Long-term space travel and planetary exploration, including missions to Moon and Mars, are the next challenging steps for crewed space missions. During these missions, exposure to space radiation might be detrimental to astronaut health as a risk factor for the development of cancer and non-cancer effects (1, 2). There are two major sources for the space radiation environment in space: (1) Galactic cosmic rays (GCR) originating beyond the solar system [energetic protons, helium nuclei and heavy ions, also called high charge and energy (HZE) nuclei] and (2) Solar energetic particles (SEP) continuously emitted by the Sun as low energetic solar wind or during coronal mass ejections as solar particle events (SPE). While GCR are composed of 98 % baryons (87 % protons, 12 % helium ions, ~1 % heavy ions) and 2 % electrons and positrons (3), SPE are composed of 89 % protons, 10 % helium ions and 1 % heavy ions. The probability of their occurrence rises during the solar maximum of the ~11-year solar cycle. As GCR cannot be completely shielded during free space voyage, a chronic low-dose rate GCR exposure of astronauts accumulates to considerable doses during a 3-year-Mars mission (~1 Sv) (4–6). Furthermore, SPE bear the risk of acute exposure to high dose rates in situations of insufficient shielding.

With this composition, space radiation differs strongly from well-characterized ionizing radiation qualities on Earth such as alpha-, beta- and gamma radiation or X-rays. High complexity of the space radiation field makes assessment of its biological effects quite difficult. GCR simulation composed of protons, helium nuclei and selected heavy ions became only recently available at the NASA Space Radiation Laboratory (NSRL) at Brookhaven National Laboratories (BNL) (7). Mostly, experiments with single mono-energetic ions are performed to assess their biological effectiveness in comparison to well-known radiation qualities such as X-rays or gamma rays. Here, the linear energy transfer (LET) of radiation is frequently used to describe ionization density and correlated biological effects. While X-rays or γ-rays are considered low-LET radiation, HZE particles of GCR are high-LET radiation.

Due to its limited repair capacity, space radiation effects on the central nervous system (CNS) are of high interest. Decreased CNS performance, but also an increased overall risk to develop a neurodegenerative disorder such as Parkinson's Disease in astronauts is suspected (2, 8). Animal experiments using space-like radiation revealed impairments in memory, deficits in processing speed, attention and cognitive flexibility, as well as elevated anxiety levels and depressive behavior in mice (8, 9). The cellular mechanisms involved in these cognitive effects are currently under investigation. Animal studies demonstrated that increased cell death (10), decreased proliferation (11), increased DNA damage (12), cell cycle changes (13) and neuroinflammation including activation of microglia and astrocytes (14–16) might be involved in the response to space-relevant radiation qualities. Since cognitive detriments and increased risk of developing neurogenerative disease are crucial factors affecting astronaut health and mission success, further investigations of the underlying cellular and molecular processes are necessary. Astrocytes are the most abundant glial cells in the CNS and despite their supportive function in the physiological processes of the brain, they react under pathophysiological conditions with cellular, molecular and functional changes. This suggests that astrocytes play a crucial role in the brain's response to radiation.

The main impact of ionizing radiation on mammalian cells is damage to deoxyribonucleic acid (DNA), inducing breakage of both DNA strands (DNA double strand break, DSB), which could subsequently lead to cell death (17). Compared to low-LET radiation, high-LET radiation leads to dense ionizations along the particle tracks and induces more complex DNA damage which is difficult to repair (18). Cellular damage induced by ionizing radiation subsequently initiates an active cellular response, the DNA damage response, comprising DNA damage repair, altered gene expression, cell cycle arrest or programmed cell death (19–21). In general, while basic principles and mechanisms of the radiation response in mammalian systems are understood nowadays, tissue- or cell-type specific effects of ionizing radiation are still under investigation. For astrocytes, knowledge of their ability to repair DNA damage induced by ionizing radiation, especially heavy ion-induced DNA damage is scarce. In a seminal study comparing murine embryonic stem cell-derived neural stem cells and corresponding terminally differentiated astrocytes, astrocytes were radioresistant and expressed non-homologous end-joining genes enabling repair of ionizing radiation-induced DNA DSB (22). The repair kinetics of these DNA damages can indicate whether a cell-type is DNA repair-proficient or -deficient. Intracellular pathways are known to be activated in response to ionizing radiation, such as the nuclear factor κB (NF-κB) pathway. NF-κB is known to transcriptionally regulate a multitude of cellular responses, like immune response, inflammation via cytokine release, proliferation, cell cycle progression, and apoptosis. In other cell types, it was shown that the NF-κB subunit p65 translocates into the nucleus upon pathway activation in response to ionizing radiation exposure (20), including heavy ion exposure (21–23).

