Importantly, we showed that neurogenesis occurs in the adult canine brain. shows the immunolabeling in the SGZ/GCL of the hippocampus in a young (A,C) and aged (B,D) animal. Only a subset of DCX-positive neurons contains BrdU (arrow). The confocal images (E,F,G) demonstrate co-localization of DCX and BrdU in the same new neurons.
Figure 1 DCX and BrdU immunohistochemical staining in the SGZ and GCL of the hippocampus in a young (A,C) and aged (B,D) Beagle. Only a subset of DCX-positive neurons colabeled with BrdU (arrow). Scale in A and B is 100 μm and scale in C and D is 20 μm. (more ...)
and show that there are significant age-related decreases in cell genesis and neurogenesis in the canine SGZ/GCL. The number of BrdU-positive nuclei, F(1, 9) = 74.00, p = .000 026 (), DCX-immunoreactive neurons F(1,9) = 125.66, p = .000 004 (), and double labeled DCX-BrdU positive neurons, F(1,9) = 60.65, p = .000 053 () were lower in aged animals compared to young, consistent with findings from other animal models. The proportion of BrdU cells expressing DCX was significantly higher in young adults compared to old animals, F(1,9) = 12.64, p = .007 ().
Estimated individual unilateral new neuron (DCX and DCX-BrdU) and new cell (BrdU) numbers in the SGZ and GCL of the hippocampus in young and aged Beagles.
Figure 2 Graphs plotting the numbers of (A) BrdU-positive nuclei, (B), DCX-positive new neurons, (C) double labeled DCX-BrdU-positive neurons, and (D) the proportion of BrdU cells expressing DCX in young, aged and treated-aged canines. Young animals had significantly (more ...)
Once we established that neurogenesis occurs in the aged canine brain, we then examined the potential to modify neurogenesis in the aged brain and whether neurogenesis correlated with cognitive function. First, we looked for differences between the discrete treatment groups to determine whether the food or enrichment paradigms were able to modify neurogenesis in the aged canine brain compared to aged controls. Second, we examined the relationship of the neuron and nuclei counts with cognitive performance to test the hypothesis that neurogenesis may be related to cognition as suggested by some rodent studies.
In the aged dogs of the treatment study, there were no statistically significant differences in the enrichment condition for total number of BrdU-positive nuclei, F(1, 21) = .007, p = .932, DCX-immunoreactive neurons, F(1, 21) = .238, p = .632, or DCX-BrdU labeled neurons, F(1, 21) = .095, p = .761, (). No statistically significant differences were obtained with the antioxidant fortified food condition for total number of BrdU-positive nuclei, F(1, 21) = .923, p = .349, DCX-immunoreactive neurons, F(1, 21) = .936, p = .346, or DCX-BrdU labeled neurons, F(1, 21) = 2.173, p = .158. These data indicate that the behavioral enrichment and antioxidant fortification did not have a clear measurable effect on the number of new nuclei or new neurons. All three treatment groups tended to have a higher proportion of BrdU nuclei express the neuronal marker DCX but this did not achieve statistical significance, F(1,21) = .678, p = .577 ().
Cognition: Annual Measures for All Animals
We then examined the data for a potential relationship between new cell counts and individual scores on cognitive tasks administered during the ~3 years of the study prior to euthanasia. Although it is unlikely that the new cells were present during the administration of the cognitive tasks, it may suggest that animals with higher rates of or potential for neurogenesis are the ones that perform superiorly on cognitive tasks. We hypothesized that there would be a link between the amount of neurogenesis and cognition.
shows the correlations between cell counts and errors scores for the baseline and treatment periods of the study. The table shows that more tasks significantly correlate with the counts as the treatment period progresses. The correlations begin to appear after the treatments were started and no significant correlations appeared with the baseline data. More significant correlations appear in the second year of the treatment period than the first year. The correlations are also only for neuron counts, DCX-positive neurons or double labeled DCX-BrdU-positive neurons, and not for BrdU-positive nuclei alone. The correlations are all negative indicating that error scores increase as new neuron numbers decrease. We see a consistent effect despite the strength of the association.
Correlations between cell counts in the SGZ/GCL of the aged canine hippocampus and error scores on the annual measures of cognitive function.
graphs the association between DCX () and DCX-BrdU () labeling with a black/white reversal learning task, the task administered closest to the time of death. also shows the relationship between DCX and a hippocampal dependent spatial memory task administered immediately prior to the black/white task.
Higher numbers of new neurons in the SGZ and GCL of the aged canine hippocampus correlated with lower error scores on the black-white discrimination reversal learning task (A, B) and the spatial memory task (C).
Cognition: Enrichment Measures for a Subset of Animals
Behaviorally enriched animals were provided with additional cognitive tasks and several significant correlations also appeared when only the animals given the enrichment tasks were considered (). We see consistent effects even when a smaller number of subjects was considered. Only DCX or DCX-BrdU labeled cells showed significant correlations. All enrichment tasks were conducted after the start of treatment.
Correlations between cell counts in the SGZ/GCL of the aged canine hippocampus and error scores on cognitive tasks administered to the behaviorally enriched animals only.