Saturday, February 15, 2020

Cells That Eat Memories

Microglia dispose of synapses that need to be forgotten.

Memories are potent, essential, abstract, sometimes maddeningly ungraspable. But they are also physical entities, stored like so many hard-drive magnetic domains, in our brains. The last decades of research have found the "engram"- the physical trace of memory in our brains; neural synapses that connect based on the convergence of neuron activities, are painstakingly consolidated through re-enactments, and judged for permanence based on their importance. Engram cells are located in many places in the brain, supporting many kinds of memory, including memories made up of all sorts of experiences, such as multiple modes of sensation. The coordinated action of neurons, combined with emotional valence, sets up an engram event, and similarly coordinated activity at some future time can prompt its readout / recall.

A recent paper described the role of microglia- the scavangers and maintainers of the brain, in cleaning out engram synapses, resulting in forgetting. The researchers set up a typical fear response training system (by electrical foot shocks) in mice, which was remembered better after 5 days than 35 days. Treating the mice with various drugs and other methods that deplete microglia caused them to remember the bad experience much better after 35 days, indeed, virtually unchanged from the 5-day time point. The cells behind these memories are, at least in part, in the dentate gyrus, a part of the hippocampus.

Engram cells. Mice with specialized genetic elements were tracked during training, when cell activation was shown by red fluorescence (neuronal activity plus tamoxifen-activated genetic recombination leading to red fluorophore expression). Later, during memory recall, another genetic system (c-Fos expression, green) was used to track neuronal cell activation. The merging and superposition of both colors locates cells that fulfill the criteria of engram cells, i.e. memory cells. The scale bar is 20 microns.

How do the microglia do this? Well, the most important question is how stale and unimportant synapses are identified for destruction, but this paper doesn't go there quite yet. They ask instead what the microglia are doing to gobble up these synapses. They use complement, which is a tagging system commonly used in the immune system to mark cells and other detritus for destruction, and is also used during early brain development to prune vast amounts of excess neurons and synapses. So this is an obvious place to look for continued dynamic pruning during adulthood. Application of inhibitors of complement in these mice caused the same enhanced remembering of their bad experiences as did the inhibition of microglia generally. Fluorescence studies showed that the initial complement cascade component, C1q, is present right at the synapses that were previously trained and presumably are elements of engrams.

But what of the selection process? The researchers devised a way to selectively inhibit the activity of the trained neurons, by adding an inhibitory protein to the genetic engineering cocktail, thus damping activity of those cells during the 35 days post-training. This treatment enhanced forgetting by one-third, while co-treatment at the same time with the drug that depleted microglia brought memory performance back to maximal levels. Thus the activity of neurons involved in engrams is important, as one would expect, to maintain memories, while the microglial disposal system is responsive to whatever system is marking inactive engram / memory synapses for removal.

This is interesting work in a very exciting and young field, starting to put meat on the bones of our knowledge of what memories are in physical terms, and how they grow and fade. Much like other tissues like muscles and bones, which are constantly regenerating and adjusting their mass and strength in response to loads, the brain dynamically responds to use as well- a lesson for those heading into older age.