During pathophysiologic processes, e.g., CNS injuries, inflammation or exposure to toxic substances, astrocytes shift their phenotypic state from a normal naive state to a reactive state, also known as astrocyte reactivity. Depending on the severity of the nervous tissue insult, astrocytes become reactive, which spreads throughout the affected area as so-called reactive astrogliosis and ultimately leads to the formation of the glial scar (23, 24). This phenotypic change upon reactivity induction is accompanied by several traits, including hyperproliferation, increased cellular maintenance, morphological alterations, increased migration rates, cytokine release, and gene expression changes. Characteristic for astrocyte reactivity is the overexpression of the intermediate filament glial fibrillary acidic protein (GFAP) (23). Furthermore, astrogliosis is a heterogeneous process (23) with a continuous spectrum of severities (25). Depending on pleiotropic factors, astrocytes may maintain damage-induced inflammatory reactions and tissue damage or promote repair of tissue after becoming reactive (23). This process can also be triggered by neuroinflammation and plays a role in neurodegenerative mechanisms. Ionizing radiation is also known to induce neuroinflammation by microglia activation or astrocyte reactivity or by induction of radiation-induced senescence further promoting chronic inflammation (26, 27). In recent studies, stress response mechanisms in astrocytes include reactivity induction and cellular senescence (25), raising the question whether astrocytes respond with a transition into a reactive state or with so-called astrosenescence after exposure to ionizing radiation. Astrosenescence is characterized, for example, by a growth arrest, a senescence-associated secretory profile (SASP) involving increased secretion of cytokines such as interleukins (IL), as well as senescence-associated β-galactosidase activity, whereas astrocyte reactivity is accompanied by increased cytokine secretion in response to CNS insults (25).

Ionizing radiation is not the only risk factor astronauts face during space missions. Microgravity affects the human body by head ward fluid shift and mechanical unloading, resulting in changes in visual acuity (8), bone (28, 29) and muscle loss (30), reductions in plasma volume, cardiovascular deconditioning and neurovestibular alterations (31–36). In 30-day 6°-head-down-tilt bedrest study, changes in white and gray matter volume and white matter tracts of the brain of healthy volunteers were shown by magnetic resonance imaging (MRI) (37). Also, in animal models, microstructural alterations were found in multiple brain regions (37). On tissue level, in a biosatellite experiment with C57BL/6N mice, myelin degeneration of the sciatic nerve (38) and transcriptome changes were observed (39), and alterations in the choroid plexus were induced by hindlimb unloading or spaceflight in rats (40). Microgravity-induced effects are also observed on a cellular level (41, 42), for example, changes of organelles and the cytoskeleton (43–45), of migration (46–48), cell cycle regulation (49), cell proliferation (50), apoptosis (51), DNA repair (52), differentiation (8, 53, 54) and T cell regulation (41, 55), and gene expression, proteome and epigenetics alterations (49, 56, 57). Interestingly, increased mechanical loading in consequence to mild hypergravity exposure (2g) yielded an attenuation of astrocyte reactivity (58). Thus, astrocytes are sensitive to changes in gravity levels, but a clear understanding of the effects of multiple space-relevant conditions including microgravity and ionizing radiation on astrocytes is still missing. As DNA repair, cell cycle arrests, apoptosis and changes in proliferation and gene expression are hallmarks of the DNA damage response, the interaction of space radiation and microgravity effects on the cellular level needs to be understood also in astrocytes.

