CINPLA-forskere: Mattis Wigestrand, Kristian Lensjø, Elise Holter Thompson, Marianne Fyhn, Torkel Hafting, Anders Malthe-Sørenssen.

De fant noe nytt i hjernen: Mattis Wigestrand, Kristian K. Lensjø, Elise Holter Thompson, Marianne Fyhn, Torkel Hafting og Anders Malthe-Sørenssen. Foto: Bjarne Røsjø, Titan / UiO.

Long-term memories protected by tiny "fishnets" in the brain

Neuroscientists from the University of Oslo are the first in the world to discover that structures on the outside of neurons play an important role in storing long-term memories. The structures look like small fishnets tightly wrapped around neurons, and long-term memories are lost if they are removed.

It is widely accepted that the brain creates and stores memories because of physical changes in synapses, the connecting points between neurons. A memory is stored through changes in a whole network of synapses; and the network is specific for each memory. These changes can remain stable for such a long time that people can remember things as far back as 80-90 years.

Perinevralt nett rundt en hjernecelle

Researchers from the Center for Integrative Neuroplasticity (CINPLA) at the University of Oslo have now shown that structures on the outside of neurons are important for memory. The last of three articles on the subject was recently published in the renowned journal Proceedings of the National Academy of Sciences and it has created quite a buzz.

You can also read this article in Norwegian

The researchers have studied proteoglycans, a type of protein structure that is connected to a large amount of sugar chains. The structure forms a net around neurons in some parts of the brain in rodents. There are small holes in the nets, giving it the “fishnet” appearance. The holes are there to make room for the synapses, which make up the connections to other neurons. The net therefore ensures the size and placement of the connections that are so important for memory.

– The correct terminology for these “fishnets” is perineuronal nets. The prefix “peri”, meaning around, is used, as these are nets that wrap around the outside of neurons. We have now shown that long-term memories are disrupted when these nets are broken down, says Marianne Fyhn, leader of CINPLA.


The researchers had an “Eureka-moment” in March 2015, when the then master student Elise Holter Thompson was looking at videos of rats going through a memory test and studying their behavior.

– The task was quite tedious actually, but suddenly I discovered something very strange. I had to rewind the video a few times to be sure, before I asked Kristian Lensjø and Mattis Wigestrand to come and have a look. They too looked at the video several times, and then we all agreed that we had found something cool, states Elise Holter Thompson – who now is a PhD student at CINPLA.

Marianne Fyhn

– The rats Elise was studying had learned to associate a light blink with an unpleasant event. This type of learning creates very robust and long-lasting memories. After learning, the rats were divided into two groups: in one group the perineuronal nets were removed, while the other group was a control were the nets were left intact. What we discovered that day in 2015 was that one group was acting very differently, explains Lensjø.

– The rats with intact perineuronal nets had a quick and distinct reaction when the light came on, because they remembered that the light signaled something unpleasant. However, the rats without this “fishnet” around their neurons, were not fazed by the light at all – it seemed as though they had forgotten what they had learned one month earlier, explains Mattis Wigestrand, who has a background working in the US with leading research groups within the field of behavioral studies.  

– To achieve setting up a behavioral study where we are testing a specific memory as late as one month after learning, is very difficult. That made it extra special when we accomplished it, says Wigestrand. 

The idea came from a Nobel Prize winner

Nobelprisvinneren Roger Tsien

– But how did you come up with the idea to remove these perineuronal “fishnets” in a rat brain, and how is it done?

– We had the idea when reading an article written by the american biochemist Roger Tsien, who received the nobel prize in chemistry in 2008. In the article he writes that these structures outside the neurons could potentially play a major role in long-term storage of memories, based on the fact that the structure is made up of molecules that are much more stable than molecules inside neurons. We wanted to explore this idea, replies Torkel Hafting.  

– I was introduced to Tsien at a conference in San Diego in March 2016. I explained to him that we had tested his idea and I showed him our results, he was very excited about our findings. I am sorry to say that he died just a few months later, Marianne Fyhn says.   

