Selective changes in thin spine density and morphology in monkey prefrontal cortex correlate with aging-related cognitive impairment

Abstract

Age-associated memory impairment (AAMI) occurs in many mammalian species, including humans. In contrast to Alzheimer's disease (AD), in which circuit disruption occurs through neuron death, AAMI is due to circuit and synapse disruption in the absence of significant neuron loss and thus may be more amenable to prevention or treatment. We have investigated the effects of aging on pyramidal neurons and synapse density in layer III of area 46 in dorsolateral prefrontal cortex of young and aged, male and female rhesus monkeys (Macaca mulatta) that were tested for cognitive status through the delayed non-matching-to-sample (DNMS) and delayed response tasks. Cognitive tests revealed an age-related decrement in both acquisition and performance on DNMS. Our morphometric analyses revealed both an age-related loss of spines (33%, p < 0.05) on pyramidal cells and decreased density of axospinous synapses (32%, p < 0.01) in layer III of area 46. In addition, there was an age-related shift in the distribution of spine types reflecting a selective vulnerability of small, thin spines, thought to be particularly plastic and linked to learning. While both synapse density and the overall spine size average of an animal were predictive of number of trials required for acquisition of DNMS (i.e., learning the task), the strongest correlate of behavior was found to be the head volume of thin spines, with no correlation between behavior and mushroom spine size or density. No synaptic index correlated with memory performance once the task was learned.

Figure 1.

Dendritic segment analysis for spine quantification. A, Projection of a raw confocal z-stack of an imaged dendrite. B, Dendritic segment after deconvolution. The image is deblurred by accounting for the point spread function in our system. Note that the spines no longer appear fuzzy, and fine morphological details are now evident. C, Automated analysis using NeuronStudio. 3D spine size measurements including head diameter, head volume, and maximum distance from the dendrite are extracted, along with dendritic length. D, Spine indices are imported into Matlab. Spines are classified based on size: a thin spine is smaller than 0.6 μm in diameter and has a maximum length at least twice the diameter; a mushroom spine is >0.6 μm in diameter; the remainder of spines are classified as “other.” Scale bar, 2 μm.

Figure 2.

Electron micrographs illustrating the disector method for synapse density. For synapse density measurements, two serial sections were used. A, B, One section was considered the reference (A) and the other the look-up (B). Only asymmetric axospinous (yellow arrows) or axodendritic (yellow arrowheads) synapses that are present in the reference but not the look-up were counted and divided by the micrograph area and section thickness. This procedure was repeated by switching the reference and the look-up. Synapses present in both the reference and the look-up are marked with purple arrows (axospinous) and arrowheads (axodendritic). Split arrows indicate perforated synapses. Scale bars, 1 μm.

Figure 3.

Cognitive performance on DNMS but not DR is impaired in aged monkeys. A, Aged monkeys perform similarly to young monkeys in the DR task. B, Aged monkeys performed worse in DNMS recognition on four of the five delay intervals tested. *p < 0.05, **p < 0.01.

Figure 4.

Spine density of layer III pyramidal neurons in area 46 of dlPFC is significantly lower in aged monkeys. A, Representative projected z-stack of a dendrite from a young animal. B, Representative projected z-stack of a dendrite from an aged animal. Scale bar, 2 μm.

Figure 5.

Synapse density in area 46 of dlPFC is significantly lower in aged monkeys and strongly correlates with spine density. A, Quantitative 3D EM counts of synapses from six young and nine aged monkeys. B, Scatter plot of EM synaptic counts versus confocal spine counts for the nine animals for which both types of analyses were performed shows that the two methods produce highly consistent results. r, Pearson correlation coefficient; **p < 0.01.

Figure 6.

Average spine head diameter (d) and volume is significantly increased in aging, which is attributable to a selective loss of thin spines. All statistics were performed based on one aggregate (i.e., average) measure per animal. A, Mean spine head diameter is significantly greater in aged monkeys. B, Mean spine head volume is significantly greater in aged monkeys. C, There is no age-related difference in the average spine maximum distance from the parent dendrite, a measure that approximates the length of the spine (l). D, The cumulative frequency plot of each spine head diameter from individual animals shows that for small spines (∼0.3–0.4 μm diameter) the group data are highly unified. E, Densities of subtypes of spines based on an unbiased classifier show that there is a selective loss of thin spines in aging. *p < 0.05, **p < 0.01, ***p < 0.001; n.s., no significance.

Figure 7.

The spine head diameter and volume of thin but not mushroom spines are increased in aging. A, Mean spine head diameter of thin spines is significantly increased in aging with no change to mushroom spine diameter. B, C, Cumulative frequency plots of individual spines show that animals segregate based on age for thin spine head diameter (B) but completely overlap for mushroom head diameter (C). D, Mean spine head volume of thin spines is significantly increased in aging with no change to mushroom spine volume. E, F, Visual inspection of the cumulative frequency plots of individual spines indicates that animals segregate based on age for thin spine head volume (E) but overlap for mushroom head volume (F). **p < 0.01, ***p < 0.001.

Figure 8.

DNMS acquisition significantly correlates with synaptic indices, in particular thin spine head volume. A, The scatter plot of DNMS acquisition (trials required to reach 90% accuracy with a 10 s delay) versus EM synaptic density shows a moderate though significant inverse correlation, meaning that higher synaptic density is predictive of faster learning. B, This inverse correlation is somewhat strengthened when the density of only thin spines is used on the subset of animals for which spine density analysis was performed (r = −0.58 in A and r = −0.69 in B; p < 0.05 for both). C, DNMS acquisition correlates most strongly with the mean volume of thin spines, where bigger volumes are predictive of slower learning. D, In contrast, no correlation is seen between learning of the DNMS task and mean mushroom head volume. r, Pearson correlation coefficient.

Figure 9.

Summary of age-related synaptic changes in layer III of area 46 of monkey PFC. There is a marked reduction of thin spines (46%, p = 0.002) in the brains of aged as compared to young subjects. In particular, it is the smallest of thin spines that are lost, the spines presumed to be the most plastic in response to learning (Kasai et al., 2003; Bourne and Harris, 2007). The decline in DNMS acquisition with age occurs in tandem with the synaptic changes and is particularly highly correlated with mean thin spine volume. In contrast, no age-related change in either density or morphology of mushroom spine has been observed.