Space Radiation: Central Nervous System Risks
The central nervous system which consists of the brain, spinal cord, and retina.
The energy absorbed per unit mass in units of gray (Gy) where 1 Gy = 1 J/kg.
A region of the brain in the temporal lobe associated with learning and memory. It has a well-defined layered anatomy that facilitates electrophysiological studies and is associated with numerous animal behavioral tests.
High charge (= atomic number, Z) energetic particles.
- LET or linear energy transfer
The energy deposited per unit length of a charged particle’s track and is proportional to Z2/v2 where Z is the charge and v is the velocity.
Long-term potentiation, a long-term elevation of synaptic strength in response to stimulation, which represents a tissue-level analog to memory trace formation.
- MeV or million electron volts
The energy acquired by an electron moving across a voltage of 106 V and, when divided by the mass of a particle in atomic mass units (amu, u or n), is the standard energy unit reported for charged particles. Particles of the same energy/n have the same velocity.
The linear pattern of energy deposition events and resultant reactive chemical species resulting from the passage of a charged particle through matter.
An emerging human health concern for space travel is the potential for deleterious effects of space radiation on the central nervous system (CNS). This might be manifest as changes in cognition, memory, vigilance or attention, reaction time, processing speed, motor function, mood and emotional control, and social interactions, all of which contribute to normal performance and behavior. Traditionally radiation biology has treated the brain as a radiation-resistant tissue. This is based largely on the fact that neurons, the principal functional units in the brain, generally do not divide in adults; therefore, replication of DNA and the possibility of mutation and mitosis-associated cell death are minimized. Also, clinical observations from radiation therapy of the head, neck, and spine show that gross tissue damage, based on traditional pathology methods, requires very large doses (e.g., >40 Gy) and is not manifested for many months after exposure (Tofilon and Fike 2000). For comparison, the cumulative background radiation dose per year at the surface of the earth (neglecting medical exposure) is about 0.004 Gy. However, over the last two decades, evidence has accumulated at an accelerating pace which shows that statistically significant structural and functional changes occur in experimental animals exposed to radiation doses similar to those expected for long-duration space missions (e.g., about 1 Gy received over a 2–3-year Mars mission). This is especially true if the radiation consists of high-energy charged particles (HZE particles) such as those that comprise galactic cosmic rays and solar particle events (Durante and Cucinotta 2011).
Estimating the radiation risks for humans based on animal models is problematic for a number of reasons. First, there is very little epidemiological data for humans exposed chronically to radiation, and there is essentially no data for charged particles. Second, the structure of the human brain is different than that of rodent models in terms of size, neuronal pathways, and cell composition. Third, human populations are genetically outbred with much genetic diversity, whereas most laboratory animals are inbred and may exhibit limited or exaggerated responses. Therefore, interindividual differences may increase the range and variability of human responses. Fourth, the behavioral repertoire of humans far exceeds that of experimental animals, and they may possess more adaptive or compensatory capacity to deal with damage. Finally, there are no accepted standards for what defines a “significant impairment,” but the high-performance levels necessary for spaceflight operations may be susceptible to radiation damage at doses well below those that rise to clinical significance, such as the “mild cognitive impairments” of early stage neurodegenerative diseases such as Alzheimer’s or Parkinson’s disease. With these caveats in mind, we will discuss what effects on the CNS have been observed with low doses of radiation, with emphasis on space-like charged particles.
The CNS consists of the brain, spinal cord, and retina and is made up of neurons organized into layers and columns, glial cells, and vasculature (Kandel et al. 2000). The cerebral cortex comprises the largest part of the human brain and contains about 20 billion neurons sustained by 100–200 billion glial cells. Neurons display a variety of sizes, shapes, and connectivity patterns and employ a variety of chemical neurotransmitters and receptors. Each neuron is organized into a cell body, a dendritic tree, and an axon. The dendritic tree and cell body receive signals from other neurons, while the axon is the transmission structure. Dendrites are highly branched processes covered with small projections or spines, the sites of most synapses – the specialized connections between neurons. Axons are thin fibers that extend from the cell bodies for long distances and branch at their termini where swellings contain synapses. Synapses are 1 μm-sized structures with a presynaptic component (from an axon) and a postsynaptic component (from a dendritic spine) separated by a thin space. Neurotransmission is the process by which an electrochemical signal (action potential) is transferred across the synaptic space by a chemical messenger that initiates electrochemical signals in the recipient cell. It involves voltage changes, ion movements, neurotransmitter release, and neurotransmitter binding to specific receptors on the postsynaptic membrane. Neurotransmitters include acetylcholine, glutamic acid, γ-aminobutyric acid, norepinephrine, serotonin, dopamine, and other substances. The glia (astrocytes, oligodendroglia, and microglia) are supporting cells which provide a scaffold, cooperate with the vasculature, regulate extracellular concentrations of neurotransmitters, and mediate inflammatory responses (Kandel et al. 2000).