Marianne Fyhn adds that the technique behind how the nets are removed was developed by Professor James Fawcett at the University of Cambridge and his Italian colleagues. Their first work showed that removing the nets made the brain more plastic, meaning more adaptable and capable of physical changes.

– The collaboration happened through Gunnar Dick, who had been a part of the proteoglycan group at the Institute for biosciences here in Oslo and was a postdoc in the Fawcett lab at Cambridge. The focus in Cambridge was on spinal cord injury in rats. It is common after this type of injury that scar tissue composed of structures similar to perineuronal nets builds up to limit the boundaries of the injury. However, this scar tissue inhibits neurons from growing past the site of injury and must be removed to allow neuronal regrowth. Here in Oslo we have studied the nets in the brain and how they impact plasticity and memory, Fyhn explains.

 – The British scientists found out that the perineuronal scar tissue could be removed by injecting an enzyme – chondroitinase – that cuts off the sugar chains that holds the nets together. After this enzyme treatment, the neurons can start to grow through the site of the injury. The scientist came quite far in developing a possible treatment for human spinal cord injury, but the pharmaceutical industry was not willing to take it through clinical trials, Fyhn says. 

Old brains turned young again

The Eureka moment in 2015 lead to a long series of control experiments and closer investigations, which confirmed that the «fishnet» structure really affects the long-term storage of memories.

- Memories of recent experiences were not affected. Such a new memory might not need stabilizing external factors. It is also important to note that newly formed memories depend on other brain areas than the ones we have studied, says Kristian K. Lensjø.

The CINPLA scientists have used rats as their brains are fairly similar to ours, so that new knowledge from the rats’ brains can provide information about how the human brain works. Marianne Fyhn explains that the perceived “holes” in the net structure is actually where synapses reside, the connections between the cell body and its connected neurons.  

– Young individuals, both in rodents and humans, are able to learn faster than adults, and the young brain has very few perineuronal nets. But at a certain age, as we go from adolescence to adulthood, the nets are formed, probably to stabilize the neurons. The disadvantage is that is becomes more demanding to learn something new, says Lensjø.

Luckily, the brain maintains a certain ability to learn even after the nets are formed. The neurons produce enzymes that are capable of shaping the structure of the perineuronal nets.

– When the brain needs to store a new and strong memory, the synapses need to grow, to accommodate this growth the brain uses its own enzymes to expand openings for the synapses, explains Torkel Hafting.

The group’s first paper about the perineuronal nets was published in February 2017, and showed that injecting chondroitinase in the brain of adult rats transformed their brains into being similar to young rats. The neurons which are normally enwrapped in a net did not function properly as they would in a normal adult brain, but the brain area that was targeted became better at learning something new.

The nets are not evenly distributed

The next paper was published in May 2017, and showed that the perineuronal nets are unevenly distributed across the brain. They are most prominent in the cortex, which is also where long-term memories are stored. In the small structure called the hippocampus, there are very few.

Hippocampus - illustrasjon fra Wikipedia

– This makes sense as the hippocampus is an area that constantly processes and combines new impressions, but is likely not responsible for storing memories over long periods of time, Fyhn notes.

The third paper was published over Christmas in 2017 and revealed the most astonishing finding: that the fishnet surrounding neurons in the cortex are crucial in stabilizing long-term memories.

– We removed something outside the neurons, which does not directly influence the communication between the neurons. We were therefore very surprised by how strong the effect was in those first experiments, says Mattis Wigestrand.  

Links to Alzheimer’s disease and schizophrenia

The CINPLA researchers hope that new knowledge about the brain will increase our understanding of diseases, and over time contribute to new treatments for conditions such as Alzheimer’s and other diseases. The perineuronal nets are also linked to schizophrenia, and the scientists are already looking into this.

– It has been shown that schizophrenia patients have abnormal perineuronal nets. Schizophrenia is a disease that often does not appear until the late teenage years, at the same time as the nets usually form. But it is important to remember that these reports are based on autopsies of patients who have usually been on a lot of medication for years, which could have influenced the nets, says Fyhn.

