May 26 2022

How Memories Are Linked

Memory research, both at the psychological and neurological level, is fascinating, partly because memories are so essential to who we are. We often don’t perceive the underlying mechanisms by which memories are formed, stored, and recalled, but they dramatically affect our mental life. Further, being aware of how our memories work is a critical part of neuropsychological humility – human memory is not perfect, it’s not like a tape recorder, it is a dynamic and flawed process. For example, different aspects of a memory and different memories are linked together, but these connections can be jumbled. We may fuse the detail of one memory to another, alter details completely, and even remember things that happened to someone else as if they happened to us.

This is partly because memory did not evolve to be a perfect recorder of our life experiences, but rather to create a meaningful and adaptive narrative of our past. One critical component of the adaptive nature of memories is that our brains can link different memories together, because they apparently have a meaningful connection. A recent study tries to elucidate one aspect of this process, and as a result may have turned up a clinically useful bit of information.

For a little background, our brains function as massively parallel processors. One of their core functions is to make associations between different things – when we remember one thing that triggers other memories, including details about the original memory but also other memories. Our brains are association machines. This is not only intrinsic to how we remember but how we think. Much of literature, storytelling, metaphor, and creativity derives from association. How does this work at a neuronal level? It’s safe to say that it’s complicated, but the researchers were trying to elucidate one tiny piece of the picture.

They studies rats who were learning how to navigate different cages, where knowledge of one cage provides an advantage in the other. This way they could study the linkage of the memories of the two cages together. They focused on the expression of the C-C chemokine receptor type 5 (CCR5), an immune receptor that is expressed in the brain. They found that an increased expression of CCR5 lead to a decrease in the ability to link the two memories in the test rats. They then knocked out the CCR5 gene in one group of rats, and they had a significantly improved ability to link the two memories. So binding of the CCR5 ligand to the receptor seems to inhibit memory linkage.

Why would an immune receptor block memory linkage? First of all, evolution is frugal, it uses what it has on hand and it repurposes biological elements rather than inventing new elements entirely (meaning that such things are more likely to happen at random). But still, why would there be a receptor in the brain that blocks memory linkage? Well, how does the brain “decide” to link memories in the first place? Not all memories are equally linked to all other memories. That would be worthless. The linkage has to be meaningful. The primary criterion that the brain uses to determine how much to link two memories is temporal association – events that happen together are linked together. This makes sense, but there has to be a mechanism. Also – how close in time do two events have to occur before they become linked, and how is that critical timing regulated?

These are all good questions that will have to be explored, but what these researchers found is that CCR5 plays a role in suppressing linkage. To a neurologist this makes perfect sense – the brain largely functions through inhibition. In order to have the correct neurons firing you have to suppress all the unwanted neurons. Here, in order to link memories that have to be linked, you also need to suppress the linkage of memories you don’t want linked. What increasing the expression of CCR5 actually did was narrow the time window in which memories would be linked.

There are a couple of additional facts important to the story. One is that expression of CCR5 increases with age. This may therefore be a mechanism by which our ability to form and link memories decreases at we get older. Therefore if we somehow suppress CCR5 activity we may be able to treat declining memory in older age, and maybe even some types of dementia. It turns out, there is already an FDA approved drug that blocks CCR5. Why? Because HIV uses the CCR5 receptor to help it infect immune and brain cells. The drug maraviroc was developed and approved as an anti-HIV drug, which works by preventing HIV from entering cells. This is extremely helpful because it’s much easier to do research on an already approved drug. You can skip all the required safety data and preliminary research on pharmacology and get right to the clinical trials looking at efficacy. If it works, the drug can then be used immediately off-label. The company can then decide if it wants to go for an additional FDA approved indication (which is essentially a marketing decision).

It’s obviously too early to tell if maraviroc or a similar drug will have clinically meaningful effects for memory in normal aging. We need to do the clinical research to tell, but it is promising (which statistically is not worth as much as you might think, but it’s something). Or perhaps researchers will elucidate further aspects of this memory linkage system that will provide further possible clinical targets. This is how we connect basic science to translational science. It’s reductionist and mechanistic, but it works.

No responses yet