Discovering taste aversion circuitry using molecular mapping tools
Date: October 5th, 2006
Speaker’s Name and Affiliation: Ilene Bernstein, Ph.D., University of Washington, Seattle Washington
Title: “Using molecular tools to map conditioned taste aversion circuitry”
Dr. Ilene Bernstein presented research on the use of molecular tools to map conditioned taste aversion (CTA) circuitry. Taste aversion learning is form of a classical conditioning in which an individual learns to associate a taste (CS) with a treatment (US) which produces nausea or illness. As a consequence of this learning, the taste becomes disliked or avoided. The process protects the organism against the ingestion of potentially harmful toxins. Taste aversion learning is different from associative conditioning of flavor preferences in that it can occur after only one trial, and there can be significant periods of time (long delays) between the CS and the US. Because of the potential for long delay learning with taste aversions, it is possible to study the molecular events that occur immediately after the taste is administered, thereby providing an insight into the neural circuitry involved with this process.
One of the tools used to study the molecular events involved with taste aversion learning in Dr. Bernstein’s lab is the Immediate Early Gene (IEG) c-fos. cFos immunostaining has several advantages over other methods. This technique allows individually activated neurons to be identified; the baseline expression is generally low; and the method is semi-quantitative and can be combined with other procedures such as immunostaining for neurochemicals. However, limitations of this method include the fact that strong and sustained stimulation of neurons is generally required before c-fos expression occurs. Further, temporal resolution of the signal can be limited, making it difficult to ascertain whether c-fos expression is a direct or indirect consequence of the stimulus. Evidence of induction of IEG proteins like c-fos in response to novel tastes would be an important step in supporting a role for these proteins in taste processing.
Taste novelty is an important issue to study, as novel tastes rapidly become the target of a conditioned taste aversion. In contrast, it is relatively difficult to elicit a CTA toward familiar tastes. Dr. Bernstein’s lab has hypothesized that the neural responses to novel and familiar tastes are also significantly different, and this relates directly to whether or not the organism will form an aversion to the taste.
There are several regions of the brain hypothesized to play a role in the neural processing of CTAs. The insular cortex, amygdala, parabrachial nucleus (PBN), and the nucleus of the solitary tract are a few such regions. Taste novelty appears to elicit the greatest response in the insular cortex and central amygdala, as evidenced by cFos staining. During taste-illness training, only training with novel stimuli will result in increased activation in the insular cortex, central amygdala, basolateral amygdala, nucleus of the solitary tract, and lateral and medial PBN. Familiar tastes do not produce activation in the above regions.
In the next portion of the talk, Dr. Bernstein made some preliminary suggestions as to the role of the insular cortex (IC) in signaling of taste novelty. Lesions of the IC interfere with an organism’s ability to learn CTAs. Because of this evidence, Dr. Bernstein asked whether IC lesions disrupt the signaling in subcortical circuitry after CS-US pairing. In one experiment, the IC was lesioned, and the subcortical circuitry was observed after novel or familiar CS-US pairing. Lesions reduced the activation normally produced by taste novelty in the lateral and central amygdala and the nucleus of the solitary tract. A different outcome was observed when IC was reversibly inactivated with muscinol taste pre-exposure. The response to familiar pairing was then strong, brought to the level of the novel paired group. This suggests a dual role for the insular cortex. First, it is involved with processing of taste memory as taste goes from novel to familiar. Second, it provides ‘on-line’ comparisons between incoming taste and stored taste memories.
There are strong comparisons between CTA and other learning models, in particular, fear conditioning. Both are associative, adaptive, and amygdala dependent. Might unique cell signaling pathways have developed to accommodate both of these mechanisms? cAMP Protein Kinase A (PKA) and cAMP response element binding protein (CREB) signaling pathways have been widely implicated in many forms of learning. Dr. Bernstein hypothesized that PKA might play a role in CTA learning. Both PKA inhibition in rats and PKA null mice show disruptions in CTA learning which appear to affect consolidation of taste learning.
Q. Are you talking about how information processing differs between familiar and novel tastes?
A. Yes, what is the difference in the neural processing between familiar and novel tastes.
Q. Is your assumption that C-Fos is somehow involved in signaling in the taste system?
A. For me, C-Fos is a functional brain imaging methodology. There may also be functional importance to C-Fos.
Q. Some researchers have used fMRI in animals. Have you considered that methodology?
A. We’ve thought about it, but C-Fos expression is more suited for our purposes. fMRI has some of the same methodological disadvantages as cFos.
Q. Do you see long term potentation (LTP) in six hours?
A. Some have looked at LTP in these pathways but this may be more likely to be involved in CS-US associations than a response to a taste.
Q. Is this why disulfuram (antiabuse) makes alcoholics sick, even though the drug doesn’t’ produce an actual taste aversion to alcohol?
A. That’s an interesting question. At first, alcoholics will often get sick on a drink, but that doesn’t result in an aversion. It might be a familiarity issue.
Q. Some have also argued that the reinforcing properties of alcohol combined with familiarity make alcoholics less likely to form a taste aversion, even though they are getting sick.
A. Interesting. Is vulnerability to alcoholism associated with more reinforcement or milder punishment?
Q. How does your data compare to work from J. Mennella’s lab, in which flavors that pass through the mother’s amniotic fluid might condition preferences in the developing fetus?
Comment: There is some indication in the animal literature to suggest that sweet taste is not actually innate, but rather learned through early exposure.
Q. Were those 1 bottle tests?
A. Yes, but if we would have given 2 bottle tests, we might have seen evidence of learning.
Q. How do you choose brain sections and quantify fos expression?
A We choose corresponding regions based on templates from an atlas, outline those areas and scoring is done blind.
Q. Is fluid consumption similar in the groups of animals?
A. We don’t measure food or fluid consumption.
Q. Why don’t you have a saline control in your experiment?
A. Saline doesn’t have much of an effect on Fos learning. We have used saline, but we’ve concluded that it’s not necessary to include it.
Q. Isn’t your experiment confounded because learning is always paired with novelty or familiarity is always paired with no learning? How do you separate these confounds?
A. We don’t view this as a confound but as the heart of our experimental question. Namely, what we are asking is the difference in patterns of brain activity when learning is or is not occurring.
Q. When are the animals sacrificed?
A. 2 hours post exposure.
Q. Why haven’t you looked at the hippocampus in conditioned taste aversions?
A. There is a literature on hippocampal lesions that suggests that it is not necessary for aversion learning.
Q. Where is muscinol infused in your animals?
A. It’s a gaba agonist and it is infused directly into the Insular cortex.
Q. If you gave muscinol to animals at the time of testing, would you predict similar results to Insular cortex lesions?
A. Yes, but we haven’t done those experiments yet.
Q. Can you get long delays with fear conditioning?
A. No, you can’t.
Q. When you give PKA inhibitors to animals, do you know they have enough time to act before you test the outcomes?
A. We given them 15 minutes prior to testing, based on literature on fear conditioning.
Q. In PKA inhibitor knock out mice, would you expect them to be disrupted in a number of learning processes?
A. They still have some PKA activity in a number of brain areas, but since amygdala is affected we would expect deficits in amygdala dependent tasks.