This study aims to characterize the response of primary murine astrocytes toward exposure to low- and high-LET radiation to further understand their role and function in the brain after radiation exposure. Primary murine astrocytes isolated from the cortex of mouse embryos are powerful tools to understand molecular pathways induced by radiation exposure and whether they secrete, e.g., cytokines (59). To determine if the DNA damage repair kinetics in astrocytes are comparable to other cell types, formation of phosphorylated H2AX (γH2AX) and p53 binding protein 1 (53BP1) foci was investigated via immunostaining after irradiation of primary murine astrocytes with different types of radiation. As arrest of cell cycle progression in response to ionizing radiation allows sufficient repair time of DNA damage, the cell cycle was analyzed and gene expression of regulators involved in different cell cycle check points (Cdkn1a, Cdkn2a) was studied via reverse transcription quantitative real-time Polymerase Chain Reaction (RT-qPCR). As the NF-κB pathway constitutes a major signaling pathway involved in inflammatory responses to ionizing radiation, its activation was investigated by immunostaining of the NF-κB subunit p65 and quantification of its nuclear localization. Astrocyte reactivity in response to ionizing radiation exposure was assessed by immunofluorescence staining of GFAP and the cell proliferation marker Ki-67. Since changes in gene expression are known to be part of astrocytes' reactivity, expression of Gfap, Il1ß, Il6, Tgfß1, and tumor necrosis factor (Tnf) was investigated by RT-qPCR. The transcriptomics profile of X-irradiated astrocytes was determined by mRNA sequencing. In order to differentiate reactivity from astrosenescence, cytokine secretion was quantified using enzyme-linked immunosorbent assays (ELISA). Furthermore, the effect of simulated microgravity on the repair of X-ray-induced DNA double strand breaks was analyzed using the principle of 2D fast clinorotation (60, 61) to gain a basic understanding of astrocytes' DNA damage response under space-like conditions.

An overview of the experiments performed in this work indicating radiation qualities, doses and time points that were investigated for the different biological endpoints in primary murine astrocytes is given in Table 1.

Experiments performed in this work showed that astrocytes repair ionizing radiation-induced DNA double strand breaks at a pace comparable to other cell types. Their radiation response is quite subdued without prominent changes in cell cycle progression and gene expression. The reaction profile trends toward senescence, and a pronounced induction of reactivity was not observed.

The radiation response of astrocytes is an important piece of the puzzle in understanding the effects of different ionizing radiation qualities on the brain as they represent the predominant cell type in the mammalian brain (22) and are crucial to normal brain function (83). Although the isolation of primary astrocytes is laborious and the results are sometimes variable, primary astrocytes from rodent embryos are recognized as a potent tool to study astrocytes, biology and mechanistic features (66, 84). As alternative cell models such as immortalized astrocytes and C6 glioma cells showed major differences in morphology, protein expression and functionality in comparison to primary astrocytes (84), we chose primary murine astrocytes for our studies. Overall, the radiation response of primary murine astrocytes was unremarkable with mostly minor changes for some doses and time points after exposure.

For coping with radiation-induced DNA damage, DNA repair is crucial. Here, it was shown that primary murine astrocytes repair DNA DSBs induced by X-rays with a kinetics that is comparable to other repair-proficient cell types (85–87). Fully functioning DNA repair was also suggested by Schneider et al. after investigation of X-rays-induced γH2AX foci in astrocytes that were differentiated from murine embryonic stem cell-derived neural stem cells (22). Also normal human astrocytes were capable to repair DNA DSBs induced by 10 Gy X-rays, and they even upregulated the expression of key proteins involved in non-homologous end joining (Ku70) and homologous recombination (RAD51) after irradiation (88) and they increased DNA repair when they were in a reactive state (89). For the radiation qualities with an LET of ~150 keV/μm, the number of 53BP1 foci was lower than the maximal number of γH2AX foci. This is in line with findings in murine neural stem/progenitor cells after γ-irradiation, in which γH2AX and 53BP1 did not fully colocalize and the number of 53BP1 foci was lower than the number of γH2AX foci in the same cell nucleus (90). After phosphorylation of H2AX, 53BP1 is recruited to the DNA DSB and it is involved in the DNA DSB repair pathway choice by controlling resection at the free DNA ends and thereby interfering with homologous recombination (91, 92). In retinal pigment epithelial RPE1 cells, 53BP1 was predominantly recruited to repair foci of cells in G0/G1 phase which was explained by its role in non-homologous end-joining, the DSB repair pathway available in mammalian cells during this cell cycle phase, while the 53BP1 foci were smaller during S phase (93). In G2 phase, 53BP1 foci might disappear faster than the γH2AX foci (94) and in general, the dynamics of radiation-induced γH2AX and 53BP1 foci disappearance can differ (95, 96). This might explain the lower number of 53BP1 foci compared to γH2AX foci that was observed in primary murine astrocytes in this work.

