<%BANNER%>

The Psychophysics of Salt Taste Transduction Pathways


PAGE 1

THE PSYCHOPHYSICS OF SALT TASTE TRANSDUCTION PATHWAYS By LAURA CLAIRE GERAN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2003

PAGE 2

Copyright 2003 by Laura C. Geran

PAGE 3

ACKNOWLEDGMENTS I would like to thank Dr. Spector and the members of his lab past and present, especially Mircea Garcea, Shachar Eylam, Stacy Kopka and Connie Colbert, for their educational and emotional support. I would also like to thank all of the faculty members and graduate students I have had the pleasure of meeting along the way, in particular Laura Tucker, Sylvia Belski, Melanie McEwen, and Cheryl Vaughan, for their advice and patience. Finally, I would like to thank Albert and my family for their unwavering support of my decision to attend graduate school. It has been a great experience. iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT.........................................................................................................................x CHAPTER 1 LITERATURE REVIEW................................................................................................1 Introduction .....................................................................................................................1 Peripheral Gustatory System...........................................................................................1 Taste Buds ...............................................................................................................1 Gustatory Nerves......................................................................................................3 Central Gustatory System...............................................................................................5 Medullary and Pontine Taste Nuclei........................................................................5 Ascending Gustatory Pathways................................................................................6 Taste Transduction..........................................................................................................8 Primary Tastes..........................................................................................................8 Transduction Mechanisms.......................................................................................9 Ion channels.......................................................................................................9 Metabotropic receptors......................................................................................9 Sodium Transduction Pathways.............................................................................10 Amiloride-sensitive (AS) pathway..................................................................10 Amiloride-insensitive (AI) pathway................................................................11 Labeled Line vs. Across-Fiber Pattern Theory......................................................12 2 ANION SIZE DOES NOT COMPROMISE SODIUM RECOGNITION BY RATS FOLLOWING ACUTE SODIUM DEPLETION.........................................................16 Background ...................................................................................................................16 Methods ...................................................................................................................18 Subjects .............................................................................................................18 Apparatus .............................................................................................................19 Training Procedure.................................................................................................20 Sodium Depletion..................................................................................................20 Brief-Access Testing..............................................................................................21 iv

PAGE 5

Urine Analysis........................................................................................................22 Data Analysis.........................................................................................................22 Results ...................................................................................................................23 Brief-Access Testing..............................................................................................23 Number of Trials Initiated......................................................................................24 Sodium Balance.....................................................................................................24 Discussion ...................................................................................................................24 3 GLOSSOPHARYNGEAL NERVE TRANSECTION DOES NOT IMPAIR POTASSIUM CHLORIDE vs. AMMONIUM CHLORIDE OR SODIUM CHLORIDE vs. AMMONIUM CHLORIDE DISCRIMINATION.............................32 Background ...................................................................................................................32 Methods ...................................................................................................................34 Subjects .............................................................................................................34 Apparatus and Trial Structure................................................................................34 Training Procedure.................................................................................................35 Presurgical Discrimination and Amiloride Testing..............................................36 Surgery .............................................................................................................37 Postsurgical Testing and the Water Control Test...................................................37 Histology .............................................................................................................38 Data Analysis.........................................................................................................38 Results ...................................................................................................................39 Presurgical Discrimination Testing........................................................................39 Postsurgical Testing and Histology........................................................................40 Discussion ...................................................................................................................41 Presurgical Discrimination Testing........................................................................41 Postsurgical Discrimination Testing......................................................................42 Potential Mechanisms Underlying KCl vs. NH 4 Cl Discrimination.......................43 Summary .............................................................................................................44 4 AMILORIDE-INSENSITIVE UNITS OF THE CHORDA TYMPANI NERVE ARE NECESSARY FOR NORMAL AMMONIUM CHLORIDE DETECTABILITY IN THE RAT......................................................................................................................57 Background ...................................................................................................................57 Methods ...................................................................................................................59 General Methods....................................................................................................59 Testing .............................................................................................................60 Surgery .............................................................................................................61 Histology .............................................................................................................61 Data Analysis.........................................................................................................62 Results ...................................................................................................................63 Presurgical Detection Threshold............................................................................63 Postsurgical Detection Threshold..........................................................................64 Water Control Test and Histology.........................................................................64 Discussion ...................................................................................................................64 v

PAGE 6

Ammonium Chloride Detectability Depends Upon an Amiloride-Insensitive Route of Transduction.........................................................................65 The Chorda Tympani Nerve is Necessary for Normal Ammonium Chloride Detection.............................................................................................67 Conclusions............................................................................................................68 5 GENERAL DISCUSSION............................................................................................79 Discrepancies Between the Electrophysiology and the Behavior In Regard to NH 4 Cl........................................................................................................79 Implications for Chloride Salt Detectability.................................................................82 Support for the Hypothesis That The Seventh Cranial Nerve Is More Important For Taste Recognition and Discrimination Than The Glossopharyngeal Nerve.........................................................................................................85 Conclusions ...................................................................................................................87 LIST OF REFERENCES...................................................................................................92 BIOGRAPHICAL SKETCH...........................................................................................105 vi

PAGE 7

LIST OF TABLES Table page 3-1. Training Schedule.......................................................................................................45 3-2. Experiment Schedule..................................................................................................46 4-1. Training Schedule.......................................................................................................70 4-2. Test Stimulus Presentation Schedule..........................................................................71 4-3. Experiment Schedule..................................................................................................71 5-1. Mean Detection Thresholds for Intact and Chorda Tympani-Transected Rats..........89 vii

PAGE 8

LIST OF FIGURES Figure page 2-1. Brief-access licking to each stimulus by sodium-depleted rats.................................29 2-2. Mean ( + SE) number of trials initiated by each group of rats....................................30 2-3. Sodium balance for each group of animals measured in mmol..................................31 3-1. Presurgical KCl vs. NH 4 Cl discrimination with and without amiloride.....................47 3-2. Presurgical NaCl vs. NH 4 Cl discrimination with and without amiloride...................48 3-3. Mean presurgical performance by concentration.......................................................49 3-4. Prevs. postsurgical performance on the KCl vs. NH 4 Cl task...................................50 3-5. Prevs. postsurgical performance on the NaCl vs. NH 4 Cl task.................................51 3-6. Mean presurgical vs. postsurgical performance by concentration.............................52 3-7. Effect of amiloride on postsurgical KCl vs. NH 4 Cl discrimination...........................53 3-8. Effect of glossopharyngeal transection on NaCl vs. NH 4 Cl performance in the presence of amiloride.............................................................................................54 3-9. Mean comparisons of postsurgical amiloride performance by concentration............55 3-10. Water control test......................................................................................................56 4-1. Effect of amiloride on NH 4 Cl detection.....................................................................72 4-2. NH 4 Cl threshold decreased again following amiloride treatment..............................73 4-3. Individual shifts in presurgical threshold...................................................................74 4-4. Individual shifts in performance with surgery............................................................75 4-5. NH 4 Cl detectability functions preand post-surgery.................................................76 4-6. Postsurgical NH 4 Cl detectability functions with and without amiloride....................77 viii

PAGE 9

4-7. Water control test........................................................................................................78 5-1. Comparison of NaCl and KCl Detectability Functions..............................................90 5-2. Comparison of Chloride Salt Detectability Functions................................................91 ix

PAGE 10

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE PSYCHOPHYSICS OF SALT TASTE TRANSDUCTION PATHWAYS By Laura Claire Geran May 2003 Chair: Alan C. Spector Major Department: Psychology Salt stimuli can activate two transduction mechanisms in the rats oral cavity. One mechanism appears to rely on taste receptor cells that contain sodium-selective ion channels on their surface. Passage through these channels can be blocked with the drug amiloride. The other mechanism is thought to be less selective, with associated fibers responding to potassium and ammonium as well as sodium salts. Most data suggest that the salt responsiveness of this population of fibers is not significantly attenuated with amiloride treatment, although some researchers have found evidence to the contrary. The two salt transduction pathways are commonly grouped according to their amiloride sensitivity as either amiloride-sensitive (AS) or amiloride-insensitive (AI). Activation of the AI pathway appears to be limited by anion size, with large anions like gluconate producing the greatest suppression. Previous research has indicated that the AS pathway is necessary and sufficient for normal sodium detection in the rat as well as necessary for sodium recognition. The current experiments were designed to determine whether the AS x

PAGE 11

pathway was also sufficient for sodium recognition as well as to elucidate possible functional roles of the AI pathway regarding the perception of nonsodium salts. This was accomplished by observing the effects of physiological manipulations like gustatory nerve transection, amiloride treatment, and acute sodium depletion on the taste-guided behavior of highly-trained rats. Briefly, the sodium-specific AS pathway appears to be sufficient for sodium recognition in acutely depleted animals. In addition, AI receptor cells innervated by the chorda tympani (CT) nerve were found to be necessary for normal detection of ammonium chloride (NH 4 Cl) and AI cells innervated by the facial nerve were both necessary and sufficient to discriminate NH 4 Cl from KCl. This finding suggests that taste receptor cells innervated by the facial nerve could use separate AI transduction mechanisms with different selectivities for ammonium and potassium. This work also supports the hypothesis that amiloride does not significantly impair the perception of nonsodium salts, as well as the contention that the facial nerve may provide unique information about taste quality in spite of innervating only about 30% of the taste buds in the oral cavity. xi

PAGE 12

CHAPTER 1 LITERATURE REVIEW Introduction The current experiments were designed to relate events at the level of the taste receptor cell due to salt stimulation with the perception experienced by the animal, in this case the Sprague-Dawley rat. To this end, performance was measured on a variety of psychophysical tasks including detection threshold, discrimination and recognition on a brief-access test using several salt stimuli. The effects of physiological manipulations such as gustatory nerve transection, oral application of the ion channel blocker amiloride, and acute sodium depletion on taste-guided behavior were also assessed. Prior to elaborating on the details of these experiments, the remainder of this chapter consists of a brief introduction to the anatomy and physiology of the mammalian gustatory system as well as taste coding theory for readers that might be unfamiliar with these concepts. Peripheral Gustatory System Taste Buds In mammals, the sensory cells of the gustatory system are found in clusters of approximately 50 epithelial cells called taste buds (Miller, 1995). Some of these modified epithelial cells are taste receptor cells (TRCs) and contain receptors on their apical membranes capable of interacting with taste compounds. Taste stimuli contact the apical region of the TRC by way of a taste pore. Taste buds also contain support cells that do not have taste receptors. Some support cells are immature TRCs, while others are thought to provide structural support to the receptor cells (Kinnamon, 1987). The majority of taste 1

PAGE 13

2 buds are found on the tongue in distinct epithelial structures called papillae and in the soft palate. In the rat, fungiform, or mushroom-shaped papillae, are located on the anterior two-thirds of the tongue, and a single, round circumvallate papilla can be found on the posterior surface of the tongue. Foliate papillae resembling gills are located on the sides of the tongue toward the back. A small number of taste buds can also be found on the larynx, epiglottis, nasoincisor ducts and esophagus. In the rat, the circumvallate papilla contains considerably more taste buds than the other papillary fields (~ 60% of the total number). The fungiform papillae contain approximately 15%, the palate 15%, and the remaining 5-10% are distributed among the other gustatory fields (Miller, 1995; Travers & Nicklas, 1990). Individual variance commonly exists both in the number of papillae per anatomic region and the number of taste buds per papilla (Miller, 1995). Dendrites from a single axon can synapse with cells from more than one taste bud, including buds located in different papillae. It is unclear whether buds that synapse with the same nerve fiber express the same receptor proteins. This system would seem to suggest convergence of information; however, individual taste buds can also be innervated by more than one axon, suggesting that peripheral coding has the potential to be somewhat more complex than it initially appears. In addition, TRCs are replaced every 10 days (Beidler & Smallman, 1965). In spite of this, response profiles from the nerves remain remarkably stable. It has been proposed that soluble factors released by the nerve or by adjacent cells affect the development of the new cell (see Miller, 1995). The number and variety of receptors found on each TRC are also a source of debate. Intracellular recording techniques have shown that most TRCs respond to a number of

PAGE 14

3 stimulus classes, suggesting that most, but not all, cells contain several different types of taste receptors (Gilbertson et al., 2001). When a stimulus interacts with a taste receptor, the TRC undergoes a change in membrane potential or intracellular calcium concentration leading to neurotransmitter release (see Herness & Gilbertson, 1999). Hence, the TRCs form chemical synapses with the innervating dendrites. When the TRC releases enough neurotransmitter, an action potential is produced in the innervating neuron. In addition to transmitting taste information to the peripheral nerves, these neurotransmitters might also bind to other TRCs in the taste bud, modifying their activity. Neurotransmitters associated with TRCs to date include serotonin, GABA and norepinephrine (Herness & Gilbertson, 1999). The presence of GABA suggests that inhibition as well as excitation may provide meaningful gustatory information. Peptides also thought to be released by TRCs with stimulation include bombesin, cholescystokinin, histidine, neuropeptide Y, and somatostatin (Norgren, 1995). Cells in a taste bud might also communicate with one another by way of gap junctions (Holland et al., 1989). Such a mechanism could lead to the excitation of adjacent cells without the use of neurotransmitters. Because not all TRCs synapse with taste afferents, electrical coupling could potentially allow information to be transmitted from a greater number of cells than chemical transmission alone (Herness & Gilbertson, 1999). Gustatory Nerves Rats have 3 main nerves that carry taste input. The chorda tympani (CT) branch of the facial nerve, or cranial nerve VII (CN VII), innervates the taste buds in the anterior two-thirds of the tongue (i.e., the fungiform papillae). The greater superficial petrosal (GSP) branch of this nerve innervates taste buds in the palate and nasoincisor ducts and the

PAGE 15

4 glossopharyngeal nerve (GL), or CN IX, carries taste information from taste buds located in the circumvallate papilla and a portion of the foliate papillae. The CT innervates the remainder of the taste buds in the foliate papillae. The superior laryngeal (SLN) branch of the vagus nerve, or CN X, also synapses with taste buds. These are located in the epiglottis, esophagus and larynx and are thought to be more important for airway protection than for the perception of taste quality (Dickman & Smith, 1988; St John & Spector, 1998; Smith & Hanamori, 1991). The somata of these gustatory cranial nerves are located in 3 ganglia. The CT and GSP somata reside in the geniculate ganglion. Glossopharyngeal somata are found in the inferior petrosal ganglion, and the cell bodies of the SLN are in the inferior nodose ganglion (Miller, 1995). These ganglia also provide some parasympathetic innervation to salivary glands (see Smith et al., 1988). In addition to differences in the receptor fields innervated by these nerves, the nerves also differ in response profile. For instance, electrophysiology has shown that the CT contains 2 classes of fibers, one that is selective for sodium and lithium salts and one that is more broadly-tuned (Frank et al., 1983). This second class of fibers is highly responsive to salts and acids and more moderately responsive to alkaloids like quinine. The GSP is highly responsive to sugars while moderately responsive to salts, acids and quinine (Nejad, 1986; Sollars & Hill, 1998). The GL contains 3 classes of fibers; one class that responds best to salts and acids, one that responds best to sugars, and one that is highly responsive to alkaloids (Boudreau et al., 1983; Frank, 1991). Fibers from the SLN do not seem to form units based on chemical sensitivity but are highly responsive to stimuli described as sour or bitter by humans (Dickman & Smith, 1988).

PAGE 16

5 Central Gustatory System Medullary and Pontine Taste Nuclei Second-order neurons of the gustatory system reside in the lateral region of the rostral nucleus of the solitary tract (NST) located in the medulla. Terminal branches from the peripheral nerves are distributed such that the nerves innervating the most rostral part of the oral cavity (i.e., the CT and GSP) synapse most rostrally in the gustatory NST. The GL is more caudally represented and the vagus nerve more caudal still. In spite of this general division, however, there is still a large degree of overlap in the terminal fields of the 3 main gustatory nerves (Hamilton & Norgren, 1984; Travers & Norgren, 1995). Trigeminal fibers also terminate in the lateral portion of the rostral NST while visceral projections from vagal afferents terminate in the caudal NST (Finger, 1987; Travers, 2002). The majority of axons from neurons in the gustatory NST project to the ventromedial subnucleus of the parabrachial complex, henceforth referred to as the parabrachial nucleus (PBN). This nucleus is located in the dorsal pons and receives both gustatory and visceral input. These inputs are mostly, but not entirely, segregated (Hermann & Rogers, 1985). Apart from the PBN, some gustatory neurons in the rostral NST project to the parvicellular reticular formation where they synapse with parasympathetic salivatory neurons and cells from the hypoglossal nucleus (see Halsell et al., 1996; Norgren, 1995). These pathways are thought to be responsible for reflexive oromotor responses to taste stimuli (see Travers & Norgren, 1983). In primates, most projections from the gustatory NST bypass the PBN and synapse directly onto thalamic neurons (Beckstead et al., 1980).

PAGE 17

6 Ascending Gustatory Pathways From the PBN, input from taste nerves is channeled into 2 separate functional pathways. The lemniscal or thalamocortical gustatory pathway projects bilaterally in the parvocellular region of the ventral posteromedial nucleus (VPMpc) of the thalamus before reaching the cortex (Kosar et al., 1986; Norgren & Leonard, 1973). Although this pathway is often touted as being responsible for learned taste associations, there is evidence to the contrary. For instance, ibotenic acid lesions of the gustatory thalamus fail to impair conditioned taste aversion acquisition (Flynn et al., 1991; Reilly, 1998). In addition, both the amygdala and PBN send some projections directly to the gustatory cortex, which could be necessary for sensory discrimination, although it is not clear whether these fibers carry information regarding taste quality. The ventral gustatory pathway, or pontolimbic system, projects mainly from the PBN to the amygdala, but also projects to the lateral hypothalamus, bed nucleus of the stria terminalis and other limbic regions of the forebrain (Norgren, 1995). This pathway is thought to be important for the hedonic responses to taste stimuli. In addition to gustatory projections from the hindbrain to the forebrain, telencephalic structures also project back onto the thalamus, PBN, NST, amygdala and hypothalamus. The amygdala and hypothalamus also send axons to the PBN and NST and the PBN reciprocally innervates the NST. These projections from limbic and cortical areas could lead to feedback onto taste pathways from other sensory modalities, as well as from areas involved with motivation and emotion (see Norgren, 1995). Recent evidence supporting this hypothesis has shown that the tuning characteristics of taste-responsive neurons in the PBN are altered with stimulation of either the lateral hypothalamus or central amygdala (Cho et al., 2002; Lundy & Norgren, 2001).

PAGE 18

7 The NST appears to be necessary for rats to respond to the taste quality of a stimulus. When this area is lesioned, concentration-response functions become flattened for all 4 prototypical stimuli (Shimura et al., 1997). While the NST appears to be necessary for the perception of taste quality, it has been suggested that the role of the PBN in taste processing involves the integration of taste quality with visceral input (see Grigson et al., 1998; Spector et al., 1992). For example, after bilateral PBNX a rat is unable to acquire a conditioned avoidance response, a sodium appetite or display successive negative contrast (Flynn et al., 1991; Grigson et al., 1994; Grigson et al., 1998; Scalera et al., 1995), but can still form associations between 2 tastants or between a somatosensory cue and malaise (Grigson et al., 1998; Reilly et al., 1993). Expression of a conditioned taste aversion acquired prior to PBN lesion does not seem to be affected by the surgery (Grigson et al., 1997). Lesions of the gustatory thalamus and cortex appear to have little or no effect on innate taste processing in the rat (Reilly & Pritchard, 1996), but may affect the acquisition and/or retention of a conditioned taste aversion (see Spector et al., 1992 for review). Grigson and colleagues (1998) have proposed that in addition to the NST and PBN, at least one forebrain structure, such as the amygdala or thalamus, must be intact to elicit conditioning to a taste stimulus. Electrophysiological recordings from neurons in the gustatory cortex have suggested that these cells might respond to the hedonic characteristics of the taste stimulus rather than to its associated taste quality (Yamamoto et al., 1989). Recent evidence using chronic recording techniques suggests that taste quality might be discerned from temporal patterns at this level recorded between 0.2 and 1 s after stimulus onset (Katz et al., 2001).

PAGE 19

8 Taste Transduction Primary Tastes There are 4 prototypical classes of tastants; salty, sweet, bitter and sour. A fifth taste known as umami that has been described as being meaty or savory is gaining acceptance among taste researchers and is typified by the taste of monosodium glutamate, or MSG. Umami compounds often modify the taste qualities of other compounds in addition to producing a taste themselves. It should be noted, however, that these categories are not definitive as they are based on the taste quality most associated with a particular stimulus, or set of stimuli, by human subjects and often only within a narrow range of effective concentrations. Sodium chloride, although often used as the prototypical salty stimulus in taste research, is said to taste sweet at very low concentrations (Bartoshuk et al., 1978), indicating that a stimulus may produce more than one taste quality depending on concentration. In addition, this salt has also been described as sour/salty at moderate concentrations, suggesting that even a prototypical stimulus like NaCl can produce significant side-band tastes (van der Klaauw & Smith, 1995). The fact that these taste qualities are based on human data also casts doubt on whether these 5 categories, or their prototypical stimuli, are stable across species. Other possible taste categories include electric, fatty acid and water although much less is known about the perceptual and physiological properties of these sensations than the others mentioned here (see Gilbertson et al., 1997; Ninomiya & Funakoshi, 1989; Shingai, 1980). The prototypical classes of tastants can be divided into 2 groups based on transduction mechanism.

PAGE 20

9 Transduction Mechanisms Ion channels Salt and acid stimuli dissociate in the saliva releasing free protons in the case of acids, and ions in the case of salts that pass through ion channels in the TRC membrane. Epithelial sodium channels (ENaCs), probably the best-characterized ion channels in the oral cavity of the rat, are voltage-sensitive and highly selective for sodium and lithium ions (see Garty & Palmer, 1997; Ye et al., 1991). Less selective, voltage-insensitive ion channels also allow sodium ions to pass into the cell (Ye et al., 1991). Acid transduction mechanisms fall into 3 main categories. Protons can pass into the TRC as a result of the concentration gradient, as observed in the hamster (Gilbertson et al., 1992), block cation channels, as observed in the frog and mudpuppy (see Kinnamon & Margolskee, 1996) or bind to sites on the basolateral membrane of TRCs (DeSimone et al., 1995). Metabotropic receptors Stimuli giving rise to sweet, bitter and umami sensations in humans have been shown to activate G-protein-coupled receptors in the oral cavity (see Kinnamon & Margolskee, 1996). A variety of second messenger pathways are reportedly affected by this coupling, including cAMP, inositol triphophate (IP 3 ) and diacyl glyceride (DAG). Both cAMP and IP 3 pathways are activated by stimuli that produce bitter or sweet taste qualities. Artificial sweeteners and amino acids are thought to use the IP 3 pathway, while natural sugars are thought to use a cAMP-dependent pathway (Herness & Gilbertson, 1999). Unfortunately, there is not a one-to-one correspondence between taste qualities and transduction pathways. For example, 2 stimuli with different taste qualities can share a common route of transduction. In the hamster, although both salts and acids use the same

PAGE 21

10 amiloride-sensitive pathway (Gilbertson et al., 1992), these stimuli result in different qualitative perceptions (Nowlis et al., 1980). In addition, stimuli that share a common taste quality like quinine and denatonium, stimuli described as bitter by humans, can affect more than one transduction pathways. Quinine is thought to be cAMP-dependent and denatonium uses an IP 3 second messenger pathway (Herness & Gilbertson, 1999), yet these 2 compounds produce a unified taste perception (Spector & Kopka, 2002). A single stimulus can also potentially activate more than one transduction pathway. Sodium ions activate 2 transduction pathways in the rat, but only one of these appears to be important for producing the characteristic taste quality associated with sodium. Sodium Transduction Pathways Amiloride-sensitive (AS) pathway The amiloride-sensitive transduction pathway in the oral cavity of the rat is selective for sodium and lithium ions (Brand et al., 1985; DeSimone & Ferrell, 1985) and activation of this pathway is significantly reduced with application of the ENaC blocker amiloride (DeSimone & Ferrell, 1985; Ninomiya & Funakoshi, 1988). Thus, functional amiloride-sensitive sodium channels, or ASSCs, are thought to be located in the apical region of taste receptor cells (DeSimone & Ferrell, 1985). Immunohistochemistry has also indicated that ASSCs are also located in the basal region of some TRCs, but these appear to be nonfunctional under normal circumstances (Lin et al., 1999). Sodium responses are only suppressed by amiloride in the CT and GSP nerves, suggesting that TRCs with functional ASSCs are predominantly, if not exclusively, innervated by the gustatory branches of the facial nerve (Kitada et al., 1998; Ninomiya & Funakoshi, 1988; Sollars & Hill, 1998).

PAGE 22

11 When a rat is depleted of body sodium, it will ingest increased quantities of sodium-containing solutions. This phenomenon is called sodium appetite (see Denton, 1982; Schulkin, 1991). With amiloride treatment both the magnitude and specificity of a sodium appetite are diminished (Bernstein & Hennessy, 1987). Amiloride treatment also impairs the rats ability to discriminate between NaCl and KCl (Spector et al., 1996). Together, these studies suggest that the amiloride-sensitive pathway is necessary for the rat to perceive the taste quality of sodium. While amiloride appears to be tasteless to the rat (Hill et al., 1990; Markison & Spector, 1995), humans report that it possesses a bitter quality (Schiffmann et al., 1983). Studies using human subjects also indicate that although amiloride decreases the perceived intensity of sodium salts, this decrease is observed only in the sour component of the stimulus, leaving saltiness unaffected (Ossebaard & Smith, 1995). Amiloride-insensitive (AI) pathway The amiloride-insensitive pathway, on the other hand, is permeable to a variety of ions including Na + Li + K + and NH 4 + This pathway also appears to use receptors located in the basolateral region of the TRC. For this reason, ions are thought to pass through tight junctions between cells before accessing receptor channels. Therefore, these ions must have a small radius. Amiloride is thought to be too large to fit through these tight junctions, making this pathway insensitive to the sodium channel blocker. Much less is known about this pathway than about the AS pathway as it is much more difficult to access. Although it does not appear to be affected by amiloride, transduction via this pathway can be partially blocked by introducing salts with large anions. Sodium passage through tight junctions depends on the electroneutral diffusion of ions. If the anion is too large to fit through these junctions, the cation will also fail to reach the receptor sites due

PAGE 23

12 to the unfavorable electrical gradient. In support of this hypothesis, the CT nerve is much less responsive to sodium gluconate (NaG) and other large anion sodium salts than to NaCl at the same concentration (Formaker & Hill, 1988; Ye et al., 1993), suggesting that perhaps the large gluconate ion keeps the sodium ions from passing through the tight junctions. This effect does not seem to be due to a loss of Cl receptor conductance, as Cl channel antagonists do not affect CT responses to NaCl (Elliot & Simon, 1990). Thus, the response to NaG appears to depend almost entirely on activation of the amiloride-sensitive pathway. Interestingly, the detectability functions for NaCl and NaG are strikingly similar (Geran & Spector, 2000b), suggesting that the AS pathway is not only necessary for sodium detection but also sufficient. It could be argued that this necessity and sufficiency qualify this pathway as a labeled line for sodium taste in the rat. Labeled Line vs. Across-Fiber Pattern Theory Briefly, labeled line theory states that activation of a particular set of neurons leads to the perception of a particular quality. When the activation is below a certain threshold, the quality is not perceived and when activation reaches this threshold, it is perceived. The across-fiber pattern (AFP) theory states that the overall pattern produced by the presence and absence of activation across all afferent taste fibers or central neurons produces the perceived taste quality (see Pfaffman, 1959). In the strictest sense, a labeled line should be both necessary and sufficient to produce a given perception. Most of the data supporting the AFP theory come from the fact that gustatory neurons at each level of the neuraxis respond to several tastant classes. How much of this activity is signal and how much is noise is a matter of debate. The breadth of tuning of neurons in the NST is often used to support the AFP theory. Although NST neurons are responsive to a variety of taste stimuli, 2 lines of evidence

PAGE 24

13 suggest that perhaps the activity of these neurons is more selective than previously thought. The first such evidence comes from immunohistochemistry. Experiments conducted by Travers and colleagues have shown that the immediate early gene c-Fos is differentially expressed in the NST in response to sucrose, citric acid and quinine (Harrer & Travers, 1996; Travers, 2002). Although overlap exists for the three stimuli, these data suggest that there may be some chemotopic arrangement within the gustatory regions of the CNS, although more experiments are required. The second line of evidence comes from electrophysiological examinations of higher brain regions in the gustatory neuraxis using either awake animals (Nakamura & Norgren, 1993) or animals in which the amygdala, gustatory cortex or lateral hypothalamus was stimulated in conjunction with tastant presentation (see Cho et al., 2002; Di Lorenzo & Monroe, 1995, Lundy & Norgren, 2001). Under these conditions, NST neurons are less broadly-tuned than in an anesthetized preparation in the absence of electrical stimulation. Hence, a feedback mechanism exists through which medullary and pontine taste responses could potentially be modified according to the motivational state or previous experience of the animal. If one takes a looser approach to the two theories they are not necessarily mutually exclusive. For instance, it is possible that one theory holds for some stimuli but not for others. For example, much of the behavioral evidence suggests that sodium taste might utilize a labeled line system, although the perception of other tastes might require activation of several fiber types or transduction mechanisms. Labeled lines have also been hypothesized to account for sweet taste in the hamster and chimpanzee (Danilova et al., 1998; Hellekant et al., 1998). It is also possible that the behavioral task in question might be more parsimoniously described by one theory than by the other. For instance,

PAGE 25

14 recognition or detection tasks might only require activation of one pathway for normal performance, while a discrimination task might require the animal to perceive more subtle differences in activity that might best be described in terms of AFP theory. Currently, it appears that AFP theory has become the accepted paradigm, although it can be argued that a somewhat labeled line-like approach has become pervasive in the form of best stimulus categories. Gustatory neurons are often classified according to the prototypical stimulus that produces the greatest magnitude of responding. For instance, a neuron that responds highly to sucrose, moderately to NaCl, somewhat to citric acid and only marginally to quinine would be termed a sucrose best cell. Activity from each of the cells in this category would be then be averaged and compared to the mean activity of other best stimulus categories. This method could potentially keep more subtle patterns of activity, perhaps due to second or third-best stimulus categories within a best stimulus classification from becoming apparent. In a strict across-fiber pattern code excitation of all cells regardless of best stimulus, as well as inactivity and inhibition of neurons with gustatory input, could potentially be meaningful and should therefore be analyzed. Another caveat to this method of analysis is the fact that researchers do not record from each cell in a particular anatomical region. Recording from every cell in the NST would, of course, be a herculean task, but recording only from those that are most accessible could potentially lead to a skewed sample. While tests of necessity and sufficiency can be performed at some levels of the gustatory system to shed light on the AFP vs. labeled line debate, we are not currently able to block responses from a particular best-stimulus category of the NST so that the same tenets might be applied to responses in the CNS. In spite of these caveats, predictions based on AFP

PAGE 26

15 theory appear to correlate well, for the most part, with the results of stimulus generalization tasks (see Erickson, 1963). As mentioned previously, gustatory coding is most likely best explained by some amalgamation of the 2 competing theories. The following chapters will examine the necessity and sufficiency of the AS and AI salt transduction pathways for a variety of taste-guided tasks. Whether these data support or challenge the existing coding theories will also be addressed.

