Life-long brain compensatory responses to galactic cosmic radiation exposure

Abstract

Galactic cosmic radiation (GCR) composed of high-energy, heavy particles (HZE) poses potentially serious hazards to long-duration crewed missions in deep space beyond earth's magnetosphere, including planned missions to Mars. Chronic effects of GCR exposure on brain structure and cognitive function are poorly understood, thereby limiting risk reduction and mitigation strategies to protect against sequelae from exposure during and after deep-space travel. Given the selective vulnerability of the hippocampus to neurotoxic insult and the importance of this brain region to learning and memory, we hypothesized that GCR-relevant HZE exposure may induce long-term alterations in adult hippocampal neurogenesis, synaptic plasticity, and hippocampal-dependent learning and memory. To test this hypothesis, we irradiated 3-month-old male and female mice with a single, whole-body dose of 10, 50, or 100 cGy ⁵⁶Fe ions (600 MeV, 181 keV/μm) at Brookhaven National Laboratory. Our data reveal complex, dynamic, time-dependent effects of HZE exposure on the hippocampus. Two months post exposure, neurogenesis, synaptic plasticity and learning were impaired compared to sham-irradiated, age-matched controls. By six months post-exposure, deficits in spatial learning were absent in irradiated mice, and synaptic potentiation was enhanced. Enhanced performance in spatial learning and facilitation of synaptic plasticity in irradiated mice persisted 12 months post-exposure, concomitant with a dramatic rebound in adult-born neurons. Synaptic plasticity and spatial learning remained enhanced 20 months post-exposure, indicating a life-long influence on plasticity and cognition from a single exposure to HZE in young adulthood. These findings suggest that GCR-exposure can persistently alter brain health and cognitive function during and after long-duration travel in deep space.

Figure 1

HZE exposure-induced suppression of adult neurogenesis in the dentate gyrus of the hippocampus is transient. Representative images (magnification 10 ×) of immunoreactivity of DCX + cells in coronal brain sections from male (a) and female (b) mice two months and 12 months post HZE exposure. Quantification of DCX + cell density in sections 2 months post exposure reveals deficits in both (c) male mice, One-way ANOVA, F(3,16) = 13.20, P = 0.0001, Tukey’s post hoc test: 0 cGy vs. 10 cGy, P = 0.070; 0 cGy vs. 50 cGy, P < 0.0001; 0 cGy vs. 100 cGy, P = 0.003, 10 cGy vs. 50 cGy, P = 0.019, 10 cGy vs. 100 cGy, P = 0.427, 50 cGy vs. 100 cGy, P = 0.312] and (e) female mice, One-way ANOVA, F(3,16) = 15.08, P = 0.0001, Tukey’s post hoc test: 0 cGy vs. 10 cGy, P = 0.009; 0 cGy vs. 50 cGy, P = 0.002; 0 cGy vs. 100 cGy, P < 0.0001, 10 cGy vs. 50 cGy, P = 0.862, 10 cGy vs. 100 cGy, P = 0.047, 50 cGy vs. 100 cGy, P = 0.190, but elevated density of DCX + cells 12 months post radiation exposure in both (d) male mice, One-way ANOVA, F(3,14) = 4.85, P = 0.01, Tukey’s post hoc test: 0 cGy vs. 10 cGy, P = 0.025; 0 cGy vs. 50 cGy, P = 0.047; 0 cGy vs. 100 cGy, P = 0.046, 10 cGy vs. 50 cGy, P = 0.988, 10 cGy vs. 100 cGy, P = 0.947, 50 cGy vs. 100 cGy, P = 0.996, and (f) female mice One-way ANOVA, F(3,12) = 9.92, P = 0.001, Tukey’s post hoc test: 0 cGy vs. 10 cGy, P = 0.002; 0 cGy vs. 50 cGy, P = 0.018; 0 cGy vs. 100 cGy, P = 0.005, 10 cGy vs. 50 cGy, P = 0.123, 10 cGy vs. 100 cGy, P = 0.798, 50 cGy vs. 100 cGy, P = 0.362, compared to age-matched sham-irradiated controls (0 cGy) is shown in bar graphs. *P < 0.05 compared to 0 cGy. Data represents mean ± SEM. n = 5 mice per dose per time point, 6 sections per mouse.