Simulated microgravity did not modulate γH2AX foci formation after exposure to X-rays and the subsequent repair of DNA DSBs by primary murine astrocytes. The question whether microgravity modulates the DNA damage response and more specifically, DNA repair, was addressed in several space experiments (97–100) and ground-based studies (52, 101) and they led to conflicting results. A growth-stimulating effect of microgravity and changes in gene expression might be contributors to microgravity effects on the DNA damage response, while DNA repair itself was mostly unaffected (52), which is in line with the results of this study. As a clear dose-dependent radiation response of astrocytes was observed only for DNA DSB induction and repair, simulated microgravity was incorporated in these experiments only and not extended to other biological endpoints investigated in this work.

The number of γH2AX foci that were induced by the same energy dose of iron ions was lower compared to X-rays. This finding is expected for high-LET radiation, as ionization occurs along tracks, resulting in a lower number of foci, which can contain more complex DNA damage and/or several DNA DSBs (85, 102). With higher LET, a lower number of average hits (and thereby, γH2AX foci) was expected after heavy ion exposure (Table 3). Furthermore, after ⁵⁶Fe ion exposure, γH2AX foci formation and removal was delayed when compared to X-rays. This delayed repair is generally explained by the complexity of the heavy ion-induced DNA damage, requiring the coordination of several repair pathways (103).

To gain time for this DNA repair, cell cycle can be arrested at different checkpoints after ionizing radiation exposure. In murine astrocytes, no significant cell cycle changes were observed after exposure to X-rays. A clear G2 arrest which is usually induced in strongly proliferating mammalian cells after ionizing radiation exposure was not observed. Due to the low proliferation rate of primary astrocytes (a cell population doubling occurred after ~180 h in passage 1), accumulation in G2 phase may be negligible or completely absent. Gene expression profiling indicated downregulation of proliferation, cell cycle and mitosis genes. In RT-qPCR of selected target genes, no significant effects on the cell cycle regulation genes Cdkn1a (encoding p21^WAF/CIP1) and Cdkn2a (encoding p16) which are involved in cell cycle arrests after ionizing radiation exposure and in senescence induction were observed after X-rays and Fe ions exposure. This is in line with the absence of a G2 arrest. Relative quantification of mRNA levels in RT-qPCR might obscure already high gene expression levels of Cdkn1a and Cdkn2a as indicators of senescence, but the presence of S-phase and G2-phase cells under all treatment conditions does not suggest a complete G1 cell cycle arrest in the primary murine astrocytes investigated in this work. A general downregulation of the DNA damage response signaling in astrocytes was described previously (22) additionally attributing to the absence of radiation-induced cell cycle arrest. The baseline cell cycle distribution in mock-irradiated cells was as follows: 40–50 % of cells were in G0/G1 and 50–60 % in S or G2 phase. For each experiment, new astrocytes were isolated and some variation for different isolates might be attributed to such isolate batch effects and to different timelines in the preparation of the beamtimes at the heavy ion accelerator. Therefore, mock-irradiated controls were generated for each experiment to account for such batch variations. The cells in G0/G1 might reside in a quiescent state (G0) that permits subsequent cell division upon stimulation (e.g., withdrawal and re-addition of FBS) or in G1 as part of the actively cycling cells (104).

Astrocyte reactivity (105) was evaluated by proliferation, GFAP and cytokine expression. Proliferation of astrocytes usually increases upon reactivity, inducing cell infiltration to damage sites in the CNS (23). As Ki-67 is the most cited proliferation marker that can be determined by immunofluorescence staining with highest levels during G2 phase and mitosis (106, 107), the fraction of Ki-67 positive cells after exposure of astrocytes to X-rays was quantified and found to be largely unaffected by X-irradiation at around 10 %. A similar low percentage of proliferating cells of 10–25 % was also observed in primary rat astrocytes (66) and in adult mouse astrocytes (108). As the proliferation rate decreased in higher passages (108), only passage 1 astrocytes were used in this work. A higher proliferation of murine astrocytes could be achieved by adding 20 % FBS instead of 10 % (108). Also, in presence of TNF-α, the percentage of Ki-67⁺ cells increased to 25–30 %, although this effect was not significant due to large standard errors. Increased astrocyte proliferation in response to the cytokines TNF-α and IL-1β increase was described previously, supporting the results of this work (79). As X-ray doses up to 8 Gy did not significantly change Ki-67 expression, this marker was not used in the heavy ion experiments.