PAGE 27

CHAPTER 2 ANION SIZE DOES NOT COMPROMISE SODIUM RECOGNITION BY RATS FOLLOWING ACUTE SODIUM DEPLETION Background When rats are in a sodium-depleted state, the apparent perception of a sodium-like taste quality will promote ingestion of the stimulus (e.g., Falk & Herman, 1961; Handal, 1965; Nachman, 1962; Richter & Eckert, 1938). This depletion-induced elevation of intake, termed sodium appetite, is specific for sodium and lithium salts (Nachman, 1962) and has been described in a number of mammals (see Denton, 1982). Researchers have used this phenomenon for several decades to test hypotheses about salt taste perception, most notably through the use of brief-access taste tests which substantially reduce the contribution of postingestive receptors (e.g., Breslin, et al., 1993; Falk & Herman, 1961; Handal, 1965; Nachman, 1962). The ability to recognize the taste of sodium when in a sodium-depleted state appears to depend upon taste receptor cells in the oral cavity that contain ion channels selective for sodium and lithium ions (Bernstein & Hennessy, 1987; McCutcheon, 1991; Roitman & Bernstein, 1999). The sodium-selective ion channels, or epithelial sodium channels (ENaCs), expressed by these cells can be blocked with the drug amiloride (see Brand et al., 1985; DeSimone & Ferrell, 1985; Doolin & Gilbertson, 1993; Heck et al., 1984; Schiffman et al., 1983). A second transduction pathway for sodium, the amiloride-insensitive (AI) pathway is not sodium-selective, but instead appears to be activated by a variety of cations including Na + K + and NH 4 + (Brand et al., 1985; DeSimone & Ferrell, 16

PAGE 28

17 1985; Kloub et al., 1997;Ye et al., 1994). Instead of ions passing directly through apical ENaCs, activation of this pathway is thought to involve the diffusion of ions across tight junctions and also perhaps through less selective ion channels in the apical membrane (DeSimone & Ferrell, 1985; DeSimone et al., 2001; Elliot & Simon, 1990; Simon, 1992; Ye et al., 1993; 1994). Activation of the AI pathway can be significantly reduced by pairing the cation with an anion of large hydrated radius, like acetate or gluconate, thus reducing the diffusion of ions across tight junctions (Elliot & Simon, 1990; Rehnberg et al., 1993; Simon, 1992; Ye et al., 1993). Although recent data support the existence of an amiloride-insensitive, nonselective cation channel in the apical membrane (DeSimone et al., 2001; Gilbertson & Zhang, 1998), large anion salts apparently do not significantly stimulate this pathway. Potassium gluconate, for example, is a very poor stimulus (Stewart et al., 1996; Ye et al., 1994) and amiloride treatment virtually eliminates the CT response to sodium acetate (NaAc) and sodium gluconate (NaGlu), especially at low to mid-range concentrations (Elliot & Simon, 1990; Formaker & Hill, 1988; Ye et al., 1993). Previously, it has been shown that the amiloride-sensitive (AS) pathway is both necessary and sufficient for normal Na + detection in the rat, at least to the extent that gluconate is capable of blocking AI transduction (Geran & Spector, 2000a; 2000b). It is possible, however, that these near threshold concentrations, although detectable, were not perceived as tasting sodium-like by the animals. For instance, humans often report that low concentrations of NaCl taste sweet rather than salty (Bartoshuk et al., 1978). Consequently, the effect on sodium recognition of reducing the contribution of the AI pathway(s) with a large anion was tested. Other researchers have shown that rats will

PAGE 29

18 ingest a variety of sodium salts when depleted, but to our knowledge all have used single stimulus (Handal, 1965; Krieckhaus & Wolf, 1968; Morrison & Young, 1971), 24-h intake (Fregly, 1958; Richter & Eckert, 1938), or 2-bottle sodium vs. nonsodium salt tests (Nachman, 1962), making comparisons among taste-guided preferences for different sodium salts impossible. Instead, we used a brief-access test so that preference for several salts could be analyzed simultaneously and without potentially confounding postingestive effects. Three sodium salts with different sized anions were used to limit AI transduction (Elliot & Simon, 1990; Formaker & Hill, 1988; Kitada et al., 1998; Ye et al., 1993). Sodium gluconate in particular was chosen based on its ineffectiveness in stimulating AI transduction (Ye et al., 1991; 1993). This experiment marks the first time to our knowledge that this salt has been presented to sodium-depleted rats. Two concentrations of each salt were chosen on the basis of mean sodium detectability functions measured previously (Geran & Spector, 2000a; 2000b). One concentration was well above detection threshold in the presence of amiloride (0.3 M), and the other below threshold when mixed with amiloride, but above threshold in the absence of amiloride (0.03 M). Potassium chloride and NH 4 Cl were included in the stimulus array for the purpose of comparison. If the AS pathway is sufficient as well as necessary for sodium recognition, sodium-depleted animals should show similar amounts of licking to all sodium salts regardless of anion. Furthermore, this increase in licking to sodium salts should be abolished with the addition of amiloride. Methods Subjects A total of 40 nave male Sprague-Dawley rats (Charles River Breeders, Wilmington, MA) were used in this experiment. Subjects were tested in 2 groups of 20. All rats

PAGE 30

19 weighed approximately 250-300g at the start of training and were placed on a 12:12 light:dark schedule with lights on at 6:00 AM. Humidity and temperature were kept constant. Rats were individually housed in hanging wire mesh cages and given ad libitum access to laboratory chow (PMI 5001 pellets. PMI Nutrition International, Brentwood, MO) except where noted. Access to distilled water, however, was restricted while in the home cage. Water bottles were removed ~24 h prior to training and replaced after the last training session 5 days later. Rats had access to water only during 30-40 min training sessions on these 5 days. Body weights were closely monitored for excessive dehydration. No rat dropped below 85% of its individual ad libitum weight during this experiment. All procedures were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Florida. Apparatus All training and testing occurred in a modified, automated taste testing apparatus called a gustometer (Spector, Andrews-Labinski, & Letterio, 1990). This apparatus was designed to deliver small volumes of fluid stimuli and record the number of licks for each stimulus presented. A narrow slot in one wall of the chamber enabled the rat to access a vertically-oriented sample spout. This spout rotated into position in front of the slot at appropriate times throughout the session. Each taste stimulus was kept in a pressurized syringe connected to the sample spout by way of Teflon tubing and a solenoid valve. Each valve was opened for a preset amount of time by the computer such that each lick delivered ~5 l of fluid after the drinking shaft was initially loaded with the stimulus (see Spector, Andrews-Labinski, & Letterio, 1990 for more details on stimulus delivery). A very low current (< 50 nA) contact circuit was used to monitor number of licks taken.

PAGE 31

20 Training Procedure On the first day of training, a drop of water was placed on the tip of the spout and on the inside wall of the chamber before each session to help in training the rat to drink from the spout. During the first 2 days of training the sample spout remained motionless in front of the access slot so that the rat had continuous access to distilled water for each 30-min session. After all animals had been trained to lick from the spout, they received 3 days of additional spout training in which they received access to distilled water and 0.1 M sucrose in 5s trials. Sucrose was used to encourage licking. The rat was required to lick the dry spout twice in 1 s to receive a stimulus presentation. When the trial was finished, the spout rotated away from the access slot and over a funnel where it was rinsed with distilled water and dried with pressurized air. This process took about 6s to complete. The spout then rotated back in front of the slot. These sessions lasted 40 min. Sodium Depletion After training was complete, rats were assigned to one of 4 groups. These groups were counterbalanced for body weight, number of trials initiated during the last 3 days of training and mean number of total licks for these 3 days. Distilled water bottles were placed on the home cages Friday following the last training session. On Monday morning the rats were moved from their standard cages to metabolism cages equipped with funnels to collect urine. They were also given a weighed amount of powdered chow at this time instead of pellets. Rats in the 2 sodium-depleted groups received Harlan Teklad 90228 sodium-deficient (0.02% NaCl) chow (Harlan Teklad, Madison, WI). Rats in the 2 non-depleted groups received the same chow mixed with 1.0% NaCl. Twenty-four hours prior to testing, each rat in the sodium-depleted groups received the first of 2 equal volumes of furosemide (total dose = 30mg/kg BW, s.c.). The second furosemide injection

PAGE 32

21 was given 2 hours later. Rats in the non-depleted groups received injections of isotonic saline (s.c.) using the same injection schedule and volume as rats given furosemide. Subjects had free access to powdered chow (with or without sodium) and distilled water during the sodium depletion phase of the experiment. Urine was collected in 100 mL flasks for 24 h immediately following the first injection. Brief-Access Testing Testing took place in the gustometers 24 h after each animals first furosemide injection. Animals were given brief access to 11 stimuli (e.g. distilled water and 0.03 & 0.3 M concentrations of NaCl, KCl, NH 4 Cl, sodium acetate (NaAc), and sodium gluconate (NaGlu)) over a 40-min period. Stimuli were presented in randomized blocks except that the first trial of each session was always 0.3 M NaCl to encourage continued sampling. All salt solutions were made fresh using reagent grade chemicals (all salts from Fisher Scientific, Orlando, FL except NaGlu from Sigma Chemical Co., St. Louis, MO) and distilled water. One liter of 100 M amiloride (Sigma Chemical Co.) stock solution was made the evening prior to testing and wrapped in aluminum foil and left on a stir pad overnight in a dark room. Two of the 4 gustometers contained the aforementioned salts dissolved in distilled water, while the remaining 2 gustometers contained salts dissolved in 100 M amiloride and 100 M amiloride in place of the distilled water stimulus. Half of the rats from each depletion group were tested with amiloride as the solvent. At the time of testing, the remaining chow was removed from the home cage and weighed to determine how much was ingested in the previous 24 h after attempting to account for spillage. Distilled water intake and urine output for each rat were measured to

PAGE 33

22 the nearest mL. A 2.0 mL sample of urine was collected for each rat and frozen in labeled plastic centrifuge tubes for later analysis. Urine Analysis Urine was analyzed using a flame photometer to determine sodium content for both sodium-depleted and non-depleted rats. Urine from sodium-depleted rats was diluted with distilled water (4 parts water: 1 part urine) prior to analysis so that the sodium concentration of each sample would fall within the range testable by the device. The values for these animals were then multiplied by a factor of 5. Data Analysis Lick data were recorded and analyzed for the entire 5s of each trial, but the main parameter of interest was the number of licks to each stimulus during the last 3s of each 5s trial (i.e., the avoidance period). This was done to minimize the number of sampling licks included in the analysis. The local lick rate for rats is approximately 7 Hz (Corbit & Luschei, 1969; Halpern, 1977), making 35 licks the ceiling in a 5s trial and 21 licks the highest performance attainable in a 3s period. The number of licks for each stimulus was then averaged for each group and compared using analyses of variance (ANOVAs) and t-tests (paired, independent and one-sample tests). The statistical rejection criterion was set at .05 for all analyses. P-values were adjusted using the Bonferroni method when a large number of t-tests were performed on the same data set. Lick data from an animal were included in the analyses only if the animal sampled all 11 stimuli in the test. Only 2 rats in the non-depleted/amiloride group and 1 rat in the non-depleted/water group passed this criterion. This sample size was too low for meaningful analysis, so all statistical tests of stimulus licking were performed on sodium-depleted rats only. Data from each rat were used, however, to analyze the number of trials initiated during testing and degree of

PAGE 34

23 sodium depletion for each of the 4 groups. Sodium balance was determined by subtracting urinary sodium output from sodium ingested. Results Brief-Access Testing The overall pattern of responsiveness for the sodium-depleted groups was the same regardless of whether the 3s avoidance period or the entire 5s period was analyzed. One-way ANOVAs of mean licks during the last 3s of each trial revealed a main effect of stimulus for both the amiloride (F(10, 90) = 35.4, p < .001) and distilled water (F(10,90) = 2.7, p < .008) groups. Intake of distilled water and amiloride did not differ significantly between groups (p > .09, 10.3 + 4 vs. 13.6 + 4 licks respectively). For rats in the distilled water condition, paired t-tests indicated that they licked significantly more to each of the 6 sodium stimuli (i.e. 3 salts, 2 concentrations) than to water (p < .005 for each t-test. Bonferroni adjusted p < .05. Figure 2-1). Furthermore, a 2-way ANOVA (anion x concentration) of responses to sodium salts revealed a significant main effect for concentration only (F(1,9) = 26.9, p < .002). One-way ANOVAs indicated no differences in performance across sodium salts at the 0.03 M concentration and a slight effect at the 0.3 M concentration (F(2,18) = 3.99, p < .04), with the greatest number of licks recorded for the salt with the intermediate-sized anion, sodium acetate (NaAc). Paired t-tests indicated significantly less licking to the NH 4 Cl solutions than to water (p < .008 for both) but this significance disappeared with a Bonferroni test (adjusted ps > .07). The KCl stimuli were not different from water (p > .34 for both, unadjusted). In the amiloride condition, .03 M NaGlu and .3 M NH 4 Cl were licked less than amiloride alone (p < .05 for both). Again, the statistical significance of this difference

PAGE 35

24 disappeared with a more conservative test (adjusted ps > .4). All other stimuli generated licking comparable to that induced by amiloride alone (all ps > .14, unadjusted). A two-way ANOVA (anion x concentration) for all sodium salts indicated that all 6 sodium stimuli were licked to similar degrees in the presence of amiloride (ps > .14). Number of Trials Initiated Not surprisingly, the animals in the sodium-depleted groups initiated a greater number of trials than rats in the non-depleted groups (p < .001). However, sodium-depleted rats in the amiloride condition took far fewer trials than sodium-depleted rats in the distilled water condition (p < .001, Figure 2-2). This suggests that although more motivated to lick than non-depleted rats, the presence of amiloride reduced the number of trials initiated by sodium-depleted rats. Sodium Balance All rats in the sodium-depleted groups were in negative sodium balance while rats in the non-depleted groups ingested more sodium than they excreted (p < .004 for each of the 4 groups. Figure 2-3). There was no difference in sodium balance between rats in the distilled water condition and those in the amiloride condition for either depletion group (both ps > .20). Discussion It is clear from the data that the sodium-specific appetite exhibited by rats in the distilled water condition was not evident in the amiloride condition. When amiloride was added to the taste stimuli, intake of the sodium salts was not greater than intake of either the nonsodium salts NH 4 Cl and KCl or of 100 M amiloride alone, strongly suggesting that the animals were unable to recognize the sodium salts and to ingest them preferentially. These findings support previous reports of impaired sodium appetite with

PAGE 36

25 amiloride (Bernstein & Hennessy, 1987; McCutcheon, 1991; Roitman & Bernstein, 1999), and extend them by showing that the taste-guided specificity of the sodium appetite is completely abolished by amiloride treatment in a brief-access test. This effect of amiloride on the specificity of the sodium appetite differentiates amiloride treatment from CT transection, which has been shown to impair but not always eradicate sodium-specific appetite (see Breslin et al., 1993; Markison et al., 1995). This difference is likely due to the fact that the greater superficial petrosal (GSP) branch of the facial nerve also contains amiloride-sensitive fibers important for the maintenance of certain taste-guided tasks involving sodium (Roitman & Bernstein, 1999; Sollars & Hill, 1998). Our results also differ from those of CT transection in certain other regards. For instance, prior studies reported increased licking to low KCl concentrations with transection (Breslin et al., 1993) that we did not observe with amiloride treatment. This is most likely due to the fact that CT transection reduces the perceived intensity of KCl in rats while amiloride does not (Geran et al., 1999). Procedural differences between this experiment and the brief-access studies with CT transection should also be taken into account when comparing performance between experiments. These include a shorter sample time (5s vs. 10s), a greater sodium : nonsodium salt ratio in the stimulus array (3:2 vs. 1:4) and a slightly lower (0.03 vs 0.05 M) concentration of KCl in the current experiment. Also of importance is the fact that sodium-depleted rats in the distilled water condition increased intake of all sodium salts (NaCl, sodium acetate and sodium gluconate) to a similar degree. Although previous studies have reported licking to a variety of sodium salts following sodium depletion in rats (see Fregly, 1958; Handal, 1965; Nachman,

PAGE 37

26 1962) the present study is the first sodium appetite experiment to our knowledge to include a variety of sodium salts in a single brief-access test. Our results strongly suggest that eliminating, or at least severely reducing, the contribution of the AI transduction pathway by changing the size of the anion does not compromise sodium recognition. Thus, the AS sodium transduction pathway appears to be both necessary and sufficient for sodium recognition in the rat just as it is for normal sodium detection. Evidence of an apically-located AI cation transduction pathway in the oral cavity of the rat (DeSimone et al., 2001; Gilbertson & Zhang, 1998) may call this sufficiency into question if this nonselective apical pathway is substantially activated by large anion sodium salts like NaGlu. However, current electrophysiology indicates that only a negligible portion of the CT response to NaAc and NaGlu remains with amiloride treatment at the stimulus concentrations used (Elliot & Simon, 1990; Formaker & Hill, 1988; Ye et al., 1993). The functional role of the AI transduction pathway to taste function is yet to be understood, but might involve the detection and/or recognition of nonsodium salts. Support for this hypothesis comes from evidence that the discrimination of 2 nonsodium salts, NH 4 Cl and KCl, is severely compromised by combined transection of the CT and GSP nerves but unaffected by amiloride treatment (Geran et al., 2002). This suggests that recognition and discrimination of these salts depends on amiloride-insensitive, salt-responsive units in these nerves. As noted above, normal KCl detectability is also compromised by CT transection but not significantly affected by amiloride treatment, suggesting that an AI pathway with input carried by the facial nerve is required for this task (Geran et al., 1999). Discrimination among sodium salts might also depend upon AI transduction. The CT response to NaCl is of greater magnitude than the response to NaAc

PAGE 38

27 or NaGlu due to a larger amiloride-insensitive response (Elliot & Simon, 1990; Formaker & Hill, 1988; Ye et al., 1993). This comparative increase in AI response does not appear to confer any particular salience to NaCl over other sodium salts at near-threshold levels (Geran & Spector, 2000b), but it could affect the taste quality or intensity of sodium salts in rodents at superthreshold levels. Humans have rated NaCl as more intense than NaGlu at equimolar concentrations (Ossebaard & Smith, 1995). Recordings from the GSP, however, indicate roughly equivalent amiloride suppression to both NaCl and NaAc at concentrations of .1 M and higher, but at .05 M the response to NaCl is more suppressed (~75%) than the response to NaAc (~50%, Sollars & Hill, 1998). This suggests that AI sodium transduction in the GSP may be unaffected or even enhanced by larger anions. Alternatively, a third pathway might exist that is independent of sodium but responsive to acetate. Responses of the GSP to NaGlu have not been reported. Clearly, more research on the salt responsiveness of this nerve is needed. Sodium-depleted rats in the distilled water condition increased their intake of both 0.03 and 0.3 M NaAc and NaGlu relative to water, suggesting that the size of the anion did not affect their ability to perceive the taste of sodium. In the amiloride condition, however, rats did not lick NaAc or NaGlu more than amiloride alone at either stimulus concentration, suggesting that sodium recognition was compromised. This is particularly noteworthy at the 0.3 M NaGlu concentration, due to the fact that rats were shown to detect this stimulus, albeit poorly, in the presence of amiloride (Geran & Spector, 2000b). Together, these results suggest that rather than responding to sodium activation of the AS transduction pathway, the behavioral detectability of high NaGlu concentrations derives from some cue related to the gluconate anion or perhaps activation of the AI pathway due

PAGE 39

28 to leakage of sodium through tight junctions, or AI apical channels, as a result of the high Na + concentration gradient. Sodium-depleted rats initiated a significantly greater number of trials than non-depleted rats regardless of whether amiloride was added to the stimuli. This indicates an increase in appetitive behavior (see Craig, 1918; Denton, 1982), even in the absence of a sodium taste cue. When the taste of sodium was present, sodium-depleted rats showed a further increase in the number of trials initiated. Thus, under these conditions, the taste of sodium in the absence of need does not produce an increase in appetitive behavior, while need in the absence of the appropriate taste cue does, although not to the same degree as need and gustatory cue combined. In summary, these data support Bernstein & Hennessys (1987) conclusion that the AS sodium transduction pathway is necessary for sodium recognition in the rat, and furthermore strongly suggest that this pathway is also sufficient (see Elliot & Simon, 1990; Formaker & Hill, 1988; Ye et al., 1993). We have also extended previous studies of the effects of amiloride on salt appetite to show that the sodium specificity of the appetite is completely abolished when the AS pathway is blocked. In addition, these findings suggest that although NaGlu concentrations higher than 0.1 M are detectable in the presence of amiloride they appear to lack the characteristic taste quality associated with sodium.

PAGE 40

29 DH2O MEAN LICKS IN LAST 3 s 05101520 AMILORIDE .03.3NaClKClNaGNaAcNH4Cl.03.03.03.03.3.3.3.3.03.3NaClKClNaGNaAcNH4Cl.03.03.03.03.3.3.3.3 Figure 2-1. Brief-access licking to each stimulus by sodium-depleted rats. Mean ( + SE) number of licks taken by sodium-depleted rats in the last 3s of each 5s bout. Stimuli were dissolved in either distilled water (left) or 100 M amiloride (right). Horizontal lines indicate mean ( + SE) number of licks to either distilled water or amiloride alone. All sodium salts (chloride, acetate and gluconate) were preferred over water (paired t-tests, p < .005 for each. p < .05 Bonferroni adjusted), while none of the salts, sodium or nonsodium (potassium chloride and ammonium chloride), were preferred over amiloride. Responses to water and amiloride alone were not significantly different.

PAGE 41

30 GROUP # TRIALS 020406080100120140160180200 Na+-DepletedNon-DepletedDH2OAMILAMILDH2O Figure 2-2. Mean ( + SE) number of trials initiated by each group of rats. Non-depleted rats took the fewest trials regardless of whether amiloride was present. Sodium-depleted rats, however, initiated considerably fewer trials in the presence of amiloride (p < .001).

PAGE 42

31 GROUP Na+ DIFFERENCE (INPUT OUTPUT) in mmol -3-2-10123 DH2OAMILDH2OAMILNa+-DepletedNon-Depleted Figure 2-3. Mean ( + SE) sodium balance for each group of animals measured in mmol. All rats in the sodium-depleted groups were in negative sodium balance prior to brief-access testing while rats in the non-depleted groups consumed more sodium than they had excreted (p < .004 for each group in a one-sample t-test).

PAGE 43

CHAPTER 3 GLOSSOPHARYNGEAL NERVE TRANSECTION DOES NOT IMPAIR POTASSIUM CHLORIDE VS. AMMONIUM CHLORIDE OR SODIUM CHLORIDE VS. AMMONIUM CHLORIDE DISCRIMINATION Background Ammonium and potassium chloride have been shown to taste very similar to rats in generalization tasks (Erickson, 1963; Hill et al., 1990; Morrison, 1967). These results have been used to support both similarities in NH 4 + and K + transduction at the receptor level and similarities in the neural response pattern at the level of the NTS (DeSimone et al., 2001; Erickson, 1963). Recently, it was reported that rats placed on a KCl vs. NH 4 Cl discrimination task with overlapping concentrations were able to consistently perform well (~90% overall performance) after a typical period of discrimination training (Geran et al., 2002). Furthermore, dissolving the salt stimuli in the epithelial sodium channel (ENaC) blocker amiloride did not impair performance, but transecting the gustatory branches of the facial nerve (i.e., the chorda tympani (CT) and greater superficial petrosal (GSP) nerves) reduced performance to chance levels for each animal (Geran et al., 2002). Rats trained and tested on a NaCl vs. NH 4 Cl discrimination task also exhibited chance levels of performance following nerve transection even though the mean presurgical discrimination performance for this group was approximately 95% (Geran et al., 2002). Amiloride significantly impaired performance on this task, although not to the extent reported previously for NaCl vs. KCl discrimination (see Kopka et al., 2000; Spector et al., 1996). Sodium chloride vs. NH 4 Cl discrimination was chosen as the comparison for KCl vs. NH 4 Cl discrimination performance because rats do not generalize between NaCl 32

PAGE 44

33 and NH 4 Cl in a conditioned taste aversion task (Hill et al., 1990). This suggests that NaCl and NH 4 Cl are easily discriminated by the rats, while KCl and NH 4 Cl are more difficult, if not impossible, for the animals to distinguish. The results of the prior discrimination experiment (Geran et al., 2002) confirmed these predictions and suggested that the amiloride-insensitive fibers of the facial nerve were necessary for discriminations involving NH 4 Cl, but the sufficiency of this input was not ascertained. Functional sufficiency of facial nerve input was tested in the current experiment by transecting the glossopharyngeal (GL) nerve. This is not an absolute test of sufficiency as the superior laryngeal branch (SLN) of the vagus nerve is still intact. However, this nerve is thought to be more important for airway protection than taste quality perception (see Dickman & Smith, 1988; St. John & Spector, 1998; Smith & Hanamori, 1991). The GL innervates about 60% of the taste buds in the oral cavity of the rat and contains fibers narrowly-tuned for salts and acids as well as fibers that are highly responsive to compounds described as bitter by human subjects (Frank, 1991). Humans have also reported that NH 4 Cl and KCl contain both bitter and salty components (van der Klaauw & Smith, 1995), making it possible that 1 or more types of narrowly-tuned fibers in the GL of the rat are activated by NH 4 Cl and/or KCl. It is also possible that discrimination could depend upon GL input due to the large number of taste buds innervated by this nerve. To date, GL transection has not produced substantial decrements in performance on discrimination or recognition tasks involving taste quality. These tasks include sucrose vs. maltose, citric acid vs. quinine, quinine vs. KCl and NaCl vs. KCl (see St. John, Markison, Guagliardo et al., 1997; St. John & Spector, 1997; St. John & Spector, 1998; Spector & Grill, 1992; Spector et al., 1997). Instead, performance

PAGE 45

34 on these tasks has been significantly affected by transection of one or both of the gustatory branches of the facial nerve (see St. John & Spector, 1998), in spite of the fact that the CT and GSP together innervate only about half the number of taste buds innervated by the GL (Miller, 1995). The remaining 5-10% of taste buds are innervated by the SLN (Travers & Nicklas, 1990). Methods Subjects A total of 32 adult male Sprague-Dawley rats (Charles River Breeders, Wilmington, MA) were used in this experiment. These rats were tested in 2 groups of 16. Each animal weighed between 250 and 300g at the start of training and was given ad libitum access to PMI 5001 pellets (PMI Nutrition International, Brentwood, MO) at all times while in the home cage. Access to distilled water was restricted to encourage performance during testing. Water bottles were removed from home cages approximately 24 hours prior to training (or testing) every Monday and replaced every Friday following the last session. Rats were able to gain access to water Monday through Friday by pressing the correct lever while in the apparatus. During the week, animals were closely monitored for excessive dehydration. Rats received supplemental water if they appeared dehydrated or if body weight decreased to 85% of the ad libitum weight calculated each week. All procedures were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Florida. Apparatus and Trial Structure Animals received training and testing in an operant chamber modified for taste research. This gustometer apparatus contains one spout for stimulus presentation and one for delivery of the reinforcer. It also contains two levers, one on either side of a spout

PAGE 46

35 access slot (see Spector, Andrews-Labinski, & Letterio, 1990 for further details). Each animal was allowed to complete as many trials as possible during a 40-min session. Each trial began when the rat made contact with the sample spout twice within 250 ms. This contact completed a low amplitude (< 50 nA) circuit and caused a solenoid valve to open with each subsequent lick so that the stimulus was delivered to the rats tongue. Each sample phase lasted 5 licks or 3 s whichever came first. This was followed by a 5-s decision phase, during which the stimulus spout rotated out of reach. If the rat pressed the correct lever (i.e., the lever associated with the stimulus during training), it received access to distilled water (20 licks or 10s) via the reinforcement spout. If the rat pressed the incorrect lever or failed to press either lever, the trial ended and the animal received a 20-s time out. An intertrial interval of 10s followed both the reinforcement and time-out phases. At this time, the lights in the chamber were extinguished and the sample spout was rotated over a funnel, rinsed with distilled water and blown dry with pressurized air. Sessions were controlled automatically by a computer and white noise was present throughout each session. Training Procedure Rats were counterbalanced for lever and stimulus. Half of the rats (i.e., 16) were trained to discriminate NaCl from NH 4 Cl and half were trained to discriminate KCl from NH 4 Cl. Training began with shaping the rats to press one lever following presentation of a single stimulus (0.2 M NH 4 Cl, 0.2 M KCl or 0.2 M NaCl). Stimuli were made each morning using reagent grade chemicals (Fisher Scientific, Orlando, FL). Once the rat had performed the target behavior for one session without aid, the animal was shaped to press the other lever in response to the other stimulus. This process took approximately 2 weeks. The animals were then switched to the alternation phase during which one

PAGE 47

36 stimulus was presented on each trial until the animal pressed the correct lever on a fixed number of trials. Once the criterion was reached, the other stimulus was presented. It was not necessary for the correct trials to be consecutive. Each rat moved to the next alternation criterion when it performed at 75% or better for the day. The alternation criterion decreased over five days from 8 correct presses to 2. Alternation was followed by discrimination training, during which stimuli were presented in randomized blocks. After two days of 75% performance or better, the limited hold was reduced and the time out increased. These parameters were systematically reduced when the performance of each rat reached the 75% criterion until the last phase of training in which 3 concentrations of each salt were added to the stimulus array (0.4, 0.1, and 0.05 M). Rats were kept on this phase of training until weekly performance was unchanged for 10 sessions (2 weeks). In the current experiment, this last phase of training required an average of 19 sessions. Rats on the KCl discrimination were moved from one phase of training to another simultaneously with rats on the NaCl task. See Table 3-1 or St. John, Markison, Guagliardo et al., 1997 for more detail concerning the training procedure. Presurgical Discrimination and Amiloride Testing After training, rats were tested for five days on the 8-stimulus discrimination array (4 concentrations of 2 salts). The following week all stimuli were dissolved in 100 M amiloride. Amiloride (Sigma Chemical, St. Louis, MO) was made each afternoon prior to testing in a flask wrapped with aluminum foil to minimize reactions with light and allowed to spin overnight. The salt solutions were then made the next morning using amiloride as the solvent instead of distilled water. Amiloride was also used in place of distilled water as the reinforcer during this phase of testing.