A little over 15 years ago, it was recognized that there are proliferating populations of cells in the adult brain that can differentiate into neurons even into old age. The linings of the ventricles and the dentate gyrus (DG) region of the hippocampus, a structure associated with learning and memory, contain stemlike cells (neural precursor cells or NPC) that can proliferate and differentiate into neurons as well as astrocytes and oligodendrocytes – this process is neurogenesis. Newly born neurons can preferentially incorporate into neuronal circuits that are highly interconnected and likely contribute to memory formation. It was found that NPCs are the most radiation-sensitive cells in the brain with respect to survival and undergo cell death after doses as low as 0.5 Gy; surviving cells may exhibit altered differentiation patterns below 0.25 Gy. Importantly, inhibition of neurogenesis is associated with cognitive impairment in animals. The metabolic status of the tissue regulates neurogenesis, and conditions of oxidative stress impair neurogenesis. Doses less than 0.1 Gy of charged particles can elicit rapid onset of oxidative stress (4 h) in cultured NPCs which can persist for many weeks. In genetically modified animals with decreased or enhanced antioxidant capacity, radiation-induced impairment of neurogenesis is accentuated or blocked, respectively, suggesting that diets or pharmaceuticals that act as antioxidants may prove to be effective mitigators or countermeasures for radiation-induced damage. Other conditions such as emotional depression and social stress can also inhibit neurogenesis, while environmental enrichment and voluntary exercise can enhance it. The magnitude of the contribution of neurogenesis to overall cognition in humans is still unresolved.
Levels of synaptic proteins such as synaptophysin and PSD95 also show alterations. High-dose experiments suggest that this may occur within less than an hour. In cultured mouse neurons, molecular signaling pathways involving Rho family GTPases (known to regulate synaptic cytoskeletal proteins such as cofilin) are modified within 4–24 h after 1 Gy of gamma ray exposure (Kempf et al. 2014) which might provide a molecular explanation for the remodeling. Alterations in microglial cell activities might also play a role, as microglia are known to regulate spine numbers on an ongoing basis. The implications of these observations are that the number of synapses and the complexity of neuron interconnection will be reduced after space-like doses of radiation which may lead to impaired computational activities in neuronal circuits.
Accompanying changes in neuronal structure are long-term alterations in microvessel numbers and connectivity such that 0.5–1 Gy exposures can reduce hippocampal capillary numbers and vessel length densities by >30% which is a level expected to produce local hypoxia and impaired tissue function. These changes are progressive for 9–12 months with recovery at later times (Mao et al. 2010). There is evidence that capillary barrier function can be compromised by low doses (0.1–0.75 Gy) of iron ions which could compromise the isolation of the brain from the peripheral immune system by the blood-brain barrier. Evidence for widespread tissue structure changes without necrotic changes comes from magnetic resonance imaging (Huang et al. 2010).
Electrophysiological experiments with low doses of charged particles have examined neuronal functional responses from weeks to months after irradiation and have found that both intrinsic properties of cells and their synaptic connections change. The main models for these measurements are acute brain slices prepared from mouse and rat cortex and hippocampus. In these preparations, freshly isolated 300–400-μm-thick slices of tissue from irradiated animals are kept in well-oxygenated artificial cerebrospinal fluid which preserves living neuronal networks for up to 0.5 day. Pairs of stimulating and recording electrodes or microelectrode arrays positioned at the slice surface then record from groups of several hundred neurons (field recordings), or, alternatively, single neurons are targeted with microelectrodes (patch clamp recordings). Batteries of stimulation-recording sequences measure voltage and current changes associated with ion movements in and out of the cells (usually Na+, K+, Ca++, and Cl−) controlled by specific ion channels whose open/closed state depends on the binding of neurotransmitters (receptor-gated ion channels).
