Science 4 September 2009:
Vol. 325. no. 5945, pp. 1214 - 1215
Neuroscience: Erasing Fear Memories
1 Istituto Neuroscienze CNR, via Moruzzi, 1 56100 Pisa, Italy.
2 Dip. Psicologia, Via S. Salvi, 12–Pad. 26 Università di Firenze, 50135 Firenze, Italy.
Events that are associated with trauma and fear often leave memories that reoccur spontaneously, leading to excessive fear, anxiety, and, in some cases, posttraumatic stress disorder. Such relapses of fear memories constitute a major clinical problem, and their elimination is a major cornerstone of psychological therapy. Many neurobiological studies are therefore focused on understanding how fear memories are controlled (1). On page 1258 of this issue, Gogolla et al. (2) take an important step in the field by determining that the extracellular environment in a particular region of the brain—the amygdala—is responsible for making fear memories erasure-resistant.
"Extinction" is a popular behavioral technique to block recurring traumatic memories. This form of learning is characterized by a decrease in a fear response when the contingent relationship—between a conditioned stimulus (e.g., a sound) and an unconditioned stimulus (e.g., an electric shock)—is compromised. This situation is most commonly implemented when the conditioned stimulus is repeatedly presented in the absence of the shock (3). It is now well accepted that extinction represents new learning and does not erase the preexisting memory (4). Indeed, the original memory can spontaneously recover, or it can be renewed, when the conditioned stimulus is presented in contexts different from that in which the extinction protocol was administered. The resilience of traumatic memories to extinction represents a serious obstacle for treating disorders characterized by abnormal fear and anxiety.
Resisting erasure. Young mice as well as adults lacking chondroitin sulfate proteoglycans (CSPGs) in their amygdalae showed a faster extinction of the fear response (solid line) (which was acquired during prior fear conditioning), compared to adult mice treated with placebo (dashed line). When fear memory was retested 1 and 4 weeks after the extinction protocol, only the young mice and adults lacking CSPGs had completely eliminated the fear memory, whereas fear response had reoccurred in the placebo group.
CREDIT: N. KEVITIYAGALA/SCIENCE
Gogolla et al. were inspired by previous work originating from fields as diverse as development of fear conditioning (when fear is associated with a neutral stimulus) and plasticity of the visual cerebral cortex. These studies demonstrated that in contrast to the inability of an extinction protocol to erase the fear memory in adult rats, extinction of acquired fear in young rats (17 days after birth) deletes the fear memory (5). Further studies showed that sensitivity to erasure of fear memories is already lost at 23 days after birth. At all ages, extinction of fear conditioning in rats implicated neuronal circuits in the amygdala, a brain region necessary for fear memory acquisition and extinction. What changes occur during amygdala development that are responsible for switching off the susceptibility of fear memory to the process of elimination?
Developmental windows during which neural plasticity is different from that of adult animals have been extensively studied in the mammalian visual cortex. Visual cortical circuits are particularly sensitive to visual deprivations during the so-called critical period. An important determinant of the closure of the critical period is maturation of the extracellular matrix that surrounds visual cortical neurons. Chondroitin sulfate proteoglycans (CSPGs) are the main constituents of the adult extracellular matrix. CSPGs condense into a netlike structure that surrounds a subclass of inhibitory interneurons (those that express the protein parvalbumin). This occurs in parallel with closure of the critical period. Eliminating CSPGs from the adult rat cortex (by injecting the enzyme chondroitinase ABC, which digests lateral chondroitin sulfate chains) can enhance plasticity almost to the level of a juvenile rat (6), suggesting that maturation of CSPGs endows the visual cortex with adultlike plasticity.
Like the visual cortex study, Gogolla et al. determined that maturation of the extracellular matrix in the amygdala is responsible for closing the developmental period during which fear memories are susceptible to extinction. The authors found that the number of CSPGs-containing perineuronal nets in the mouse amygdala increases sharply between 16 and 23 days after birth, in correlation with the developmental switch of fear memories from erasure-prone to erasure-resistant. Adult mice were injected in the amygdala with chondroitinase ABC to eliminate CSPGs, and then subjected to fear conditioning by pairing a sound with an electric shock. Then, mice were subjected to an extinction protocol consisting of repeated presentation of the sound without the shock. To examine context specificity, the authors administered the extinction protocol in a context different from the one in which the animal had been previously conditioned. During the extinction protocol, the sound elicited progressively less fear response (cessation of body movement, or "freezing") both in mice treated with chondroitinase ABC and in control mice treated with placebo (saline solution). However, the rate at which the fear response decayed was much faster in chondroitinase ABC–treated mice than in control mice (see the figure). Strikingly, when response to the conditioned sound was retested 1 and 4 weeks after the extinction protocol, the fear response was not observed in animals treated with chondroitinase ABC—neither in the same context in which the extinction protocol had been administered, nor in the context in which the animals had originally been conditioned. This indicates complete loss of the fear memory. Thus, treating adult mice with chondroitinase ABC restores fear memory acquisition to the erasure-prone modality that typifies young rodents. By contrast, substantial spontaneous recovery and fear renewal occurred in control adult mice. Interestingly, chondroitinase ABC was effective if injected before fear conditioning, but not if injected before the extinction protocol. Thus, CSPGs are not directly regulating the processes occurring during extinction, but they are important for coding an erasure-resistant memory during memory acquisition.
The results of Gogolla et al., together with previous experiments on the visual cortex, suggest that maturation of the extracellular matrix could be a mechanism used by different brain circuits to change from a malleable to a more crystallized state during development. The presence of a high concentration of CSPGs in the perineuronal nets surrounding inhibitory neurons suggests that inhibitory circuits could play an important role in the developmental control of plasticity. Evidence supports this role for parvalbumin-expressing cells in the rodent visual cortex (7), but much less is known about other circuits. The mechanistic aspects of how CSPGs control plasticity are also still obscure: Is their organization in perineuronal nets important? What is the function of this structure at the molecular level? Further work is needed to understand how general this mechanism is and whether it applies to the human brain as well.
• 1. I. Ehrlich et al., Neuron 62, 757 (2009). [CrossRef] [Web of Science] [Medline]
• 2. N. Gogolla, P. Caroni, A. Lüthi, C. Herry, Science 325, 1258 (2009).[Abstract/Free Full Text]
• 3. K. M. Myers, M. Davis, Neuron 36, 567 (2002). [CrossRef] [Web of Science] [Medline]
• 4. M. E. Bouton, Learn. Mem. 11, 485 (2004).[Abstract/Free Full Text]
• 5. J. H. Kim, R. Richardson, J. Neurosci. 28, 1282 (2008).[Abstract/Free Full Text]
• 6. T. Pizzorusso et al., Science 298, 1248 (2002).[Abstract/Free Full Text]
• 7. Z. J. Huang, G. Di Cristo, Neuron 59, 355 (2008). [CrossRef] [Web of Science] [Medline]