– What we find intriguing is that the nets are mainly found around the type of neuron which generates a specific type of high-frequent brain oscillation – this oscillation is abnormal in schizophrenia patients.

The CINPLA group has, together with collaborators at Oslo University Hospital and others, started a large new project called DigiBrain, which will study the possible causes to the differences in brain oscillations by using EEG measurements in healthy and diseased humans and in animal models. By combining genetics, measurements in humans and animals, and mathematical models they hope to uncover some of the mechanisms underlying the disease.

The perineuronal nets also appear to be linked to Alzheimer’s disease. The neurons which contains plaques, typically associated with the disease, have more nets in some animal models. Recently published work has shown that removing the nets in mice with plaque formation has a positive influence on memory function.

– This could be a dead end, as a lot of work indicates that the plaques are not the cause of Alzheimer’s disease. But there is still reason to believe that nets could be involved in some of the processes that negatively affect the neurons, adds Fyhn.

Surprised by how much neuroscientist already know

Anders Malthe-Sørenssen is a professor of physics who, together with Marianne Fyhn, established CINPLA in 2014. He has previously done pioneering work in the field of geophysics, but is now also using his experience from computational physics to study the brain.

– When I entered this field a few years ago, I was surprised by how much was already known – for instance the detailed knowledge that exists about how neurons communicate through electrical signals. But there is still a long way to go before we understand all the molecular processes that occur inside the neurons. And perhaps an even greater challenge is to understand how the vast networks of neurons function.

Several large projects in the making

COBRA-forskere: Anders Malthe-Sørenssen, Marianne Fyhn, Gaute Einevoll

The CINPLA researchers have several large ongoing projects, including Brain Matrix, an interdisciplinary project funded by the Research Council of Norway (RCN). The project is novel in the sense that physicists and mathematicians work together with biologists to investigate how the brain works. In the project they aim to investigate what the nets are actually doing and how they are regulated, all the way from single molecules to during learning tasks. While most brain researchers have a background in biology or medicine, a surprising number of physicists are world leading in neuroscience and other parts of the life science disciplines.

Together with another physics professor, Gaute Einevoll, they have also started the RCN funded project Computing BRAin signals (COBRA). In the COBRA project, Einevoll, Malthe-Sørenssen and Fyhn will build a digital model of a small piece of a mouse brain, and simulate the activity in a network of neurons in this region. The aim is to calculate what kind of EEG signals the neurons’ activity gives rise to on the outside of the brain, in order to understand better what one actually measures with EEG recordings which are widely used in clinical investigations of patients.

New research on mental disorders

The researchers at UiO also has a fourth project, named DigiBrain. Here, Fyhn and her colleagues at IBV collaborate with several groups including professor Ole Andreassen at the Norwegian Center for Mental Disorders Research, who has extensive experience and resources in relation to diseases such as schizophrenia and bipolar disorder.

– Our recent findings are important early steps to understand how the brain processes memories, but it’s important to remember that so we have only described a phenomenon. The next step will be to search for the molecular mechanisms that underlie this phenomenon. We also have high expectations for the work that aims to model the processes that go on in larger parts of the brain, which complements the experimental work that we are doing, says Fyhn in conclusion.


Scientific papers:

E.H. Thompson, K.K. Lensjø, M.B. Brænne Wigestrand, A. Malthe-Sørenssen, T. Hafting and M. Fyhn: Removal of perineuronal nets disrupts recall of a remote fear memory. PNAS, December 26, 2017.

K.K. Lensjø, A.C. Christensen, S. Tennøe, M. Fyhn and T. Hafting: Differential Expression and Cell-Type Specificity of Perineuronal Nets in Hippocampus, Medial Entorhinal Cortex, and Visual Cortex Examined in the Rat and Mouse. eNeuro 25 May 2017.

K.K. Lensjø, M.E. Lepperød, G. Dick, T. Hafting and M. Fyhn: Removal of Perineuronal Nets Unlocks Juvenile Plasticity Through Network Mechanisms of Decreased Inhibition and Increased Gamma Activity. Journal of Neuroscience, 1 February 2017.

More on