Minutes of the Columbia University Seminar on Appetitive Behavior (#529)

Chairman:Harry R. Kissileff, Ph.D.
Rapporteur:Janet L. Guss
Date:Thursday, April 19th, 2001
Title:"Chemical Senses: Goad to Appetite"
Speaker:Thomas R. Scott, San Diego State University, San Diego, CA

Presentation:

     Animals negotiate their physical environments using vision, audition and somesthesis; they negotiate their chemical environments using olfaction and taste. The acquisition of nutrients and avoidance of toxins is the most fundamental behavior organisms perform.
     The chemical selections humans make are guided by familiarity, custom, vision and olfaction from a distance, but the final decision to swallow-to make "other" into "self"-is mediated by taste. The behavior is guided first by the reflexes of the medulla, and subsequently by the hedonic assessment of the forebrain. What begins as an analysis of chemical structure at the receptor level is transformed to an assessment of flavor.
     The taste system operates through at least nine synaptic levels: receptors, peripheral nerves, medulla, pons, thalamus, primary cortex, secondary cortex, amygdala, and hypothalamus. Its role is to separate nutrients from toxins. Since chemicals that impact the body come in such variety, multiple transduction mechanisms are required to recognize them. These mechanisms are not unique to taste. Sodium passing through amiloride-sensitive channels on the tongue produces a salty taste; in the kidney, sodium resorption. Taste cells and pancreatic b cells probably recognize glucose by the same mechanism, the one eliciting a perception of sweetness, the other a release of insulin. The umami taste of MSG and the recognition of glutamate throughout the nervous system are most likely managed by the same receptor, though tuned a thousand-fold less sensitive in the mouth. Thus, what distinguishes taste is not its ability to recognize these chemicals, but to do so before the decision to swallow is made.

Taste Receptors

     Humans have a mean of about 300,000 taste receptor cells, collected in groups of about 50 into some 6000 taste buds. Two-thirds of the buds are housed in fungiform (N=200), foliate (N=12) or circumvallate (N=9) papillae on the tongue. The remainder are embedded in the epithelium of the soft palate, pharynx, larynx and other oral regions. Taste receptors are modified epithelial cells, and live for about 12 days, born of trophic factors at the fringes of their bud, and migrating inward to their demise in its center.
     The variety of gustatory transduction mechanisms demands that taste cells be imbued with a range of ion channels, and they are: sodium channels, inward and outward potassium channels, calcium and chloride channels. Yet the receptor currents generated by tastants-typically only several picoamperes-were deemed insufficient to generate action potentials in what intracellular studies concluded were leaky cells. When these were superseded by patch clamp experiments, the impedance of taste cells was revealed to be up to 10 GOmega, such that a 3 pA current would cause a depolarization of 30 mV and generate action potentials. Why action potentials in such a closed space? Probably because only one third of taste receptors are coupled to afferent nerve fibers. The cells within a bud communicate via gap junctions, and action potentials in one receptor could elicit neurotransmitter release from its neighbor. Immunocytochemistry indicates that the transmitter is occasionally GABA, but more often serotonin or norepinephrine.
     Taste is the only sensory system to use a variety of transduction mechanisms, both ionotropic and metabotropic.
     Sodium simply passes into the receptor cell along its concentration gradient to generate salty taste.
     Sweet chemicals use two metabotropic pathways. Sugars work through adenylate cyclase (AC) and cAMP to close K+ channels on the basolateral membrane, causing depolarization with increased membrane resistance. Non-sugars operate through the phospholipase C-inositol triphosphate (PLC-IP3) system to release intracellular calcium, and so neurotransmitter.
     Monosodium glutamate serves as a ligand for the metabotropic glutamate receptor, type 4 (mGluR4), whose activation may inhibit adenylyl cyclase, and so on to cAMP. The loss of cAMP may close K+ channels in the membrane, depolarizing the cell.
     Fats are broken down to free fatty acids by lingual lipase. When transported to receptor cells, fatty acids inhibit K+ channels, causing depolarization, increased intracellular Ca2+ and neurotransmitter release. The perception of fats is also mediated by texture.
     The protons responsible for sour taste are small and mobile and may affect the electrical character of the receptor through a number of mechanisms. Only one is established; another likely. Protons block leaky potassium channels, causing depolarization with increased membrane resistance. Secondly, they elicit an inward current that is prevented by the sodium channel blocker amiloride. Therefore, protons may use the same channels to elicit sourness that sodium ions use to evoke saltiness. The issue is how the distinction between these two taste qualities is maintained when both use the same channels in the same cells for transduction.
     Bitter chemicals have diverse molecular structures and a multiplicity of physiological effects, with the common feature of toxicity. To protect against this multifaceted threat, the taste system is imbued with a family of protein receptors, numbering perhaps 60. Each is thought to be sensitive to one, or a small group of related chemicals. When a receptor cell expresses one member of the family, as about 15% of taste cells do, they express many, perhaps all members. Thus, if a receptor cell is assigned the role of vigilance against toxins, it may be invested with the full array of receptor proteins to accomplish that function. After binding, at least two second messenger systems are involved. First, the PLC-IP3 system causes the release of Ca2+ from intracellular stores; secondly, phosphodiesterase (PDE) decreases cAMP activity, and so may impede the egress of cations or disinhibit closed inward channels, in each case depolarizing the cell.