PAGE 48

37 Surgery After amiloride testing, the rats in each discrimination group were divided into 2 groups counterbalanced for body weight, mean number of trials initiated during discrimination testing and overall proportion correct. Half of the animals in each discrimination (n = 8) received bilateral GL transection (GLX) and the other half received sham transection (SHAM). All rats were anesthetized with an intramuscular injection of ketamine hydrochloride (125 mg/kg body weight) mixed with xylazine hydrochloride (5 mg/kg). An incision was made down the midline of the ventral neck and the musculature and salivary glands retracted until the GL could be blunt-dissected from the hypoglossal and vagus nerves with glass rods. The GL nerve was stretched with curved glass rods and cut with microscissors where it met the hypoglossal and vagus nerves such that approximately 5-10 mm of the nerve was removed. Rats in the sham-transected group received only midline incision followed by retraction of the musculature and salivary glands until the GL was exposed. A small amount of sterile saline was introduced into the wound of each animal before it was closed with nylon sutures. Animals received subcutaneous injections of penicillin (30,000 units Flocillin) and analgesic (2 mg/kg Ketorolac) immediately after surgery and for the next 3 days. Animals were allowed 9 (SHAM) or 10 (GLX) days to recover from surgery. One rat in each of the two GLX groups (KCl vs. NH 4 Cl and NaCl vs. NH 4 Cl) died the night of surgery. It is suspected that the vagal nerve of these animals was accidentally damaged during surgery. Postsurgical Testing and the Water Control Test After the recovery period, animals were tested on the original discrimination for 5 sessions followed by amiloride testing for an additional 5 sessions. These tests were performed exactly as they were prior to surgery (see Table 3-2 for experiment schedule).

PAGE 49

38 Rats in the second testing group (n = 15) received a 1-day water control test upon the completion of amiloride testing. Each fluid reservoir was filled with distilled water with half of these assigned to the left lever and half to the right lever. Thus, any performance significantly better than chance could be attributed to an extraneous cue, such as noise or temperature associated with a particular reservoir. Histology Upon completion of the water control test, rats were deeply anesthetized with sodium pentobarbital and transcardially perfused with physiological saline and 10% buffered formalin. The oral tissue of each animal was removed and stored in formalin until it could be analyzed. At this time, the tongue was allowed to soak in distilled water followed by dissection of the circumvallate papilla from both the anterior tongue and the underlying connective tissue. The papilla was embedded in paraffin. It was then sliced into 10 m sections using a microtome, placed on slides, stained using hematoxylin and eosin and coverslipped. Taste buds were counted under a light microscope by an observer blind to surgical group. The presence of a taste pore or the characteristic fusiform cells in the absence of a taste pore was taken as indication of an intact taste bud. Previous research has shown that taste buds degenerate in the absence of afferent innervation, allowing the use of this assay in testing the completeness of nerve transections (Cheal & Oakley, 1977; Hard af Segerstad et al., 1989; St. John et al., 1995). Data Analysis Overall discrimination performance for each rat was based on the number of trials with a correct response divided by the total number of trials with a press for each phase of the experiment collapsed across salt and concentration. Two-way analyses of variance (ANOVAs) were also performed to determine the effects of condition and concentration

PAGE 50

39 on performance to each salt. In cases where the ANOVA revealed a main effect of condition, paired t-tests were performed to determine concentrations for which performance was significantly altered. Paired t-tests were also used to compare overall proportion correct between conditions and independent t-tests were performed in some cases to determine whether between-group differences were significant. Finally, the normal approximation of the binomial distribution was applied to the results of the water control test to determine whether performance was significantly different from chance (Brown & Hollander, 1977). A more conservative Bonferroni procedure was applied to these results in order to adjust for the number of t-tests performed. Statistical significance was set at p < 0.05 for all tests. Results Presurgical Discrimination Testing Overall mean presurgical discrimination performance was greater than 90% regardless of stimuli. Furthermore, the performance of rats in the KCl vs. NH 4 Cl discrimination group did not differ from that of rats in the NaCl vs. NH 4 Cl group (p > .08). When the salts were dissolved in 100 M amiloride, KCl vs. NH 4 Cl discrimination remained unchanged, while NaCl vs. NH 4 Cl discrimination decreased significantly (p > .62 and p < .001 respectively. Figures 3-1 and 3-2). Two-way ANOVAs (condition x concentration) for each salt in the latter condition indicated an effect of condition for both NaCl and NH 4 Cl (F(1,15) > 47, p < .001 for both). Furthermore, paired t-tests indicated that performance on each concentration of each salt (NaCl and NH 4 Cl) declined significantly with amiloride treatment (p < .004 for each. Figure 3-3).

PAGE 51

40 Postsurgical Testing and Histology Mean KCl vs. NH 4 Cl performance was significantly lower (p < .03, paired t-test) for GLX rats after surgery than before surgery (post-GLX mean = 88% + 2 vs. pre-GLX mean = 92% + 2. Figure 3-4.). Performance of sham-transected rats also dropped with surgery but was not significant. Both presurgical and postsurgical performance values for the GLX group, however, failed to differ from the values obtained for rats in the SHAM group for the same condition (p > .88 for both). Two-way (condition x concentration) ANOVAs for the prevs. post-GLX performance of each salt indicated a condition effect for NH 4 Cl, but not KCl (F(1,6) = 6.5, p < .05 and F(1,6) = 5.4, p > .05, respectively). A post-hoc analysis indicated a significant change in performance with transection for 0.2 M NH 4 Cl only (p < .04). Mean NaCl vs. NH 4 Cl performance declined significantly following sham-surgery but not GL transection (p < .03 vs. p > .07, respectively. Figure 3-5.). Two-way (condition x concentration) ANOVAs, however, failed to indicate significant effects of condition or the interaction of condition and concentration for either salt in sham-transected rats (F(1,6) < 5.7, p > .05 for both salts and F(3, 18) < .5, p > 68 for both interactions. Figure 3-6). In addition, amiloride treatment did not significantly affect the postsurgical performance of GLX rats trained to discriminate KCl from NH 4 Cl (Figure 3-7). Nor did GL transection significantly impair NaCl vs. NH 4 Cl discrimination in the presence of amiloride (Figures 3-8 & 3-9). One rat performed significantly better than chance (i.e., 50%) on the water control test (60% performance, p < .03 using a one-tailed test). After performing a Bonferroni adjustment, this performance was no longer significant (z = 1.99, p > .34. Figure 3-10).

PAGE 52

41 Rats in the GLX group had significantly fewer (p < .001) circumvallate taste buds than rats in the SHAM group (0.14 + 0.14 and 422 + 19 respectively). Discussion Presurgical Discrimination Testing This study replicated the results of our previous NH 4 Cl discrimination experiment (Geran et al., 2002) in that rats on both the KCl vs. NH 4 Cl and the NaCl vs. NH 4 Cl tasks were clearly able to discriminate between the 2 salts. The prior study was also replicated with regard to the effect of amiloride on these discriminations. Sodium chloride vs. NH 4 Cl performance was compromised (11% decline in mean performance, p < .001) by the addition of 100 M amiloride, while KCl vs. NH 4 Cl was not significantly affected. Additionally, this impairment was observed at each concentration of both NaCl and NH 4 Cl, and performance remained above chance at each concentration (Figure 3-3). This concentration-response pattern is unlike the more one-sided impairment in NaCl vs. KCl discrimination observed with amiloride and less substantial (Spector et al., 1996). In the NaCl vs. KCl task, performance on NaCl trials dropped below chance with amiloride while performance on KCl trials was not significantly altered. Thus, the rats pressed the lever associated with KCl on NaCl trials in the presence of amiloride, suggesting that the taste qualities of KCl and NaCl + amiloride are very similar. This conclusion is supported by additional data from conditioned taste aversion tests and sham-intake experiments in sodium-depleted rats (Hill et al., 1990; Roitman & Bernstein, 1999). The prior NH 4 Cl discrimination experiment indicated similar concentration-response patterns for both the KCl vs. NH 4 Cl task and the NaCl vs. NH 4 Cl tasks with amiloride (Geran et al., 2002). Although performance to each concentration was well above chance for both the previous and current experiments, in general there was a slight nonsignificant

PAGE 53

42 decline in performance at the 0.05 M NH 4 Cl concentration when tested against NaCl + amiloride. This decline mirrors that observed in the KCl vs. NH 4 Cl discrimination group both in the presence and absence of amiloride (Figure 3-3; Geran et al., 2002). The fact that this decline did not result in near-chance performance and that the concentrations chosen span almost an order of magnitude suggest that the animals were most likely discriminating based on taste quality rather than intensity, although this has not been tested directly. Detection threshold tasks have indicated that all 3 stimuli (NaCl dissolved in amiloride, KCl and NH 4 Cl) are detected by rats at concentrations lower than those tested in the discrimination task (Geran et al., 1999; Geran & Spector, 2000a; Chapter 4). Although the concentrations tested are detectable to the rat, it is not clear how the stimuli compare in terms of suprathreshold intensity. Postsurgical Discrimination Testing Potassium chloride vs. NH 4 Cl discrimination performance declined slightly but significantly (p < .03) following GL transection, while sham transection did not significantly affect performance (Figure 3-4). Interestingly, the decline in performance with surgery was very similar for both GL and sham-transected rats (4% vs. 3%, respectively) and 2-group t-tests failed to indicate significance for either preor postsurgical performance. This suggests that perhaps the small number of animals in this group contributed to significance and that the data might simply have been less variable in the GLX group. In support of this interpretation, the difference in standard deviation for the SHAM rats (.065) is almost twice that of the GLX rats (.035). Regardless of statistical significance, a decline in performance from 92% to 88% is meager at best, particularly when compared to a decline from 90% or better to approximately 50% with combined transection of the CT and GSP nerves (Geran et al., 2002).

PAGE 54

43 The performance of rats in the NaCl vs. NH 4 Cl discrimination group was not affected by GL transection, but was affected by sham-transection (p < .03). Again, this appears to be the result of a small number of subjects combined with low variability (Figure 3-5). Together, these findings suggest that the GL is not necessary for rats to discriminate between NH 4 Cl and either NaCl or KCl. In order to focus on the sufficiency of amiloride-insensitive units within the facial nerve, the rats were tested again after surgery in the presence of amiloride. As expected, this drug had no discernable effect on the postsurgical KCl vs. NH 4 Cl performance of GL-transected rats (Figure 3-7) and the combination of GLX and amiloride did not impair NaCl vs. NH 4 Cl performance beyond that seen with amiloride alone (Figure 3-8). Thus, every rat tested remained able to perform at presurgical levels even after removal of essentially all gustatory input but that of the amiloride-insensitive fibers of the facial nerve. Every rat except 1 failed to perform better than chance on the water control test, suggesting that these animals were under stimulus control. This rat scored only 60% on the water control test but consistently performed at over 95%, even after GLX, on all other phases of the experiment. This suggests that significance on the water control test might have been due more to the number of t-tests performed than to the perception of an extraneous cue. This interpretation is supported by the Bonferroni adjustment. Potential Mechanisms Underlying KCl vs. NH 4 Cl Discrimination Although, the animals in this experiment were clearly able to discriminate between KCl and NH 4 Cl, the basis for this discrimination is not obvious. One possibility is that a population of taste receptor cells innervated by the facial nerve exists that contains ion channels selectively permeable to either K + or NH 4 + Recent whole cell recordings have shown that although the vast majority of TRCs responsive to 0.1 M KCl or 0.1 M NH 4 Cl

PAGE 55

44 are responsive to both stimuli, a few cells respond to only one of these stimuli (Gilbertson et al., 2001). Thus, although unlikely, there is some evidence that such a mechanism might lead to KCl vs. NH 4 Cl discrimination in the rat. It is also possible that a general cation receptor, like that described by DeSimone and colleagues (2001), might exhibit different kinetics for K + and NH 4 + that could result in discrimination. Nishijo and Norgren (1997) reported that a group of neurons in the parabrachial nucleus (PBN) was highly responsive to NH 4 Cl, but not KCl. Most of the neurons in the PBN that responded to NH 4 Cl, however, also responded to KCl. It is not known whether the selectivity of TRC or PBN responses to NH 4 Cl changes with stimulus concentration. Regardless of the basis for KCl vs. NH 4 Cl discrimination, it is clear that this discrimination exists and depends upon input from the gustatory branches of the facial nerve. The implications of this experiment in regard to taste coding are given further consideration in the final chapter. Summary These data support the contention that the facial nerve is more important than the GL nerve for tasks involving taste quality (see St. John & Spector, 1998) in spite of the fact that the GL innervates roughly twice the number of taste buds (Miller, 1995). The possibility remains that vagal, olfactory or somatosensory afferents may provide input that could be useful in KCl or NaCl vs. NH 4 Cl performance, making the input from the amiloride-insensitive units of the facial nerve necessary but not entirely sufficient for one or both of these tasks. At present, there is no evidence of such input affecting the perception of salt taste.

PAGE 56

45 Table 3-1. Training Schedule # of Days Phase Time Out (s) Limited Hold a (s) Stimuli Stimulus Presentation Schedule 6 Shaping I none 180 0.2 M NH 4 Cl, KCl or NaCl constant 3 Shaping II b none 180 Same as above constant 8 Alternation c 10 15 0.2 M NH 4 Cl and either KCl or NaCl alternated after x correct responses 2 Discrimination Training I 10 10 Same as above semi-random d 3 Discrimination Training II 20 10 Same as above semi-random 19 Discrimination Training III 20 5 0.05, 0.1, 0.2, & 0.4 M NH 4 Cl and either KCl or NaCl semi-random a Limited hold refers to the amount of time the rat is given to make a response. b During Shaping II the rat is trained on the opposite stimulus and lever as in Shaping I. c A stimulus is presented repeatedly until a certain number of correct responses are made. This required number of responses, known as the alternation criterion, decreases with each session. Eight for the first session, six for the second, four for the third, three for the fourth, and two for the fifth. It is not necessary that the correct responses be consecutive. d Stimuli were presented in randomized blocks.

PAGE 57

46 Table 3-2. Experiment Schedule. Phase # of Sessions (or Days) Training 41 Presurgical Discrimination Testing 5 Presurgical Amiloride Testing 5 Surgery 2 Recovery 9-10 Postsurgical Discrimination Testing 5 Postsurgical Amiloride Testing 5 Water Control Test 1

PAGE 58

47 RATS 123456789101112131415M OVERALL PROPORTION CORRECT 0.50.60.70.80.91.0 Figure 3-1. Presurgical KCl vs. NH 4 Cl discrimination with and without amiloride. Overall proportion correct for each individual with (white) and without (gray) 100 M amiloride and followed by the mean (M). Note that 50% correct performance (chance) is used as the origin of the y-axis. Amiloride did not significantly impair discrimination performance.

PAGE 59

48 RATS 123456789101112131415M OVERALL PROPORTION CORRECT 0.50.60.70.80.91.0 Figure 3-2. Presurgical NaCl vs. NH 4 Cl discrimination with and without amiloride. Overall proportion correct for each individual with (white) and without (gray) 100 M amiloride and followed by the mean (M). Note that 50% correct performance (chance) is used as the origin of the y-axis. Amiloride significantly impaired mean performance (p < .001).

PAGE 60

49 NaCl (M) 0.40.20.10.05 PERFORMANCE 0.60.81.0 NH4Cl (M) 0.40.20.10.05 NaCl vs. NH4Cl Group*******KCl vs. NH4Cl GroupKCl (M) 0.40.20.10.05 PERFORMANCE 0.60.81.0 NO AMILORIDE AMILORIDE NH4Cl (M) 0.40.20.10.05 Figure 3-3. Mean presurgical performance by concentration. Performance is separated according to task (KCl vs. NH 4 Cl: top, NaCl vs. NH 4 Cl: bottom) and salt (NaCl or KCl: left, NH 4 Cl: right). Asterisks indicate significant differences between discrimination testing in the presence (white) and absence (black) of 100 M amiloride (p < .004 for each). Note that 50% performance (chance) is used as the origin of each y-axis.

PAGE 61

50 SHAM RATS 12345678M OVERALL PROPORTION CORRECT 0.50.60.70.80.91.0 GLX RATS 1234567M 0.50.60.70.80.91.0 Figure 3-4. Prevs. postsurgical performance on the KCl vs. NH 4 Cl task. Rats that underwent bilateral glossopharyngeal nerve transection (GLX) are shown on the right and sham-transected (SHAM) rats on the left. Each graph contains presurgical (gray) and postsurgical (white) performance for each rat followed by the mean (M) for each surgical group. Discrimination performance dropped slightly with transection (p < .03), but performance of GLX rats was not significantly different from sham-operated controls either before or after surgery (p > .88 for both).

PAGE 62

51 SHAM RATS 12345678M OVERALL PROPORTION CORRECT 0.50.60.70.80.91.0 GLX RATS 1234567M 0.50.60.70.80.91.0 Figure 3-5. Prevs. postsurgical performance on the NaCl vs. NH 4 Cl task. Rats that underwent bilateral glossopharyngeal nerve transection (GLX) are shown on the right and sham-transected (SHAM) rats on the left. Each graph contains presurgical (gray) and postsurgical (white) performance for each rat followed by the mean (M) for each surgical group. Discrimination performance dropped slightly following sham transection only (p < .03).

PAGE 63

52 NaCl (M) 0.40.20.10.05 PERFORMANCE 0.60.81.0 NH4Cl (M) 0.40.20.10.05 NaCl vs. NH4Cl GroupKCl (M) 0.40.20.10.05 PERFORMANCE 0.60.81.0 POST-GLX PRE-GLX NH4Cl (M) 0.40.20.10.05 KCl vs. NH4Cl Group* Figure 3-6. Mean presurgical vs. postsurgical performance by concentration. All data points are from rats in the GL transection group. Performance is separated according to task (KCl vs. NH 4 Cl: top, NaCl vs. NH 4 Cl: bottom) and salt (NaCl or KCl: left, NH 4 Cl: right). Asterisks indicate significant differences between presurgical (white) and postsurgical (black) discrimination (p < .04). Note that 50% performance (chance) is used as the origin of each y-axis.

PAGE 64

53 GLX RATS 1234567M OVERALL PROPORTION CORRECT 0.50.60.70.80.91.0 Figure 3-7. Effect of amiloride on postsurgical KCl vs. NH 4 Cl discrimination. All data are from GL-transected (GLX) rats. Amiloride (white) did not further impair discrimination performance following surgery (gray).

PAGE 65

54 GLX RATS 1234567M OVERALL PROPORTION CORRECT 0.50.60.70.80.91.0 Figure 3-8. Effect of glossopharyngeal transection on NaCl vs. NH 4 Cl performance in the presence of amiloride. All data are from rats in the GL transection (GLX) group. Transection (white) did not further impair discrimination performance in the presence of amiloride. Presurgical amiloride performance is shown in gray.

PAGE 66

55 NaCl (M) 0.40.20.10.05 PERFORMANCE 0.60.81.0 POST-GLX AMIL PRE-GLX AMIL NH4Cl (M) 0.40.20.10.05 NaCl vs. NH4Cl GroupKCl (M) 0.40.20.10.05 PERFORMANCE 0.60.81.0 POST-GLX POST-GLX + AMIL NH4Cl (M) 0.40.20.10.05 KCl vs. NH4Cl Group Figure 3-9. Mean comparisons of postsurgical amiloride performance by concentration. All data are from rats in the GL transection group. Performance is separated according to task (KCl vs. NH 4 Cl: top, NaCl vs. NH 4 Cl: bottom) and salt (NaCl or KCl: left, NH 4 Cl: right). In the top graphs, there is no difference between postsurgical performance (black) and postsurgical performance in the presence of amiloride (white). In the bottom graphs, there is no difference between presurgical (white) and postsurgical (black) performance in the presence of amiloride. Note that 50% performance (chance) is used as the origin of each y-axis.

PAGE 67

56 RATS 123456789101112131415M OVERALL PROPORTION CORRECT 0.00.20.40.60.81.0 Figure 3-10. Water control test. After testing was complete, all fluid reservoirs were filled with water and assigned to either the left or right lever to assess each rats ability to respond to extraneous cues. One rat performed significantly better than chance using a one-tailed test (p < .03). This significance disappeared when corrected for the number of t-tests performed (p > .34). A dashed line represents 50% (chance) performance.

PAGE 68

CHAPTER 4 AMILORIDE-INSENSITIVE UNITS OF THE CHORDA TYMPANI NERVE ARE NECESSARY FOR NORMAL AMMONIUM CHLORIDE DETECTABILITY IN THE RAT Background The 3 main gustatory nerves of the rat are highly responsive to NH 4 Cl at mid to high concentrations (Frank 1991; Frank et al., 1983; Kitada et al, 1998; Nejad, 1986; Sollars & Hill, 1998). For this reason, NH 4 Cl is often used as the standard when recording from taste afferents, but very little is known about the perceptual characteristics of this salt. Human subjects have reported that NH 4 Cl contains both salty and bitter components (van der Klaauw & Smith, 1995). Because direct quality scaling and magnitude estimation procedures cannot be used when working with animal subjects, analyses of the perceived taste quality of NH 4 Cl in rodents have used conditioned shock avoidance and conditioned taste aversion to measure generalization to salts and other stimuli. All 3 of the published studies using rats have concluded that a conditioned avoidance response or aversion to NH 4 Cl generalizes strongly to KCl, and weakly to NaCl (Erickson, 1963; Hill et al., 1990; Morrison, 1967). These results have been interpreted as evidence that NH 4 Cl shares a common taste quality with KCl, but not NaCl. This supports the theory that NH 4 + and K + activate at least one common transduction mechanism in the oral cavity (DeSimone et al., 2001; Kloub et al., 1997), while a separate transduction pathway exists for Na + in the rat (Brand et al., 1985; DeSimone & Ferrell, 1985; Ninomiya & Funakoshi, 1988). Additional transduction sites might also exist that are more specific for either 57

PAGE 69

58 NH 4 + or K + as rats have been shown to discriminate easily between KCl and NH 4 Cl in spite of evident similarities in taste quality (Geran et al., 2002). There is some controversy as to whether the epithelial sodium channel (ENaC) blocker amiloride inhibits only sodium and lithium responses in the CT and GSP nerves, or whether its action is more general, reducing neural responding to potassium and ammonium salts as well (see Lundy & Contreras, 1999; Minear et al., 1996). Researchers have demonstrated maximal whole-nerve CT inhibition of up to 48% at .05 M NH 4 Cl with the addition of 100 M amiloride (Kloub et al., 1997. See also Lundy & Contreras, 1997; Lundy et al., 1997). Inhibition was less pronounced at higher NH 4 Cl concentrations, losing significance at approximately .3 M (Kloub et al., 1997). The magnitude of this suppression (48%) is greater than that observed for the same concentration of NaCl (30% at .05 M) with 100 M amiloride, but much less than the 70-80% maximal inhibition observed at higher NaCl concentrations (Brand et al., 1985; DeSimone & Ferrell, 1985). Other researchers, meanwhile, failed to find any appreciable effect of amiloride on CT responding to either KCl or NH 4 Cl (Brand et al., 1985; Formaker & Hill, 1988; Hill & Bour, 1985; Hill et al., 1982). These results are supported by the observation that oral amiloride treatment did not inhibit either the potassium or ammonium response of CT and GSP somata in the geniculate ganglion (Lundy & Contreras, 1999) or activity in the subsets of NTS neurons most responsive to NH 4 Cl (Giza & Scott, 1991). At the behavioral level, amiloride pretreatment did not affect generalization following acquisition of a conditioned taste aversion to NH 4 Cl (Hill et al., 1990). Also, amiloride failed to compromise either KCl detection threshold or NH 4 Cl vs. KCl discrimination in

PAGE 70

59 rats (Geran et al., 1999; Geran et al., 2002). Together, these findings suggest that although the effect of amiloride on nonsodium salt responsiveness is perhaps at times statistically significant at the level of the peripheral nervous system, it does not appear to reach significance at higher levels of the gustatory system or lead to behavioral significance. Taste sensitivity to NH 4 Cl was tested in the presence and absence of amiloride to determine whether any inhibition in the periphery might affect the taste-guided behavioral response of the animal. Amiloride suppression of the CT response to NH 4 Cl was greatest at low concentrations (Kloub et al., 1997) suggesting that impairment would be most noticeable at the limits of detectability. Detection threshold was also chosen for measurement because amiloride was previously shown to significantly raise the thresholds of both NaCl and Na-gluconate using a similar procedure (Geran & Spector 2000a; 2000b). Earlier detection threshold experiments have also reported that NaCl and KCl sensitivities were compromised by bilateral transection of the CT (CTX) suggesting that input from this nerve is necessary for normal salt detectability (Geran et al, 1999; Kopka & Spector, 2001; Slotnick et al., 1991; Spector, Schwartz & Grill, 1990). To further test this hypothesis NH 4 Cl detection threshold was also measured after severing the CT nerves. Methods General Methods All procedures were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Florida. Ten adult male Sprague-Dawley rats were placed on a water-restriction schedule with water bottles removed about 24h prior to the start of training or testing on Monday and replaced after the last session Friday. During the

PAGE 71

60 week, animals worked for water access in the gustometer apparatus (see Spector, Andrews-Labinski & Letterio, 1990). Subjects were cared for exactly like subjects in Chapter 3, except where noted. The animals were trained to press one lever immediately after a presentation of 0.2 M NH 4 Cl and the other lever immediately after a distilled water presentation. As subjects grew more proficient, more concentrations were added to the stimulus array and the parameters of the trial structure were altered such that the decision phase was reduced and the time out increased. Subjects moved to the next phase of alternation or training when they reached at least a 70% overall performance criterion. See Table 4-1 for the training schedule. Each session lasted 40 min with the rats allowed to complete as many trials as possible during this time. In the last phase of training, stimuli were presented in randomized blocks of 10 consisting of 5 NH 4 Cl concentrations and 5 distilled water presentations. This final training phase was exactly like the first week of testing. As performance did not change over the last week of the final training phase (Discrimination Training III), this period was retroactively defined as the first week of presurgical threshold testing. Testing Detection threshold was tested over the course of 4 weeks using a total of 8 NH 4 Cl concentrations ranging from .00325 to .4 M. Each Monday rats received the standard stimulus array (i.e., 0.4, .2, .1, .05 & .025 M NH 4 Cl). A different array was presented Tuesday through Friday containing at least 2 concentrations from the standard array and several lower concentrations. At least one stimulus in the test array was replaced with a lower concentration each week until performance approached a minimum asymptote. Rats were given a second week of testing with the lowest concentration array to increase the number of trials for these stimuli. See Table 4-2 for test stimulus presentation

PAGE 72

61 schedule. During presurgical and postsurgical amiloride testing, 100 M amiloride was used as the solvent for all NH 4 Cl stimuli and in place of water for both stimulus presentations and reinforcement. A water control test was performed following the final postsurgical test to assess whether the rats were capable of responding to extraneous cues unrelated to the chemosensory properties of the NH 4 Cl stimuli. This was accomplished by filling each reservoir with distilled water and assigning half of these to the left lever and half to the right. See Table 4-3 for phases of the experiment. Surgery Rats were divided into 2 groups counterbalanced for weight, overall performance and number of trials initiated. All rats were anesthetized with a mixture of ketamine (125 mg/kg body weight) and xylazine hydrochloride (5 mg/kg) injected intramuscularly. Five of these rats received bilateral CT nerve cauterization. This was accomplished by retracting the external ear canal to expose the tympanic membrane. The membrane, along with the rim of the ear canal, the malleus and the CT nerve were then cauterized. This procedure stimulates the production of cerumen, which keeps the CT from reinnervating taste buds in the anterior tongue for at least 118 days (Kopka & Spector, 2001). The 5 rats in the sham-transected group had each ear retracted and the tympanic membrane punctured with microforceps. All animals received subcuteneous injections of penicillin (30,000 units Flocillin) and an analgesic (Ketorolac, 2 mg/kg body weight) immediately following surgery and for the next 3 days. Rats were given 6-7 days to recover from surgery before testing resumed. Histology After postsurgical testing, rats were deeply anesthetized with sodium pentobarbital (i.p.) and rapidly perfused with saline and 10% buffered formalin. The tongue and palate

PAGE 73

62 were removed and stored in formalin. Staining was accomplished by placing the anterior portion of the tongue from the tip to the intermolar eminence in distilled water for 30 min then dipping it in 0.5% methylene blue until dark and rinsing the tissue with water to remove excess stain. The epithelium of the anterior tongue was then pressed between 2 glass slides and examined under a light microscope. The number of intact fungiform papillae and taste pores were counted on each tongue. Taste pores appeared as small round dots surrounded by blue circles under the microscope (Parks & Whitehead, 1998; St. John et al., 1995). The presence of a discernable blue dot was counted as a pore for the purposes of this experiment. Results using this method correspond with those from a hematoxylin and eosin stain and also correlate with the degeneration and regeneration of taste buds concomitant with denervation and reinnervation by the CT nerve (St John et al., 1995). Histology was performed blind to the rats surgical treatment. Data Analysis The percentage of correct responses on NH 4 Cl trials was adjusted for false alarm probability (see Gescheider, 1997). This was accomplished using the following equation for corrected hit rate, or P(Hit) c : )(1)()(FAPFAPHitPHitPc where P(Hit) was the proportion of NH 4 Cl trials on which the rat pressed the correct lever and P(FA) was the proportion of water trials on which the animal pressed the wrong lever. Hit rates were corrected for each rat at each NH 4 Cl concentration. The following logistic function was then used to fit curves to the corrected hit rate values for each animal:

PAGE 74

63 -b(x-c)axf101 where a = maximum asymptote of performance, b = slope, x = NH 4 Cl log 10 concentration and c = threshold. Threshold was defined as the NH 4 Cl concentration at one-half the maximum asymptote of performance. Analyses of variance (ANOVAs), paired and independent t-tests and the normal approximation of the binomial distribution (Brown & Hollander, 1977) were used to assess statistical significance. Alpha was set at the conventional .05 level. Results Presurgical Detection Threshold Mean detection threshold for the first 20-session threshold was .012 M NH 4 Cl + .001. With the addition of 100 M amiloride, mean threshold decreased significantly (p < .002, Figure 4-1). In other words, the rats performed better at near threshold NH 4 Cl concentrations with amiloride than prior to treatment. When NH 4 Cl threshold was again measured without amiloride, sensitivity improved further still (p < .006, mean threshold = .009 M + .001. Figure 4-2). Two-way ANOVAs (phase x concentration) of corrected hit rates indicated main effects for both phase: (F(1,9) > 17, p < .003) and concentration: (F(7,63) > 468, p < .001) as well as for their interaction (F(7, 63) > 4, p < .002) when each of the 2 presurgical tests without amiloride was compared to the amiloride test. From the first presurgical measurement to the last, mean threshold decreased by .25 log 10 units + .03 (p < .001; see Figure 4-3 for individual shifts in threshold).

