The primary finding of the current study is that while EC-SOD deficiency is associated with some hippocampus-dependent impairments prior to irradiation, a finding also shown by others (Levin et al., 1998
), after irradiation, hippocampus-dependent cognitive measures are enhanced in mice lacking EC-SOD. This paradoxical effect supports the idea that ROS can have both positive and negative effects, depending upon the circumstances (Kamsler and Segal, 2003
; Levin et al., 1998
; Valko et al., 2007
). While the precise mechanism(s) responsible for the ‘protective’ type of response seen here is (are) not yet known, in a general sense this effect resembles a preconditioning (Gori and Forconi, 2005
; Pespeni et al., 2005
), adaptive (reviewed in (Yu and Chung, 2006
)), or inducible-like radioprotective response (Qutob et al., 2006
), where a sublethal or potentially injurious stimulus (e.g. oxidative stress) induces tolerance to a subsequent and potentially more damaging insult (e.g. irradiation). For instance, ischemic/hypoxic preconditioning has been shown to protect the brain under some conditions (Murry et al., 1986
; Kitagawa et al., 1991
; Chen et al., 1996; for review Liu at al., 2009
) but those effects can change with age (for review Schaller, 2007
). Additionally, the same EC-SOD KO mice were shown to be more sensitive to focal cerebral ischemia injury (Sheng et al., 1999
). These and other data highlight the complexity of preconditioning or adaptive responses and highlight the fact that such responses may be context-dependent.
The relationship between EC-SOD and processes involved in learning and memory is not simple, with both over and under expression reported to impair learning, (Hu et al., 2006
; Levin et al., 1998
). Whether this represents differences in superoxide regulation, differing levels of hydrogen peroxide, components of the nitric oxide (NO) pathway and bio-availability of NO, or alterations in stress signaling pathways, needs to be determined. In the present study we used EC-SOD KO mice, which show indications of persistent oxidative stress but which do not display any compensatory changes in levels or activities of the other SOD isoforms, catalase or glutathione peroxidase (Rola et al., 2007
). Prior to irradiation these mice showed some cognitive impairments, i.e. hippocampus-dependent novel location recognition, but for other cognitive measures there were no differences observed between sham-irradiated WT and sham-irradiated KO mice. This indicated that compared to WT mice, the EC-SOD phenotype was relatively subtle, which might relate to the fact that ROS might have both positive and negative effects on brain function.
The nitrotyrosine analyses revealed two predominant nitrated proteins, of about 26 and 52 kDa, respectively. As increased levels of nitrated proteins have been shown in various neurodegenerative disorders (Castegna et al., 2003
; Sacksteder et al., 2006
), future efforts are warranted to identify these two proteins.
In this study we used a battery of cognitive tests to address hippocampus dependent function. Inclusion of multiple hippocampus-dependent cognitive tests is particularly important in comparative studies of wild-type and mutant mice, because the tests differ in the amount of training, complexity, and motivation, and might differ in their sensitivity for detecting potential detrimental or beneficial effects of irradiation. Additionally, the ability to detect potential effects of irradiation may depend upon the specific level of functioning under baseline conditions. That is, detrimental effects of irradiation may be more likely revealed when the cognitive function at baseline is intact while beneficial effects of irradiation may be more likely revealed when the cognitive function at baseline is impaired.
In the present study, WT mice showed novel location recognition both before and after irradiation. This suggested that the novel location recognition test was not very sensitive in detecting detrimental effects of gamma irradiation in male mice. The situation is different for female mice, however, where it was shown that 10 Gy of 137
Cs irradiation impaired novel location recognition in WT mice and those lacking apolipoprotein E (Acevedo et al., 2008a
). On the other hand, the novel location recognition paradigm used here was sensitive enough to detect the beneficial effects of irradiation in mutant male mice not showing novel location recognition under baseline conditions (). That is, sham-irradiated KO mice showed impairments in this test, which is consistent with impairments in hippocampus-dependent learning and memory of EC-SOD KO mice reported by others (Levin et al., 1998
), but KO mice did show novel location recognition after irradiation. Thus, this measure of hippocampal function was fully recovered following radiation exposure. These data indicate that the sensitivity of the novel location recognition test to detect effects of cranial irradiation is critically influenced by genetic and environmental factors.