Peripheral Nerves

     The three cranial nerves that serve taste may play different roles in the control of feeding. Receptors in the nasoincisor ducts of rodents are particularly sensitive to sweet stimuli, and send their impulses through the greater superficial petrosal branch of cranial nerve VII. Receptors on fungiform papillae are most sensitive to salt, and send their signals through the chorda tympani branch of CN VII. Therefore, CN VII appears to serve the appetitive taste qualities.
     Surprisingly, however, it appears to serve the aversive tastes as well. Sectioning CN IX has no impact on the rat's capacity to discriminate between quinine and KCl, but cutting CN VII causes a significant loss. The implication is that the fibers in CN VII are responsible for most taste discriminations. Those of CN IX offer chemical protection by driving rejection reflexes to toxins, and CN X may serve to protect the airways rather than being chemosensory.

Nucleus of the Solitary Tract

     NST is the engine of taste. Responses to basic stimuli are at least double those of any other synaptic level in the system. The basis for this energetically expensive amplification may be the number of targets that NST neurons must drive. First, clusters of cells in ventral NST send projections to bulbar motor nuclei through which acceptance-rejection reflexes are governed. Secondly, axons are sent from NST to DMNX to influence parasympathetic reflexes associated with ingestion. Finally, a subset of NST neurons projects rostrally to the parabrachial nucleus and thalamus to process taste quality and intensity. Lesions in gustatory NST cause a greater loss of discriminative capacity than damage to any other relay.

Parabrachial Nucleus

     While electrophysiology supports a role for the NST in managing the interactions between taste and physiological state, anatomical, lesion and immunocytochemical studies favor PBN. Viscerosensory information from caudal NST projects to lateral PBN, while taste drives cells more medially. PBN lesions block the formation of CTAs, not from a loss of taste or visceral sensitivity, but from an inability to form an association between them. C-fos studies of CTAs reveal that the expression elicited by the CS shifts from subnuclei associated with positive hedonics to one normally activated by aversive tastes following conditioning. PBN lesions also block sodium appetite, and sodium depletion reduces responsiveness to NaCl in PBN, as it does in CT and NST. Finally, satiating loads of intralipid in the duodenum reduce the responsiveness of sucrose-oriented taste cells, just as hyperglycemia does in NST.

Thalamic Taste Area

     Thalamic taste cells are overwhelmingly multimodal, responding to touch and temperature stimulation as well as to taste. Lesions in this small, obligatory gustatory relay have little effect on taste discrimination, CTAs, sodium appetite, or preference-aversion functions. The only known behavioral deficit is in gustatory memory, where rats cannot learn an anticipatory contrast.