Figure 2

Impairments in Schaffer collateral-CA1 LTP and spatial learning two months after exposure to ⁵⁶Fe particle radiation (a, b) Time course and magnitude (inset bar graph) of LTP in slices from mice exposed to 10 cGy, 50 cGy, or 100 cGy radiation, compared to sham-irradiated controls (0 cGy). After a 15 min baseline, LTP was elicited by two high-frequency TBS stimulus trains (arrows), and magnitude of LTP between 35 and 40 min post TBS (perforated box) was compared across doses in (a) male mice, One-way RM ANOVA, F(1.84,18.35) = 292.8, P < 0.0001, Tukey’s post hoc test: 0 cGy vs. 10 cGy, P < 0.0001; 0 cGy vs. 50 cGy, P < 0.0001; 0 cGy vs. 100 cGy, P < 0.0001, 10 cGy vs. 50 cGy, P = 0.008, 10 cGy vs. 100 cGy, P < 0.0001, 50 cGy vs. 100 cGy, P = 0.041, and (b) female mice, One-way RM ANOVA, F(2.06,20.59) = 1169.0, P < 0.001, Tukey’s post hoc test: 0 cGy vs. 10 cGy, P < 0.0001; 0 cGy vs. 50 cGy, P < 0.001; 0 cGy vs. 100 cGy, P < 0.0001, 10 cGy vs. 50 cGy, P < 0.0001, 10 cGy vs. 100 cGy, P < 0.0001, 50 cGy vs. 100 cGy, P < 0.0001. *P < 0.05 compared to 0 cGy, n = 12–16 slices per dose per sex. (c, d) Learning curves, including pretraining, training days 1–3, and conflict training days 1–2, are shown for (c) male mice, Two-way RM ANOVA, dose: F(3,16) = 4.51, P = 0.017; trial: F(3.59,57.51) = 7.22, P = 0.002; interaction: F(15,80) = 2.29, P = 0.009, Tukey’s post hoc test: Conflict Day 2, 0 cGy vs. 50 cGy, P = 0.01, 0 cGy vs. 100 cGy, P = 0.043 and (d) female mice, Two-way RM ANOVA, dose: F(3,16) = 3.58, P = 0.037; trial: F(4.29,68.72) = 6.86, P < 0.001; interaction: F(15,80) = 1.15, P = 0.32, Tukey’s post hoc test: Training Day 1, 0 cGy vs. 100 cGy, P = 0.043, Conflict Day 1, 0 cGy vs. 100 cGy, P = 0.027, Conflict Day 2, 0 cGy vs. 100 cGy, P = 0.029 as a function of normalized number of entries into the stationary shock zone (Errors). *P < 0.05, n = 5 mice per dose per sex. Each point represents mean errors normalized to pre-training entries ± SEM.

Figure 3

Enhancement in Schaffer collateral-CA1 LTP and recovery of spatial learning deficits six months after exposure to ⁵⁶Fe particle radiation (a, b) Time course and magnitude (inset bar graph) of LTP in slices from mice exposed to 10 cGy, 50 cGy, or 100 cGy radiation, compared to sham-irradiated controls (0 cGy). After a 15 min baseline, LTP was elicited by two high-frequency TBS stimulus trains (arrows), and magnitude of LTP between 35–40 min post TBS (perforated box) was compared across doses in (a) male mice, One-way RM ANOVA, F(1.95,19.45) = 820.8, P < 0.0001, Tukey’s post hoc test: 0 cGy vs. 50 cGy, P < 0.0001; 0 cGy vs. 100 cGy, P < 0.0001, 50 cGy vs. 100 cGy, P < 0.0001 and (b) female mice, One-way RM ANOVA, F(1.30,12.95) = 393.0, P < 0.0001, Tukey’s post hoc test: 0 cGy vs. 50 cGy, P < 0.0001; 0 cGy vs. 100 cGy, P < 0.0001, 50 cGy vs. 100 cGy, P < 0.0001. *P < 0.05 compared to 0 cGy, n = 12–16 slices per dose per sex. (c) Representative paired-pulse responses of population compound action potentials elicited in slices from a male sham-irradiated mouse, taken at 10 and 500 ms inter-stimulus interval. (d) Mean paired-pulse profiles across inter-stimulus intervals in all slices from control (0 cGy) and irradiated (100 cGy) mice. Each point represents mean ± SEM. n = 12–16 slices per dose per sex. (e, f) Learning curves, including pretraining, training days 1–3, and conflict training days 1–2, are shown for (e) male mice, Two-way RM ANOVA, dose: F(1,8) = 0.59, P = 0.59, trial: F(2.43,19.43) = 18.88, P < 0.0001; interaction: F(5,40) = 0.993, P = 0.434, (Tukey’s post hoc test did not reveal significant within treatment effect) and (f) female mice, Two-way RM ANOVA, dose: F(1,8) = 0.99, P = 0.99; trial: F(1.96,15.66) = 31.54, P < 0.0001; interaction: F(5,40) = 1.896, P = 0.117, (Tukey’s post hoc test did not reveal significant within treatment effect), as a function of normalized number of entries into the stationary shock zone (Errors). n = 5 mice per dose per sex. Each point represents mean errors normalized to pre-training entries ± SEM.