In this work, astrocytes basally expressed GFAP without significant changes after exposure to low–LET radiation up to a dose of 8 Gy. In the context of research to improve radiotherapy of brain tumors, reactive gliosis after irradiation with higher doses (10 Gy) and in systemic context (whole-body X-irradiation of mice) was observed based on upregulation of GFAP in the brain (109). Increased GFAP expression was also documented 6 h and 24 h after head-only X-irradiation (15 Gy) of rats (15) and of mice (20 Gy) (110). As absence of serum response factor (SRF) resulted in increased GFAP expression in SRF knockout mice, the isolated culture of primary astrocytes in presence of serum-containing medium could be an explanation for suppressed GFAP expression and absence of increase after irradiation (111). While isolated culture of astrocytes offers the advantage that the specific radiation response of this cell type can be analyzed by methods not suitable for co-cultures or brain slices, responses that occur only in the multicellular context of brain tissue cannot be addressed. Here, it has to be considered that astrocytes' responses to CNS insults are a multicellular process in which the reaction of all cell types in the brain including microglia, oligodendrocytes or neurons and their release of signaling molecules is integrated to activate astrocytes (23). For increased proliferation of astrocytes in response to injury, crosstalk of astrocytes with macrophages is required (112). In cultured rat brain astrocytes incubated with media supernatants from X-irradiated (2 and 10 Gy) microglia, GFAP expression was increased after 24 h (15), indicating an important role of microglia in astrogliosis induction by ionizing radiation exposure. Also, in co-culture with endothelial cells in an organ-on-a-chip-model, exposure to 0.3 Gy and 0.82 Gy ⁵⁶Fe ions (600 MeV/n, LET 170 keV/μm) increased GFAP expression in astrocytes 3 days after exposure (113).

Immunostaining of p65 and IL-6 ELISA revealed that in primary murine astrocytes, basal activity of the NF-κB pathway was present resulting in continuous IL-6 expression and secretion which was not further enhanced by exposure to X-rays or C ions.

The absence of radiated-induced NF-κB activation correlates well with the absence of radiation-induced GFAP expression as GFAP expression is regulated by NF-κB by a κB binding site in the GFAP promoter region (114). In this context, the previously observed downregulation of DNA damage response signaling including ATM activity in astrocytes (22) is of interest as ATM is a key player in ionizing radiation-induced NF-κB pathway activation (70), possibly explaining the absence of NF-κB activation by the radiation qualities investigated in this work.

In the CNS, the proinflammatory cytokine IL-6 is predominantly produced by astrocytes and is involved in cell-cell communication and reactivity of astrocytes. IL-6 could exert some autocrine actions in astrocytes (74, 115, 116). Both, proliferative (117) and antiproliferative actions (118) of IL-6 in astrocytes have been described. Transient local cytokine secretion might promote the brain's recovery after injury, but long-term upregulation might result in damage (76). IL-6 expression is expected to be found in a senescence-associated secretory profile (SASP), a phenotypic shift leading to premature or stress-induced cellular senescence (119–121). An irreversible cell cycle arrest is a major characteristic of cellular senescence (49, 50). In human fibroblasts, cell populations with < 10 % proliferative cells were designated as non-dividing senescent cultures (51). Here, no significant reductions in Ki-67⁺ astrocytes were observed. Therefore, this finding of basal IL-6 secretion suggests that these cells quickly adopt an only partially senescent phenotype in isolated culture, which is not further enhanced by in vitro exposure to ionizing radiation. TP53 was reported to regulate cellular senescence in astrocytes induced by ionizing radiation exposure (119). The absence of a decided TP53 gene expression signature (including e.g., the expression of Cdkn1a—p21) in the RNA sequencing data after X-rays exposure might explain why the radiation response of primary astrocytes was so reluctant. While TP53 in astrocytes was related to various disease processes (122), its role in the DNA damage response of astrocytes remains elusive. In astrocytes that were differentiated from murine embryonic stem cell-derived neural stem cells and exposed to 10 Gy or even 50 Gy X-rays, no TP53 activation occurred 1 h and 24 h after irradiation, and the expression levels of the TP53 target genes GADD45a, BAX and PUMA remained largely unchanged, only CDKN1A expression was upregulated (22). This absence of a TP53-mediated transcriptional response after exposure to X-rays was also observed in cortical astroglia cell cultures from 1-d old mouse pups (123) and might be an important factor for the observed radioresistance of astrocytes (22). However, in activated proliferating astrocytes, exposure to 4 Gy X-rays activated TP53 (124), indicating that the reactivity status might influence the TP53 response of astrocytes.