Primary Taste Cortex

     The available data on taste cortex derives primarily from macaques. In this insular-opercuilar (IO) region, only 6% of neurons are responsive to taste. Oral-related sensory and motor activity defines another 30% of the neurons, but the sensitivities of some 64% remain undetermined. Anatomical connections imply that the area could be involved in an integrated set of activities related to the selection and digestion of foods.
     Nevertheless, activity in IO is not influenced by feeding macaques to satiety. Rather, IO neurons provide what is apparently the most precise representation of taste quality. Human reports of taste qualities are nearly identical to what is implied by an analysis of macaque neural data derived from IO.

Secondary Taste Cortex

     In orbitofrontal cortex (OFC) there is a shift toward greater specificity to tastants, but increased breadth across modalities. Visual, somatosensory, and olfactory inputs may converge with taste, presumably to construct a representation of flavor. When cells respond to stimuli from more than one modality, they are in register, such as the sight, smell and taste of fruit.
     Accordingly, OFC neurons are sensitive to physiological condition. Responses given to an appetitive stimulus when the macaque is deprived decline and disappear as he is sated. Therefore, the cognitive analysis of quality that characterizes IO is transformed to an analysis of hedonic value in OFC.

Amygdala

     The amygdala receives projections from both primary and secondary taste cortex. Its representation of taste stimuli is rather crude, and does not provide an adequate basis for our discriminative capacity. Its cells are variably sensitive to physiological condition: some totally suppressed by satiety and others unaffected. In rodents, amygdaloid lesions disrupt CTA learning, reinforcing its presumed function of relating internal needs to external resources.

Hypothalamus

     The lateral hypothalamic area receives projections from OFC and amygdala, and its activity is clearly related to hedonic evaluation. Appetitive tastes tend to arouse excitation, and aversive stimuli, inhibition. Responses to rewarding tastes are suppressed to baseline as the macaque is fed to satiety.
     The progression we find across nine synaptic levels of taste is from molecular recognition through increasingly refined analyses to integration, both with other senses to form flavor and with visceral input to relate flavor to the body's needs. Appropriate behaviors, such as reflexes, are carried out as the analysis reaches the point of enabling them.

Discussion:

Q: Since you believe that gustatory coding is analogous to the visual system, are the movements of the tongue analogous to movements of the eye in the visual system?
A: No; vision is a location sense-it is exquisitely sensitive to place. Taste is quite different: the location of a taste stimulus in the mouth is given by touch more than by taste. Tongue movements are important in opening the channels that permit taste molecules to reach receptors, so an active tongue yields larger responses than a passive one. But tongue movements do not scan the taste world in the way eye movements scan the visual world.

Q: What about within the gustatory cortex; is there ever cross-talk that links vision with taste?
A: Rolls's lab reports that gustatory and visual information converge onto individual cells in the orbitofrontal cortex. We occasionally find neurons in insular-opercular cortex that respond to the sight of the syringe from which the monkey will be fed. The sight of non-food objects does not activate these cells. At the receptor level, both vision and taste use similar g-proteins: transducin in the visual system and gusducin in the gustatory system.

Q: Continuing with the analogy to the visual system, are the basic tastes- salt, sweet, sour, bitter- analogous to 'primary colors'?
A: There are analogies. We conclude that there are information channels in taste that might correspond to wavelength channels in vision (r, y, g, b). There are reports of aftereffects in taste, where a neutral stimulus (distilled water) tastes sweet after bitter adaptation and vice-versa. This is analogous to r-g and b-y color aftereffects, implying an opponent process system in taste. But there is no evidence that tastes are as intimately compared to synthesize new tastes, as primary colors are to generate secondary colors.

Q: Are the reciprocal connections from the forebrain to the hindbrain in primates very dense?
A: There are rich projections, from each forebrain site to the nucleus of the solitary tract, and from related visceral areas to the parabrachial nucleus.

Q: Can taste cells generate action potentials?
A: Yes; but initially the channels were thought to be too leaky to produce action potentials. The presence of action potentials was subsequently CONFIRMED when it was shown that channel blockade with amiloride (a sodium channel blocker) altered membrane potential. It turns out that the presumed leakage actually resulted from damage to the cell membrane during penetration by an intracellular electrode. When this approach was replaced by patch clamping, the membrane impedance was found to be 10-100 gigaohms, sufficient to permit a receptor current of a few picoamps to generate the tens of millivolts necessary to depolarize the cell to threshold. Taste cells have a full range of ion channels, and so the biophysical machinery necessary to generate action potentials.