Evidence suggests that both pre- and postsynaptic remodelings occur. Taken together electrophysiology measurements point to impairment of neuronal excitability as well as synaptic remodeling and stable strengthening which could lead to impaired information processing and altered behaviors. Other experiments find that effects on electrophysiological parameters are radiation species dependent with complex relationships to LET, and responses from different ions at equal doses can even be opposing. Finally, behavioral training and peripheral immune function can influence the effects of radiation exposure.
A variety of biochemical measurements have shown that the metabolic status of the brain changes after irradiation. Multiple observations indicate that there is a persistent elevation of oxidative stress with alterations in enzymes such as superoxide dismutases, catalase, and glutathione peroxidases. There is an elevation of pro-inflammatory cytokines such as tissue necrosis factor-α, interleukin 1β, and interleukin 6 (Tofilon and Fike 2000; Morganti et al. 2014). This is accompanied by increases in levels in microglia and astrocyte activation markers (e.g., CD68, GFAP). Infiltration of the brain by peripheral monocytes and T lymphocytes occurs at higher doses, and there can be a decrease in microvascular adhesion molecules such as ICAM-1 (Sweet et al. 2014). Glutamate-gated ion channel levels are altered, while GABA-gated channels remain stable (Machida et al. 2010; Marty et al. 2014). Acetylcholine and dopamine pathway enzymes and metabolites show alterations. Transcription profiling and proteomic analysis indicate changes in a multitude of genes associated with neurotrophic functions, cell survival, synaptic composition, DNA repair, regulation of ion channels, and behavioral activation (e.g., Kempf et al. 2014; Lowe et al. 2009). These changes have been observed for gamma rays and a number of ions at doses well below 1 Gy and are dependent on dose, dose rate, and radiation species. Thus radiation exposure is implicated in a wide number of brain tissue functions and regulatory networks that collectively will determine its functional status.
In addition to radiation effects that may manifest at time intervals corresponding to space missions (acute effects), there may be radiation-induced changes that can exacerbate or accelerate the development of late neurodegenerative conditions such as Alzheimer’s or Parkinson’s disease. Several investigators have used genetically modified mouse strains expressing mutant human genes associated with disease processes to test this notion. Thus, in mice expressing human amyloid precursor protein, Alzheimer-like pathology develops with time and may include amyloid and tau protein deposits, decrements in electrophysiological parameters, and cognitive decline in excess of normal aging. Cherry et al. (2012) demonstrated that iron ion exposures of 0.1 or 1 Gy led to cognitive decline (novel object recognition and contextual fear conditioning, see below) in the APP/PS1 mouse model after 6 months. Male but not female animals showed an accelerated accumulation of amyloid deposits in the cortex. Vlkolinsky et al. (2010), using the APP23 mouse, found early onset of synaptic transmission defects 9 months after iron ion exposure. This suggests that radiation may initiate or promote conditions leading to pathological events related to Alzheimer’s disease.
A wide variety of behavioral tests have been employed with mice and rats to interrogate the effects of low-dose radiation on the CNS. Behavior represents the output of the CNS in response to environmental stimuli and integrates all of the interactions of cells, their structure, connectivity, microenvironment, metabolic status, and status of molecular pathways. Because information processing is distributed between many regions of the brain, it is difficult to assign behaviors to particular anatomical sites, but ablation studies have allowed mapping of regions required (but not sufficient) for the expression of behaviors. Thus, crudely, the frontal cortex is associated with attentional status and executive functions, the hippocampus is implicated in memory and learning, the amygdala is associated with fear and anxiety, and the cerebellum is associated with motor control. Behavioral effects are difficult to quantify and, in addition to the test treatment, are dependent on animal species, genotype, sex, age of exposure, specific test used, time of analysis post treatment, and any sources of stress. Despite these potential sources of interference, whole body or head-only irradiation reliably elicits quantifiable behavioral changes in rodents at doses ≥0.25 Gy which may appear acutely or develop over many months. The most sensitive tests reportedly are capable of detecting statistically significant effects at doses below 0.05 Gy. Many different types of radiation have been investigated including x-rays and gamma rays, neutrons, and charged particles of Z = 1, 6, 8, 14, 20, and 26 over a broad range of energies similar to those comprising the space environment.