We also included a hippocampus-independent version of the object recognition test. In the present investigation, both sham-irradiated WT and sham-irradiated KO mice showed hippocampus-independent novel object recognition (), that is, they spent more time exploring the novel object. In both genotypes, this behavior was impaired by irradiation, indicating that the effects of radiation on cognitive function were not limited to the hippocampus but also involved the cortex, and that pre-existing and persistent oxidative stress did not impact this response. To the best of our knowledge, this is the first time that effects of gamma irradiation on hippocampus-independent novel object recognition have been reported. These data, together with the water maze data discussed below, also highlight the importance of including both hippocampus-dependent and non-hippocampus-dependent versions of cognitive tests in the evaluation of radiation effects on brain function.
This study involved the assessment of spatial learning and memory using the Morris water maze, a test we have used previously after irradiation of adult WT (Raber et al., 2004
) and mutant mice (Villasana et al., 2006
). In our earlier study in WT male mice irradiated at 2 months of age, this test involved a single probe trial performed after 3 days of testing, and under that testing paradigm, radiation did not have any discernible effects (Raber et al., 2004
). The current results confirmed this finding in that there were no differences between sham-irradiated and irradiated WT mice when the probe trial was done after the third day of hidden platform training (). Similar results were seen in EC-SOD KO mice. However, when probe trial performance was analyzed after the first and second days of testing, there were significant differences between sham-irradiated and irradiated KO mice, and in the probe trial following the first day of hidden platform training the irradiated KO mice clearly outperformed sham-irradiated WT mice (). These data, together with our studies in mutant and WT female mice (Villasana et al., 2006
; Acevedo et al., 2008b
), highlight the strength of including multiple probe trials in the design of the water maze to detect effects of irradiation, particularly in mutant mice. Thus, a water maze paradigm including only one probe trial at the end of 3 days of hidden platform training would not have revealed the genotype-dependent effects of irradiation in the current study.
Given the complexities associated with cognitive function and the differential sensitivities of cognitive tests to detect effects of irradiation, we also used contextual fear conditioning to assess hippocampus-dependent emotional learning and memory. Sham-irradiated and irradiated mice showed genotype-dependent contextual fear conditioning. While contextual fear conditioning was impaired in WT mice following irradiation, it was enhanced in KO mice (). These data show that the contextual fear conditioning test is particularly sensitive to detect effects of cranial irradiation in WT and mutant male mice. Sham-irradiated and irradiated WT and KO both showed robust and comparable hippocampus-independent cued fear conditioning, suggesting that the effects of irradiation on contextual fear conditioning are hippocampus-dependent and not due to general impairments in fear conditioning, which is more amygdala based.
With regard to the overall radiation response, it was particularly striking that in contrast to what is seen in WT mice, performance in hippocampus-dependent novel location recognition, spatial memory retention in the first and second water maze probe trials, and contextual fear conditioning was enhanced in animals deficient in EC-SOD. This improvement was associated with genotype-dependent effects on measures of oxidative stress. In cortex and hippocampus, nitrotyrosine levels were chronically elevated in KO mice but irradiation caused a greater increase in oxidative stress in wild-type mice than in mice lacking extracellular superoxide dismutase. In addition to measures of oxidative stress, genotype-dependent effects of hippocampal neurogenesis following irradiation might have contributed to the cognitive performance. In our earlier neurogenesis study that did not involve behavioral assessments, baseline neurogenesis was lower in KO than WT mice but following irradiation, neurogenesis was strongly reduced in WT, but not in KO mice (Rola et al., 2007
). Thus the net results are that measures of oxidative stress are lower and hippocampal neurogenesis is higher in irradiated KO than WT mice. Whether or not measures of oxidative stress and/or neurogenesis play a causal or contributory role in the “protection” against radiation-induced cognitive impairments or the enhanced cognitive performance of irradiated mice when compared to sham-irradiated mice needs to be established. A new animal model that involves the conditional expression of EC-SOD now exists that may provide a novel way to address this idea (Zou et al., 2009
). Given that radiation affected the hippocampus-independent novel object recognition cognitive measure in WT as well as KO mice, it seems unlikely that neurogenesis alone is causally responsible for the improved cognitive performance seen here. Therefore, to understand the functional effects seen here, other avenues also need to be explored, including, perhaps, how specific molecular markers associated with learning and memory (Rosi et al., 2008
) are affected in animals deficient in EC-SOD. While the molecular mechanisms underlying the opposing genotype-dependent effects of irradiation reported here are not yet known, increased efforts are warranted to study potential beneficial effects associated with chronic oxidative stress in the context of a secondary insult such as irradiation. Ultimately understanding how such an effect develops may provide new insight into the evolution of cognitive injury after irradiation and provide information useful for the development of new approaches for the management of radiation-induced brain injury.