–1.0 Gy radiation can originate from morphological and functional changes within the neuronal and vascular networks (14, 23, 26, 117, 118 ). A severe form of delayed neuropathy, myelopathy, neuroinflammation, vascular impairment, and decline in cognitive capacity occurs after exposure to doses ranging from 1–5 Gy (i.e., LD50/60 for human total-body irradiation) followed by supportive care (14, 23, 26, 117, 118, 141 ). Furthermore, single doses of radiation over 20 Gy produces the acute cerebrovascular syndrome characterized by very short prodromal and latent phases followed by neurological problems, including headache, seizures, cerebral edema, abnormal cognition, neurological deficits and loss of consciousness, as well as death (4, 26, 46, 47, 146, 147 ). Given that radiation oncology treatments of the head and neck employ high doses (i.e., cumulative dose ∼60 Gy or more with 2–6 Gy per fraction), the brain tissue is regarded as a dose-limiting organ, with considerable concerns about radiation-induced vascular injury (1, 2, 25, 26, 28, 112, 118–122, 148–151 ). Most importantly, the brain microvasculature represents the largest vascular network in humans and animals, and up to 20% of the total oxygen consumed in the body is provided to the brain (152–154 ).
Based on the clinical sequelae, radiation-induced brain injury can be characterized as protracted acute, early delayed and late delayed injury (25, 26, 118, 155–157 ). Protracted acute cerebrovascular injury develops in hours to days depending on radiation dose to the cranium (147, 157 ). Early delayed brain injury occurs 1–4 months postirradiation and can involve alterations in the architecture of neuronal network, cerebral edema, transient micro-symptoms of organic damage to the nervous system and demyelination followed by perturbations in the functional activity of the nervous system (118, 121, 157 ). Although both of these early injuries can result in severe reactions, they are considered to be resolvable.
In contrast, late delayed brain injury is characterized histopathologically by gray matter organic damage, ataxia, vascular abnormalities, demyelination and, ultimately, white matter necrosis, which usually begins to occur 4–6 months postirradiation and can develop over years thereafter (118 ). The late delayed injuries have been considered irreversible and progressive (13, 14, 16, 22, 25, 26, 112, 114, 116, 118, 120, 122, 155–158 ). Thus, radiation-induced volumetric changes related to cortical atrophy (i.e., temporal and limbic cortex) have been reported in these patients a year after radiotherapy. This volumetric effect occurs in a dose-dependent manner at a dose range of 1–60 Gy with greater effects at higher doses (159 ). It has also been found that doses above 28.6 Gy resulted in a greater than 20% probability of cortical atrophy, and the estimated decrease in cortical thickness is –0.0033 mm (P < 0.001) for every 1 Gy increase in the dose (159 ).
The explanation for this pathology has been exclusively based on alterations in mitotic activity, impairment of the proliferating capacity and apoptotic responses of either glial cells (e.g., astrocytes and oligodendrocytes in the forebrain subependyma) or vascular endothelial cells (6, 22, 112, 114, 122, 129, 141–143, 158, 160, 161 ). Evidently, irradiation of the subependymal zone can also dramatically affect a population of the adult neural stem cells that sustain brain tissue remodeling and cognitive functions (7, 23, 39, 116, 117, 127, 149, 160 ). Moreover, the radiation parenchymal effects can be exacerbated by the radiation-induced profound loss of capillary density within the hippocampus region leading to perfusion impairment (121, 161 ). Overall, the extensive damage to the parenchymal and/or to the stromal cells as well as a progressive loss of their progenitors can lead to deterioration by neuroinflammation along with breach of the blood-brain barrier (BBB) that can ultimately lead to white matter lesions (5, 7, 25, 26, 115, 116, 128, 144, 158, 162 ). With the above considerations, there are two conceptually different hypotheses proposed to explain the mechanisms underlying the radiation-induced temporal and structural alterations of brain tissues and the radiation brain disease. One of these is based on parenchyma-associated pathogenic mechanisms, and the other is focused on the vascular etiology of the disease (5, 6, 39, 116, 118, 120–122, 138, 162, 163 ).