Another possibility is that a senescent phenotype is present because of high basal expression of Cdkn1a and Cdkn2a, the major mediators of senescence-associated proliferation arrest (121), that did not further increase after treatment. In vivo, cellular senescence describes a state in which astrocytes do not replicate but remain alive in the tissue, producing pro-inflammatory and neurotoxic factors, and contributing to CNS damage (119, 125). Such aging of astrocytes was associated with a smaller pool of synaptic vesicles in co-cultured neurons and decreased neuroprotective capacity (126, 127).

The data acquired in this work provide an indication that exposure of primary murine astrocytes in isolated culture to X-rays or heavy ions did not result in astrocyte reactivity. Therefore, further experiments were not performed to characterize reactivity in more detail by including various markers (vimentin, leucine zipper kinase, and nestin) (23, 111, 128) and pathways involved in reactivity such as STAT3, cyclic adenosine monophosphate (cAMP) or C-Jun-N-terminal kinase (JNK) (23, 129).

Sequencing of mRNA isolated from X-irradiated murine astrocytes revealed that a low dose of 0.1 Gy had no effect on global gene expression 6 and 24 h after irradiation. Gene expression changes were observed only at the late time point 24 h after irradiation, and for the higher dose of 2 Gy. The number of affected genes was low, the majority being downregulated. This downregulation was moderate and affected mostly genes involved in proliferation and DNA repair. As both processes were not significantly affected after irradiation of astrocytes, the biological role of these downregulations remains unclear. The downregulation was not observed at the 6 h time point while DNA DSB repair was still ongoing. The majority of the nine downregulated DNA repair genes is involved in homologous recombination (Fignl1, Brip1, Fancd2, Gen1, Brca1, Blm) which is not expected to be the prominent DNA DSB repair pathway in astrocytes. Therefore, at the current stage, no strong indications of a high functional relevance of the small reduction in expression of DNA repair genes can be derived. No overlap of the profiles was observed when comparing this gene expression profile to the expression profile of reactive astrocytes in two mouse brain injury models, in which several hundreds of genes were upregulated (105). This suggests that the expression profile observed in this work is not indicative of reactive astrogliosis, although it has to be considered that the gene expression profile of reactive astrocytes was described to be specific for a given injury (105).

Only two genes were upregulated in response to 2 Gy X-rays: Svop and Abhd11os. The synaptic vesicle protein SVOP is described as a 548-aa protein of ~60 kDa (130) with 12 transmembrane regions (131) capable of binding nucleotides [e.g., nicotinamide adenine dinucleotide (NAD)] (132) and of transporting nicotinate (133). Its expression was described to be limited to the CNS (134); in the adult mouse brain it was predominantly found in hippocampus and cerebellum (130). Based on its structure as a transporter-like protein (135) and functional studies performed so far, a possible role in synaptic vesicle uptake/transport is assumed that is not required for survival under normal conditions (136). In humans, abnormal methylation of the SVOP gene located on chromosome 12 (137) was correlated with prognosis of glioblastoma (138). Not much is known on the function of long non-coding RNA (lncRNA) Abhd11os (139) in astrocytes. The human homolog ABHD11 antisense RNA 1 (ABHD11-AS1) was found to be highly expressed in gastric, lung, breast, colorectal, thyroid, pancreas, ovary, endometrium, cervix, and bladder cancer and was therefore suggested as biomarker for diagnosis and prognosis (140). In mouse models of Huntington's disease, expression of Abhd11os was reduced (139) or dysregulated (141) and its overexpression had neuroprotective effects in mice against mutant huntingtin-induced toxicity (139). Besides these findings in cancer and Huntington's disease models, the role of Abhd11os expression in myocardial infarction was addressed: Increased Abhd11os expression was found in a rat myocardial ischemia/reperfusion injury model and hypoxia/reoxygenation-treated cardiomyocytes (142). This upregulation of Abhd11os inhibited proliferation of cardiomyocytes but promoted cell apoptosis, while downregulation of Abhd11os inhibited apoptosis of cardiomyocytes thereby attenuating the injury (142). Interestingly, after whole-body irradiation of mice, the expression of the Abhd11os increased dose-dependently in heart tissue (143), in line with the X-rays-induced upregulation observed in murine astrocytes in this study.

