Share this article on:

Astrocytes are the Key to Understanding Marijuana's Effect on Memory

Robinson, Richard

doi: 10.1097/

Marijuana's effect on short-term memory is mediated through cannabinoid receptors on astrocytes, not neurons, according to a March 2 paper in the journal Cell. The new research both answers a key question in the biology of the drug's effects, and highlights a novel and central role for astrocytes in one of the brain's most important functions.

“We've known for a long time that marijuana has a strong impact on certain forms of memory,” explained Giovanni Marsicano, PhD, researcher and team leader at the Neurocentre Magendie (INSERM) at the University of Bordeaux in France, “but we didn't know the mechanism.” The worsening of short-term memory is a significant effect, he said, and is one barrier to therapeutic applications of cannabinoids in neurologic conditions such as pain.

Receptors for the body's own endocannabinoids exist on multiple cell types in the brain, including on hippocampal neurons and astrocytes, but their concentration on astrocytes is much lower than on neurons. THC, the main psychoactive ingredient in marijuana, activates cannabinoid receptors, including the type-1 receptor (CB1R). Activation of CB1R at hippocampal synapses induces long-term depression (LTD) of synaptic firing, which has been thought to be responsible for the memory impairment. But through which cells does THC exert its effect?



Back to Top | Article Outline


To determine the cellular site of THC's effect, Dr. Marsicano and colleagues began by knocking out in a rat model CB1R in the glutamatergic neurons within the hippocampus. They found that the ability of THC, or a synthetic cannabinoid HU210, to cause LTD in rats was unchanged; the same occurred when he knocked out CB1R in GABAergic neurons.

But then he knocked out CB1R in astrocytes, achieving a 79 percent reduction in the number of CA1 astrocytes expressing the protein, without affecting expression in neurons. While THC induced LTD in control animals, it was unable to do so in those missing CB1R on their astrocytes.

One effect of cannabinoid receptor stimulation on astrocytes is to increase extracellular glutamate. The investigators reasoned that if THC was inducing LTD through this pathway, then LTD should also be triggered by a glutamate reuptake inhibitor, and they next confirmed this was so.

What was the site of glutamate's effect? It was not at post-synaptic metabotropic glutamate receptors, because blocking them had no effect on CB1R-induced LTD. But LTD was blocked by an antagonist of glutamatergic NMDA [N-Methyl-D-aspartate] receptors, specifically those receptors containing the NR2B subunit.

NMDA receptors have been known to have a role in facilitating LTD, Dr. Marsicano explained, through their effect on post-synaptic AMPA (alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) receptors. Activation of NMDA receptors causes endocytosis of AMPA receptors, and by blocking that endocytosis, Dr. Marsicano blocked CB1R-induced LTD in rats.

Long-term depression is a cellular electrophysiological phenomenon, while memory impairment is a behavioral one. To show that the cellular pathway he uncovered was responsible for the memory impairment, the investigators examined rats in a T-maze that required them to remember where a reward was hidden on the last trial, and then make the opposite choice to get their next reward. Wild-type rats perform worse on this test after exposure to THC, reflecting the decline in short-term memory. But when the team knocked out CB1R on astrocytes, THC was incapable of impairing short-term memory. Similarly, the effect of cannabinoids on memory in wild-type rats was eliminated when they received either a drug that antagonized NR2B/NMDA receptors, or blocked internalization of the AMPA receptor.

Thus, the investigators concluded, THC's effect on short-term memory is due to its activation of CB1R on astrocytes, which then trigger an increase in extracellular glutamate and endocytosis of AMPA receptors, causing long-term synaptic depression in hippocampal neurons.

“We think these results are important for two reasons. First, they indicate how marijuana affects short-term or working memory,” Dr. Marsicano said. Whether cannabinoid receptors play roles in other types of memory remains to be seen, although the ubiquity of the receptors argues for more widespread functions.

“We are also now trying to understand whether the other effects of marijuana work with the same mechanism,” Dr. Marsicano said. If a cannabinoid could be found that avoided THC's memory effects, while still delivering its beneficial effects elsewhere in the brain, it may allow cannabinoid receptor targeting to become more valuable therapeutically. “We don't yet have such drugs, but I think this is the direction for pharmacology to go.”

“Second, the results also point out the importance of astrocytes in memory. Until a few years ago we thought astrocytes were mainly important for support, protection, and providing nutrition to neurons. This paper is one of the first showing that they play a very active role in the processing of memory,” Dr. Marsicano said.

Ultimately, he said, the goal is to understand the function of the brain in all its complexity. Due to limitations of current experimental approaches, “we've been more or less treating the brain as a liver, in the sense that we think of it as composed of only a limited number of types of cells,” Dr. Marsicano said. But, due to plasticity, “the brain is made up of billions of different types of cells,” each with its unique history and behavioral repertoire. “The brain's complexity is driven by the enormous number of kinds of cells involved,” he added.

Back to Top | Article Outline


Cristina Alberini, PhD, professor of neural science at New York University in Manhattan, echoed both points. “This is an interesting new mechanism, both in how cannabinoids are involved in plasticity and memory, and in how the presence of their receptors on astrocytes is required,” Dr. Alberini said. “We are still biased in thinking that whatever happens in the brain occurs in neurons and neural networks. And we forget glia are a very important part of learning. I think this opens a new direction to explore a number of mechanisms in astrocytes.”

Dr. Alberini reported last year in Cell that long-term memory requires the movement of lactate from astrocytes to neurons. Studies in dissociated neurons or dissociated astrocytes “are not going to give a complete picture of what they do together,” she said. “It's very important to isolate the single components to gain a better understanding of their individual functions, but then it's very important to look in vivo at how they work together.

“Fortunately we are seeing more and more that astrocytes can contribute to complex functions, including memory,” a realization that is likely to become even more prominent as more is learned about these cells. “I think we are becoming more aware of their complexity. We are talking about systems biology.”

Listen here as Dr. Giovanni Marsico describes his work on marijuana's effect on short-term memory as mediated through cannabinoid receptors:

Back to Top | Article Outline


• Han J, Kesner P, Zhang X, et al. Acute cannabinoids impair working memory through astroglial CB(1) receptor modulation of hippocampal LTD. Cell 2012;148(5):1039-50.
    • Suzuki A, Stern SA, Alberini CM, et al. Astrocyte-neuron lactate transport is required for long-term memory formation. Cell 2011 Mar 4;144(5):810-23.
      ©2012 American Academy of Neurology