The parenchymal hypothesis of radiation-induced brain injury is focused on the neural progenitor cells (NPCs) and oligodendrocyte precursors (OP) that are required for adult neurogenesis and oligodendrogenesis that occur in the specific microenvironments, i.e., the angiogenic and astroglial niches of the subventricular zone and the subependymal zone at the lateral ventricles as well as in the dentate gyrus at the hippocampus (Fig. 1). Thus, endothelial cells and astrocytes situated in the niches control, promote and mediate NPC renewal and/or further differentiation to neurons and oligodendrocytes (5–7, 9, 13–16, 39, 114–117, 122, 124, 144, 158, 160, 162–164 ). Cranial irradiation is shown to induce: 1. Apoptosis in the subependymal zone and the dentate gyrus; and 2. A dose-dependent loss of the neuronal stem cells as well as suppression of proliferation and differentiation of the surviving NPCs. Moreover, radiation toxicity affects reproductive capacity of oligodendrocyte type-2 astrocyte (O-2A) progenitor cells essential for the oligodendrocyte maturation (10, 13, 14, 15, 21, 22, 39, 122, 130, 156, 162–164 ).
The vascular hypothesis is supported by a large body of data that has described the radiation-induced vascular structural changes, including vessel dilation, vasculopathy and depletion of endothelial progenitors, appearance of senescent and aberrant endothelial cell, as well as activation of endothelial apoptosis and vascular inflammation (7, 120, 121, 128, 136, 154 ). Quantitative studies of irradiated animal brains have also demonstrated time- and dose-dependent reductions in the number of endothelial cell nuclei and blood vessel lengths (112, 119, 121, 161 ).
The vascular tissue displays a biphasic mode in response to radiation. The first acute phase can occur within 24 h postirradiation where the underlying mechanisms are associated with the radiation-induced apoptosis of endothelial cells (133 ). The second late phase of radiation effects on vasculature beds requires month(s) for development and features capillary collapse, thickening of basement membrane and a loss of the endothelial clonogenic activity (133 ). Notably, the radiosensitivity of vascular tissues can vary based on the vessel caliber, i.e., macrovessels >300 microns, microvessels <300 microns, arterioles and venules >10 microns and capillaries <10 microns (137 ). Evidently, capillaries are the most radiosensitive components of the vasculature (165 ). Thus, Dimitrievich et al. (166, 167 ) have demonstrated that sensitivity of the capillaries to single X-ray doses ranging from 200 to 2,000 rad has been significantly greater than that of larger vessels. They have also shown that the major features of radiation injury have been represented by capillary disruption, and extravasation of blood components as well as proinflammatory alterations (166 ). There is reported evidence that the high vulnerability of capillary beds to radiation is due specifically to high radiation sensitivity of their endothelium, which is the major structural constituent of capillary walls (136, 168, 169 ). Referring to the radiation-induced brain disease, the microvasculature represents the principal histo-hematic interface in the CNS that sustains and controls the nervous tissue homeostasis (10 ). Therefore, the radiation-induced damage to the normal capillary endothelium can result in a breach of the brain-blood barrier leading to severe health outcomes.
Overall, while it is widely accepted that radiation does not produce pathognomonic morphologic features, that is, radiation-induced alterations may occur as a result of injuries by other factors, nonetheless, it is evident that microvascular injury may drive some unique self-sustaining mechanisms of chronic radiation diseases (6, 7, 119, 137 ). Neurological disorders resulting from brain irradiation are listed in Table 1.
Endothelial Cell Response to Radiation Exposure and the Vascular Patterns of CNS Effects after Radiotherapy and Radiation Combined Injury
Microvascular Endothelium in Integrity of Brain Vascular Barrier. The endothelium of the microvascular beds comprises specific cell phenotypes and represents a part of (BBB) that tightly controls the brain immunochemical homeostasis (10 ). The BBB is a highly organized multicellular gate-keeping structure, which sustains the brain tissue immunochemical homeostasis by regulating the molecular trans-endothelial transport between the parenchyma and the systemic circulation and by restricting translocation of the peripheral immune cells. Importantly, there is a lack or low level of constitutive expression of major histocompatibility I molecules on neurons and oligodendrocytes of the adult brain. The limited permeability of the CNS microvascular endothelium is mostly attributed to its intrinsic low pinocytic activity and a high level of efflux transporters. This BBB-specific cell layer is tightly bound by the tight junction (TJ) and the adherent junction (AJ) molecules that are situated in the intercellular space between the adjacent endothelial cells.