PAGE 75

64 Postsurgical Detection Threshold Mean NH 4 Cl threshold increased following CT transection by .54 log 10 units + .09 to .04 M (p < .004. Figure 4-4). This value differs significantly from that of the sham-operated rats (p < .001, see Figure 4-5). Threshold did not change significantly with surgery for sham-operated rats compared with the last presurgical threshold assessment (p > .08). Nor did amiloride alter NH 4 Cl threshold in either the CT-transected or sham-transected rats after surgery (p > .29 for both. Figure 4-6). Water Control Test and Histology No rat scored better than chance when all stimuli were replaced with distilled water (one-tailed t-tests using the normal approximation of the binomial distribution, p > .17 for each. Figure 4-7). Histological analysis indicated no significant difference in the number of fungiform papillae for the 2 surgical groups (CTX: 106 + 34 SHAM: 139 + 23. p > .11). Rats in the CT-transected group, however, had far fewer taste pores on the anterior tongue than rats in the sham-transected group (6 + 8 and 132 + 23 respectively, p < .001). Discussion It is clear from these data that the rats were responding to the chemical properties of the stimuli rather than to extraneous cues such as temperature or sound, as none of the rats scored better than chance on the water control test (Figure 4-7). The lowest mean detection threshold for NH 4 Cl (.009 M) recorded in this experiment was slightly higher than that found previously for NaCl (.005 to.006 M) using the same 2-lever operant procedure (Geran & Spector, 2000a; 2000b; Kopka & Spector, 2001). In contrast, the detection threshold for KCl (.04 M) was considerably higher than for either of the other chloride salts tested to date (Geran et al., 1999). The detectability functions for these 3 salts will be compared in greater detail in the final chapter.

PAGE 76

65 Ammonium Chloride Detectability Depends Upon an Amiloride-Insensitive Route of Transduction A subset of electrophysiological studies has reported that amiloride significantly suppressed CT responding to low and mid-range concentrations of NH 4 Cl (Kloub et al., 1997; Lundy & Contreras., 1997; Lundy et al., 1997), suggesting that amiloride might also compromise NH 4 Cl detection threshold. Instead, threshold decreased in the presence of the ENaC blocker. Amiloride did not seem to be the cause of the increase in NH 4 Cl sensitivity, however, as threshold decreased further still when the rats were tested a second time in the absence of amiloride (Figure 4-2). In addition, amiloride failed to decrease postsurgical threshold for either surgery group. It is unclear whether this decrease in threshold with repeated testing represents an aspect unique to NH 4 Cl such as an upregulation of ammonium-sensitive receptor elements. Alternatively, the decrease in threshold could be due to increased performance resulting from previous experience with low concentrations of the stimulus. Gilbertson and colleagues (1993), reported that vasopressin, a hormone involved in fluid homeostasis, increases the inward Na + current through amiloride-sensitive channels in the taste receptor cells of hamsters. Although the amiloride-sensitive transduction pathway of the hamster is different in significant ways from that of rats, it is conceivable that such a mechanism could potentially have been activated by the restricted fluid access experienced by the rats in the current experiment. One would expect such a mechanism to have a greater effect on NaCl threshold than NH 4 Cl threshold, however, as this task is more dependent on amiloride-sensitive units. Sodium chloride threshold increases significantly in the presence of amiloride, and returns to the pre-amiloride threshold when tested following amiloride treatment (Geran & Spector, 2000a; 2000b). This suggests that

PAGE 77

66 the decrease in NH 4 Cl threshold observed during and after amiloride treatment was not due to an upregulation of amiloride-sensitive receptors, although amiloride-insensitive receptors permeable to NH 4 + could have been affected. There is currently no evidence to support such a hypothesis, however. In fact, NH 4 Cl vs. KCl discrimination performance was unchanged with amiloride treatment (Geran et al., 2002). Thus, the decrease in NH 4 Cl threshold with each presurgical test is most likely the result of continued experience with the task. In addition to having little to no effect on the perceived intensity of NH 4 Cl at low concentrations, amiloride also seems to be without effect on the taste quality of this salt. For instance, amiloride treatment failed to compromise performance on a KCl vs. NH 4 Cl task (Geran et al., 2002). In addition, taste aversions conditioned to NH 4 Cl resulted in the same pattern of generalization regardless of whether the animals were treated with amiloride at the time of conditioning (Hill et al., 1990). Thus, amiloride does not seem to appreciably affect the perceived intensity of NH 4 Cl at low concentrations, or its perceived taste quality at mid-range to high concentrations (i.e., .05 to .4 M. Geran et al., 2002; Hill et al., 1990). It is unlikely that the drug compromises NH 4 Cl intensity at higher concentrations as the electrophysiology indicates very little suppression, if any, at concentrations above .3 M (Kloub et al., 1997; Lundy et al., 1997). Therefore, it is more likely that amiloride affects the taste quality of NH 4 Cl at low concentrations, if it has any effect at all on taste-guided performance to NH 4 Cl. The current procedure only measures detection, not recognition (see Gescheider, 1997). Thus, although the rats in this study were able to discriminate NH 4 Cl dissolved in amiloride from amiloride alone, they may

PAGE 78

67 have nevertheless been unable to recognize the stimulus as NH 4 Cl at near threshold concentrations. The Chorda Tympani Nerve is Necessary for Normal Ammonium Chloride Detection Bilateral transection of the CT significantly increased NH 4 Cl detection threshold (Figure 4-4), suggesting that the information carried by this nerve is necessary for normal detection of this salt. It is also possible that the GSP and/or GL are necessary, although insufficient, for normal NH 4 Cl detection. The result of this experiment, along with the increases previously reported for both KCl and NaCl threshold with CT transection (Geran et al., 1999; Kopka & Spector, 2001; Slotnick et al., 1991; Spector, Schwartz & Grill, 1990), support the hypothesis that the CT nerve is important for normal salt detection. This nerve also appears to be important for the recognition of salt stimuli as CTX impairs performance on a NaCl vs. KCl discrimination task (Kopka et al., 2000; St John, Markison, Guagliardo et al., 1997; Spector & Grill, 1992). It is unclear whether the role of the CT in salt recognition extends to NH 4 Cl, as CTX does not alter unconditioned licking to this salt in a 2-bottle preference test (Sollars & Bernstein, 1996). Of course, factors other than taste quality, such as hedonic value or the postingestive consequences of NH 4 Cl consumption, could have influenced performance on the 2-bottle test (see Spector, 2000). Like the CT, the GSP also appears to carry behaviorally-relevant information about salt stimuli (Kopka et al., 2000; Roitman & Bernstein, 1999; St. John, Markison & Spector, 1997). For instance, amiloride abolishes NaCl vs. KCl discrimination while CTX merely impairs it, suggesting that residual discrimination after CTX relies upon amiloride-sensitive, sodium-selective receptors innervated by the GSP (see Kopka et al.,

PAGE 79

68 2000; Roitman & Bernstein, 1999; St. John, Markison & Spector, 1997). The GL is not thought to contain AS taste receptor cells (Doolin & Gilbertson, 1993; Kitada et al., 1998). Thus, both the CT and GSP appear to carry information important for NaCl recognition, and could perhaps also be necessary for normal NH 4 Cl recognition. It is not clear at present whether the CT, GSP, or the combined input of the two is required for the rat to accurately perceive the taste quality normally associated with NH 4 Cl. If the GSP is required for this task, one would expect CTX to impair NH 4 Cl vs. KCl discrimination performance significantly less than combined transection of the CT and GSP. In contrast, input from the GL does not appear to be necessary for the perception of salt taste quality (Markison et al., 1995; Spector & Grill, 1992; Chapter 3). Instead, the GL appears to be more important for the stimulation of unconditioned aversive gustatory reflexes, like gaping upon contact with quinine, than for the perception of taste quality (St. John & Spector, 1998; Travers et al., 1987). For instance, combined transection of the CT and GSP nerves drops performance on a KCl vs. NH 4 Cl discrimination task to chance while GL transection is without effect (Geran et al., 2002; Chapter 3). Conclusions Like KCl detection, normal NH 4 Cl detectability appears to depend upon amiloride-insensitive receptors innervated by the CT nerve. It is unclear whether the same population of receptors is responsible for both NH 4 Cl and KCl detection. The fact that KCl and NH 4 Cl are easily discriminated by the rat (Geran et al., 2002) suggests that differences in activation of the NTS exist for the 2 salts but have not yet been found. The increase in detection threshold for NH 4 Cl with CT transection lends further support to the hypothesis that this nerve is necessary for the normal detection of salt stimuli. This manipulation has also impaired both NaCl and KCl detectability in previous

PAGE 80

69 tests (Geran et al., 1999; Kopka & Spector, 2001; Slotnick et al., 1991; Spector, Schwartz & Grill, 1990). Finally, the current experiment was also useful in that it provides a detection threshold for NH 4 Cl that can be used to determine testing concentrations for future psychophysical studies.

PAGE 81

70 Table 4-1. Training Schedule # of Days Phase Time Out (s) Limited Hold a (s) Stimuli Stimulus Presentation Schedule 9 Shaping I none 180 0.2 M NH 4 Cl or DH 2 O constant 4 Shaping II b none 180 0.2 M NH 4 Cl or DH 2 O constant 7-14 Alternation c 10 15 0.2 M NH 4 Cl and DH 2 O alternated after x correct responses 2 Discrimination Training I 10 10 0.2 M NH 4 Cl and DH 2 O semi-random d 3 Discrimination Training II 20 10 0.2 M NH 4 Cl and DH 2 O semi-random 8-15 Discrimination Training III e 20 5 0.025, 0.05, 0.1, 0.2, 0.4 M NH 4 Cl and DH 2 O semi-random a Limited hold refers to the amount of time the rat is given to make a response. b During Shaping II the rat is trained on the opposite stimulus and lever as in Shaping I. c A stimulus is presented repeatedly until a certain number of correct responses are made. This required number of responses, known as the alternation criterion, decreases with each session. Eight for the first session, six for the second, four for the third, three for the fourth, and two for the fifth. It is not necessary that the correct responses be consecutive. d Stimuli were presented in randomized blocks. e The number of days shown here includes the first week of presurgical threshold testing.

PAGE 82

71 Table 4-2. Test Stimulus Presentation Schedule NH 4 Cl Concentrations (M) in Test Array .4 .2 .1 .05 .025 .013 .0065 .00325 Week 1 & Mondays: Standard Array Week 2 (Tue to Fri) Week 3 (Tue to Fri) Week 4 (Tue to Fri) Table 4-3. Experiment Schedule Phase # of Sessions (or Days) Training 35 Presurgical Threshold Testing 1 20 Presurgical Amiloride Testing 20 Presurgical Threshold Testing 2 20 Surgery 2 Recovery 6-7 Postsurgical Threshold Testing 20 Postsurgical Amiloride Testing 20 Water Control Test 1

PAGE 83

72 NH4Cl CONCENTRATION (M) 0.010.1 CORRECTED HIT RATE 020406080100 Figure 4-1. Effect of amiloride on NH 4 Cl detection. Mean performance with and without amiloride (100 M). Threshold decreased significantly (i.e. sensitivity increased) with amiloride (p < .002). Amiloride performance is represented with open symbols.

PAGE 84

73 NH4Cl CONCENTRATION (M) 0.010.1 CORRECTED HIT RATE 020406080100 Figure 4-2. NH 4 Cl threshold decreased again following amiloride treatment. Threshold decreased significantly following amiloride testing (p < .006). Open symbols represent amiloride performance.

PAGE 85

74 RAT NUMBER 14569M DIFFERENCE IN THRESHOLD (PRESURGERY 2 PRESURGERY 1) -0.4-0.20.00.20.40.60.81.0 237810M CTXSHAM Figure 4-3. Individual shifts in presurgical threshold. Shifts for each animal followed by the mean for each surgical group. All values are before surgery. Each animal showed a decrease in threshold between the first and second measurement of NH 4 Cl detectability in the absence of amiloride. Animals on the left underwent transection of the chorda tympani nerve (CT) after these shifts in threshold were measured while animals on the right received sham surgery.

PAGE 86

75 RAT NUMBER 14569M DIFFERENCE IN THRESHOLD (POST-SURGERY PRE-SURGERY 2) -0.4-0.20.00.20.40.60.81.0 237810M CTXSHAM Figure 4-4. Individual shifts in performance with surgery followed by the mean for each group. Animals that received bilateral chorda tympani transection (left) showed a significant increase in threshold (p < .004), while sham-operated rats (right) did not.

PAGE 87

76 NH4Cl CONCENTRATION (M) 0.010.1 CORRECTED HIT RATE 020406080100 PRE-CTX POST-CTX 0.010.1 PRE-SHAM POST-SHAM CTXSHAM Figure 4-5. NH 4 Cl detectability functions preand post-surgery. Mean performance for each surgical group. Rats in the chorda tympani transection group (left) had significantly higher thresholds (p < .004) following surgery while rats in the sham-operated group (right) did not. Presurgical results are from the second determination of threshold in the absence of amiloride.

PAGE 88

77 NH4Cl CONCENTRATION (M) 0.010.1 CORRECTED HIT RATE 020406080100 POST-CTX CTX + AMIL 0.010.1 POST-SHAM SHAM + AMIL CTXSHAM Figure 4-6. Postsurgical NH 4 Cl detectability functions with and without amiloride (100 M). Neither rats in the chorda tympani transected group (left) nor rats in the sham-operated group (right) showed any significant change in NH 4 Cl threshold with amiloride.

PAGE 89

78 RAT NUMBER 14569237810 OVERALL PERFORMANCE 0.00.20.40.60.81.0 CTXSHAM Figure 4-7. Water control test. Individual overall performance when all stimuli were replaced with distilled water. Neither rats in the chorda tympani transected group (left), nor rats in the sham-operated group (right) performed significantly better than chance.

PAGE 90

CHAPTER 5 GENERAL DISCUSSION Discrepancies Between the Electrophysiology and the Behavior In Regard to NH 4 Cl Regardless of whether NH 4 Cl vs. KCl discrimination is due to different receptor subtypes or different temporal properties within the same receptor subtype, significant differences in activity at the level of the CNS should be noticeable. At present, no such differences have been observed, although this might be due to the paucity of experiments that include both NH 4 Cl and KCl as stimuli. The overall patterns of activity are very similar for NH 4 Cl and KCl in the NTS (Erickson, 1963; Nakamura & Norgren, 1993). This similarity in neural responding in conjunction with similar generalization patterns for the two salts in conditioned aversion and avoidance tasks has historically been used to support both the across-fiber pattern theory and the best-stimulus method of classification (see Erickson, 1963; Smith et al., 2000). Although these salts have been shown to taste similar to the rat (Erickson, 1963; Hill et al., 1990; Morrison, 1967), they are also easily discriminated suggesting that coding theorists must also provide an explanation for this phenomenon. In addition to the small number of CNS experiments using both KCl and NH 4 Cl, it is also possible that differences between the NTS electrophysiology and the behavior exist because the gustatory system was altered over the course of behavioral testing. Perhaps prolonged experience with NH 4 Cl amplified signals from afferents that synapsed with cells containing ammonium-sensitive receptors. Such a mechanism would not be observed in electrophysiological tests in which the rats first experience with NH 4 Cl was 79

PAGE 91

80 on the test day. Furthermore, the subjects in these preparations are usually deeply anesthetized, allowing for little influence from higher brain regions on gustatory activity, unlike in an awake animal. Along these lines, stimulation of the central amygdala, thought to be important for attention and memory, has been shown to affect responses in the NTS and PBN (Li et al., 2002; Lundy & Norgren, 2001). Another difference between the electrophysiological literature and the majority of the behavioral literature involves the depletion state at the time of testing. In order to motivate animals to perform behavioral tasks, researchers must often restrict food or water access, or deplete subjects of some other commodity like sodium. These methods are not necessary for electrophysiological recordings. The few electrophysiological experiments that have been reported using depleted animals have suggested that such a state can affect neural responses in some areas of the gustatory system. For example, sodium depletion, or hormones associated with depletion, affect both NTS and CT responses to NaCl (Contreras & Frank, 1979; Herness, 1992; Lundy, 1998; McCaughey & Scott, 2000; Tamura & Norgren, 1997). Likewise, vasopressin, a hormone important for regulating fluid homeostasis, can also significantly increase inward Na + and proton currents in hamster TRCs (see Gilbertson et al., 1993). A paper by Lundy & Contreras (1999) is one of the few studies to claim that KCl and NH 4 Cl could potentially produce discriminable perceptions. The central thesis of this paper is that NaCl, KCl and NH 4 Cl could perhaps be discriminated by the rat due to across fiber patterns shaped predominantly by the activity of 2 classes of neurons in the geniculate ganglion. Sodium chloride (NaCl)-specialists appear to synapse with taste cells containing amiloride-sensitive, apically-located ENaCs that are highly selective for

PAGE 92

81 Na + but also permit passage of small amounts of K + and to a lesser extent NH 4 + Hydrochloric acid (HCl)-generalists, on the other hand, respond well to all 3 salts, but show a more robust response to NH 4 Cl than to the others. The remaining types of ganglion cells respond equally well to all 3 salts. This hypothesis is plausible, but given that the difference in NaCl-specialist activation with KCl and NH 4 Cl stimulation is very slight, does not sufficiently address how high concentrations of KCl would be distinguished either from low concentrations of NH 4 Cl or from other nonsodium chloride salts and acids which produce moderate activation of HCl-generalist cells. Additionally, the finding that NH 4 Cl and KCl are discriminable even in the presence of amiloride (Geran et al., 2002), when responses from NaCl-specialists should be suppressed, suggests that we must look elsewhere for the underlying cause of this discrimination. As for the discrepancy among electrophysiological tests as to whether amiloride impairs nonsodium salt responsiveness of the CT nerve, this could be due to important differences in the methods employed. For instance, most of the experiments reporting inhibition with amiloride deposited a salt solution on the tongue immediately followed by either a mixture of salt and amiloride (Lundy & Contreras, 1997; Lundy et al., 1997), or amiloride alone (Minear et al., 1996). Thus, the nerve was adapted to the salt prior to amiloride treatment. When the reverse was attempted, Minear and colleagues (1996) noted that adapting the tongue to amiloride prior to salt treatment had no effect on KCl responding. This has also been shown to hold true for NH 4 Cl responsiveness (Formaker & Hill, 1988; Hill & Bour, 1985; Hill et al., 1982). Also without effect were experiments in which the tongue was adapted to amiloride prior to the presentation of salt dissolved in amiloride (Brand et al., 1985). This method more closely resembles that of the current

PAGE 93

82 detection threshold procedure. Rats ingest amiloride when they are reinforced for pressing the correct lever. This serves to adapt the tongue to amiloride prior to a stimulus trial consisting of salt dissolved in amiloride or amiloride alone. Differences among electrophysiological procedures also exist for duration of stimulus delivery, stimulus concentration, time constant, and whether the integrated response or the number of action potentials is used as the dependent variable, making comparisons difficult. Implications for Chloride Salt Detectability When the parameters from salt detectability functions measured to date were compared, we found that KCl was more similar to NaCl dissolved in amiloride than to NaCl dissolved in distilled water both in slope and threshold (Figure 5-1). Although, these functions are based on data from separate groups of subjects, this finding is interesting because NaCl + amiloride and KCl also share a similar taste quality (Hill et al., 1990; Kopka et al., 2000; Spector et al., 1996) and route of transduction at higher concentrations (Ye et al., 1994). Taken together, these results suggest that NaCl + amiloride and KCl are indiscriminable to the rat at a variety of concentrations due to activation of a common pathway. One caveat to this interpretation is that although the slope and threshold for KCl and NaCl + amiloride are similar, the asymptotic performance is significantly different. This is most likely due to the fact that rats in the amiloride experiment were trained on NaCl dissolved in distilled water, and therefore might have performed less well at the higher concentrations due to a change in perceived intensity or quality with amiloride at suprathreshold concentrations. Rats in the KCl experiment, however, were trained on KCl and therefore, presumably did not experience any changes in intensity or quality with testing. It would be interesting to test this

PAGE 94

83 hypothesis by measuring NaCl + amiloride detectability in animals trained on a KCl vs. water task. Given the results of the NaCl + amiloride experiment, it was hypothesized that the detectability function for NH 4 Cl might also fit this pattern and support data from electrophysiological and taste quality experiments. Ammonium chloride shares a similar taste quality (Erickson, 1963; Hill et al., 1990; Morrison, 1967) and perhaps at least one transduction mechanism with KCl and NaCl + amiloride (DeSimone et al., 2001), but produces a detectability function that is different from that of KCl and instead very similar to that of NaCl (Figure 5-2), a stimulus that is easily discriminated from nonsodium salts at higher concentrations (Geran et al., 2002; Hill et al., 1990; Spector & Grill, 1992). In addition to evidence that NH 4 Cl and KCl activate the same transduction pathway, or pathways (DeSimone et al., 2001), there is also evidence that NH 4 Cl might utilize a separate AS pathway at lower concentrations (Kloub et al., 1997). If NH 4 Cl transduction does depend on an AS route, this pathway cannot be blocked sufficiently by micromolar doses of amiloride to impair detectability. Larger doses of amiloride have not been tested as they might lead to nonspecific effects (see DeSimone & Ferrell, 1985). Sodium chloride detectability appears to depend heavily on CT input. Precisely how heavily depends on the procedure, with values ranging from 1-2 log 10 units for the shift following transection (Kopka & Spector, 2001; Slotnick et al., 1991; Spector, Schwartz & Grill, 1990). Kopka & Spector (2001) reported that the shift in threshold with CTX was not further impaired with amiloride treatment, suggesting that AS taste receptor cells innervated by the GSP are not important for normal NaCl detectability in the rat. Performance to concentrations greater than the mean postsurgical threshold of 0.1 M

PAGE 95

84 NaCl are presumably due to combined activation of AI cells in the GL and GSP (Kopka & Spector, 2001). Potassium chloride detectability also appears to be largely dependent upon input from the CT. Chorda tympani transection increases KCl threshold by about 0.6 log 10 units (Geran et al., 1999) to approximately 0.1 M KCl. This shift is similar to that observed for NH 4 Cl, but much less than that reported for NaCl after CTX (Kopka & Spector, 2001; Slotnick et al., 1991; Spector, Schwartz & Grill, 1990). The KCl threshold for intact animals, however, is markedly (~ 0.5 to 0.7 log 10 unit) higher than that of either NaCl or NH 4 Cl. Thus, KCl and NaCl result in approximately the same threshold following CTX while the threshold for NH 4 Cl is slightly lower (Table 5-1). This suggests that the remaining taste receptor cells innervated by the GL and/or GSP are more sensitive to NH 4 Cl than to KCl or NaCl. This finding is supported by the electrophysiology literature in that each of the 3 main gustatory nerves responds more robustly to mid-range and high concentrations of NH 4 Cl than to similar concentrations of KCl or NaCl (see Frank et al., 1983; Kitada et al., 1998; Sollars & Hill, 1998). It is not apparent, however, if this comparatively large response to ammonium is also observed at lower, near-threshold concentrations. It is not known what effect, if any, the anion might have on salt detectability. To date, detectability has been measured for only 1 nonchloride salt (NaG) in the rat and it was found to be very similar to NaCl (Geran & Spector, 2000b). It is unclear whether other nonchloride salts like potassium gluconate or ammonium hippurate would also produce detectability functions that mimic those of their halogenated counterparts. Likewise, the detection thresholds of divalent salts like calcium and magnesium chloride have not been tested. In the future, it might be worthwhile to expand the number of stimuli tested and

PAGE 96

85 nerve transections performed, including GSP and GL neurotomies, to better assess processes necessary for normal salt detectability in the rat. Support for the Hypothesis That The Seventh Cranial Nerve Is More Important For Taste Recognition and Discrimination Than The Glossopharyngeal Nerve A variety of nerve transection experiments have suggested that the gustatory branches of the 7 th cranial nerve appear to be involved in taste recognition and discrimination, while the 9 th cranial nerve, or GL, is more important for the expression of oromotor reflexes to aversive stimuli (see St. John & Spector, 1998). The discrimination experiments have involved tastants from each of the 4 main taste categories. These discriminations have included sucrose vs. maltose, citric acid vs. quinine, KCl vs. quinine, NaCl vs. KCl (St. John, Markison, Guagliardo et al., 1997; St. John & Spector, 1997; St. John & Spector, 1998; Spector & Grill, 1992; Spector et al., 1997) and now KCl vs. NH 4 Cl and NaCl vs. NH 4 Cl. Furthermore, although the GL is the only gustatory nerve that contains fibers highly responsive to quinine (Frank, 1991) and innervates approximately 60% of the taste buds in the oral cavity (Miller, 1995), GL transection has little to no effect on quinine performance in a variety of tasks. These include concentration-dependent avoidance in a 2-bottle preference test and detection threshold (Akaike et al., 1965; St. John & Spector, 1996). Additionally, in their seminal paper, St. John & Spector (1998) reported that GL transection likewise had no effect on KCl vs. quinine discrimination, while combined transection of the CT and GSP nerves did. This was surprising given that the GL contains fibers that are differentially responsive to these 2 stimuli while neither the CT nor the GSP are thought to contain a substantial number of quinine-sensitive units (Frank et al., 1983; Sollars & Hill, 1998), although a single fiber analysis of the GSP has not yet been performed. These results were interpreted as

PAGE 97

86 evidence that input from the CT could potentially be important for the perception of a wide variety of taste qualities, not just the salt and acid stimuli to which it is most responsive. This apparent dichotomy between the functional roles of the facial and GL nerves is more pronounced in the catfish (Finger & Morita, 1987). For these animals, the facial nerve is necessary for locating food while the GL is necessary for initiation of the appropriate oromotor response once a food stimulus has been located (see Caprio et al., 1993). In rodents, the role of the GL is not quite so obvious. Deficits in the sensory/discriminative and hedonic domains of taste function (see Spector, 2000), although minimal, can be produced with GL transection in addition to changes in oromotor reflexes. For instance, GL transection has produced a modest impairment in the quinine avoidance of nave rats (Markison et al., 1999). Combined transection of the CT and GL increases the detection threshold for quinine and compromises pre-trained quinine avoidance in a brief-access test, although GL or CT transection alone is without effect on these tasks (St. John et al., 1994; St. John & Spector, 1996). This suggests that perhaps the GL and CT carry information that is redundant for the maintenance of these tasks in pre-trained rats. Deficits in aversive oromotor reflexes, on the other hand, are quite pronounced with GL transection (Travers et al., 1987), suggesting that the GL is more important for unconditioned reflexes to aversive stimuli (see Eylam et al, 2000), than for taste perception or hedonic responses. While it is easy to appreciate the utility of fibers narrowly tuned for bitter compounds or salts and acids in performing such a function, it is less obvious why this nerve would contain fibers that are narrowly-tuned for sugars (Frank, 1991). It has been

PAGE 98

87 hypothesized (St. John & Spector, 1998) that although the GL is not sufficient to maintain a KCl vs. quinine discrimination for rats trained while intact, this input may be sufficient for rats to learn this or other discriminations over time based on cues mediated by the GL. A second hypothesis concerning the role of the GL, is that this nerve could provide information important for determining suprathreshold stimulus intensity. Humans report an increase in perceived intensity of quinine following bilateral anesthesia of the CT nerve that appears to be due to a release of inhibition (see Catalanotto et al., 1993). It is possible that such a mechanism also exists in the rat, but has not been apparent due to obstacles inherent in measuring suprathreshold intensity in animal subjects. Clearly more research is necessary to determine the role of the GL in the taste-guided behavior of rodents. Conclusions Overall, this series of experiments supports the possibility of a labeled line for sodium taste quality in the rat, as the AS transduction pathway appears to be both necessary and sufficient for the recognition and detection of sodium salts (see Bernstein & Hennessy, 1987; Geran & Spector, 2000b). The fact that amiloride does not seem to affect the taste quality or intensity of KCl or NH 4 Cl lends credence to the hypothesis that this AS pathway is appreciably activated only by sodium and lithium ions. In contrast, the perceived intensities of all three salts tested to date (i. e., NaCl, KCl and NH 4 Cl) have been compromised by CT transection, suggesting that this nerve might be highly sensitive to low concentrations of both sodium and nonsodium chloride salts and/or provide input to regions of the gustatory CNS important for salt detection. It would be interesting to test this hypothesis using a wider array of salt stimuli and nerve transections.