Figure 4

Enhancements in Schaffer collateral-CA1 LTP and spatial learning 12 months after exposure to ⁵⁶Fe particle radiation (a, b) Time course and magnitude (inset bar graph) of LTP in slices from mice exposed to 10 cGy, 50 cGy, or 100 cGy radiation, compared to sham-irradiated controls (0 cGy). After a 15 min baseline, LTP was elicited by two high-frequency TBS stimulus trains (arrows), and magnitude of LTP between 35 and 40 min post TBS (perforated box) was compared across doses in (a) male mice, One-way RM ANOVA, F(1.73,17.31) = 3137.0, P < 0.001, Tukey’s post hoc test: 0 cGy vs. 10 cGy, P < 0.0001; 0 cGy vs. 50 cGy, P < 0.0001; 0 cGy vs. 100 cGy, P < 0.0001, 10 cGy vs. 50 cGy, P < 0.0001, 10 cGy vs. 100 cGy, P < 0.0001, 50 cGy vs. 100 cGy, P < 0.0001 and (b) female mice, One-way RM ANOVA, F(2.48,24.77) = 449.1, P < 0.0001, Tukey’s post hoc test: 0 cGy vs. 10 cGy, P < 0.0001; 0 cGy vs. 50 cGy, P < 0.0001; 0 cGy vs. 100 cGy, P < 0.0001, 10 cGy vs. 50 cGy, P < 0.0001, 10 cGy vs. 100 cGy, P < 0.0001, 50 cGy vs. 100 cGy, P < 0.0001. *P < 0.05 compared to 0 cGy, n = 12–16 slices per dose per sex. (c, d) Learning curves, including pretraining, training days 1–3, and conflict training days 1–2, are shown for (c) male mice Two-way RM ANOVA, dose: F(3,16) = 2.78, P = 0.08; trial: F(1.933,30.93) = 54.98, P < 0.0001; interaction: F(15,80) = 1.31, P = 0.21, Tukey’s post hoc test: Training Day 1, 0 cGy vs. 50 cGy, P = 0.008, 0 cGy vs. 100 cGy, P = 0.045, Training Day 2, 0 cGy vs. 50 cGy, P = 0.019 and (d) female mice Two-way RM ANOVA, dose: F(3,16) = 2.57, P = 0.09; trial: F(2.445,39.12) = 47.93, P < 0.0001; interaction: F(15,80) = 1.23, P = 0.266, Tukey’s post hoc test: Training Day 2, 0 cGy vs. 50 cGy, P = 0.021, Training Day 3, 0 cGy vs. 50 cGy, P = 0.044, Conflict Day 2, 0 cGy vs. 100 cGy, P = 0.038, as a function of normalized number of entries into the stationary shock zone (Errors). *P < 0.05, n = 5 mice per dose per sex. Each point represents mean errors normalized to pre-training entries ± SEM.

Figure 5

Enhancements in Schaffer collateral-CA1 LTP and spatial learning 20 months after exposure to ⁵⁶Fe particle radiation (a) Time course and magnitude (inset bar graph) of stimulus-evoked LTP in slices from male mice exposed to 100 cGy radiation, compared to sham-irradiated controls (0 cGy). After a 15 min baseline, LTP was elicited by two TBS stimulus trains (arrows), and the magnitude of LTP between 35–40 min post TBS or bath application (perforated box) compared across doses. *P < 0.0001, Two-tailed t test, n = 12–16 slices per dose. (b) Time course and magnitude (inset bar graph) of chemically-evoked LTP elicited by bath application (bar) of 10 µM forskolin plus 10 µM rolipram to slices from male mice exposed to 100 cGy radiation compared to sham-irradiated controls (0 cGy). Each point represents mean ± SEM. *P < 0.0001, Two-tailed t test, n = 12–16 slices per dose. (c, d) Learning curves depicting (c) latency for entry into an escape box in the Barnes Maze, Two-way RM ANOVA, dose: F(1,90) = 30.73, P < 0.0001; trial: F(4,90) = 2.22, P = 0.073; interaction: F(4,90) = 0.553, P = 0.698, Tukey’s post hoc test: Day 2, 0 cGy vs. 100 cGy, P = 0.006, Day 4, 0 cGy vs. 100 cGy, P = 0.027, or (d) the fraction of successful trials per day, Two-way RM ANOVA, dose: F(1,90) = 10.51, P = 0.002, trial: F(4,90) = 3.06, P = 0.020; interaction: F(4,90) = 0.849, P = 0.497, (Tukey’s post hoc test did not reveal significant within treatment effect), are shown for male mice exposed to 100 cGy radiation compared to sham-irradiated controls (0 cGy). Each point represents mean latency ± sem. *P < 0.05, n = 10 mice per dose.