There is a large body of evidence to indicate that the vascular cellular components along with the parenchymal components constitute structural/functional modules, named as the multicellular neurovascular units (NVUs). NVUs can be defined as complex functional and anatomical structures composed of endothelial cells with their TJs/AJs, a basal lamina covered with pericytes and smooth muscular and parenchymal cells, including astrocytes, neurons, interneurons and adjacent perivascular microglia (144, 178, 179 ). Moreover, the vascular endothelial monolayer is embedded in a complex meshwork of interacting proteins, glycoproteins, proteoglycans, glycolipids and extracellular vesicles that constitute glycocalyx and extracellular matrix. Thus, the NVU architecture depicts the influence of parenchymal, mural, extracellular components and paracrine factors in the unique function of the brain microvascular endothelium (10, 24, 178 ). Moreover, NVUs can represent functional platforms for integrating responses of pro- and anti-inflammatory pathways under normal and pathological conditions (41 ).
Radiation Induces Endothelial Remodeling and Changes Endothelial Function. Evidently, the vascular endothelial cells in the NVU construct are vulnerable to direct radiation impacts, indirect bystander effects and the secondary inflammatory factors (6, 9, 16, 79, 121, 133, 144, 151 ). This makes the microvascular tissue extremely susceptible to radiation injury. Indeed, radiation-induced biochemical alterations and stressogenic stimuli can drive morphological and functional alterations in endothelial cells, namely expression of pro-inflammatory phenotypes, remodeling of tight junctions and NVU interactions, increases in permeability of endothelial lining of microvasculature, cell death and detachment from the basement membrane (136, 144, 180, 181 ). These vascular effects are often exacerbated by neuroinflammatory responses associated with radiation, such as decreases in expression of protein kinase B (Akt) and anti-inflammatory cytokines, and increases in expression of pro-inflammatory mitogen-activated protein kinases (MAPK) (182 ). Ultimately, this microvascular remodeling can result in a BBB breach and an intraparenchymal hemorrhage.
There is some evidence that high-dose (>2 Gy) radiation is associated with development of aberrant endothelial cells, whereby the affected cells can either transfer from the quiescent state to a pro-inflammatory phenotype or proceed to mitotic or apoptotic cell death, depending on the delivered radiation dose (11, 76, 120, 121, 128, 151, 181, 183, 184 ).
The endothelial pro-inflammatory phenotypes are characterized by the expression of cytokines, chemokines, and adhesion molecules that facilitate the recruitment and homing immune cells to sites of infection or tissue injury (9, 11, 118, 121, 142, 144 ). Naturally, endothelial cell activation is a normal part of the vascular defense mechanisms. In the event of radiation exposure, there are several pathways, including DAMP-activating pathways, which can trigger pro-inflammatory endothelial response (185, 186 ). Evidently, DAMP implicates signaling cascades activated via toll-like receptors, purinergic receptors and inflammasomes. Thus, exposure of endothelial cells to the radiation-produced DAMP, dual-function alarmins, such as HMGB1, and exosomes, can upregulate pro-inflammatory responses mediated by NF-κB, MAPK and interferon regulatory factor 3 (IRF3) (187, 188 ). Ultimately, these reactions can result in expression of several adhesion molecules such as intercellular adhesion molecule (ICAM)-1, vascular cell adhesion molecule (VCAM)-1, and E-selectin (184 ).
Apoptosis of microvascular endothelial cells can occur within 24 h postirradiation to >5 Gy of gamma photons or X rays in a dose-dependent manner (123, 181, 183, 184, 189 ). The data accumulated on this topic over the past 30 years indicate that endothelial apoptosis drives the early microvascular radiation toxicity (190 ). These acute vascular effects in different tissues, including brain, can appear within days to weeks postirradiation. Chronic effects are associated with the endothelial cell senescence in cerebral vascular beds and these transformations may occur within months to years (137 ). Thus, the applied cumulative radiation dose, implemented fraction size, intrinsic nature of the vascular bed and specificity of brain areas subjected to irradiation, together can determine radiation outcomes. Moreover, another factor that defines the endothelial response to ionizing radiation is the degree of differentiation of endothelial cells. Notably, endothelial progenitor cells (EPCs) are not only attributed to embryonic tissues, but the pool of peripheral blood EPCs originates from bone marrow and resident sources of many other tissues. It appears that this reservoir of EPCs can actively contribute to vascular remodeling, whereas the radiation-induced suppression of proliferation and ablation of EPCs causes vascular dysfunction (10, 191 ).