PAGE 99

88 Also of interest is the finding that KCl and NH 4 Cl, stimuli that produce similar patterns of activity in the NST (Erickson, 1963), are discriminated by the rat. This result raises questions about the effect of training on the tuning characteristics of the NST, as well as the possibility that across-fiber pattern coding, as currently applied, might not always be the best method for taste quality classification. Furthermore, this finding underscores the importance of using several different behavioral tasks to describe the perception associated with a particular taste stimulus and suggests that perhaps small differences at the neural level might produce significant differences at the behavioral level. Perhaps focusing on neural and behavioral differences within a prototypical taste quality, such as saltiness, might enable researchers to make interesting predictions about the coding of a particular class of taste stimuli or even taste coding in general.

PAGE 100

89 Table 5-1. Mean Detection Thresholds for Intact and Chorda Tympani-Transected Rats Stimulus Mean Threshold for Intact Rats Shift in Threshold with CTX Mean Threshold for CTX Rats NaCl a ~ 0.006 M 1.0 -1.2 log 10 units ~ 0.1 M NaCl + 100 M amiloride a ~ 0.04 M 0.4 log 10 units ~ 0.1 M KCl b ~ 0.03 M 0.6 log 10 units ~ 0.1 M NH 4 Cl ~ 0.009 0.01 M 0.5 log 10 units ~ 0.04 M a Values from Kopka & Spector, 2001. b Values from Geran et al., 1999.

PAGE 101

90 CONCENTRATION ( M ) 0.00010.0010.010.11 CORRECTED HIT RATE 020406080100 NaCl KCl NaCl + AMIL Figure 5-1. Comparison of NaCl and KCl Detectability Functions. Note that the slopes and inflection points (thresholds) are very similar for KCl (open squares) and NaCl + amiloride (closed circles), but not for NaCl (closed triangles). Both of the NaCl functions are from Geran & Spector, 2000a while the KCl function is from a separate group of rats (Geran et al., 1999).

PAGE 102

91 CONCENTRATION ( M ) 0.00010.0010.010.11 CORRECTED HIT RATE 020406080100 NaCl KCl NH4Cl Figure 5-2. Comparison of Chloride Salt Detectability Functions. Note that the slopes and inflection points (thresholds) are very similar for NaCl (closed triangles) and NH 4 Cl (open circles), but different for KCl (open squares). The NaCl and KCl functions are from Geran & Spector, 2000a & Geran et al., 1999.

PAGE 103

LIST OF REFERENCES Akaike, N., Hiji, Y., & Yamada, K. (1965). Taste preference and aversion in rats following denervation of the chorda tympani and the IX th nerve. Kumamoto Medical Journal, 18, 108-109. Bartoshuk, L. M., Murphy, C., & Cleveland, C. T. (1978). Sweet taste of dilute NaCl: Psychophysical evidence for a sweet stimulus. Physiology and Behavior, 21, 609-613. Beckstead, R. M., Morse, J. R., & Norgren, R. (1980). The nucleus of the solitary tract in the monkey: projections to the thalamus and brainstem nuclei. Journal of Comparative Neurology, 190, 259-282. Beidler, L. M., & Smallman, R. L. (1965). Renewal of cells within taste buds. Journal of Cell Biology, 27, 263-272. Bernstein, I. L., & Hennessy, C. J. (1987). Amiloride-sensitive sodium channels and expression of sodium appetite in rats. American Journal of Physiology Regulatory Integrative and Comparative Physiology, 253, R371-R374. Boudreau, J. C., Hoang, J., Oravec, J., & Do, L. T. (1983). Rat neurophysiological taste responses to salt solutions. Chemical Senses, 8, 131-150. Brand, J. G., Teeter, J. H., & Silver, W. L. (1985). Inhibition by amiloride of chorda tympani responses evoked by monovalent salts. Brain Research, 334, 207-214. Breslin, P. A. S., Spector, A. C., & Grill, H. J. (1993). Chorda tympani section decreases the cation specificity of depletion-induced sodium appetite in rats. American Journal of Physiology Regulatory Integrative and Comparative Physiology, 264, R319-R323. Brown, B. J., & Hollander, M. (1977). Statistics: a biomedical introduction. New York: Wiley. Caprio, J., Brand, J. G., Teeter, J. H., Valentincic, T., Kalinoski, D. L., Kohbara, J., Kumazawa, T., & Wegert, S. (1993). The taste system of the channel catfish: from biophysics to behavior. Trends in Neurosciences, 16, 192-197. 92

PAGE 104

93 Catalanotto, F. A., Bartoshuk, L. M., Ostrum, K. M., Gent, J. F., & Fast, K. (1993). Effects of anesthesia of the facial nerve on taste. Chemical Senses, 18, 461-470. Cheal, M., & Oakley, B. (1977). Regeneration of fungiform taste buds: temporal and spatial characteristics. Journal of Comparative Neurology, 172, 609-626. Cho, Y. K., Li, C. S., & Smith, D. V. (2002). Taste responses of neurons of the hamster solitary nucleus are enhanced by lateral hypothalamic stimulation. Journal of Neurophysiology, 87, 1981-92. Contreras, R. J., & Frank, M. E. (1979). Sodium deprivation alters neural responses to gustatory stimuli. Journal of General Physiology, 73, 569-94. Corbit, J. D., & Luschei, E. S. (1969). Invariance of the rats rate of drinking. Journal of Comparative and Physiological Psychology, 60, 119-125. Craig, W. C. (1918). Appetites and aversions as constituents of instincts. Biological Bulletin 34, 91-107. Danilova, V., Hellekant, G., Tinti, J. M., & Nofre, C. (1998). Gustatory responses of the hamster mesocricetus auratus to various compounds considered sweet by humans. Journal of Neurophysiology, 80, 2102-2112. Denton, D. A. (1982). The hunger for salt. Berlin: Springer-Verlag. DeSimone, J. A., Callaham, E. M., & Heck, G. L. (1995). Chorda tympani taste response of rat to hydrochloric acid subject to voltage-clamped lingual receptive field. American Journal of Physiology, 37, 1295-1300. DeSimone, J. A., & Ferrell, F. (1985). Analysis of amiloride inhibition of chorda tympani taste response of rat to NaCl. American Journal of Physiology Regulatory Integrative and Comparative Physiology, 249, R52-R61. DeSimone, J. A., Lyall, V., Heck, G. L., Phan, T. T., Alam, R. I., Feldman, G. M., & Buch, R. M. (2001). A novel pharmacological probe links the amiloride-insensitive NaCl, KCl and NH 4 Cl chorda tympani responses. Journal of Neurophysiology, 86, 2638-2641. Dickman, J. D., & Smith, D. V. (1988). Response properties of fibers in the hamster superior laryngeal nerve. Brain Research, 450, 25-38. DiLorenzo, P. M., & Monroe, S. (1995). Corticofugal influence on taste responses in the nucleus of the solitary tract in the rat. Journal of Neurophysiology, 74, 258-272.

PAGE 105

94 Doolin, R., & Gilbertson, T. A. (1993). Distribution and characterization of functional amiloride-sensitive sodium channels in rat tongue. Journal of General Physiology, 107, 545-554. Elliot, E. J., & Simon, S. A. (1990). The anion in salt taste: a possible role for paracellular pathways. Brain Research, 535, 9-17. Erickson, R. P. (1963). Sensory neural patterns in gustation. In Y. Zotterman (Ed.), Olfaction and taste (pp. 205-213). Oxford, UK: Pergamon. Eylam, S., Garcea, M., & Spector, A. C. (2000). Glossopharyngeal nerve transection does not alter taste reactivity to sucrose conditioned to be aversive. Chemical Senses, 25, 423-428. Falk, J. L., & Herman, T. S. (1961). Specific appetite for NaCl without postingestional repletion. Journal of Comparative and Physiological Psychology, 54, 405-408. Finger, T. E. (1987). Gustatory nuclei and pathways in the central nervous system. In T. E. Finger & W. L. Silver (Eds.), Neurobiology of taste and smell (pp. 331-354). New York: Wiley. Finger, T. E., & Morita, Y. (1985). Two gustatory systems: facial and vagal gustatory nuclei have different brainstem connections. Science, 227, 776-778. Flynn, F. W., Grill, H. J., Schulkin, J., & Norgren, R. (1991). Central gustatory lesions: II. Effects on sodium appetite, taste aversion learning, and feeding behavior. Behavioral Neuroscience, 105, 944-954. Frank, M. E. (1991). Taste-responsive neurons of the glossopharyngeal nerve of the rat. Journal of Neurophysiology, 65, 1452-1463. Frank, M. E., Contreras, R. J., & Hettinger, T. P. (1983). Nerve fibers sensitive to ionic taste stimuli in chorda tympani of the rat. Journal of Neurophysiology, 50, 941-960. Fregly, M. J. (1958). Specificity of the sodium chloride appetite of adrenalectomized rats: substitution of lithium chloride for sodium chloride. American Journal of Physiology, 195, 645-653. Formaker, B. K., & Hill, D. L. (1988). An analysis of residual NaCl taste response after amiloride. American Journal of Physiology Regulatory Integrative and Comparative Physiology, 255, R1002-R1007. Garty, H., & Palmer, L. G. (1997). Epithelial sodium channels: function, structure and regulation. Physiological Reviews, 77, 359-396.

PAGE 106

95 Geran, L. C., Garcea, M., & Spector, A. C. (2002). Transecting the gustatory branches of the facial nerve impairs NH 4 Cl vs. KCl discrimination in rats. American Journal of Physiology Regulatory Integrative and Comparative Physiology, 283, R739-R747. Geran, L. C., Guagliardo, N. A., & Spector, A. C. (1999). Chorda tympani nerve transection, but not amiloride, increases the KCl taste detection threshold in rats. Behavioral Neuroscience, 113, 185-195. Geran, L. C., & Spector, A. C. (2000a). Amiloride increases sodium chloride taste detection threshold in rats. Behavioral Neuroscience, 114, 623-634. Geran, L. C., & Spector, A. C. (2000b). Sodium taste detectability in rats is independent of anion size: the psychophysical characteristics of the transcellular taste transduction pathway. Behavioral Neuroscience, 114, 1229-1238. Gescheider, G. A. (1997). Psychophysics: the fundamentals. Mahwah, NJ: Erlbaum. Gilbertson, T. A., Avenet, P., Kinnamon, S. C., Roper, S. D. (1992). Proton currents through amiloride-sensitive Na + channels in hamster taste cells. Role in acid transduction. Journal of General Physiology, 100, 803-824. Gilbertson, T. A., Boughter, J. D. Jr., Zhang, H., & Smith, D. V. (2001). Distribution of gustatory sensitivities in rat taste cells: whole-cell responses to apical chemical stimulation. Journal of Neuroscience, 21, 4931-4941. Gilbertson, T. A., Fontenot, D. T., Liu, L., Zhang, H., & Monroe, W. T. (1997). Fatty acid modulation of K+ channels in taste receptor cells: gustatory cues for dietary fat. American Journal of Physiology, 272, C1203-C1210. Gilbertson, T. A., Roper, S. D., & Kinnamon, S. C. (1993). Proton currents through amiloride-sensitive Na + channels in isolated hamster taste cells: enhancement by vasopressin and cAMP. Neuron, 10, 931-42. Gilbertson, T. A., & Zhang, H. (1998). Characterization of sodium transport in gustatory epithelia from the hamster and rat. Chemical Senses, 23, 283-293. Giza, B. K., & Scott, T. R. (1991). The effect of amiloride on taste-evoked activity in the nucleus tractus solitarius of the rat. Brain Research, 550, 247-256. Grigson, P. S., Reilly, S., Shimura, T., & Norgren, R. (1998). Ibotenic acid lesions of the parabrachial nucleus and conditioned taste aversion: further evidence for an associative deficit in rats. Behavioral Neuroscience, 112, 160-171. Grigson, P. S., Shimura, T., & Norgren, R. (1997). Brainstem lesions and gustatory function: III. The role of the nucleus of the solitary tract and the parabrachial

PAGE 107

96 nucleus in the retention of a conditioned taste aversion in rats. Behavioral Neuroscience, 111, 180-187. Grigson, P. S., Spector, A. C., & Norgren, R. (1994). Lesions of the pontine parabrachial nuclei eliminate successive negative contrast effects in rats. Behavioral Neuroscience, 108, 714-723. Halpern, B. P. (1977). Functional anatomy of the tongue and mouth of mammals. In J. A. W. M. Weijnen & J. Mendelson (Eds.), Drinking behavior (pp. 1-92). New York: Plenum Press. Halsell, C. B., Travers, S. P., & Travers, J. B. (1996). Ascending and descending projections from the rostral nucleus of the solitary tract originate from separate neuronal populations. Neuroscience, 72, 185-197. Hamilton, R. B., & Norgren, R. (1984). Central projections of gustatory nerves. Journal of Comparative Neurology, 222, 560-577. Handal, P. J. (1965). Immediate acceptance of sodium salts by sodium deficient rats. Psychonomic Science, 3, 315-316. Hard af Segerstad, C., Hellekant, G., & Farbman, A. I. (1989). Changes in number and morphology of fungiform taste buds in rat after transection of the chorda tympani or chorda-lingual nerve. Chemical Senses, 14, 335-348. Harrer, M. I., & Travers, S. P. (1996). Topographic organization of Fos-like immunoreactivity in the rostral nucleus of the solitary tract evoked by gustatory stimulation with sucrose and quinine. Brain Research, 4, 125-137. Heck, G. L., Mierson, S., & DeSimone, J. A. (1984). Salt taste transduction occurs through an amiloride-sensitive sodium transport pathway. Science, 223, 403-405. Hellekant, G., Ninomiya, Y., & Danilova, V. (1998). Taste in chimpanzees: III. Labeled-line coding in sweet taste. Physiology and Behavior, 65, 191-200. Hermann, G. E., & Rogers, R. C. (1985). Convergence of vagal and gustatory afferent input within the parabrachial nucleus of the rat. Journal of the Autonomic Nervous System, 13, 1-17. Herness, M. S. (1992). Aldosterone increases the amiloride-sensitivity of the rat gustatory neural response to NaCl. Comparative Biochemistry and Physiology, 103, 269-273. Herness, M. S., & Gilbertson, T. A. (1999). Cellular mechanisms of taste transduction. Annual Review of Physiology, 61, 873-900.

PAGE 108

97 Hill, D. L., & Bour, T. C. (1985). Addition of functional amiloride-sensitive components to the receptor membrane: a possible mechanism for altered taste responses during development. Brain Research, 352, 310-313. Hill, D. L., Formaker, B. K., & White, K. S. (1990). Perceptual characteristics of the amiloride-suppressed sodium chloride taste response in the rat. Behavioral Neuroscience, 104, 734-741. Hill, D. L., Mistretta, C. M., & Bradley, R. M. (1982). Developmental changes in taste response characteristics of rat single chorda tympani fibers. Journal of Neuroscience, 2, 782-790. Holland, V. F., Zamphigi, G. A., & Simon, S. A. (1989). Morphology of fungiform papillae in canine lingual epithelium: location of intercellular junctions in the epithelium. Journal of Comparative Neurology, 279, 1-15. Katz, D. B., Simon, S. A., & Nicolelis, M. A. (2001). Dynamic and multimodal responses of gustatory cortical neurons in awake rats. Journal of Neuroscience, 21, 4478-4489. Kinnamon, J. C. (1987). Organization and innervation of taste buds. In T. E. Finger & W. L. Silver (Eds.), Neurobiology of taste and smell (pp. 277-299). New York: Wiley. Kinnamon, S. C., & Margolskee, R. F. (1996). Mechanisms of taste transduction. Current Opinions in Neurobiology, 6, 506-513. Kitada, Y., Mitoh, Y., & Hill, D. L. (1998). Salt taste responses of the IXth nerve in Sprague-Dawley rats: lack of sensitivity to amiloride. Physiology & Behavior, 63, 945-949. Kloub, M. A., Heck, G. L., & DeSimone, J. A. (1997). Chorda tympani responses under lingual voltage clamp: implications for NH 4 salt taste transduction. Journal of Neurophysiology, 77, 1393-1406. Kopka, S. L., Geran, L. C., & Spector, A. C. (2000). The effects of time versus regeneration on the recovery of salt discrimination function following chorda tympani transection. American Journal of Physiology Regulatory Integrative and Comparative Physiology, 278, R720-R731. Kopka, S. L., & Spector, A. C. (2001). Functional recovery of taste sensitivity to sodium chloride depends on regeneration of the chorda tympani nerve after transection in the rat. Behavioral Neuroscience, 115, 1073-1085. Kosar, E., Grill, H. J., & Norgren, R. (1986). Gustatory cortex in the rat. II. Thalamocortical projections. Brain Research, 379, 342-352.

PAGE 109

98 Krieckhaus, E. E., & Wolf, G. (1968). Acquisition of sodium by rats: Interaction of innate mechanisms and latent learning. Journal of Comparative and Physiological Psychology, 65, 197-201. Li, C. S., Cho, Y. K., & Smith, D. V. (2002). Taste responses of neurons in the hamster solitary nucleus are modulated by the central nucleus of the amygdala. Journal of Neuroscience, 88, 2979-2992. Lin, W., Finger, T. E., Rossier, B. C., Kinnamon, S. C. (1999). Epithelial Na + channel subunits in rat taste cells: localization and regulation by aldosterone. Journal of Comparative Neurology, 405, 406-420. Lundy, R. F. Jr. (1998). Potential mechanisms for functional changes in taste receptor cells following sodium deficiency in mammals. Neuroscience and Biobehavioral Reviews, 23, 103-109. Lundy, R. F. Jr., & Contreras, R. J. (1997). Temperature and amiloride alter taste nerve responses to Na + K + and NH 4 + salts in rats. Brain Research, 744, 309-317. Lundy, R. F. Jr., & Contreras, R. J. (1999). Gustatory neuron types in rat geniculate ganglion. Journal of Neurophysiology, 82, 2970-2988. Lundy, R. F. Jr., & Norgren, R. (2001). Pontine gustatory activity is altered by electrical stimulation in the central nucleus of the amygdala. Journal of Neurophysiology, 85, 770-83. Lundy, R. F. Jr., Pittman, D. W., & Contreras, R. J. (1997). Role for passive Na + channels and putative Na + /H + exchange in salt taste transduction in rats. American Journal of Physiology Regulative Integrative and Comparative Physiology, 23, R1923-R1931. Markison, S., St. John, S. J., & Spector, A. C. (1995). Glossopharyngeal nerve transection does not compromise the specificity of taste-guided sodium appetite in rats. American Journal of Physiology Regulatory Integrative and Comparative Physiology, 269, R215-R221. Markison, S., St. John, S. J., & Spector, A. C. (1999). Glossopharyngeal nerve transection reduces quinine avoidance in rats not given presurgical stimulus exposure. Physiology and Behavior, 65, 773-778. Markison, S., & Spector, A. C. (1995). Amiloride is an ineffective conditioned stimulus in taste aversion learning. Chemical Senses, 20, 559-563. McCaughey, S. A., & Scott, T. A. (2000). Rapid induction of sodium appetite modifies taste-evoked activity in the rat nucleus of the solitary tract. American Journal of Physiology Regulatory Integrative and Comparative Physiology, 279, R1121-31.

PAGE 110

99 McCutcheon, N. B. (1991). Sodium deficient rats are unmotivated by sodium chloride solutions mixed with the sodium channel blocker amiloride. Behavioral Neuroscience, 105, 764-766. Miller, I. J. (1995). Anatomy of the peripheral gustatory system. In R. L. Doty (Ed.), Handbook of olfaction and gustation (pp. 521-547). New York: Marcel Dekker, Inc. Minear, M. M., Hammack, S. E., Lundy, R. F. Jr., & Contreras, R. J. (1996). Amiloride inhibits taste nerve responses to NaCl and KCl in Sprague-Dawley and Fischer 344 rats. Physiology and Behavior, 60, 507-516. Morrison, G. R. (1967). Behavioural response patterns to salt stimuli in the rat. Canadian Journal of Psychology, 21, 141-152. Morrison, G. R., & Young, J. C. (1971). Taste control over sodium intake in sodium deficient rats. Physiology & Behavior, 8, 29-32. Nachman, M. (1962). Taste preferences for sodium salts by adrenalectomized rats. Journal of Comparative and Physiological Psychology, 55, 1124-1129. Nakamura, K. & Norgren, R. (1993). Taste responses of neurons in the nucleus of the solitary tract of awake rats: an extended stimulus array. Journal of Neurophysiology, 70, 879-891. Nejad, M. S. (1986). The neural activities of the greater superficial petrosal nerve of the rat in response to chemical stimulation of the palate. Chemical Senses, 11, 283-293. Ninomiya, Y., & Funakoshi, M. (1988). Amiloride inhibition of responses of rat single chorda tympani fibers to chemical and electrical tongue stimulations. Brain Research, 451, 319-325. Ninomiya, Y., & Funakoshi, M. (1989). Selective procaine inhibition of rat chorda tympani responses to electric taste stimulation. Comparative Biochemistry and Physiology A, 92, 185-188. Nishijo, H., & Norgren, R. (1997). Parabrachial neural coding of taste stimuli in awake rats. Journal of Neurophysiology, 78, 2254-2268. Norgren, R. (1995). Gustatory system. In G. Paxinos (Ed.), The rat nervous system (pp.751-771). Academic Press, Inc. Norgren, R., & Leonard, C. M. (1973). Ascending central gustatory pathways. Journal of Comparative Neurology, 150, 217-238.

PAGE 111

100 Nowlis, G. H., Frank, M. E., & Pfaffman, C. (1980). Specificity of acquired aversions to taste qualities in hamsters and rats. Journal of Comparative and Physiological Psychology, 94, 932-942. Ossebaard, C. A., & Smith, D. V. (1995). Effect of amiloride on the taste of NaCl, Na-gluconate and KCl in humans: Implications for Na + receptor mechanisms. Chemical Senses, 20, 37-46. Parks, J. D., & Whitehead, M. C. (1998). Scanning electron microscopy of denervated taste buds in hamster: Morphology of fungiform taste pores. Anatomical Record, 251, 230-239. Pfaffman, C. (1959). The afferent code for sensory quality. American Psychology, 14, 226-232. Rehnberg, B. G., Mackinnon, B. I., Hettinger, T. P., & Frank, M. E. (1993). Anion modulation of taste responses in sodium-sensitive neurons of the hamster chorda tympani nerve. Journal of General Physiology, 101, 453-465. Reilly, S. (1998). The role of the gustatory thalamus in taste-guided behavior. Neuroscience and Biobehavioral Reviews, 22, 883-901. Reilly, S., Grigson, P. S., & Norgren, R. (1993). Parabrachial nucleus lesions and conditioned taste aversion: evidence supporting an associative deficit. Behavioral Neuroscience, 107, 1005-1017. Reilly, S., & Pritchard, T. C. (1996). Gustatory thalamus lesions in the rat: II. Aversive and appetitive taste conditioning. Behavioral Neuroscience, 110, 746-759. Richter, C. P., & Eckert, J. F. (1938). Mineral metabolism of adrenalectomized rats studied by the appetite method. Endocrinology, 22, 214-224. Roitman, M. F., & Bernstein, I. L. (1999). Amiloride-sensitive sodium signals and salt appetite: multiple gustatory pathways. American Journal of Physiology Regulatory Integrative and Comparative Physiology, 276, R1732-R1738. St. John, S. J., Garcea, M., & Spector, A. C. (1994). Combined, but not single, gustatory nerve transection substantially alters taste-guided licking behavior to quinine in rats. Behavioral Neuroscience, 108, 131-140. St. John, S. J., Markison, S., Guagliardo, N. A., Hackenberg, T. D. & Spector, A. C. (1997). Chorda tympani transection and selective desalivation differentially disrupt two-lever salt discrimination performance in rats. Behavioral Neuroscience, 111, 450-459.

PAGE 112

101 St. John, S. J., Markison, S., & Spector, A. C. (1995). Salt discriminability is related to number of regenerated taste buds after chorda tympani nerve section in rats. American Journal of Physiology: Regulatory, Integrative and Comparative Physiology, 269, R141-R153. St. John, S. J., Markison, S., & Spector, A. C. (1997). Chorda tympani nerve transection disrupts taste aversion learning to potassium chloride, but not sodium chloride. Behavioral Neuroscience, 111, 188-194. St. John, S. J., & Spector, A. C. (1996). Combined glossopharyngeal and chorda tympani nerve transection elevates quinine detection thresholds in rats (Rattus norvegicus). Behavioral Neursocience, 110, 1456-1468. St. John, S. J., & Spector, A. C. (1997). The effect of gustatory nerve transections on taste discriminations involving quinine in rats. Chemical Senses, 22, 800. St. John, S. J., & Spector, A. C. (1998). Behavioral discrimination between quinine and KCl is dependent on input from the seventh cranial nerve: implications for the functional roles of the gustatory nerves in rats. Journal of Neuroscience, 18, 4353-4362. Scalera, G., Spector, A. C., & Norgren, R. (1995). Excitotoxic lesions of the parabrachial nuclei prevent conditioned taste aversions and sodium appetite in rats. Behavioral Neuroscience, 109, 997-1008. Schiffman, S. S., Lockhead, E., & Maes, F. W. (1983). Amiloride reduces taste intensity of Na + and Li + sweeteners. Proceedings of the National Academy of Sciences, USA, 80, 6136-6140. Schulkin, J. (1991). Sodium hunger: the search for a salty taste. Cambridge, England: Cambridge University Press. Shimura, T., Grigson, P. S. & Norgren, R. (1997). Brainstem lesions and gustatory function: I. The role of the nucleus of the solitary tract during a brief intake test in rats. Behavioral Neuroscience, 111, 155-168. Shingai, T. (1980). Water fibers in the superior laryngeal nerve of the rat. Japanese Journal of Physiology, 30, 305-307. Simon, S. A. (1992). Influence of tight junctions on the interaction of salts with lingual epithelia: responses of chorda tympani and lingual nerves. Molecular Cellular Biochemistry, 114, 43-48. Slotnick, B. M., Sheelar, S., & Rentmeister-Bryant, H. (1991). Transection of the chorda tympani and insertion of ear pins for stereotaxic surgery: equivalent effects on taste sensitivity. Physiology and Behavior, 50, 1123-1127.

PAGE 113

102 Smith, D. V., & Hanamori, T. (1991). Organization of gustatory sensitivities in hamster superior laryngeal nerve fibers. Journal of Neurophysiology, 65, 1098-1114. Smith, D. V., St. John, S. J., & Boughter, J. D. (2000). Neuronal cell types and taste quality coding. Physiology and Behavior, 69, 77-85. Smith, J. C., Miller, I. J., Krimm, R. F., Nejad, M. S., & Beidler, L. M. (1988). A comparison of the effects of bilateral sections of the chorda tympani nerve and extirpation of the submaxillary and sublingual salivary glands on the eating and drinking patterns of the rat. Physiology and Behavior, 44, 435-444. Sollars, S. I., & Bernstein, I. L. (1996). Neonatal chorda tympani transection alters adult preference for ammonium chloride in the rat. Behavioral Neuroscience, 110, 551-558. Sollars, S. I., & Hill, D. L. (1998). Taste responses in the greater superficial petrosal nerve: substantial sodium salt and amiloride sensitivities demonstrated in two rat strains. Behavioral Neuroscience, 112, 991-1000. Spector, A. C. (2000). Linking gustatory neurobiology to behavior in vertebrates. Neuroscience and Biobehavioral Reviews, 24, 391-416. Spector, A. C., Andrews-Labinski, J., & Letterio, F. C. (1990). A new gustometer for psychophysical taste testing in the rat. Physiology and Behavior, 47, 795-803. Spector, A. C., & Grill, H. J. (1992). Salt taste discrimination after bilateral section of the chorda tympani or glossopharyngeal nerves. American Journal of Physiology Regulatory Integrative and Comparative Physiology, 263, R169-R176. Spector, A. C., Guagliardo, N. A., & St. John, S. J. (1996). Amiloride disrupts NaCl vs. KCl discrimination performance: implications for salt taste coding in the rat. Journal of Neuroscience,16, 8115-8122. Spector, A. C., & Kopka, S. L. (2002). Rats fail to discriminate quinine from denatonium: implications for the neural coding of bitter-tasting compounds. Journal of Neuroscience, 22, 1937-1941. Spector, A. C., Markison, S., St. John, S. J., & Garcea, M. (1997). Sucrose vs. maltose taste discrimination by rats depends on the input of the seventh cranial nerve. American Journal of Physiology, 41, R1210-R1228. Spector, A. C., Norgren, R., & Grill, H. J. (1992). Parabrachial gustatory lesions impair taste aversion learning in rats. Behavioral Neuroscience, 106, 147-161.

PAGE 114

103 Spector, A. C., Schwartz, G. J., & Grill, H. J. (1990). Chemospecific deficits in taste detection after selective gustatory deafferentation in rats. American Journal of Physiology, 258, R820-R826. Stewart, R. E., Heck, G. L., & DeSimone, J. A. (1996). Taste-mixture suppression: functional dissection of cellular and paracellular origins. Journal of Neurophysiology, 75, 2124-2128. Tamura, R., & Norgren, R. (1997). Repeated sodium depletion affects gustatory neural responses in the nucleus of the solitary tract of rats. American Journal of Physiology, 273, R1381-1391. Travers, J. B., Grill, H. J., & Norgren, R. (1987). The effects of glossopharyngeal and chorda tympani nerve cuts on the ingestion and rejection of sapid stimuli: an electromyographic analysis in the rat. Behavioral Brain Research, 25, 233-246. Travers, J. B., & Norgren, R. (1983). Afferent connections to the oral motor nuclei in the rat. Journal of Comparative Neurology, 220, 280-293. Travers, S. P. (2002). Quinine and citric acid elicit distinctive Fos-like immunoreactivity in the rat nucleus of the solitary tract. American Journal of Physiology Regulatory Integrative and Comparative Physiology, 282, R1798-R1810. Travers, S. P., & Nicklas, K. (1990). Taste bud distribution in the rat pharynx and larynx. Anatomical Record, 227, 373-379. Travers, S. P., & Norgren, R. (1995). Organization of orosensory responses in the nucleus of the solitary tract of the rat. Journal of Neurophysiology, 73, 2144-2162. van der Klaauw, N. J., & Smith, D. V. (1995). Taste quality profiles for fifteen organic and inorganic salts. Physiology and Behavior, 58, 295-306. Yamamoto, T., Matsuo, R., Kiyomitsu, Y., & Kitamura, R. (1989). Taste responses of cortical neurons in freely ingesting rats. Journal of Neurophysiology, 61, 1244-1258. Ye, Q., Heck, G. L., & DeSimone, J. A. (1991). The anion paradox in sodium taste reception: resolution by voltage clamp studies. Science, 254, 724-726. Ye, Q., Heck, G. L., & DeSimone, J. A. (1993). Voltage dependence of the rat chorda tympani response to Na + salts: implications for the functional organization of taste receptor cells. Journal of Neurophysiology, 70, 167-178. Ye, Q., Heck, G. L., & DeSimone, J. A. (1994). Effects of voltage perturbation of the lingual receptive field on chorda tympani nerve responses to Na + and K + salts in the

PAGE 115

104 rat: implications for gustatory transduction. Journal of General Physiology, 104, 885-907.