Q: Can animals be re-trained to a CTA after one has been extinguished, or is the time course constant?
A: I do not know if there are data on the time course of retraining of a CTA after extinction. Certainly the continued existence of a neural signal of the former CTA would imply that there is some record of that event, and this could underlie rapid relearning. The CTA is learned so quickly under normal circumstances that one might have to use uncommon approaches to see a difference, such as a very mild UCS.

Q: Is there any evidence that learning of any sort can take place in the NTS?
A: Certainly in the hindbrain. Grill's lab has reported that decerebrate rats alter their reactions to tastes according to primary signals of satiety, such as blood glucose levels. Specifically in the NTS there is electrophysiological evidence of changes brought about by conditioning, but no indication that these changes drive altered behavior.

Q: Regarding your extinction work, have you found the results to correlate with changes in c-fos expression?
A: Tom Houpt or Ilene Bernstein may have data on the fate of c-fos expression after extinction of a CTA, but I am not aware of it. We have used c-fos to mark the sites of involvement in generating a CTA, but have not used it after extinction.

Q: How do you deal with the methodological problems that arise when comparing firing rates in deprived versus replete states? One cannot know if he is recording from the same neuron!
A: That's a crucial issue. When we see changes between the taste responses in sodium deprived and sodium replete rats, how can we tell whether salt cells have turned to sweet cells during the 2 weeks of deprivation, or whether salt cells just quiet down and sweet cells become more active? Using separate groups of rats, and so separate populations of neurons, we cannot. Therefore, Stuart Mccaughey took on the heroic task of recording from the same cell before and after the sodium appetite was generated. The time frame of creating the sodium appetite is brought within the bounds of the recording time by using the technique developed in Epstein's lab. A sodium appetite is created quickly by mimicking the effects of the neurotransmitters and neuromodulators that underlie it. Mccaughey primed each rat with daily injections of aldosterone to increase AII receptors. Then, during recording, he isolated the activity of an NTS taste cell, characterized its response profile in the absence of a sodium appetite, then gave a pulse injection of renin intracerebroventricularly, to generate AII, and the sodium appetite. He then reapplied all stimuli to re-characterize the cell's response profile (yoked behavioral controls provided confidence that there was no sodium appetite before the renin and a robust appetite 5 minutes after its administration). Mccaughey's results indicate that taste cells do not change identity. Salt cells stayed salt cells, but their responses to sodium declined. Sugar cells stayed sugar cells, but their responses to sodium were enhanced.

Q: How can a cell be isolated for study, while you are simultaneously generating a sodium appetite in it?
A: One can use the technique pioneered by Alan Epstein and described above.

Q: Can you offer sodium, and then sucrose, and eventually get the animal to be unable to distinguish the two tastes?
A: While it is unlikely ever to come to inability to distinguish, there is a suggestion from frankmann's lab that feeding rats sucrose partially allays the sodium appetite, as if the tastes were confounded. We performed a behavioral study to see if rats would confuse sodium with sucrose when they were sodium deprived. They did not, but this was a simple 2-choice test with moderate concentrations of both stimuli, and the distinction should have been very clear. Perhaps a subtler test of whether sugar can substitute for sodium in deprived rats would be successful.

Q: Is there any lateral inhibition?
A: Inglis Miller showed lateral inhibition among fungiform papillae in 1971. There is also an indication that the 7th and 9th nerve mutually inhibit, such that cutting either one leads to hypergeusia from the disinhibition of the other.

Q: What about olfactory responses in the thalamus?
A: I do not know of any data on olfactory overlap with taste in the thalamus. However, olfactory input has been reported in the NTS by Rick Van Buskirk in Erickson's lab and independently by Pat Dilorenzo. Certainly there are olfactory-gustatory interactions in orbitofrontal cortex of the macaque monkey, and in the central nucleus of the amygdala and the lateral hypothalamus.
Top

SEMINARS AND CONFERENCES || OBESITY RESEARCH CENTER