In addition to the above, there is a large body of published research with a focus on elucidating the molecular pathways leading to endothelial apoptosis. In this respect, of particular interest are the intrinsic mechanisms activated by radiation-induced damage to DNA, mitochondria and plasma membrane. The apoptosis mechanism activated by DNA damage implicates the p53-dependent pathway in which p53 activates transcription of the pro-apoptotic BH3-only proteins PUMA, NOXA or Bax, while the “mitochondrial damage associated” mechanism is mediated by the subsequently released cytochrome C (62, 132, 134 ).
Moreover, much of the literature based on in vitro and in vivo research have also suggested that the radiation-induced endothelial cell apoptosis is largely mediated by the lipid second messenger ceramide upon activation of acid/neutral sphingomyelinases (ASMase/NSMase) hydrolyzing sphingomyelin and releasing ceramide (133, 136, 192 ). One crucial target of ceramide is the RAC1/MEKK1 pathway, which interacts with the protein kinase MAPK8 regulating apoptosis through effector caspases, i.e., caspase-1, caspase-3 and caspase-6, as well as the autocrine stimulation of the death receptor pathway. Interestingly, this protein kinase is also implicated in the extrinsic apoptotic mechanism activated by external cues such as TNF-α (193 ).
The Radiation-Induced Endothelial Damage and Microvascular Impairment. As discussed above, the recent concepts of cellular mechanisms of the radiation-induced impairment of brain microvasculature regard the vascular endothelium as one of the main targets in radiation exposure. Thus, a high-dose radiation exposure or repeated fractional radiotherapy can exceed the adaptive physiological response and intrinsic resilience of the endothelium, thereby leading to endothelial dysfunction. This pathological condition can first appear in the form of insufficient responses to paracrine/endocrine/physiological stimuli and a failure of the endothelium to perform its normal, physiologic functions. Then, the radiation-induced massive formation of aberrant endothelial cells would lead to reduction of the cell density in the microvasculature that culminates in deterioration of the vascular tone, and vascular inflammation and declining integrity of BBB. This string of events, along with evident development of coagulopathy associated with declining the levels of platelets in the peripheral blood, often culminates in the parenchymal hemorrhage as a part of the sequelae of the acute disease (47, 194 ).
It is widely accepted that upon nuclear/radiological accidents or acts of nuclear detonation, the ionizing radiation would, with high-percentage estimates, be confounded by physical trauma (including burns) or exposure to toxic chemicals; this could also occur with infection with endemic, environmental or weaponized pathogens. A combination of these factors results in combined injury, which is more severe compared to due exposure to the same radiation dose/dose rate alone (45 ). Although the mechanisms of this synergistic interaction are not clearly understood, evidently a combination of factors can, while damaging/deranging biological barriers, also synergistically induce immune suppression and upregulate cascades of the systemic and local reactive responses in the injured parenchymal and vascular tissues. Ultimately, these alterations increase bacterial translocation, susceptibility to sepsis and the multiple organ failure outcomes (45 ).
Since Lawrence et al. reported Enterobacteriaceae bacteremia and subsequent sepsis as major factors of animal mortality after irradiation, the crucial role of bacterial breach, translocation and sepsis in radiation sequelae has been broadly documented in clinical observations and a variety of animal models of radiation combined injury (45, 195 ).
It is worth noting that the development of immunosuppression in irradiated animals can increase susceptibility to bacterial inflammagens and septicemia in order of magnitude compared to nonirradiated controls (45, 195 ). This phenomenon also suggests implication of the “secondary” septic responses in vasculature and parenchymal tissues sensitized by “the primary” radiation exposure. Moreover, these secondary responses can align with coagulopathy and the generalized hemorrhagic Shwartzman-like reaction, which leads to brain hemorrhage documented in animal models in the early delayed phase postirradiation (194 ).
Taken together, based on the information above, it is reasonable to suggest that acute and/or delayed encephalopathy due to nuclear accidents (196 ) can be aggravated by the infection-induced impairment of endothelium in the cerebral vasculature. Implication of neurovascular pattern in the encephalopathy due to combined impacts would be even more evident when such combination constitutes radiation and acute or sub-acute traumatic brain injuries characterized by mechanical damages to vasculature and intracranial hemorrhage.
Overall, a decline in brain function due to radiation combined injury is a challenge to civilian and military physicians. Development of new therapeutic modalities for managing the associated neurovascular injury could be a crucial step to address this challenge.