Here increased time to find the escape hole (escape latency) indicates poorer performance. The surprising result found by the investigators was that as much as 40-fold higher doses of x-rays were required to impair performance (≥8 Gy) illustrating the difference between more effective structured versus less effective unstructured energy deposition.
Another test which combines spatial memory and fear motivation is contextual fear conditioning in which animals are trained to expect a foot shock after a sound cue when placed in a particular visual and olfactory environment. After training, when the sound cue is delivered in either the same or a different test chamber, the “freezing” fear response assesses the association of the environment (context) with the cue. Raber and collaborators have demonstrated decrements in contextual freezing after ≥0.5 Gy of iron ions and a peak effectiveness of 0.25 Gy with silicon ions (Raber et al. 2014). They have also used this test to demonstrate dependence of radiation effects on mouse strain, sex, age, and an association with behaviorally induced Arc gene expression. Contextual freezing is dependent on the hippocampus, amygdala, and frontal cortex.
Novel object recognition (NOR) or novel object location (NOL) tests measure the relative time an animal spends exploring a new object placed in a test arena, or an old object placed in a new location, after learning the identities and positions of two identical objects in fixed locations. The tests require association of the object identities with distant visual cues and involve the hippocampus and perirhinal cortex. The tests have detected impairments at doses as low as 0.1 Gy in many independent laboratories. The lowest reported effective doses were by Parihar et al. (2015) using 0.05–0.3 Gy of accelerated oxygen and titanium particles.
Two complex tasks of executive functioning in rats have been used by Britten et al. (2014) and Davis et al. (2014). These are the attentional set shift and psychomotor vigilance tests, respectively, which measure abilities of animals to maintain attention on a task, their reaction times, and impulsivity. These tests demonstrated impairments at doses ≥0.15 or ≥0.25 Gy. Importantly the experiments demonstrated that animals naturally sorted into high- and low-performing groups and high- and low-sensitivity groups. This is important as astronauts are high-performing individuals and tests based on population averages may miss important behavioral changes due to interindividual differences.
Human and Nonhuman Primate Data
Controlled experiments with nonhuman primates would be valuable in extrapolating results from rodent models to animals with brain structure and function more closely resembling humans. Unfortunately there is a paucity of such data for low-dose radiation exposures. Many high-dose tests with macaques and chimpanzees were conducted in the 1950s and 1960s to understand effects of radiation in the context of nuclear weapons (Mickley et al. 1989). Behavioral impairments were seen in these test subjects but the confounding health effects limit their value. Recent experiments with rhesus macaques designed to understand side effects of radiotherapy have shown behavioral impairments after cumulative fractionated head-only irradiation doses of ≈40 Gy (Robbins et al. 2011). However, none of the tests used charged particles.
Human exposures to radiation rely in large part on observations of atomic bomb victims or radiation accidents and are largely limited to gamma ray exposures. No increases in dementia associated with A-bomb radiation exposure to adults have been detected (Yamada et al. 2009), but children exposed in utero have exhibited mental retardation. Armstrong et al. (2013) have determined that adult survivors of childhood irradiation for acute lymphocytic leukemia exhibit impairments in cognitive processing speed, memory, attention, and learning. However, developing brains with massive cell proliferation do not reflect the condition of adults. Workers at the Chernobyl power plant accident have been reported to exhibit elevated incidence of psychiatric disorders and altered electroencephalograph patterns (Loganovsky and Yuryev 2001).
While these observations indicate a susceptibility of humans to radiation exposure for CNS function, they are mostly uncontrolled exposures with confounding environmental stresses. Modeling of pathophysiological processes common to humans and experimental animals and impacted by radiation will be needed to extrapolate experimental results to humans. Carefully conducted, humane experiments with nonhuman primates exposed to low doses of charged particles would also facilitate understanding the results from rodent models in the context of space-like radiation exposures in humans.
