Mounting evidence suggests that immune system molecules can wear different hats in the central nervous system—as mediators of learning, memory, and synaptic plasticity. In support of this, scientists led by Mark Mattson at the National Institute on Aging report in the August 16 PNAS online that toll-like receptor 3 (TLR3), one of a family of immune-related receptors, appears to dampen hippocampal-dependent learning and memory in normal mouse brains. Mattson and colleagues examined a TLR3 knockout mouse, and found that the animals demonstrate enhanced memory retention and working memory in several behavioral tests. Surprisingly, TLR3 appeared to mediate these effects, not through its known, immune-related signaling pathway, but through extracellular signal-regulated kinase (ERK) and cAMP response element binding protein (CREB), molecules known to be involved in memory. The results suggest a previously unknown role for TLR3 in inhibiting synaptic plasticity and memory, and add yet another candidate to the list of possible therapeutic targets in memory disorders.
This paper provides “another piece of evidence that immune-like molecules can be involved in behavior and in learning,” said Paul Patterson at the California Institute of Technology in Pasadena. He investigates interactions between the nervous and immune systems, and said that other examples of immune molecules with roles in brain function include major histocompatibility complex proteins, interleukins 1 and 6, and some complement proteins (see ARF related news story). Patterson said the new findings suggest that TLR3 may have a constitutive function in the brain that is entirely different from its role in the immune system.
The known function of the membrane-bound toll-like receptors is to bind molecules found in invading microorganisms and damaged tissue, and help jumpstart the local immune response. In the central nervous system, toll-like receptors promote inflammation and help recruit circulating leukocytes into the brain (see Okun et al., 2008). TLR2, TLR3, and TLR4 are expressed in neurons, microglia, and astrocytes, and can worsen damage due to stroke (see Tang et al., 2007). Other studies have shown an Alzheimer connection: In microglia, TLR2 and TLR4 bind fibrillar Aβ and are essential for microglial activation, which leads to inflammation and production of reactive oxygen species (see Jana et al., 2008 and ARF related news story on Reed-Geaghan et al., 2009). TLR4 in neurons also contributes to neuronal apoptosis due to Aβ (see Tang et al., 2008).
Rather than looking at the role of TLR3 in inflammation, however, Mattson and colleagues used healthy adult TLR3-knockout mice to evaluate the role of TLR3 in learning and memory. First author Eitan Okun found that the knockouts showed improved memory abilities over wild-type mice in several tests—they had slightly longer memory retention in the Morris water maze, higher preference for the novel object in a novel object recognition test, and better memory retention in a fear-conditioning test, in which mice freeze when placed in a chamber where they previously received a shock. The TLR3 knockouts also demonstrated better working memory, a type of situational memory in which the organism has to react to environmental cues, in this case by finding the changed location of a hidden platform in the Morris water maze test more quickly than wild-type mice did.
Some amygdala-dependent behaviors were reduced in the TLR3 knockouts, however. They showed less anxiety in open field and elevated maze tests than did wild-type mice, and in a cued fear-conditioning paradigm, where a shock was paired with a tone, the knockout mice froze less in response to the tone in a new environment than wild-type mice did. It is not entirely clear why the knockouts responded differently in two different fear-conditioning paradigms, Mattson said, but he noted that the cued fear test measures working memory, which requires the simultaneous storing and processing of information. In the cued fear-conditioning test, the animals need to react to new cues in their environment to extinguish the memory of the shock. Mattson speculated that the knockouts might be better at extinguishing old memories than wild-type mice and therefore not associate the new environment with the cued shock. That could also relate to their better working memory in the Morris water maze, where animals must extinguish the memory of the original platform location in order to find the new location. It appears that the knockouts extinguish old memories better when presented with a new context, aiding them in learning.
Though these results are intriguing, they say nothing about whether these memory effects occur acutely in the adult, or might be a result of different developmental processes in the TLR3 knockouts. To get at this question, the authors used polyinosinic:polycytidylic acid [Poly(I:C)], which Mattson said is a synthetic molecule with a structure similar to a type of double-stranded RNA present in many viruses that has been shown to selectively activate TLR3. When Okun and colleagues infused Poly(I:C) into the brains of wild-type mice, they demonstrated poorer working memory by having more difficulty in finding the changed location of the hidden platform in the Morris water maze. This showed that direct activation of TLR3 could impair working memory, in agreement with the better working memory exhibited by TLR3 knockouts.
Patterson points out, however, that additional behavioral tests with Poly(I:C) might be needed to more fully address the question of whether all of the effects of TLR3 on behavior are acute, or whether some effects might be due to developmental changes. TLR3 is known to have developmental effects on brain function: For example, research from Patterson’s lab shows that treating pregnant rodents with Poly(I:C) leads to abnormalities in information processing in the offspring (see Ito et al., 2010). In addition, Mattson’s lab previously showed that neural progenitor cells proliferate more during development in the TLR3 knockouts than in wild-type mice (see Lathia et al., 2008). In their current paper, Mattson and colleagues also saw some effects of TLR3 on adult neurogenesis: Although the total number of newborn cells in the hippocampus did not change, as shown by BrdU labeling, significantly more of them expressed a mature neuronal marker, which the authors speculate is due to increased neuronal differentiation or survival. The TLR3 knockouts also had greater volume in the dentate gyrus and CA1 areas of the hippocampus, although it is not clear what caused this.
Finally, Okun and colleagues sought to characterize the mechanisms by which TLR3 affects memory. In the TLR3 knockouts, they found no change in activation of the canonical immune signaling pathway downstream of the receptor; however, they saw increased activation of ERK1, ERK2, and CREB in the hippocampi, adding to the evidence that TLR3 is acting through a distinct mechanism to affect learning. ERKs and CREB are believed to play critical roles in both synaptic plasticity and neurogenesis. In the knockouts, the authors also saw increased levels of the AMPA glutamate receptor GluR1, which is known to have a role in spatial learning and working memory.
One unanswered question, Mattson said, is whether the TLR effects require glial cells or are cell-autonomous in neurons. One way his lab will address this is by using purified cell cultures and neuron-glial co-cultures, using cells from both wild-type and various TLR knockout mice. They will also follow up by doing some electrophysiology studies to look at synaptic plasticity in slice cultures, Mattson said, as well as looking in vivo to see if other toll-like receptors modify learning and memory.
Another intriguing question is whether there are endogenous ligands in the brain for the TLRs, Mattson said. “In the absence of an infectious agent, what’s activating the receptors?” Because such ligands might be produced in an activity-dependent way, his lab will search for them by comparing media from stimulated and unstimulated neuronal cultures, Mattson said. Mattson points out that “an animal may go its whole life without having any infection or major injury to the brain, so it makes sense, intuitively, that these immune receptors would have other functions in the brain.”
From the standpoint of Alzheimer disease, Mattson said, it would be worthwhile to look at mouse models of AD and examine human postmortem AD brain tissue to see if they show alterations in the level of expression of the TLRs. In the long term, Mattson said, his lab plans to cross TLR-knockout mice with AD mice and look at Aβ and tau pathology as well as learning and memory in the offspring. Because the suppression of TLR3 appears to enhance memory, Mattson also speculated that there could be some potential for AD drug development down the road, if a useful antagonist of TLR3 could be identified.—Madolyn Bowman Rogers