Perspectives on Mitigation of Microvascular Injury due to Radiation Therapy and Radiation Combined Impacts
Over time, several approaches and strategies have been considered for reducing radiation effects in the endothelium and brain cells, which are based on recruiting intrinsic resources for repair/resistance capacity and on the reduction/modulation of responses that drive pro-apoptotic pathways or senescence (12 , 65 , 133, 136, 197 ). These strategies include: Increase in capacities of the EPC-clonogen-generating tissues and the growth factor-producing cells; 2. Development of recombinant growth factors (e.g., VEGF); and 3. Molecular interference approaches using small molecules such as antioxidants, transcriptional modulators, inhibitors of the central renin-angiotensin system, inhibitors/modulators of the protein kinase pathways, among others. There is a large body of recently-published literature on these topics (12, 197–199 ).
In this respect, remarkable advances have been made based on investigations into the radioprotective effects of the basic fibroblast growth factor (bFGF) on the endothelium (133, 200 ). Here, it was suggested that the radioprotection resulted from suppression of the ceramide-related apoptosis (discussed above) where the PKC system operates as a bFGF effector (200 ). Moreover, implementation of the emerging cell therapy techniques for reconstitution of damaged microvasculature using donor EPCs and mesenchymal stromal cells presents a new opportunity for managing cerebrovascular impairment caused by radiation, trauma, sepsis or combination thereof (201–204 ).
A growing body of translational data suggests that the ghrelin peptide, an endogenous ligand of the growth hormone secretagogue receptor, can ameliorate vascular impairment under different pathological conditions (205–207 ). Ghrelin originally was reported to induce growth hormone release through pituitary GHSR-1a stimulation. However, recently published studies have indicated multiple paracrine, autocrine and endocrine roles of ghrelin, reflecting the ubiquitous expression of GHSR-1α in a variety of tissues and organs. Thus, in addition to the established effects on food intake and growth hormone release, ghrelin has emerged as a potent immunoregulatory and anti-inflammatory agent. Moreover, ghrelin can also ameliorate neuronal apoptotic transformations in models of brain injury, making it an effective intrinsic neuroprotector (206, 208, 209 ). There is evidence to suggest that ghrelin can inhibit neuronal and endothelial apoptosis by activating the extracellular-signaling-regulated-kinase (ERK)1/2, mitogen-activated protein kinase (MAPK) (180, 210 ), protein kinase A (PKA) and protein kinase C pathways (PKC). The activation of these pathways is associated with reduced activation of BAX, an improved Bcl2/BAX ratio and suppression of apoptosis/improved cell survival (210, 211 ).
There are promising published studies that have shown the pro-survival effects of ghrelin in rodent models of radiation injury and the combined radiation injury (radiation accompanied by sepsis or trauma) (180, 192 ). Using a mouse model of acute radiation injury and combined radiation injury (skin trauma), Kiang et al. has previously reported on the mitigating effects of ghrelin in hematopoietic tissue (180 ) and cerebral vasculature (180 ).
Radiation and radiation combined injury induce severe brain injury in a dose- and trauma score-dependent manner and, in part, are associated with neurovascular impairment. Whether the combined injury was the result of nuclear industrial accidents, detonation of nuclear/radiological devices or radiation therapy, it can nevertheless lead to devastating human health outcomes including severe brain injury. Research and development of new therapeutic modalities for radiation- and radiation combined brain injury should include the integrative radiation systems biology methodology, and should employ analyses of big data obtained from translational research, combined radiotherapy (169 ) and clinical data from nuclear and radiological accidents and incidents.
The writing of this review was supported by AFRRI RBB34363, RAB 33529, RAB310934, NIH/NIAID IAA YI-AI-5045-04 and JPC-7 VP000276-01 to JGK. This review has been cleared and approved by AFRRI and USUHS leadership management. The views, opinions and findings contained in this report are those of the authors and do not reflect official policy or positions of the Armed Forces Radiobiology Research Institute, the Uniformed Services University of the Health Sciences, the Department of Defense, or the U.S. government. The commercial products identified in this article do not necessarily imply recommendation or endorsement by the federal government and do not imply that the products identified are necessarily the best available for the purpose.
1 Previously employed at Armed Forces Radiobiology Research Institute, Uniformed Services University, Bethesda, MD.