Doses of radiation at or below 1 Gy, commensurate with exposures expected on long-duration space missions outside LEO, have been demonstrated to cause a number of structural and functional changes in mammalian central nervous systems. The structural changes include inhibition of neurogenesis, reduction of dendrite complexity and dendritic spine numbers, reversible reductions in the number of microvessel segments and endothelial cells, and widespread altered contrast features detectable by magnetic resonance imaging. Inflammatory responses and persistent oxidative stress alter the tissue microenvironment and may present nonpermissive conditions for many processes. Pro-inflammatory cytokine levels are elevated, while enzymes involved in free radical detoxification and DNA repair show transient increases. Electrochemical properties of neuronal membranes show alterations, and the capacity of synapses to adjust to stimuli (plasticity) is impaired. Receptor ion channel levels show alterations, as do pre- and postsynaptic components such as synaptophysin, PSD95, and neural cell adhesion molecule. Gene expression and protein expression profiles indicate dose- and radiation-type-dependent changes for genes involved in cell survival, differentiation, and regulation of cytoskeleton and ion channels. A number of behavioral tests interrogating cognition, memory, learning, anxiety, reaction time, and ability to maintain attention can be impaired. These responses are modulated by genotype, sex, and age and show important interindividual variations. Finally, using transgenic mouse models, radiation exposure was observed to accelerate the development of pathophysiological changes associated with Alzheimer’s disease. Together, these observations speak to impairments in functional components at molecular, cellular, and system levels resulting in cognitive and performance deficits and make a credible argument for potential changes in humans as well, but there are many caveats.
Laboratory and accelerator experiments have delivered acute radiation exposures of a few single radiation types, while, in space, radiation exposure is highly protracted and consists of complex radiation mixtures. Dose rates are known to have large influence on outcome measures following irradiation. Animal experiments have been limited to a few strains of mice and rats, and, for practical reasons, most work has been done with young adults. By contrast, the age of astronauts at the time of their first mission is in the late 40s. Human and primate brains are larger and more complex and differ in overall structure from rodent brains while sharing many basic features. Human populations are genetically heterogeneous, while most laboratory animals are inbred so that the contribution of interindividual differences has not been adequately assessed in the animal models. Extrapolation from rodent-based data to humans is limited by the dearth of information on nonhuman primates and human epidemiology and the near complete absence of data from charged particles.
The unique properties of the CNS and the space radiation environment together make risk estimation a formidable task. Cancer risks are based largely on mortality criteria and ultimately are linked to well-understood DNA damage in cell nuclei leading to cell death or uncontrolled growth. At low doses, cell death and proliferation play only a minor role in CNS responses, and the CNS exhibits a gradation of functional impairments rather than simple survival changes. Therefore, morbidity, quality of life, or performance criteria apply rather than mortality. At the present time, there is no widely accepted criteria for significant impairment as pertains to adequate performance in a complex work environment such as spaceflight, but these levels are probably well below those approximating clinical significance as associated with neurological diseases. Future work in estimating CNS risks from space radiation will need to organize adverse outcome findings from radiation exposure according to pathophysiological processes and pathways shared by humans and animals supported by modern systems biology approaches and modeling.
- Armstrong GT, Reddick WE, Petersen RC, Santucci A, Zhang N, Srivastava D, Ogg RJ, Hillenbrand CM, Sabin N, Krasin MJ, Kun L, Pui CH, Hudson MM, Robison LL, Krull KR (2013) Evaluation of memory impairment in aging adult survivors of childhood acute lymphoblastic leukemia treated with cranial radiotherapy. J Natl Cancer Inst 105:899–907CrossRefGoogle Scholar
- Kandel ER, Schwartz JH, Jessel TM (2000) Principles of neural science, 4th edn. McGraw-Hill Companies, New YorkGoogle Scholar
- Kempf SJ, Casciati A, Buratovic S, Janik D, von Toerne C, Ueffing M, Neff F, Moertl S, Stenerlöw B, Saran A, Atkinson MJ, Eriksson P, Pazzaglia S, Tapio S (2014) The cognitive defects of neonatally irradiated mice are accompanied by changed synaptic plasticity, adult neurogenesis and neuroinflammation. Mol Neurodegener 9:57–73CrossRefGoogle Scholar
- Mickley G, Bogo V, West B (1989) Behavioral and neurophysiological changes with exposure to ionizing radiation, Ch. 7. In: Walker RI, Cerveny TJ (eds) Textbook of military medicine. Zajtchuk R (series eds) Part I. Vol. 2. Medical consequences of nuclear war. Textbook of Military Medicine Publications, Office of the Surgeon General, Falls ChurchGoogle Scholar
- Nelson GA (2009) Neurological effects of space radiation. Gravitational Space Biol 22(2):33–38Google Scholar