To assess whether gene expression might be more modulated after exposure to higher doses or other radiation qualities (Fe and C ions), or only transiently affected, several genes of interest were analyzed by RT-qPCR: Cdkn1a and Cdkn2a, that are involved in cell cycle progression Gfap as marker for astrocyte reactivity; the cytokines Tnf , Il1β and Il6, being involved in inflammation, proliferation and apoptosis; Tgfβ1, which acts in anti-inflammatory and anti-apoptotic manner. For these genes, no clear dose dependence of up- or downregulation was observed, and some regulations were transient. For example, X-irradiation with 8 Gy caused only a transient upregulation of Il6 at the time point 6 h.

Due to the limited availability of beamtimes at heavy ion accelerators, the heavy ion experiments could not be repeated in independent beamtimes, but biological replicates isolated from different animals were included and the sample size of each biological replicate contained several thousand cells. Batch effects of different isolates were observed for example in terms of the higher basal IL-6 secretion in the C ion (7 MeV/n) experiments at GSI compared to the experiments with X-rays at DLR. Furthermore, not all biological endpoints could be analyzed for all four radiation qualities due to beamtime time restrictions. Nonetheless, the results of the heavy ion experiments were interpreted only as possible trends if statistical tests were not possible. Also, due to the low energy of the carbon ions, the UNILAC beamtime required a different experimental setting where astrocytes were seeded in petri dishes and kept in a reservoir with cell culture medium for irradiation.

The choice of radiation doses used in this work was based on the average mission doses on ISS [6 months ~ 90–150 mSv (144, 145)] and a 1,000-days Mars mission [~340–1,000 mSv, depending on solar activity and shielding (146)]. Higher doses up to 8 Gy were added to generate dose response curves. As the relative biological effectiveness (RBE) for the investigated biological endpoints in astrocytes is not known, we used mostly the same dose range for X-rays and heavy ions to generate data from which the RBE could be derived. As in most of the space radiobiological in vitro experiments, the effects of an acute radiation exposure were investigated—a protracted radiation exposure of cultured cells with heavy ions over 6 months to 3 years, imitating the mission durations, is simply not feasible. The extrapolation of the effects of this acute high-dose rate exposure to chronic low-dose rate exposure requires some assumptions that are usually considered in terms of a dose- and dose-rate reduction factor (DDREF) (147–150), which, in worst case where a lower dose rate does not alleviate the damage, is 1.

In conclusion, primary murine astrocytes were shown to be fully repair-proficient for DNA DSBs induced by low- and high-LET radiation. They seemed to be quite radioresistant and a comprehensive DNA damage response also including cell cycle arrests and cell death was absent. In isolated culture, they did not shift toward astrocyte reactivity but the indicators of a senescent phenotype warrant further investigation. Also, based on the findings of this work, it seems that the question of astrogliosis or astrosenescence cannot be answered when astrocytes are isolated from their natural microenvironment in the brain. More complex systems such as co-cultures, multicellular culture models (59), organ-on-a-chip models (113), brain slices or brain organoids could be interesting models to study astrocytes' response to space-relevant radiation qualities embedded in the cellular crosstalk within the brain.

The RNA Seq datasets presented in this study can be found in online repositories. The name of the repository and accession number is NCBI Gene Expression Omnibus (GEO) GSE215383 - Astrocytes' Radiation Response to be found under the link https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE215383.

The animal study was reviewed and approved by Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen, LANUV, Germany on 4 December, 2017, Nr. 84-02.04.2017.A319.

Conceptualization and writing—original draft: CEH, JK, and MDR. Data curation: MDR, SH, and EW. Formal analysis and visualization: CEH, MDR, SH, EW, and JK. Funding acquisition: CEH and SD. Investigation: CEH, BK, MDR, SH, EW, JK, SD, and HN. Methodology: CEH, MDR, SH, EW, JK, and SD. Project administration: CEH and JK. Resources: CEH. Supervision: CEH, JK, SD, and CL. Writing- review and editing: HN, CEH, MDR, BK, SH, and CL. All authors have read, agreed to the published version of the manuscript, approved the submitted version, agree to be personally accountable for the author's own contributions and for ensuring that questions related to the accuracy or integrity of any part of the work, even ones in which the author was not personally involved, are appropriately investigated, resolved, and documented in the literature.