As I mentioned in my summary of the meat and potatoes of Perez-Cruz and friends’ paper, this group largely reproduced previous findings, but they also added a new dimension of depth to those findings. Another attractive aspect of the paper was its simplicity, and although this was partially its downfall, it is nice to see important findings made elegantly and without extremely complicated procedures and analysis.
There were two main instances in which I was impressed by the the simplicity of the paper. The first is that it used very basic techniques to describe a very basic effect of the two mutations on spine density. Perez-Cruz and friends used a technique for staining neurons that was discovered by Camillo Golgi in 1873. Although this technique has been modified over the ensuing century and a half, and is now the Golgi-Cox method, the gist remains – it is a staining protocol that allows for strong labeling of between 1 and 10% of neurons in a histological slice. Cellular neuroscience started with Golgi staining almost 150 years ago, and since I am nostalgic, it’s nice to see that it is still useful. It is also nice to see that it was useful in identifying such a pronounced and important, but again, basic effect; namely, the loss of spines. Perez-Cruz and friends used Golgi staining to assess spine loss on pyramidal neurons in area CA1 of the hippocampus. Rather than simply looking at one area of the dendritic tree of these cells though, they assess 4 different dendritic subregions: subregions of basal proximal and distal to the soma, and subregions apical dendrites proximal and distal to the soma. The importance of considering these different subregions came out in the results, as they found that both mutants showed spine loss specifically on proximal basal dendrites (35-45% loss), and around 30% spine loss on proximal apical dendrites only in the London mutant. While many people, including myself, tend to overlook the differences between dendritic subregions, it is clear that different dendritic subregions in the hippocampal receive inputs from separate brain areas and neuronal subpopulations, so kudos to this group going the extra mile.
The second simplistic point comes when the group links the spine loss phenotype to memory deficits, albeit in an unavoidably correlative way. Perez-Cruz and friends find that the Swedish mutant has a deficit in contextual fear condition, while the London mutant has a deficit in the Morris Water Maze. Wild type mice froze much more often in the fear conditioned context than the Swedish mutants, indicating that the mutants have a deficit for contextual fear conditioning. On the other hand, the London mutants didn’t have a contextual fear deficit. They did, however, have a deficit in the Morris Water Maze test, exhibiting a significantly longer latency to find a submerged platform in a small pool of water after 9 days of training. These data indicate a memory deficit as a result of the mutations. One issue with these experiments is that although each mutant strain showed a deficit in one task, they didn’t have a deficit in the other. Perez-Cruz and friends chalk this up to the fact that the mutants are in different genetic backgrounds, so it would have been nice to see the tests done in the same backgrounds – although I appreciate the time it takes to back-cross a mutation into an alternative background. However, I also have a soft spot for the mutation- and strain-specific discrepancies of these memory tests because it emphasizes the complexity of memory to the point where people just have to be truthful and say we don’t understand why this is the case. Honesty in science is sometimes a triumph in itself. Regardless though, both mutants show some type of memory deficit associated with the spine loss.
After examining these basic phenotypes, Perez-Cruz and friends dig a bit deeper. It is well known that over-excitation in neural circuits is one cause of spine loss. For instance, decreased spine density is a well characterized consequence of epileptic seizures. So, given the spine loss that they see, this group wanted to look at activity levels in their mutant mice. Rather than assessing activity directly though, they look at the expression levels of a protein called Arc (activity-regulate cytoskeletal-associated protein) , which as you can tell from the name, is regulated by activity levels. Specifically, the higher the activity, the more Arc that is expressed. Indeed, Perez-Cruz and friends found that both the Swedish and London mutants expressed higher levels of Arc in their hippocampi, regardless of whether environmental enrichment and memory testing done on them. This indicates a higher level of activity in the mutant hippocampi, and likely overly high activity since the Arc increase is concomitant with spine loss. One problem with their assessment of Arc expression though, was that they only performed low magnification analysis of Arc immunostaining, quantifying their DAB staining over an entire section of CA1. This is not the most accurate staining analysis, and it would have been more interesting to see the subcellular localization of Arc, since it is located at individual synapses.
However, Perez-Cruz and friends go on to look for an explanation of the over-activity that likely led to the spine loss and Arc upregulation. They posit that since inhibitory interneurons are known to limit the level of excitation in the hippocampus, if some of them died, you would expect an increase in activity levels. Indeed, some groups have shown that by pharmacologicaly inhibiting these interneurons, you can induce a robust decrease in spine density. So, the group assessed the number of inhibitory interneurons in the hippocampi of their Swedish mutants (oddly they didn’t assess the London mutants) and found that compared to controls, the mutants had fewer interneurons in the area surrounding the basal dendrites of the CA1 pyramidal cells (recall, this is where the group saw the majority of the spine loss).
So in the end Perez-Cruz and friends present a pretty complete story. They first show that both the Swedish and London mutant models of Alzheimer’s have spine loss in specific dendritic subregions of CA1 pyramidal cells. They link this spine loss to memory deficits. They then find that the spine loss is likely correlated with increased activity levels in the hippocampus, and finally indicate that this increased activity level may be due to a loss of inhibitory interneurons. Based on this last point, the group suggests that by protecting this interneuron population we may be able to mitigate the spine loss that likely gives rise to cognitive deficits associated with Alzheimer’s, but I must admit I can’t figure out exactly how we might do that, particularly since we don’t seem to know why there are fewer of these interneurons in the mutant mice. Discovering more targets for potential therapeutics certainly isn’t a bad thing though.
Although the experiments and results were fairly simplistic, Perez-Cruz and friends’ discussion of their findings is decidedly more complex. The discussion goes in depth into how these findings might possibly fit into the expansive field of Alzheimer’s research, so if you are interested I would highly recommend checking out the paper itself.