PAGE 116

BIOGRAPHICAL SKETCH Laura received bachelors degrees in psychology and biological science from Florida State University and is looking forward to moving up north with her boyfriend Albert and their cat to study central gustatory processes at Ohio State. 105


Permanent Link: http://ufdc.ufl.edu/UFE0000685/00001

Material Information

Title: The Psychophysics of Salt Taste Transduction Pathways
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0000685:00001

Permanent Link: http://ufdc.ufl.edu/UFE0000685/00001

Material Information

Title: The Psychophysics of Salt Taste Transduction Pathways
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0000685:00001


This item has the following downloads:


Full Text











THE PSYCHOPHYSICS OF SALT TASTE TRANSDUCTION PATHWAYS


By

LAURA CLAIRE GERAN


















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2003





























Copyright 2003

by

Laura C. Geran















ACKNOWLEDGMENTS

I would like to thank Dr. Spector and the members of his lab past and present,

especially Mircea Garcea, Shachar Eylam, Stacy Kopka and Connie Colbert, for their

educational and emotional support. I would also like to thank all of the faculty members

and graduate students I have had the pleasure of meeting along the way, in particular

Laura Tucker, Sylvia Belski, Melanie McEwen, and Cheryl Vaughan, for their advice and

patience. Finally, I would like to thank Albert and my family for their unwavering

support of my decision to attend graduate school. It has been a great experience.
















TABLE OF CONTENTS
page

A C K N O W L E D G M E N T S ......... .................................................................................... iii

L IST O F T A B L E S ......... ............................... ......... .... ........... .. vii

LIST OF FIGURES ............. .. ..... ...... ........ ....... .......................... viii

A B STR A C T ................................................. ..................................... .. x

CHAPTER

1 LITERATURE REVIEW ............................ ........................................ 1

In tro d u ctio n .............................................................................................. 1
Peripheral Gustatory System.......... .................. ... ........................ 1
Taste B uds ........................................................ .............. 1
G u statory N erv es.......................................... .......................... 3
C central G u statory Sy stem ............................................................................................... 5
M edullary and Pontine Taste Nuclei.............. .................. ............. .............. 5
A scending Gustatory Pathw ays....................................... ............................. 6
Taste Transduction.................................... .............. 8
Prim ary Tastes.................................................................... 8
Transduction M echanism s ............................................... .............. .............. 9
Ion channels ............................................................... 9
M etabotropic receptors .................................. ...................................... 9
Sodium Transduction Pathw ays........................ ........................... ........ ....... 10
Am iloride-sensitive (A S) pathw ay ...................................... ............... 10
A m iloride-insensitive (A I) pathw ay ................................................................ 11
Labeled Line vs. Across-Fiber Pattern Theory .............................................. 12

2 ANION SIZE DOES NOT COMPROMISE SODIUM RECOGNITION BY RATS
FOLLOWING ACUTE SODIUM DEPLETION.................... ................16

Background ........................................................... ..... ...... ........ 16
M e th o d s ..................................................................................... 1 8
S u b j e c ts .................................................................................. 1 8
Apparatus ......................................... 19
Training Procedure ...................................................... ............................. 20
Sodium D epletion ................................................................... 20
Brief-Access Testing........................................................ 21









U rine A nalysis............................................ 22
D ata A analysis .......................................... .............. 22
R results ................. ................................... .......... .... 23
B rief-A access Testing................... ......................................... .......................... 23
N um ber of Trials Initiated......... .............. ...... ........... ................. .............. 24
Sodium Balance ......................................... 24
D iscu ssio n .................................................................... 2 4

3 GLOSSOPHARYNGEAL NERVE TRANSACTION DOES NOT IMPAIR
POTASSIUM CHLORIDE vs. AMMONIUM CHLORIDE OR SODIUM
CHLORIDE vs. AMMONIUM CHLORIDE DISCRIMINATION .............................32

B a c k g ro u n d ...................................................................................................... 3 2
M e th o d s .................................................................................... 3 4
Subjects ........................................ .............. 34
A pparatus and Trial Structure .......................................................... ......... .... 34
Training Procedure....................... .. .... .. ...................... 35
Presurgical Discrimination and Amiloride Testing ......................................... 36
Surgery ............... .. ... .... .. .... .... ..... ....................... 37
Postsurgical Testing and the W ater Control Test.................................... .... ... 37
H isto lo g y ............................................................................... 3 8
D ata Analysis ......................................... 38
R results ................. .................................. .......... ...... 39
Presurgical Discrimination Testing................................. .................... 39
Postsurgical Testing and Histology ................................. ................ 40
D iscu ssion ................... ..................... 4 1
Presurgical Discrimination Testing ................................................... ........ 41
Postsurgical D iscrim nation Testing ..................... ........................................ 42
Potential Mechanisms Underlying KC1 vs. NH4C1 Discrimination.................... 43
S u m m ary .................................. ........................................ 4 4

4 AMILORIDE-INSENSITIVE UNITS OF THE CHORDA TYMPANI NERVE ARE
NECESSARY FOR NORMAL AMMONIUM CHLORIDE DETECTABILITY IN
T H E R A T ............................................................................ 5 7

B a c k g ro u n d ...................................................................................................... 5 7
M e th o d s ................................................................................ 5 9
G general M methods .................. .............................. .. .. .......... .............. .. 59
Testing ........................... ........ ............. 60
Surgery ...................................................................................... 6 1
H isto lo g y ............................................................................... 6 1
D ata Analysis ......................................... 62
R e su lts ................. ..... ................................... ................................... ..... 6 3
Presurgical D election Threshold .................................................................... 63
Postsurgical Detection Threshold ........................................ .. .......... 64
Water Control Test and Histology ..................................... .............. 64
D isc u ssio n ....................................................................................................... 6 4


v









Ammonium Chloride Detectability Depends Upon an Amiloride-Insensitive
R oute of Transduction........................................... ........................... 65
The Chorda Tympani Nerve is Necessary for Normal Ammonium Chloride
D election ........... ......... ......... ........... ............... .......... 67
Conclusions .............. ...... .. ................ ........... ....... ......... 68

5 G E N E R A L D ISC U SSIO N ......................................... .............................................79

Discrepancies Between the Electrophysiology and the Behavior In Regard to
NH4C1 ................................................... ...... ....... ........... 79
Implications for Chloride Salt Detectability ............................................................... 82
Support for the Hypothesis That The Seventh Cranial Nerve Is More Important
For Taste Recognition and Discrimination Than The Glossopharyngeal
Nerve............................................ .......... 85
C onclusions............................... ........... .......... 87

L IST O F R E FE R E N C E S ....................................................................... ... ................... 92

BIOGRAPHICAL SKETCH ............................................................. ............... 105
















LIST OF TABLES

Table p

3-1. Training Schedule....................................................................45

3-2. Experim ent Schedule ............................ ........................................... ............... 46

4-1. Training Schedule....................................................................70

4-2. Test Stim ulus Presentation Schedule...................................... ........................ 71

4-3. Experiment Schedule ................................................................ 71

5-1. Mean Detection Thresholds for Intact and Chorda Tympani-Transected Rats..........89
















LIST OF FIGURES


Figure p

2-1. Brief-access licking to each stimulus by sodium-depleted rats ............ ...............29

2-2. Mean (+ SE) number of trials initiated by each group of rats .................................30

2-3. Sodium balance for each group of animals measured in mmol............... ...............31

3-1. Presurgical KC1 vs. NH4Cl discrimination with and without amiloride...................47

3-2. Presurgical NaCl vs. NH4Cl discrimination with and without amiloride...................48

3-3. Mean presurgical performance by concentration.. ............ ...................................49

3-4. Pre- vs. postsurgical performance on the KC1 vs. NH4Cl task................................50

3-5. Pre- vs. postsurgical performance on the NaCl vs. NH4C1 task...............................51

3-6. Mean presurgical vs. postsurgical performance by concentration...........................52

3-7. Effect of amiloride on postsurgical KC1 vs. NH4Cl discrimination .........................53

3-8. Effect of glossopharyngeal transaction on NaCl vs. NH4C1 performance in the
presence of am iloride........................... .... .............. ........................... 54

3-9. Mean comparisons of postsurgical amiloride performance by concentration............55

3-10. W after control test ............... ................. .......... ............ ......... 56

4-1. Effect of amiloride on NH4Cl detection.............. ........... ...................72

4-2. NH4Cl threshold decreased again following amiloride treatment...........................73

4-3. Individual shifts in presurgical threshold.. ............................................................... 74

4-4. Individual shifts in performance with surgery ......................................................75

4-5. NH4Cl detectability functions pre- and post-surgery.. ............................................76

4-6. Postsurgical NH4Cl detectability functions with and without amiloride ..................77









4-7. W after control test ................. ......... .. .................. .. ...... 78

5-1. Comparison of NaCl and KCl Detectability Functions...........................................90

5-2. Comparison of Chloride Salt Detectability Functions....................................91















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

THE PSYCHOPHYSICS OF SALT TASTE TRANSDUCTION PATHWAYS


By

Laura Claire Geran

May 2003


Chair: Alan C. Spector
Major Department: Psychology

Salt stimuli can activate two transduction mechanisms in the rat's oral cavity. One

mechanism appears to rely on taste receptor cells that contain sodium-selective ion

channels on their surface. Passage through these channels can be blocked with the drug

amiloride. The other mechanism is thought to be less selective, with associated fibers

responding to potassium and ammonium as well as sodium salts. Most data suggest that

the salt responsiveness of this population of fibers is not significantly attenuated with

amiloride treatment, although some researchers have found evidence to the contrary. The

two salt transduction pathways are commonly grouped according to their amiloride

sensitivity as either amiloride-sensitive (AS) or amiloride-insensitive (AI). Activation of

the AI pathway appears to be limited by anion size, with large anions like gluconate

producing the greatest suppression. Previous research has indicated that the AS pathway

is necessary and sufficient for normal sodium detection in the rat as well as necessary for

sodium recognition. The current experiments were designed to determine whether the AS









pathway was also sufficient for sodium recognition as well as to elucidate possible

functional roles of the AI pathway regarding the perception of nonsodium salts. This was

accomplished by observing the effects of physiological manipulations like gustatory

nerve transaction, amiloride treatment, and acute sodium depletion on the taste-guided

behavior of highly-trained rats. Briefly, the sodium-specific AS pathway appears to be

sufficient for sodium recognition in acutely depleted animals. In addition, AI receptor

cells innervated by the chorda tympani (CT) nerve were found to be necessary for normal

detection of ammonium chloride (NH4C1) and AI cells innervated by the facial nerve

were both necessary and sufficient to discriminate NH4C1 from KC1. This finding

suggests that taste receptor cells innervated by the facial nerve could use separate AI

transduction mechanisms with different selectivities for ammonium and potassium. This

work also supports the hypothesis that amiloride does not significantly impair the

perception of nonsodium salts, as well as the contention that the facial nerve may provide

unique information about taste quality in spite of innervating only about 30% of the taste

buds in the oral cavity.














CHAPTER 1
LITERATURE REVIEW

Introduction

The current experiments were designed to relate events at the level of the taste

receptor cell due to salt stimulation with the perception experienced by the animal, in this

case the Sprague-Dawley rat. To this end, performance was measured on a variety of

psychophysical tasks including detection threshold, discrimination and recognition on a

brief-access test using several salt stimuli. The effects of physiological manipulations

such as gustatory nerve transaction, oral application of the ion channel blocker amiloride,

and acute sodium depletion on taste-guided behavior were also assessed. Prior to

elaborating on the details of these experiments, the remainder of this chapter consists of a

brief introduction to the anatomy and physiology of the mammalian gustatory system as

well as taste coding theory for readers that might be unfamiliar with these concepts.

Peripheral Gustatory System

Taste Buds

In mammals, the sensory cells of the gustatory system are found in clusters of

approximately 50 epithelial cells called taste buds (Miller, 1995). Some of these modified

epithelial cells are taste receptor cells (TRCs) and contain receptors on their apical

membranes capable of interacting with taste compounds. Taste stimuli contact the apical

region of the TRC by way of a taste pore. Taste buds also contain support cells that do

not have taste receptors. Some support cells are immature TRCs, while others are thought

to provide structural support to the receptor cells (Kinnamon, 1987). The majority of taste









buds are found on the tongue in distinct epithelial structures called papillae and in the soft

palate. In the rat, fungiform, or mushroom-shaped papillae, are located on the anterior

two-thirds of the tongue, and a single, round circumvallate papilla can be found on the

posterior surface of the tongue. Foliate papillae resembling gills are located on the sides

of the tongue toward the back. A small number of taste buds can also be found on the

larynx, epiglottis, nasoincisor ducts and esophagus. In the rat, the circumvallate papilla

contains considerably more taste buds than the other papillary fields (- 60% of the total

number). The fungiform papillae contain approximately 15%, the palate 15%, and the

remaining 5-10% are distributed among the other gustatory fields (Miller, 1995; Travers

& Nicklas, 1990). Individual variance commonly exists both in the number of papillae

per anatomic region and the number of taste buds per papilla (Miller, 1995).

Dendrites from a single axon can synapse with cells from more than one taste bud,

including buds located in different papillae. It is unclear whether buds that synapse with

the same nerve fiber express the same receptor proteins. This system would seem to

suggest convergence of information; however, individual taste buds can also be

innervated by more than one axon, suggesting that peripheral coding has the potential to

be somewhat more complex than it initially appears. In addition, TRCs are replaced every

10 days (Beidler & Smallman, 1965). In spite of this, response profiles from the nerves

remain remarkably stable. It has been proposed that soluble factors released by the nerve

or by adjacent cells affect the development of the new cell (see Miller, 1995). The

number and variety of receptors found on each TRC are also a source of debate.

Intracellular recording techniques have shown that most TRCs respond to a number of









stimulus classes, suggesting that most, but not all, cells contain several different types of

taste receptors (Gilbertson et al., 2001).

When a stimulus interacts with a taste receptor, the TRC undergoes a change in

membrane potential or intracellular calcium concentration leading to neurotransmitter

release (see Herness & Gilbertson, 1999). Hence, the TRCs form chemical synapses with

the innervating dendrites. When the TRC releases enough neurotransmitter, an action

potential is produced in the innervating neuron. In addition to transmitting taste

information to the peripheral nerves, these neurotransmitters might also bind to other

TRCs in the taste bud, modifying their activity. Neurotransmitters associated with TRCs

to date include serotonin, GABA and norepinephrine (Herness & Gilbertson, 1999). The

presence of GABA suggests that inhibition as well as excitation may provide meaningful

gustatory information. Peptides also thought to be released by TRCs with stimulation

include bombesin, cholescystokinin, histidine, neuropeptide Y, and somatostatin

(Norgren, 1995). Cells in a taste bud might also communicate with one another by way of

gap junctions (Holland et al., 1989). Such a mechanism could lead to the excitation of

adjacent cells without the use of neurotransmitters. Because not all TRCs synapse with

taste afferents, electrical coupling could potentially allow information to be transmitted

from a greater number of cells than chemical transmission alone (Herness & Gilbertson,

1999).

Gustatory Nerves

Rats have 3 main nerves that carry taste input. The chorda tympani (CT) branch of the

facial nerve, or cranial nerve VII (CN VII), innervates the taste buds in the anterior two-

thirds of the tongue (i.e., the fungiform papillae). The greater superficial petrosal (GSP)

branch of this nerve innervates taste buds in the palate and nasoincisor ducts and the









glossopharyngeal nerve (GL), or CN IX, carries taste information from taste buds located

in the circumvallate papilla and a portion of the foliate papillae. The CT innervates the

remainder of the taste buds in the foliate papillae. The superior laryngeal (SLN) branch of

the vagus nerve, or CN X, also synapses with taste buds. These are located in the

epiglottis, esophagus and larynx and are thought to be more important for airway

protection than for the perception of taste quality (Dickman & Smith, 1988; St John &

Spector, 1998; Smith & Hanamori, 1991). The somata of these gustatory cranial nerves

are located in 3 ganglia. The CT and GSP somata reside in the geniculate ganglion.

Glossopharyngeal somata are found in the inferior petrosal ganglion, and the cell bodies

of the SLN are in the inferior nodose ganglion (Miller, 1995). These ganglia also provide

some parasympathetic innervation to salivary glands (see Smith et al., 1988).

In addition to differences in the receptor fields innervated by these nerves, the nerves

also differ in response profile. For instance, electrophysiology has shown that the CT

contains 2 classes of fibers, one that is selective for sodium and lithium salts and one that

is more broadly-tuned (Frank et al., 1983). This second class of fibers is highly

responsive to salts and acids and more moderately responsive to alkaloids like quinine.

The GSP is highly responsive to sugars while moderately responsive to salts, acids and

quinine (Nejad, 1986; Sollars & Hill, 1998). The GL contains 3 classes of fibers; one

class that responds best to salts and acids, one that responds best to sugars, and one that is

highly responsive to alkaloids (Boudreau et al., 1983; Frank, 1991). Fibers from the SLN

do not seem to form units based on chemical sensitivity but are highly responsive to

stimuli described as "sour" or "bitter" by humans (Dickman & Smith, 1988).









Central Gustatory System

Medullary and Pontine Taste Nuclei

Second-order neurons of the gustatory system reside in the lateral region of the rostral

nucleus of the solitary tract (NST) located in the medulla. Terminal branches from the

peripheral nerves are distributed such that the nerves innervating the most rostral part of

the oral cavity (i.e., the CT and GSP) synapse most rostrally in the gustatory NST. The

GL is more caudally represented and the vagus nerve more caudal still. In spite of this

general division, however, there is still a large degree of overlap in the terminal fields of

the 3 main gustatory nerves (Hamilton & Norgren, 1984; Travers & Norgren, 1995).

Trigeminal fibers also terminate in the lateral portion of the rostral NST while visceral

projections from vagal afferents terminate in the caudal NST (Finger, 1987; Travers,

2002). The majority of axons from neurons in the gustatory NST project to the

ventromedial subnucleus of the parabrachial complex, henceforth referred to as the

parabrachial nucleus (PBN). This nucleus is located in the dorsal pons and receives both

gustatory and visceral input. These inputs are mostly, but not entirely, segregated

(Hermann & Rogers, 1985). Apart from the PBN, some gustatory neurons in the rostral

NST project to the parvicellular reticular formation where they synapse with

parasympathetic salivatory neurons and cells from the hypoglossal nucleus (see Halsell et

al., 1996; Norgren, 1995). These pathways are thought to be responsible for reflexive

oromotor responses to taste stimuli (see Travers & Norgren, 1983). In primates, most

projections from the gustatory NST bypass the PBN and synapse directly onto thalamic

neurons (Beckstead et al., 1980).









Ascending Gustatory Pathways

From the PBN, input from taste nerves is channeled into 2 separate functional

pathways. The lemniscal or thalamocortical gustatory pathway projects bilaterally in the

parvocellular region of the ventral posteromedial nucleus (VPMpc) of the thalamus

before reaching the cortex (Kosar et al., 1986; Norgren & Leonard, 1973). Although this

pathway is often touted as being responsible for learned taste associations, there is

evidence to the contrary. For instance, ibotenic acid lesions of the gustatory thalamus fail

to impair conditioned taste aversion acquisition (Flynn et al., 1991; Reilly, 1998). In

addition, both the amygdala and PBN send some projections directly to the gustatory

cortex, which could be necessary for sensory discrimination, although it is not clear

whether these fibers carry information regarding taste quality. The ventral gustatory

pathway, or pontolimbic system, projects mainly from the PBN to the amygdala, but also

projects to the lateral hypothalamus, bed nucleus of the stria terminalis and other limbic

regions of the forebrain (Norgren, 1995). This pathway is thought to be important for the

hedonic responses to taste stimuli.

In addition to gustatory projections from the hindbrain to the forebrain, telencephalic

structures also project back onto the thalamus, PBN, NST, amygdala and hypothalamus.

The amygdala and hypothalamus also send axons to the PBN and NST and the PBN

reciprocally innervates the NST. These projections from limbic and cortical areas could

lead to feedback onto taste pathways from other sensory modalities, as well as from areas

involved with motivation and emotion (see Norgren, 1995). Recent evidence supporting

this hypothesis has shown that the tuning characteristics of taste-responsive neurons in

the PBN are altered with stimulation of either the lateral hypothalamus or central

amygdala (Cho et al., 2002; Lundy & Norgren, 2001).









The NST appears to be necessary for rats to respond to the taste quality of a stimulus.

When this area is lesioned, concentration-response functions become flattened for all 4

prototypical stimuli (Shimura et al., 1997). While the NST appears to be necessary for the

perception of taste quality, it has been suggested that the role of the PBN in taste

processing involves the integration of taste quality with visceral input (see Grigson et al.,

1998; Spector et al., 1992). For example, after bilateral PBNX a rat is unable to acquire a

conditioned avoidance response, a sodium appetite or display successive negative

contrast (Flynn et al., 1991; Grigson et al., 1994; Grigson et al., 1998; Scalera et al.,

1995), but can still form associations between 2 tastants or between a somatosensory cue

and malaise (Grigson et al., 1998; Reilly et al., 1993). Expression of a conditioned taste

aversion acquired prior to PBN lesion does not seem to be affected by the surgery

(Grigson et al., 1997).

Lesions of the gustatory thalamus and cortex appear to have little or no effect on

innate taste processing in the rat (Reilly & Pritchard, 1996), but may affect the

acquisition and/or retention of a conditioned taste aversion (see Spector et al., 1992 for

review). Grigson and colleagues (1998) have proposed that in addition to the NST and

PBN, at least one forebrain structure, such as the amygdala or thalamus, must be intact to

elicit conditioning to a taste stimulus. Electrophysiological recordings from neurons in

the gustatory cortex have suggested that these cells might respond to the hedonic

characteristics of the taste stimulus rather than to its associated taste quality (Yamamoto

et al., 1989). Recent evidence using chronic recording techniques suggests that taste

quality might be discerned from temporal patterns at this level recorded between 0.2 and

1 s after stimulus onset (Katz et al., 2001).









Taste Transduction

Primary Tastes

There are 4 prototypical classes of tastants; "salty," "sweet," "bitter" and "sour." A

fifth taste known as "umami" that has been described as being "meaty" or "savory" is

gaining acceptance among taste researchers and is typified by the taste of monosodium

glutamate, or MSG. Umami compounds often modify the taste qualities of other

compounds in addition to producing a taste themselves. It should be noted, however, that

these categories are not definitive as they are based on the taste quality most associated

with a particular stimulus, or set of stimuli, by human subjects and often only within a

narrow range of effective concentrations. Sodium chloride, although often used as the

prototypical "salty" stimulus in taste research, is said to taste "sweet" at very low

concentrations (Bartoshuk et al., 1978), indicating that a stimulus may produce more than

one taste quality depending on concentration. In addition, this salt has also been

described as "sour/salty" at moderate concentrations, suggesting that even a prototypical

stimulus like NaCl can produce significant side-band tastes (van der Klaauw & Smith,

1995). The fact that these taste qualities are based on human data also casts doubt on

whether these 5 categories, or their prototypical stimuli, are stable across species. Other

possible taste categories include "electric," "fatty acid" and "water" although much less

is known about the perceptual and physiological properties of these sensations than the

others mentioned here (see Gilbertson et al., 1997; Ninomiya & Funakoshi, 1989;

Shingai, 1980). The prototypical classes of tastants can be divided into 2 groups based on

transduction mechanism.









Transduction Mechanisms

Ion channels

Salt and acid stimuli dissociate in the saliva releasing free protons in the case of acids,

and ions in the case of salts that pass through ion channels in the TRC membrane.

Epithelial sodium channels (ENaCs), probably the best-characterized ion channels in the

oral cavity of the rat, are voltage-sensitive and highly selective for sodium and lithium

ions (see Garty & Palmer, 1997; Ye et al., 1991). Less selective, voltage-insensitive ion

channels also allow sodium ions to pass into the cell (Ye et al., 1991). Acid transduction

mechanisms fall into 3 main categories. Protons can pass into the TRC as a result of the

concentration gradient, as observed in the hamster (Gilbertson et al., 1992), block cation

channels, as observed in the frog and mudpuppy (see Kinnamon & Margolskee, 1996) or

bind to sites on the basolateral membrane of TRCs (DeSimone et al., 1995).

Metabotropic receptors

Stimuli giving rise to "sweet," "bitter" and "umami" sensations in humans have been

shown to activate G-protein-coupled receptors in the oral cavity (see Kinnamon &

Margolskee, 1996). A variety of second messenger pathways are reportedly affected by

this coupling, including cAMP, inositol triphophate (IP3) and diacyl glyceride (DAG).

Both cAMP and IP3 pathways are activated by stimuli that produce "bitter" or "sweet"

taste qualities. Artificial sweeteners and amino acids are thought to use the IP3 pathway,

while natural sugars are thought to use a cAMP-dependent pathway (Herness &

Gilbertson, 1999).

Unfortunately, there is not a one-to-one correspondence between taste qualities and

transduction pathways. For example, 2 stimuli with different taste qualities can share a

common route of transduction. In the hamster, although both salts and acids use the same









amiloride-sensitive pathway (Gilbertson et al., 1992), these stimuli result in different

qualitative perceptions (Nowlis et al., 1980). In addition, stimuli that share a common

taste quality like quinine and denatonium, stimuli described as "bitter" by humans, can

affect more than one transduction pathways. Quinine is thought to be cAMP-dependent

and denatonium uses an IP3 second messenger pathway (Herness & Gilbertson, 1999),

yet these 2 compounds produce a unified taste perception (Spector & Kopka, 2002). A

single stimulus can also potentially activate more than one transduction pathway. Sodium

ions activate 2 transduction pathways in the rat, but only one of these appears to be

important for producing the characteristic taste quality associated with sodium.

Sodium Transduction Pathways

Amiloride-sensitive (AS) pathway

The amiloride-sensitive transduction pathway in the oral cavity of the rat is selective

for sodium and lithium ions (Brand et al., 1985; DeSimone & Ferrell, 1985) and

activation of this pathway is significantly reduced with application of the ENaC blocker

amiloride (DeSimone & Ferrell, 1985; Ninomiya & Funakoshi, 1988). Thus, functional

amiloride-sensitive sodium channels, or ASSCs, are thought to be located in the apical

region of taste receptor cells (DeSimone & Ferrell, 1985). Immunohistochemistry has

also indicated that ASSCs are also located in the basal region of some TRCs, but these

appear to be nonfunctional under normal circumstances (Lin et al., 1999). Sodium

responses are only suppressed by amiloride in the CT and GSP nerves, suggesting that

TRCs with functional ASSCs are predominantly, if not exclusively, innervated by the

gustatory branches of the facial nerve (Kitada et al., 1998; Ninomiya & Funakoshi, 1988;

Sollars & Hill, 1998).









When a rat is depleted of body sodium, it will ingest increased quantities of sodium-

containing solutions. This phenomenon is called sodium appetite (see Denton, 1982;

Schulkin, 1991). With amiloride treatment both the magnitude and specificity of a

sodium appetite are diminished (Bernstein & Hennessy, 1987). Amiloride treatment also

impairs the rat's ability to discriminate between NaCl and KC1 (Spector et al., 1996).

Together, these studies suggest that the amiloride-sensitive pathway is necessary for the

rat to perceive the taste quality of sodium. While amiloride appears to be tasteless to the

rat (Hill et al., 1990; Markison & Spector, 1995), humans report that it possesses a

"bitter" quality (Schiffmann et al., 1983). Studies using human subjects also indicate that

although amiloride decreases the perceived intensity of sodium salts, this decrease is

observed only in the "sour" component of the stimulus, leaving "saltiness" unaffected

(Ossebaard & Smith, 1995).

Amiloride-insensitive (AI) pathway

The amiloride-insensitive pathway, on the other hand, is permeable to a variety of ions

including Na Li K+ and NH4 This pathway also appears to use receptors located in

the basolateral region of the TRC. For this reason, ions are thought to pass through tight

junctions between cells before accessing receptor channels. Therefore, these ions must

have a small radius. Amiloride is thought to be too large to fit through these tight

junctions, making this pathway insensitive to the sodium channel blocker. Much less is

known about this pathway than about the AS pathway as it is much more difficult to

access. Although it does not appear to be affected by amiloride, transduction via this

pathway can be partially blocked by introducing salts with large anions. Sodium passage

through tight junctions depends on the electroneutral diffusion of ions. If the anion is too

large to fit through these junctions, the cation will also fail to reach the receptor sites due









to the unfavorable electrical gradient. In support of this hypothesis, the CT nerve is much

less responsive to sodium gluconate (NaG) and other large anion sodium salts than to

NaCl at the same concentration (Formaker & Hill, 1988; Ye et al., 1993), suggesting that

perhaps the large gluconate ion keeps the sodium ions from passing through the tight

junctions. This effect does not seem to be due to a loss of C1 receptor conductance, as C1-

channel antagonists do not affect CT responses to NaCl (Elliot & Simon, 1990). Thus, the

response to NaG appears to depend almost entirely on activation of the amiloride-

sensitive pathway. Interestingly, the detectability functions for NaCl and NaG are

strikingly similar (Geran & Spector, 2000b), suggesting that the AS pathway is not only

necessary for sodium detection but also sufficient. It could be argued that this necessity

and sufficiency qualify this pathway as a labeled line for sodium taste in the rat.

Labeled Line vs. Across-Fiber Pattern Theory

Briefly, labeled line theory states that activation of a particular set of neurons leads to

the perception of a particular quality. When the activation is below a certain threshold,

the quality is not perceived and when activation reaches this threshold, it is perceived.

The across-fiber pattern (AFP) theory states that the overall pattern produced by the

presence and absence of activation across all afferent taste fibers or central neurons

produces the perceived taste quality (see Pfaffman, 1959). In the strictest sense, a labeled

line should be both necessary and sufficient to produce a given perception. Most of the

data supporting the AFP theory come from the fact that gustatory neurons at each level of

the neuraxis respond to several tastant classes. How much of this activity is "signal" and

how much is "noise" is a matter of debate.

The breadth of tuning of neurons in the NST is often used to support the AFP theory.

Although NST neurons are responsive to a variety of taste stimuli, 2 lines of evidence









suggest that perhaps the activity of these neurons is more selective than previously

thought. The first such evidence comes from immunohistochemistry. Experiments

conducted by Travers and colleagues have shown that the immediate early gene c-Fos is

differentially expressed in the NST in response to sucrose, citric acid and quinine (Harrer

& Travers, 1996; Travers, 2002). Although overlap exists for the three stimuli, these data

suggest that there may be some chemotopic arrangement within the gustatory regions of

the CNS, although more experiments are required. The second line of evidence comes

from electrophysiological examinations of higher brain regions in the gustatory neuraxis

using either awake animals (Nakamura & Norgren, 1993) or animals in which the

amygdala, gustatory cortex or lateral hypothalamus was stimulated in conjunction with

tastant presentation (see Cho et al., 2002; Di Lorenzo & Monroe, 1995, Lundy &

Norgren, 2001). Under these conditions, NST neurons are less broadly-tuned than in an

anesthetized preparation in the absence of electrical stimulation. Hence, a feedback

mechanism exists through which medullary and pontine taste responses could potentially

be modified according to the motivational state or previous experience of the animal.

If one takes a looser approach to the two theories they are not necessarily mutually

exclusive. For instance, it is possible that one theory holds for some stimuli but not for

others. For example, much of the behavioral evidence suggests that sodium taste might

utilize a labeled line system, although the perception of other tastes might require

activation of several fiber types or transduction mechanisms. Labeled lines have also

been hypothesized to account for "sweet" taste in the hamster and chimpanzee (Danilova

et al., 1998; Hellekant et al., 1998). It is also possible that the behavioral task in question

might be more parsimoniously described by one theory than by the other. For instance,









recognition or detection tasks might only require activation of one pathway for normal

performance, while a discrimination task might require the animal to perceive more

subtle differences in activity that might best be described in terms of AFP theory.

Currently, it appears that AFP theory has become the accepted paradigm, although it

can be argued that a somewhat labeled line-like approach has become pervasive in the

form of "best stimulus" categories. Gustatory neurons are often classified according to

the prototypical stimulus that produces the greatest magnitude of responding. For

instance, a neuron that responds highly to sucrose, moderately to NaC1, somewhat to

citric acid and only marginally to quinine would be termed a "sucrose best" cell. Activity

from each of the cells in this category would be then be averaged and compared to the

mean activity of other "best stimulus" categories. This method could potentially keep

more subtle patterns of activity, perhaps due to second or third-best stimulus categories

within a "best stimulus" classification from becoming apparent. In a strict across-fiber

pattern code excitation of all cells regardless of best stimulus, as well as inactivity and

inhibition of neurons with gustatory input, could potentially be meaningful and should

therefore be analyzed. Another caveat to this method of analysis is the fact that

researchers do not record from each cell in a particular anatomical region. Recording

from every cell in the NST would, of course, be a herculean task, but recording only from

those that are most accessible could potentially lead to a skewed sample. While tests of

necessity and sufficiency can be performed at some levels of the gustatory system to shed

light on the AFP vs. labeled line debate, we are not currently able to block responses

from a particular best-stimulus category of the NST so that the same tenets might be

applied to responses in the CNS. In spite of these caveats, predictions based on AFP






15


theory appear to correlate well, for the most part, with the results of stimulus

generalization tasks (see Erickson, 1963). As mentioned previously, gustatory coding is

most likely best explained by some amalgamation of the 2 competing theories. The

following chapters will examine the necessity and sufficiency of the AS and AI salt

transduction pathways for a variety of taste-guided tasks. Whether these data support or

challenge the existing coding theories will also be addressed.














CHAPTER 2
ANION SIZE DOES NOT COMPROMISE SODIUM RECOGNITION BY RATS
FOLLOWING ACUTE SODIUM DEPLETION

Background

When rats are in a sodium-depleted state, the apparent perception of a "sodium-like"

taste quality will promote ingestion of the stimulus (e.g., Falk & Herman, 1961; Handal,

1965; Nachman, 1962; Richter & Eckert, 1938). This depletion-induced elevation of

intake, termed sodium appetite, is specific for sodium and lithium salts (Nachman, 1962)

and has been described in a number of mammals (see Denton, 1982). Researchers have

used this phenomenon for several decades to test hypotheses about salt taste perception,

most notably through the use of brief-access taste tests which substantially reduce the

contribution of postingestive receptors (e.g., Breslin, et al., 1993; Falk & Herman, 1961;

Handal, 1965; Nachman, 1962).

The ability to recognize the taste of sodium when in a sodium-depleted state appears

to depend upon taste receptor cells in the oral cavity that contain ion channels selective

for sodium and lithium ions (Bernstein & Hennessy, 1987; McCutcheon, 1991; Roitman

& Bernstein, 1999). The sodium-selective ion channels, or epithelial sodium channels

(ENaCs), expressed by these cells can be blocked with the drug amiloride (see Brand et

al., 1985; DeSimone & Ferrell, 1985; Doolin & Gilbertson, 1993; Heck et al., 1984;

Schiffman et al., 1983). A second transduction pathway for sodium, the amiloride-

insensitive (AI) pathway is not sodium-selective, but instead appears to be activated by a

variety of cations including Na K+, and NH4+ (Brand et al., 1985; DeSimone & Ferrell,









1985; Kloub et al., 1997;Ye et al., 1994). Instead of ions passing directly through apical

ENaCs, activation of this pathway is thought to involve the diffusion of ions across tight

junctions and also perhaps through less selective ion channels in the apical membrane

(DeSimone & Ferrell, 1985; DeSimone et al., 2001; Elliot & Simon, 1990; Simon, 1992;

Ye et al., 1993; 1994). Activation of the AI pathway can be significantly reduced by

pairing the cation with an anion of large hydrated radius, like acetate or gluconate, thus

reducing the diffusion of ions across tight junctions (Elliot & Simon, 1990; Rehnberg et

al., 1993; Simon, 1992; Ye et al., 1993). Although recent data support the existence of an

amiloride-insensitive, nonselective cation channel in the apical membrane (DeSimone et

al., 2001; Gilbertson & Zhang, 1998), large anion salts apparently do not significantly

stimulate this pathway. Potassium gluconate, for example, is a very poor stimulus

(Stewart et al., 1996; Ye et al., 1994) and amiloride treatment virtually eliminates the CT

response to sodium acetate (NaAc) and sodium gluconate (NaGlu), especially at low to

mid-range concentrations (Elliot & Simon, 1990; Formaker & Hill, 1988; Ye et al.,

1993).

Previously, it has been shown that the amiloride-sensitive (AS) pathway is both

necessary and sufficient for normal Na+ detection in the rat, at least to the extent that

gluconate is capable of blocking AI transduction (Geran & Spector, 2000a; 2000b). It is

possible, however, that these near threshold concentrations, although detectable, were not

perceived as tasting "sodium-like" by the animals. For instance, humans often report that

low concentrations of NaCl taste "sweet" rather than "salty" (Bartoshuk et al., 1978).

Consequently, the effect on sodium recognition of reducing the contribution of the AI

pathway(s) with a large anion was tested. Other researchers have shown that rats will









ingest a variety of sodium salts when depleted, but to our knowledge all have used single

stimulus (Handal, 1965; Krieckhaus & Wolf, 1968; Morrison & Young, 1971), 24-h

intake (Fregly, 1958; Richter & Eckert, 1938), or 2-bottle sodium vs. nonsodium salt tests

(Nachman, 1962), making comparisons among taste-guided preferences for different

sodium salts impossible. Instead, we used a brief-access test so that preference for several

salts could be analyzed simultaneously and without potentially confounding postingestive

effects. Three sodium salts with different sized anions were used to limit AI transduction

(Elliot & Simon, 1990; Formaker & Hill, 1988; Kitada et al., 1998; Ye et al., 1993).

Sodium gluconate in particular was chosen based on its ineffectiveness in stimulating AI

transduction (Ye et al., 1991; 1993). This experiment marks the first time to our

knowledge that this salt has been presented to sodium-depleted rats. Two concentrations

of each salt were chosen on the basis of mean sodium detectability functions measured

previously (Geran & Spector, 2000a; 2000b). One concentration was well above

detection threshold in the presence of amiloride (0.3 M), and the other below threshold

when mixed with amiloride, but above threshold in the absence of amiloride (0.03 M).

Potassium chloride and NH4Cl were included in the stimulus array for the purpose of

comparison. If the AS pathway is sufficient as well as necessary for sodium recognition,

sodium-depleted animals should show similar amounts of licking to all sodium salts

regardless of anion. Furthermore, this increase in licking to sodium salts should be

abolished with the addition of amiloride.

Methods

Subjects

A total of 40 naive male Sprague-Dawley rats (Charles River Breeders, Wilmington,

MA) were used in this experiment. Subjects were tested in 2 groups of 20. All rats









weighed approximately 250-300g at the start of training and were placed on a 12:12

light:dark schedule with lights on at 6:00 AM. Humidity and temperature were kept

constant. Rats were individually housed in hanging wire mesh cages and given ad libitum

access to laboratory chow (PMI 5001 pellets. PMI Nutrition International, Brentwood,

MO) except where noted. Access to distilled water, however, was restricted while in the

home cage. Water bottles were removed -24 h prior to training and replaced after the last

training session 5 days later. Rats had access to water only during 30-40 min training

sessions on these 5 days. Body weights were closely monitored for excessive

dehydration. No rat dropped below 85% of its individual ad libitum weight during this

experiment. All procedures were approved by the Institutional Animal Care and Use

Committee (IACUC) at the University of Florida.

Apparatus

All training and testing occurred in a modified, automated taste testing apparatus

called a gustometer (Spector, Andrews-Labinski, & Letterio, 1990). This apparatus was

designed to deliver small volumes of fluid stimuli and record the number of licks for each

stimulus presented. A narrow slot in one wall of the chamber enabled the rat to access a

vertically-oriented sample spout. This spout rotated into position in front of the slot at

appropriate times throughout the session. Each taste stimulus was kept in a pressurized

syringe connected to the sample spout by way of Teflon tubing and a solenoid valve.

Each valve was opened for a preset amount of time by the computer such that each lick

delivered -5 [tl of fluid after the drinking shaft was initially loaded with the stimulus (see

Spector, Andrews-Labinski, & Letterio, 1990 for more details on stimulus delivery). A

very low current (< 50 nA) contact circuit was used to monitor number of licks taken.









Training Procedure

On the first day of training, a drop of water was placed on the tip of the spout and on

the inside wall of the chamber before each session to help in training the rat to drink from

the spout. During the first 2 days of training the sample spout remained motionless in

front of the access slot so that the rat had continuous access to distilled water for each 30-

min session. After all animals had been trained to lick from the spout, they received 3

days of additional spout training in which they received access to distilled water and 0.1

M sucrose in 5s trials. Sucrose was used to encourage licking. The rat was required to

lick the dry spout twice in 1 s to receive a stimulus presentation. When the trial was

finished, the spout rotated away from the access slot and over a funnel where it was

rinsed with distilled water and dried with pressurized air. This process took about 6s to

complete. The spout then rotated back in front of the slot. These sessions lasted 40 min.

Sodium Depletion

After training was complete, rats were assigned to one of 4 groups. These groups were

counterbalanced for body weight, number of trials initiated during the last 3 days of

training and mean number of total licks for these 3 days. Distilled water bottles were

placed on the home cages Friday following the last training session. On Monday morning

the rats were moved from their standard cages to metabolism cages equipped with

funnels to collect urine. They were also given a weighed amount of powdered chow at

this time instead of pellets. Rats in the 2 sodium-depleted groups received Harlan Teklad

90228 sodium-deficient (0.02% NaC1) chow (Harlan Teklad, Madison, WI). Rats in the 2

non-depleted groups received the same chow mixed with 1.0% NaC1. Twenty-four hours

prior to testing, each rat in the sodium-depleted groups received the first of 2 equal

volumes of furosemide (total dose = 30mg/kg BW, s.c.). The second furosemide injection









was given 2 hours later. Rats in the non-depleted groups received injections of isotonic

saline (s.c.) using the same injection schedule and volume as rats given furosemide.

Subjects had free access to powdered chow (with or without sodium) and distilled water

during the sodium depletion phase of the experiment. Urine was collected in 100 mL

flasks for 24 h immediately following the first injection.

Brief-Access Testing

Testing took place in the gustometers 24 h after each animal's first furosemide

injection. Animals were given brief access to 11 stimuli (e.g. distilled water and 0.03 &

0.3 M concentrations ofNaC1, KC1, NH4C1, sodium acetate (NaAc), and sodium

gluconate (NaGlu)) over a 40-min period. Stimuli were presented in randomized blocks

except that the first trial of each session was always 0.3 M NaCl to encourage continued

sampling. All salt solutions were made fresh using reagent grade chemicals (all salts from

Fisher Scientific, Orlando, FL except NaGlu from Sigma Chemical Co., St. Louis, MO)

and distilled water. One liter of 100 [LM amiloride (Sigma Chemical Co.) stock solution

was made the evening prior to testing and wrapped in aluminum foil and left on a stir pad

overnight in a dark room. Two of the 4 gustometers contained the aforementioned salts

dissolved in distilled water, while the remaining 2 gustometers contained salts dissolved

in 100 [LM amiloride and 100 [LM amiloride in place of the distilled water stimulus. Half

of the rats from each depletion group were tested with amiloride as the solvent.

At the time of testing, the remaining chow was removed from the home cage and

weighed to determine how much was ingested in the previous 24 h after attempting to

account for spillage. Distilled water intake and urine output for each rat were measured to









the nearest mL. A 2.0 mL sample of urine was collected for each rat and frozen in labeled

plastic centrifuge tubes for later analysis.

Urine Analysis

Urine was analyzed using a flame photometer to determine sodium content for both

sodium-depleted and non-depleted rats. Urine from sodium-depleted rats was diluted

with distilled water (4 parts water: 1 part urine) prior to analysis so that the sodium

concentration of each sample would fall within the range testable by the device. The

values for these animals were then multiplied by a factor of 5.

Data Analysis

Lick data were recorded and analyzed for the entire 5s of each trial, but the main

parameter of interest was the number of licks to each stimulus during the last 3s of each

5s trial (i.e., the avoidance period). This was done to minimize the number of sampling

licks included in the analysis. The local lick rate for rats is approximately 7 Hz (Corbit &

Luschei, 1969; Halpem, 1977), making 35 licks the ceiling in a 5s trial and 21 licks the

highest performance attainable in a 3s period. The number of licks for each stimulus was

then averaged for each group and compared using analyses of variance (ANOVAs) and t-

tests (paired, independent and one-sample tests). The statistical rejection criterion was set

at .05 for all analyses. P-values were adjusted using the Bonferroni method when a large

number oft-tests were performed on the same data set. Lick data from an animal were

included in the analyses only if the animal sampled all 11 stimuli in the test. Only 2 rats

in the non-depleted/amiloride group and 1 rat in the non-depleted/water group passed this

criterion. This sample size was too low for meaningful analysis, so all statistical tests of

stimulus licking were performed on sodium-depleted rats only. Data from each rat were

used, however, to analyze the number of trials initiated during testing and degree of









sodium depletion for each of the 4 groups. Sodium balance was determined by

subtracting urinary sodium output from sodium ingested.

Results

Brief-Access Testing

The overall pattern of responsiveness for the sodium-depleted groups was the same

regardless of whether the 3s avoidance period or the entire 5s period was analyzed. One-

way ANOVAs of mean licks during the last 3s of each trial revealed a main effect of

stimulus for both the amiloride (F(10, 90) = 35.4, p < .001) and distilled water (F(10,90)

= 2.7, p < .008) groups. Intake of distilled water and amiloride did not differ significantly

between groups (p > .09, 10.3 + 4 vs. 13.6 + 4 licks respectively).

For rats in the distilled water condition, paired t-tests indicated that they licked

significantly more to each of the 6 sodium stimuli (i.e. 3 salts, 2 concentrations) than to

water (p < .005 for each t-test. Bonferroni adjusted p < .05. Figure 2-1). Furthermore, a 2-

way ANOVA anionn x concentration) of responses to sodium salts revealed a significant

main effect for concentration only (F(1,9) = 26.9, p < .002). One-way ANOVAs

indicated no differences in performance across sodium salts at the 0.03 M concentration

and a slight effect at the 0.3 M concentration (F(2,18) = 3.99, p < .04), with the greatest

number of licks recorded for the salt with the intermediate-sized anion, sodium acetate

(NaAc). Paired t-tests indicated significantly less licking to the NH4C1 solutions than to

water (p < .008 for both) but this significance disappeared with a Bonferroni test

(adjusted ps > .07). The KC1 stimuli were not different from water (p > .34 for both,

unadjusted).

In the amiloride condition, .03 M NaGlu and .3 M NH4Cl were licked less than

amiloride alone (p < .05 for both). Again, the statistical significance of this difference









disappeared with a more conservative test (adjusted ps > .4). All other stimuli generated

licking comparable to that induced by amiloride alone (all ps > .14, unadjusted). A two-

way ANOVA anionn x concentration) for all sodium salts indicated that all 6 sodium

stimuli were licked to similar degrees in the presence of amiloride (ps > .14).

Number of Trials Initiated

Not surprisingly, the animals in the sodium-depleted groups initiated a greater number

of trials than rats in the non-depleted groups (p < .001). However, sodium-depleted rats in

the amiloride condition took far fewer trials than sodium-depleted rats in the distilled

water condition (p < .001, Figure 2-2). This suggests that although more motivated to lick

than non-depleted rats, the presence of amiloride reduced the number of trials initiated by

sodium-depleted rats.

Sodium Balance

All rats in the sodium-depleted groups were in negative sodium balance while rats in

the non-depleted groups ingested more sodium than they excreted (p < .004 for each of

the 4 groups. Figure 2-3). There was no difference in sodium balance between rats in the

distilled water condition and those in the amiloride condition for either depletion group

(both ps > .20).

Discussion

It is clear from the data that the sodium-specific appetite exhibited by rats in the

distilled water condition was not evident in the amiloride condition. When amiloride was

added to the taste stimuli, intake of the sodium salts was not greater than intake of either

the nonsodium salts NH4C1 and KC1 or of 100 [tM amiloride alone, strongly suggesting

that the animals were unable to recognize the sodium salts and to ingest them

preferentially. These findings support previous reports of impaired sodium appetite with









amiloride (Bernstein & Hennessy, 1987; McCutcheon, 1991; Roitman & Bernstein,

1999), and extend them by showing that the taste-guided specificity of the sodium

appetite is completely abolished by amiloride treatment in a brief-access test.

This effect of amiloride on the specificity of the sodium appetite differentiates

amiloride treatment from CT transaction, which has been shown to impair but not always

eradicate sodium-specific appetite (see Breslin et al., 1993; Markison et al., 1995). This

difference is likely due to the fact that the greater superficial petrosal (GSP) branch of the

facial nerve also contains amiloride-sensitive fibers important for the maintenance of

certain taste-guided tasks involving sodium (Roitman & Bernstein, 1999; Sollars & Hill,

1998). Our results also differ from those of CT transaction in certain other regards. For

instance, prior studies reported increased licking to low KC1 concentrations with

transaction (Breslin et al., 1993) that we did not observe with amiloride treatment. This is

most likely due to the fact that CT transaction reduces the perceived intensity of KC1 in

rats while amiloride does not (Geran et al., 1999). Procedural differences between this

experiment and the brief-access studies with CT transaction should also be taken into

account when comparing performance between experiments. These include a shorter

sample time (5s vs. 10s), a greater sodium : nonsodium salt ratio in the stimulus array

(3:2 vs. 1:4) and a slightly lower (0.03 vs 0.05 M) concentration of KC1 in the current

experiment.

Also of importance is the fact that sodium-depleted rats in the distilled water condition

increased intake of all sodium salts (NaC1, sodium acetate and sodium gluconate) to a

similar degree. Although previous studies have reported licking to a variety of sodium

salts following sodium depletion in rats (see Fregly, 1958; Handal, 1965; Nachman,









1962) the present study is the first sodium appetite experiment to our knowledge to

include a variety of sodium salts in a single brief-access test. Our results strongly suggest

that eliminating, or at least severely reducing, the contribution of the AI transduction

pathway by changing the size of the anion does not compromise sodium recognition.

Thus, the AS sodium transduction pathway appears to be both necessary and sufficient

for sodium recognition in the rat just as it is for normal sodium detection. Evidence of an

apically-located AI cation transduction pathway in the oral cavity of the rat (DeSimone et

al., 2001; Gilbertson & Zhang, 1998) may call this sufficiency into question if this

nonselective apical pathway is substantially activated by large anion sodium salts like

NaGlu. However, current electrophysiology indicates that only a negligible portion of the

CT response to NaAc and NaGlu remains with amiloride treatment at the stimulus

concentrations used (Elliot & Simon, 1990; Formaker & Hill, 1988; Ye et al., 1993).

The functional role of the AI transduction pathway to taste function is yet to be

understood, but might involve the detection and/or recognition of nonsodium salts.

Support for this hypothesis comes from evidence that the discrimination of 2 nonsodium

salts, NH4C1 and KC1, is severely compromised by combined transaction of the CT and

GSP nerves but unaffected by amiloride treatment (Geran et al., 2002). This suggests that

recognition and discrimination of these salts depends on amiloride-insensitive, salt-

responsive units in these nerves. As noted above, normal KC1 detectability is also

compromised by CT transaction but not significantly affected by amiloride treatment,

suggesting that an AI pathway with input carried by the facial nerve is required for this

task (Geran et al., 1999). Discrimination among sodium salts might also depend upon AI

transduction. The CT response to NaCl is of greater magnitude than the response to NaAc









or NaGlu due to a larger amiloride-insensitive response (Elliot & Simon, 1990; Formaker

& Hill, 1988; Ye et al., 1993). This comparative increase in AI response does not appear

to confer any particular salience to NaCl over other sodium salts at near-threshold levels

(Geran & Spector, 2000b), but it could affect the taste quality or intensity of sodium salts

in rodents at superthreshold levels. Humans have rated NaCl as more intense than NaGlu

at equimolar concentrations (Ossebaard & Smith, 1995). Recordings from the GSP,

however, indicate roughly equivalent amiloride suppression to both NaCl and NaAc at

concentrations of .1 M and higher, but at .05 M the response to NaCl is more suppressed

(-75%) than the response to NaAc (-50%, Sollars & Hill, 1998). This suggests that AI

sodium transduction in the GSP may be unaffected or even enhanced by larger anions.

Alternatively, a third pathway might exist that is independent of sodium but responsive to

acetate. Responses of the GSP to NaGlu have not been reported. Clearly, more research

on the salt responsiveness of this nerve is needed.

Sodium-depleted rats in the distilled water condition increased their intake of both

0.03 and 0.3 M NaAc and NaGlu relative to water, suggesting that the size of the anion

did not affect their ability to perceive the taste of sodium. In the amiloride condition,

however, rats did not lick NaAc or NaGlu more than amiloride alone at either stimulus

concentration, suggesting that sodium recognition was compromised. This is particularly

noteworthy at the 0.3 M NaGlu concentration, due to the fact that rats were shown to

detect this stimulus, albeit poorly, in the presence of amiloride (Geran & Spector, 2000b).

Together, these results suggest that rather than responding to sodium activation of the AS

transduction pathway, the behavioral detectability of high NaGlu concentrations derives

from some cue related to the gluconate anion or perhaps activation of the AI pathway due









to leakage of sodium through tight junctions, or AI apical channels, as a result of the high

Na+ concentration gradient.

Sodium-depleted rats initiated a significantly greater number of trials than non-

depleted rats regardless of whether amiloride was added to the stimuli. This indicates an

increase in appetitive behavior (see Craig, 1918; Denton, 1982), even in the absence of a

sodium taste cue. When the taste of sodium was present, sodium-depleted rats showed a

further increase in the number of trials initiated. Thus, under these conditions, the taste of

sodium in the absence of need does not produce an increase in appetitive behavior, while

need in the absence of the appropriate taste cue does, although not to the same degree as

need and gustatory cue combined.

In summary, these data support Bernstein & Hennessy's (1987) conclusion that the AS

sodium transduction pathway is necessary for sodium recognition in the rat, and

furthermore strongly suggest that this pathway is also sufficient (see Elliot & Simon,

1990; Formaker & Hill, 1988; Ye et al., 1993). We have also extended previous studies

of the effects of amiloride on salt appetite to show that the sodium specificity of the

appetite is completely abolished when the AS pathway is blocked. In addition, these

findings suggest that although NaGlu concentrations higher than 0.1 M are detectable in

the presence of amiloride they appear to lack the characteristic taste quality associated

with sodium.














20

CO
-15
_J
z
cn 10
_0

S5
w

0


AMILORIDE


.03 .3.03 .3.03 .3.03.3.03.3 .03 .3.03 .3.03 .3.03.3.03.3
NaCI NaAc NaG KCI NH4CI NaCI NaAc NaG KCI NH4CI


Figure 2-1. Brief-access licking to each stimulus by sodium-depleted rats. Mean (+ SE)
number of licks taken by sodium-depleted rats in the last 3s of each 5s bout. Stimuli were
dissolved in either distilled water (left) or 100 ptM amiloride (right). Horizontal lines
indicate mean (+ SE) number of licks to either distilled water or amiloride alone. All
sodium salts (chloride, acetate and gluconate) were preferred over water (paired t-tests, p
< .005 for each. p < .05 Bonferroni adjusted), while none of the salts, sodium or
nonsodium (potassium chloride and ammonium chloride), were preferred over amiloride.
Responses to water and amiloride alone were not significantly different.


DH20














200
Na+-Depleted Non-Depleted
180
160
140
0 120
100
I-8
80
60
40 -
20 -
20/

DH20 AMIL DH20 AMIL
GROUP


Figure 2-2. Mean (+ SE) number of trials initiated by each group of rats. Non-depleted
rats took the fewest trials regardless of whether amiloride was present. Sodium-depleted
rats, however, initiated considerably fewer trials in the presence of amiloride (p < .001).
















E 2
E
- 1
I
0

I-

S-2
z -2


- Na+-Depleted Non-Depleted

T


-^

//
-/_

-^
:_ ^ ^ _//_


DH20


AMIL DH20 AMIL
GROUP


Figure 2-3. Mean (+ SE) sodium balance for each group of animals measured in mmol.
All rats in the sodium-depleted groups were in negative sodium balance prior to brief-
access testing while rats in the non-depleted groups consumed more sodium than they
had excreted (p < .004 for each group in a one-sample t-test).














CHAPTER 3
GLOSSOPHARYNGEAL NERVE TRANSACTION DOES NOT IMPAIR
POTASSIUM CHLORIDE VS. AMMONIUM CHLORIDE OR SODIUM CHLORIDE
VS. AMMONIUM CHLORIDE DISCRIMINATION

Background

Ammonium and potassium chloride have been shown to taste very similar to rats in

generalization tasks (Erickson, 1963; Hill et al., 1990; Morrison, 1967). These results

have been used to support both similarities in NH4+ and K+ transduction at the receptor

level and similarities in the neural response pattern at the level of the NTS (DeSimone et

al., 2001; Erickson, 1963). Recently, it was reported that rats placed on a KC1 vs. NH4C1

discrimination task with overlapping concentrations were able to consistently perform

well (-90% overall performance) after a typical period of discrimination training (Geran

et al., 2002). Furthermore, dissolving the salt stimuli in the epithelial sodium channel

(ENaC) blocker amiloride did not impair performance, but transecting the gustatory

branches of the facial nerve (i.e., the chorda tympani (CT) and greater superficial petrosal

(GSP) nerves) reduced performance to chance levels for each animal (Geran et al., 2002).

Rats trained and tested on a NaCl vs. NH4C1 discrimination task also exhibited chance

levels of performance following nerve transaction even though the mean presurgical

discrimination performance for this group was approximately 95% (Geran et al., 2002).

Amiloride significantly impaired performance on this task, although not to the extent

reported previously for NaCl vs. KC1 discrimination (see Kopka et al., 2000; Spector et

al., 1996). Sodium chloride vs. NH4C1 discrimination was chosen as the comparison for

KC1 vs. NH4C1 discrimination performance because rats do not generalize between NaCl









and NH4C1 in a conditioned taste aversion task (Hill et al., 1990). This suggests that NaCI

and NH4C1 are easily discriminated by the rats, while KC1 and NH4C1 are more difficult,

if not impossible, for the animals to distinguish. The results of the prior discrimination

experiment (Geran et al., 2002) confirmed these predictions and suggested that the

amiloride-insensitive fibers of the facial nerve were necessary for discrimination

involving NH4C1, but the sufficiency of this input was not ascertained.

Functional sufficiency of facial nerve input was tested in the current experiment by

transecting the glossopharyngeal (GL) nerve. This is not an absolute test of sufficiency as

the superior laryngeal branch (SLN) of the vagus nerve is still intact. However, this nerve

is thought to be more important for airway protection than taste quality perception (see

Dickman & Smith, 1988; St. John & Spector, 1998; Smith & Hanamori, 1991). The GL

innervates about 60% of the taste buds in the oral cavity of the rat and contains fibers

narrowly-tuned for salts and acids as well as fibers that are highly responsive to

compounds described as "bitter" by human subjects (Frank, 1991). Humans have also

reported that NH4C1 and KC1 contain both "bitter" and "salty" components (van der

Klaauw & Smith, 1995), making it possible that 1 or more types of narrowly-tuned fibers

in the GL of the rat are activated by NH4C1 and/or KC1. It is also possible that

discrimination could depend upon GL input due to the large number of taste buds

innervated by this nerve. To date, GL transaction has not produced substantial

decrements in performance on discrimination or recognition tasks involving taste quality.

These tasks include sucrose vs. maltose, citric acid vs. quinine, quinine vs. KC1 and NaCl

vs. KC1 (see St. John, Markison, Guagliardo et al., 1997; St. John & Spector, 1997; St.

John & Spector, 1998; Spector & Grill, 1992; Spector et al., 1997). Instead, performance









on these tasks has been significantly affected by transaction of one or both of the

gustatory branches of the facial nerve (see St. John & Spector, 1998), in spite of the fact

that the CT and GSP together innervate only about half the number of taste buds

innervated by the GL (Miller, 1995). The remaining 5-10% of taste buds are innervated

by the SLN (Travers & Nicklas, 1990).

Methods

Subjects

A total of 32 adult male Sprague-Dawley rats (Charles River Breeders, Wilmington,

MA) were used in this experiment. These rats were tested in 2 groups of 16. Each animal

weighed between 250 and 300g at the start of training and was given ad libitum access to

PMI 5001 pellets (PMI Nutrition International, Brentwood, MO) at all times while in the

home cage. Access to distilled water was restricted to encourage performance during

testing. Water bottles were removed from home cages approximately 24 hours prior to

training (or testing) every Monday and replaced every Friday following the last session.

Rats were able to gain access to water Monday through Friday by pressing the correct

lever while in the apparatus. During the week, animals were closely monitored for

excessive dehydration. Rats received supplemental water if they appeared dehydrated or

if body weight decreased to 85% of the ad libitum weight calculated each week. All

procedures were approved by the Institutional Animal Care and Use Committee (IACUC)

at the University of Florida.

Apparatus and Trial Structure

Animals received training and testing in an operant chamber modified for taste

research. This gustometer apparatus contains one spout for stimulus presentation and one

for delivery of the reinforcer. It also contains two levers, one on either side of a spout-









access slot (see Spector, Andrews-Labinski, & Letterio, 1990 for further details). Each

animal was allowed to complete as many trials as possible during a 40-min session. Each

trial began when the rat made contact with the sample spout twice within 250 ms. This

contact completed a low amplitude (< 50 nA) circuit and caused a solenoid valve to open

with each subsequent lick so that the stimulus was delivered to the rat's tongue. Each

sample phase lasted 5 licks or 3 s whichever came first. This was followed by a 5-s

decision phase, during which the stimulus spout rotated out of reach. If the rat pressed the

correct lever (i.e., the lever associated with the stimulus during training), it received

access to distilled water (20 licks or 10s) via the reinforcement spout. If the rat pressed

the incorrect lever or failed to press either lever, the trial ended and the animal received a

20-s time out. An intertrial interval of 10s followed both the reinforcement and time-out

phases. At this time, the lights in the chamber were extinguished and the sample spout

was rotated over a funnel, rinsed with distilled water and blown dry with pressurized air.

Sessions were controlled automatically by a computer and white noise was present

throughout each session.

Training Procedure

Rats were counterbalanced for lever and stimulus. Half of the rats (i.e., 16) were

trained to discriminate NaCl from NH4C1 and half were trained to discriminate KC1 from

NH4C1. Training began with shaping the rats to press one lever following presentation of

a single stimulus (0.2 M NH4C1, 0.2 M KC1 or 0.2 M NaC1). Stimuli were made each

morning using reagent grade chemicals (Fisher Scientific, Orlando, FL). Once the rat had

performed the target behavior for one session without aid, the animal was shaped to press

the other lever in response to the other stimulus. This process took approximately 2

weeks. The animals were then switched to the alternation phase during which one









stimulus was presented on each trial until the animal pressed the correct lever on a fixed

number of trials. Once the criterion was reached, the other stimulus was presented. It was

not necessary for the correct trials to be consecutive. Each rat moved to the next

alternation criterion when it performed at 75% or better for the day. The alternation

criterion decreased over five days from 8 correct presses to 2. Alternation was followed

by discrimination training, during which stimuli were presented in randomized blocks.

After two days of 75% performance or better, the limited hold was reduced and the time

out increased. These parameters were systematically reduced when the performance of

each rat reached the 75% criterion until the last phase of training in which 3

concentrations of each salt were added to the stimulus array (0.4, 0.1, and 0.05 M). Rats

were kept on this phase of training until weekly performance was unchanged for 10

sessions (2 weeks). In the current experiment, this last phase of training required an

average of 19 sessions. Rats on the KC1 discrimination were moved from one phase of

training to another simultaneously with rats on the NaCl task. See Table 3-1 or St. John,

Markison, Guagliardo et al., 1997 for more detail concerning the training procedure.

Presurgical Discrimination and Amiloride Testing

After training, rats were tested for five days on the 8-stimulus discrimination array (4

concentrations of 2 salts). The following week all stimuli were dissolved in 100 [LM

amiloride. Amiloride (Sigma Chemical, St. Louis, MO) was made each afternoon prior to

testing in a flask wrapped with aluminum foil to minimize reactions with light and

allowed to spin overnight. The salt solutions were then made the next morning using

amiloride as the solvent instead of distilled water. Amiloride was also used in place of

distilled water as the reinforcer during this phase of testing.









Surgery

After amiloride testing, the rats in each discrimination group were divided into 2

groups counterbalanced for body weight, mean number of trials initiated during

discrimination testing and overall proportion correct. Half of the animals in each

discrimination (n = 8) received bilateral GL transaction (GLX) and the other half

received sham transaction (SHAM). All rats were anesthetized with an intramuscular

injection of ketamine hydrochloride (125 mg/kg body weight) mixed with xylazine

hydrochloride (5 mg/kg). An incision was made down the midline of the ventral neck and

the musculature and salivary glands retracted until the GL could be blunt-dissected from

the hypoglossal and vagus nerves with glass rods. The GL nerve was stretched with

curved glass rods and cut with microscissors where it met the hypoglossal and vagus

nerves such that approximately 5-10 mm of the nerve was removed. Rats in the sham-

transected group received only midline incision followed by retraction of the musculature

and salivary glands until the GL was exposed. A small amount of sterile saline was

introduced into the wound of each animal before it was closed with nylon sutures.

Animals received subcutaneous injections of penicillin (30,000 units Flocillin) and

analgesic (2 mg/kg Ketorolac) immediately after surgery and for the next 3 days. Animals

were allowed 9 (SHAM) or 10 (GLX) days to recover from surgery. One rat in each of

the two GLX groups (KC1 vs. NH4Cl and NaCl vs. NH4C1) died the night of surgery. It is

suspected that the vagal nerve of these animals was accidentally damaged during surgery.

Postsurgical Testing and the Water Control Test

After the recovery period, animals were tested on the original discrimination for 5

sessions followed by amiloride testing for an additional 5 sessions. These tests were

performed exactly as they were prior to surgery (see Table 3-2 for experiment schedule).









Rats in the second testing group (n = 15) received a 1-day water control test upon the

completion of amiloride testing. Each fluid reservoir was filled with distilled water with

half of these assigned to the left lever and half to the right lever. Thus, any performance

significantly better than chance could be attributed to an extraneous cue, such as noise or

temperature associated with a particular reservoir.

Histology

Upon completion of the water control test, rats were deeply anesthetized with sodium

pentobarbital and transcardially perfused with physiological saline and 10% buffered

formalin. The oral tissue of each animal was removed and stored in formalin until it could

be analyzed. At this time, the tongue was allowed to soak in distilled water followed by

dissection of the circumvallate papilla from both the anterior tongue and the underlying

connective tissue. The papilla was embedded in paraffin. It was then sliced into 10 rtm

sections using a microtome, placed on slides, stained using hematoxylin and eosin and

coverslipped. Taste buds were counted under a light microscope by an observer blind to

surgical group. The presence of a taste pore or the characteristic fusiform cells in the

absence of a taste pore was taken as indication of an intact taste bud. Previous research

has shown that taste buds degenerate in the absence of afferent innervation, allowing the

use of this assay in testing the completeness of nerve transactions (Cheal & Oakley,

1977; Hard af Segerstad et al., 1989; St. John et al., 1995).

Data Analysis

Overall discrimination performance for each rat was based on the number of trials

with a correct response divided by the total number of trials with a press for each phase

of the experiment collapsed across salt and concentration. Two-way analyses of variance

(ANOVAs) were also performed to determine the effects of condition and concentration









on performance to each salt. In cases where the ANOVA revealed a main effect of

condition, paired t-tests were performed to determine concentrations for which

performance was significantly altered. Paired t-tests were also used to compare overall

proportion correct between conditions and independent t-tests were performed in some

cases to determine whether between-group differences were significant. Finally, the

normal approximation of the binomial distribution was applied to the results of the water

control test to determine whether performance was significantly different from chance

(Brown & Hollander, 1977). A more conservative Bonferroni procedure was applied to

these results in order to adjust for the number of t-tests performed. Statistical significance

was set at p < 0.05 for all tests.

Results

Presurgical Discrimination Testing

Overall mean presurgical discrimination performance was greater than 90% regardless

of stimuli. Furthermore, the performance of rats in the KC1 vs. NH4C1 discrimination

group did not differ from that of rats in the NaCl vs. NH4C1 group (p > .08). When the

salts were dissolved in 100 [LM amiloride, KC1 vs. NH4C1 discrimination remained

unchanged, while NaCl vs. NH4C1 discrimination decreased significantly (p > .62 and p <

.001 respectively. Figures 3-1 and 3-2). Two-way ANOVAs (condition x concentration)

for each salt in the latter condition indicated an effect of condition for both NaCl and

NH4C1 (F(1,15) > 47, p < .001 for both). Furthermore, paired t-tests indicated that

performance on each concentration of each salt (NaCl and NH4C1) declined significantly

with amiloride treatment (p < .004 for each. Figure 3-3).









Postsurgical Testing and Histology

Mean KC1 vs. NH4C1 performance was significantly lower (p < .03, paired t-test) for

GLX rats after surgery than before surgery (post-GLX mean = 88% + 2 vs. pre-GLX

mean = 92% + 2. Figure 3-4.). Performance of sham-transected rats also dropped with

surgery but was not significant. Both presurgical and postsurgical performance values for

the GLX group, however, failed to differ from the values obtained for rats in the SHAM

group for the same condition (p > .88 for both). Two-way (condition x concentration)

ANOVAs for the pre- vs. post-GLX performance of each salt indicated a condition effect

for NH4C1, but not KC1 (F(1,6) = 6.5, p < .05 and F(1,6) = 5.4, p > .05, respectively). A

post-hoc analysis indicated a significant change in performance with transaction for 0.2

M NH4C1 only (p < .04).

Mean NaCl vs. NH4C1 performance declined significantly following sham-surgery but

not GL transaction (p < .03 vs. p > .07, respectively. Figure 3-5.). Two-way (condition x

concentration) ANOVAs, however, failed to indicate significant effects of condition or

the interaction of condition and concentration for either salt in sham-transected rats

(F(1,6) < 5.7, p > .05 for both salts and F(3, 18) < .5, p > 68 for both interactions. Figure

3-6). In addition, amiloride treatment did not significantly affect the postsurgical

performance of GLX rats trained to discriminate KC1 from NH4Cl (Figure 3-7). Nor did

GL transaction significantly impair NaCl vs. NH4Cl discrimination in the presence of

amiloride (Figures 3-8 & 3-9).

One rat performed significantly better than chance (i.e., 50%) on the water control test

(60% performance, p < .03 using a one-tailed test). After performing a Bonferroni

adjustment, this performance was no longer significant (z = 1.99, p > .34. Figure 3-10).









Rats in the GLX group had significantly fewer (p < .001) circumvallate taste buds than

rats in the SHAM group (0.14 + 0.14 and 422 + 19 respectively).

Discussion

Presurgical Discrimination Testing

This study replicated the results of our previous NH4C1 discrimination experiment

(Geran et al., 2002) in that rats on both the KC1 vs. NH4C1 and the NaCl vs. NH4C1 tasks

were clearly able to discriminate between the 2 salts. The prior study was also replicated

with regard to the effect of amiloride on these discrimination. Sodium chloride vs.

NH4C1 performance was compromised (11% decline in mean performance, p < .001) by

the addition of 100 [tM amiloride, while KC1 vs. NH4C1 was not significantly affected.

Additionally, this impairment was observed at each concentration of both NaCl and

NH4C1, and performance remained above chance at each concentration (Figure 3-3). This

concentration-response pattern is unlike the more one-sided impairment in NaCl vs. KCl

discrimination observed with amiloride and less substantial (Spector et al., 1996). In the

NaCl vs. KC1 task, performance on NaCl trials dropped below chance with amiloride

while performance on KC1 trials was not significantly altered. Thus, the rats pressed the

lever associated with KC1 on NaCl trials in the presence of amiloride, suggesting that the

taste qualities of KC1 and NaCl + amiloride are very similar. This conclusion is supported

by additional data from conditioned taste aversion tests and sham-intake experiments in

sodium-depleted rats (Hill et al., 1990; Roitman & Bernstein, 1999).

The prior NH4C1 discrimination experiment indicated similar concentration-response

patterns for both the KC1 vs. NH4C1 task and the NaCl vs. NH4C1 tasks with amiloride

(Geran et al., 2002). Although performance to each concentration was well above chance

for both the previous and current experiments, in general there was a slight nonsignificant









decline in performance at the 0.05 M NH4C1 concentration when tested against NaC1 +

amiloride. This decline mirrors that observed in the KC1 vs. NH4C1 discrimination group

both in the presence and absence of amiloride (Figure 3-3; Geran et al., 2002). The fact

that this decline did not result in near-chance performance and that the concentrations

chosen span almost an order of magnitude suggest that the animals were most likely

discriminating based on taste quality rather than intensity, although this has not been

tested directly. Detection threshold tasks have indicated that all 3 stimuli (NaCl dissolved

in amiloride, KC1 and NH4C1) are detected by rats at concentrations lower than those

tested in the discrimination task (Geran et al., 1999; Geran & Spector, 2000a; Chapter 4).

Although the concentrations tested are detectable to the rat, it is not clear how the stimuli

compare in terms of suprathreshold intensity.

Postsurgical Discrimination Testing

Potassium chloride vs. NH4C1 discrimination performance declined slightly but

significantly (p < .03) following GL transaction, while sham transaction did not

significantly affect performance (Figure 3-4). Interestingly, the decline in performance

with surgery was very similar for both GL and sham-transected rats (4% vs. 3%,

respectively) and 2-group t-tests failed to indicate significance for either pre- or

postsurgical performance. This suggests that perhaps the small number of animals in this

group contributed to significance and that the data might simply have been less variable

in the GLX group. In support of this interpretation, the difference in standard deviation

for the SHAM rats (.065) is almost twice that of the GLX rats (.035). Regardless of

statistical significance, a decline in performance from 92% to 88% is meager at best,

particularly when compared to a decline from 90% or better to approximately 50% with

combined transaction of the CT and GSP nerves (Geran et al., 2002).









The performance of rats in the NaCl vs. NH4C1 discrimination group was not affected

by GL transaction, but was affected by sham-transection (p < .03). Again, this appears to

be the result of a small number of subjects combined with low variability (Figure 3-5).

Together, these findings suggest that the GL is not necessary for rats to discriminate

between NH4C1 and either NaCl or KC1. In order to focus on the sufficiency of amiloride-

insensitive units within the facial nerve, the rats were tested again after surgery in the

presence of amiloride. As expected, this drug had no discernable effect on the

postsurgical KC1 vs. NH4C1 performance of GL-transected rats (Figure 3-7) and the

combination of GLX and amiloride did not impair NaCl vs. NH4C1 performance beyond

that seen with amiloride alone (Figure 3-8). Thus, every rat tested remained able to

perform at presurgical levels even after removal of essentially all gustatory input but that

of the amiloride-insensitive fibers of the facial nerve.

Every rat except 1 failed to perform better than chance on the water control test,

suggesting that these animals were under stimulus control. This rat scored only 60% on

the water control test but consistently performed at over 95%, even after GLX, on all

other phases of the experiment. This suggests that significance on the water control test

might have been due more to the number of t-tests performed than to the perception of an

extraneous cue. This interpretation is supported by the Bonferroni adjustment.

Potential Mechanisms Underlying KCI vs. NH4CI Discrimination

Although, the animals in this experiment were clearly able to discriminate between

KC1 and NH4C1, the basis for this discrimination is not obvious. One possibility is that a

population of taste receptor cells innervated by the facial nerve exists that contains ion

channels selectively permeable to either K+ or NH4+. Recent whole cell recordings have

shown that although the vast majority of TRCs responsive to 0.1 M KC1 or 0.1 M NH4C1









are responsive to both stimuli, a few cells respond to only one of these stimuli (Gilbertson

et al., 2001). Thus, although unlikely, there is some evidence that such a mechanism

might lead to KC1 vs. NH4C1 discrimination in the rat. It is also possible that a general

cation receptor, like that described by DeSimone and colleagues (2001), might exhibit

different kinetics for K+ and NH4+ that could result in discrimination. Nishijo and

Norgren (1997) reported that a group of neurons in the parabrachial nucleus (PBN) was

highly responsive to NH4C1, but not KC1. Most of the neurons in the PBN that responded

to NH4C1, however, also responded to KC1. It is not known whether the selectivity of

TRC or PBN responses to NH4C1 changes with stimulus concentration. Regardless of the

basis for KC1 vs. NH4C1 discrimination, it is clear that this discrimination exists and

depends upon input from the gustatory branches of the facial nerve. The implications of

this experiment in regard to taste coding are given further consideration in the final

chapter.

Summary

These data support the contention that the facial nerve is more important than the GL

nerve for tasks involving taste quality (see St. John & Spector, 1998) in spite of the fact

that the GL innervates roughly twice the number of taste buds (Miller, 1995). The

possibility remains that vagal, olfactory or somatosensory afferents may provide input

that could be useful in KC1 or NaCl vs. NH4C1 performance, making the input from the

amiloride-insensitive units of the facial nerve necessary but not entirely sufficient for one

or both of these tasks. At present, there is no evidence of such input affecting the

perception of salt taste.











Table 3-1. Training Schedule
Time Limited Stimulus
S Phase Out Holda Stimuli Presentation
Days (s) (s) Schedule
0.2 M NH4CI,
6 Shaping I none 180 0.2 M NH4CI constant
KCI or NaCI
3 Shaping IIb none 180 Same as constant
above
alternated
0.2 M NH4CI after "
8 Alternation c 10 15 and either
correct
KCI or NaCI correct
responses
2 Discrimination 10 10 Same as semi-random d
Training I above
Discrimination Same as
3 20 10 semi-random
Training II above
0.05, 0.1, 0.2,
.& 0.4 M
Discrimination & 0.4 M
19 20 5 NH4CI and semi-random
Training III either KCI or
NaCI

a Limited hold refers to the amount of time the rat is given to make a response.
b During Shaping II the rat is trained on the opposite stimulus and lever as in Shaping I.
SA stimulus is presented repeatedly until a certain number of correct responses are
made. This required number of responses, known as the alternation criterion, decreases
with each session. Eight for the first session, six for the second, four for the third, three
for the fourth, and two for the fifth. It is not necessary that the correct responses be
consecutive.
d Stimuli were presented in randomized blocks.











Table 3-2. Experiment Schedule.
Phase # of Sessions
(or Days)
Training 41
Presurgical Discrimination Testing 5
Presurgical Amiloride Testing 5
Surgery 2
Recovery 9-10
Postsurgical Discrimination Testing 5
Postsurgical Amiloride Testing 5
Water Control Test 1







47






- 1.0
I.J

o 0.9
0

0.8
0 >



_ 0.6
LU >
0 0.5 L >>.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 M
RATS


Figure 3-1. Presurgical KC1 vs. NH4C1 discrimination with and without amiloride.
Overall proportion correct for each individual with (white) and without (gray) 100 pM
amiloride and followed by the mean (M). Note that 50% correct performance (chance) is
used as the origin of the y-axis. Amiloride did not significantly impair discrimination
performance.
















H--
o 1.0
LU

0 0.9
z
0
S0.8
00)
o 0.7
a.
-J<
S0.6


0 0.5
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 M
RATS




Figure 3-2. Presurgical NaC1 vs. NH4C1 discrimination with and without amiloride.
Overall proportion correct for each individual with (white) and without (gray) 100 [tM
amiloride and followed by the mean (M). Note that 50% correct performance (chance) is
used as the origin of the y-axis. Amiloride significantly impaired mean performance (p <
.001).











KCI vs. NH4CI Group
1.0

0.8-_
-- NO AMILORIDE
0.6 -o- AMILORIDE

0.05 0.1 0.2 0.4 0.05 0.1 0.2 0.4
KCI (M) NH4CI (M)


NaCI vs. NH4CI Group
1.0
1 .0 -- ------* -*

0.8 -
0. *
*- *
0.6 *


0.05


0.2


0.4 0.05


NaCI (M)


0.1


0.2


0.4


NH4CI (M)


Figure 3-3. Mean presurgical performance by concentration. Performance is separated
according to task (KCl vs. NH4C1: top, NaCl vs. NH4C1: bottom) and salt (NaCl or KCI:
left, NH4Cl: right). Asterisks indicate significant differences between discrimination
testing in the presence (white) and absence (black) of 100 p[M amiloride (p < .004 for
each). Note that 50% performance (chance) is used as the origin of each y-axis.
















- 1.0 1.0


0 0.9 0.9-
z
P 0.8 0.8
0
a.
0 0.7 -0.7
_1
S0.6 -0.6
>
0 0.5 L1 0.5
1 2 3 4 5 6 7 8 M 1 2 3 4 5 6 7 M

SHAM RATS GLX RATS





Figure 3-4. Pre- vs. postsurgical performance on the KC1 vs. NH4C1 task. Rats that
underwent bilateral glossopharyngeal nerve transaction (GLX) are shown on the right and
sham-transected (SHAM) rats on the left. Each graph contains presurgical (gray) and
postsurgical (white) performance for each rat followed by the mean (M) for each surgical
group. Discrimination performance dropped slightly with transaction (p < .03), but
performance of GLX rats was not significantly different from sham-operated controls
either before or after surgery (p > .88 for both).




















00.9- 0.9-
1- 0.0 1.0-
0


0 0. 0.9-





S0.58 0.5
0 0.7 0.7"


p0.6 0.6-

uJ


0 0.5 0. 0.5 .
1 2 3 4 5 6 7 8 M 1 2 3 4 5 6 7 M

SHAM RATS GLX RATS


Figure 3-5. Pre- vs. postsurgical performance on the NaC1 vs. NH4C1 task. Rats that
underwent bilateral glossopharyngeal nerve transaction (GLX) are shown on the right and
sham-transected (SHAM) rats on the left. Each graph contains presurgical (gray) and
postsurgical (white) performance for each rat followed by the mean (M) for each surgical
group. Discrimination performance dropped slightly following sham transaction only (p <
.03).












KCI vs. NH4CI Group


0.05 0.1 0.2 0.40.05 0.1 0.2 0.4


NH4CI (M)


NaCI vs. NH4CI Group


0.4 0.05


0.2


0.4


NH4CI (M)


Figure 3-6. Mean presurgical vs. postsurgical performance by concentration. All data
points are from rats in the GL transaction group. Performance is separated according to
task (KC1 vs. NH4Cl: top, NaC1 vs. NH4Cl: bottom) and salt (NaCl or KCl: left, NH4C1:
right). Asterisks indicate significant differences between presurgical (white) and
postsurgical (black) discrimination (p < .04). Note that 50% performance (chance) is used
as the origin of each y-axis.


1.0

0.8

0.6


KCI (M)


1.0 -

0.8 -

0.6 -


0.05


0.2


NaCI (M)














1.0
LU

00.9




O 0.7
z
0





-0.6



0 0.5
1 2 3 4 5 6 7 M

GLX RATS


Figure 3-7. Effect of amiloride on postsurgical KCl vs. NH4C1 discrimination. All data
are from GL-transected (GLX) rats. Amiloride (white) did not further impair
discrimination performance following surgery (gray).














1.0
to


0.




o0.5 7
S0. -



w
0 0.5
1 2 3 4 5 6 7

GLX RATS


Figure 3-8. Effect of glossopharyngeal transaction on NaCl vs. NH4C1 performance in the
presence of amiloride. All data are from rats in the GL transaction (GLX) group.
Transection (white) did not further impair discrimination performance in the presence of
amiloride. Presurgical amiloride performance is shown in gray.













KCI vs. NH4CI Group




-- POST-GLX
-0- POST-GLX+AMIL


0.05 0.1 0.2 0.4 0.05 0.1 0.2 0.4
KCI (M) NH4CI (M)


NaCI vs. NH4CI Group




-e- POST-GLX AMIL
--- PRE-GLXAMIL


0.05


0.2


0.4 0.05


NaCI (M)


0.1


0.2


0.4


NH4CI (M)


Figure 3-9. Mean comparisons of postsurgical amiloride performance by concentration.
All data are from rats in the GL transaction group. Performance is separated according to
task (KCl vs. NH4C1: top, NaC1 vs. NH4C1: bottom) and salt (NaC1 or KCI: left, NH4C1:
right). In the top graphs, there is no difference between postsurgical performance (black)
and postsurgical performance in the presence of amiloride (white). In the bottom graphs,
there is no difference between presurgical (white) and postsurgical (black) performance
in the presence of amiloride. Note that 50% performance (chance) is used as the origin of
each y-axis.


1.0

0.8

0.6








1.0

0.8

0.6






56






1.0
I-
0
LU
a" 0.8
0
0

0 0.6
0


O 0.4






O
Q-
-J
< 0.2
LU

0.0 ,, ,, ,
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 M

RATS




Figure 3-10. Water control test. After testing was complete, all fluid reservoirs were filled
with water and assigned to either the left or right lever to assess each rat's ability to
respond to extraneous cues. One rat performed significantly better than chance using a
one-tailed test (p < .03). This significance disappeared when corrected for the number of
t-tests performed (p > .34). A dashed line represents 50% (chance) performance.














CHAPTER 4
AMILORIDE-INSENSITIVE UNITS OF THE CHORDA TYMPANI NERVE ARE
NECESSARY FOR NORMAL AMMONIUM CHLORIDE DETECTABILITY IN THE
RAT

Background

The 3 main gustatory nerves of the rat are highly responsive to NH4C1 at mid to high

concentrations (Frank 1991; Frank et al., 1983; Kitada et al, 1998; Nejad, 1986; Sollars &

Hill, 1998). For this reason, NH4C1 is often used as the standard when recording from

taste afferents, but very little is known about the perceptual characteristics of this salt.

Human subjects have reported that NH4C1 contains both "salty" and "bitter" components

(van der Klaauw & Smith, 1995). Because direct quality scaling and magnitude

estimation procedures cannot be used when working with animal subjects, analyses of the

perceived taste quality of NH4C1 in rodents have used conditioned shock avoidance and

conditioned taste aversion to measure generalization to salts and other stimuli. All 3 of

the published studies using rats have concluded that a conditioned avoidance response or

aversion to NH4C1 generalizes strongly to KC1, and weakly to NaCl (Erickson, 1963; Hill

et al., 1990; Morrison, 1967). These results have been interpreted as evidence that NH4C1

shares a common taste quality with KC1, but not NaC1. This supports the theory that

NH4+ and K+ activate at least one common transduction mechanism in the oral cavity

(DeSimone et al., 2001; Kloub et al., 1997), while a separate transduction pathway exists

for Na+ in the rat (Brand et al., 1985; DeSimone & Ferrell, 1985; Ninomiya & Funakoshi,

1988). Additional transduction sites might also exist that are more specific for either









NH4+ or K as rats have been shown to discriminate easily between KC1 and NH4C1 in

spite of evident similarities in taste quality (Geran et al., 2002).

There is some controversy as to whether the epithelial sodium channel (ENaC) blocker

amiloride inhibits only sodium and lithium responses in the CT and GSP nerves, or

whether its action is more general, reducing neural responding to potassium and

ammonium salts as well (see Lundy & Contreras, 1999; Minear et al., 1996). Researchers

have demonstrated maximal whole-nerve CT inhibition of up to 48% at .05 M NH4C1

with the addition of 100 [LM amiloride (Kloub et al., 1997. See also Lundy & Contreras,

1997; Lundy et al., 1997). Inhibition was less pronounced at higher NH4C1

concentrations, losing significance at approximately .3 M (Kloub et al., 1997). The

magnitude of this suppression (48%) is greater than that observed for the same

concentration of NaCl (30% at .05 M) with 100 [LM amiloride, but much less than the 70-

80% maximal inhibition observed at higher NaCl concentrations (Brand et al., 1985;

DeSimone & Ferrell, 1985).

Other researchers, meanwhile, failed to find any appreciable effect of amiloride on CT

responding to either KC1 or NH4C1 (Brand et al., 1985; Formaker & Hill, 1988; Hill &

Bour, 1985; Hill et al., 1982). These results are supported by the observation that oral

amiloride treatment did not inhibit either the potassium or ammonium response of CT

and GSP somata in the geniculate ganglion (Lundy & Contreras, 1999) or activity in the

subsets of NTS neurons most responsive to NH4C1 (Giza & Scott, 1991). At the

behavioral level, amiloride pretreatment did not affect generalization following

acquisition of a conditioned taste aversion to NH4C1 (Hill et al., 1990). Also, amiloride

failed to compromise either KC1 detection threshold or NH4C1 vs. KC1 discrimination in









rats (Geran et al., 1999; Geran et al., 2002). Together, these findings suggest that

although the effect of amiloride on nonsodium salt responsiveness is perhaps at times

statistically significant at the level of the peripheral nervous system, it does not appear to

reach significance at higher levels of the gustatory system or lead to behavioral

significance.

Taste sensitivity to NH4C1 was tested in the presence and absence of amiloride to

determine whether any inhibition in the periphery might affect the taste-guided

behavioral response of the animal. Amiloride suppression of the CT response to NH4C1

was greatest at low concentrations (Kloub et al., 1997) suggesting that impairment would

be most noticeable at the limits of detectability. Detection threshold was also chosen for

measurement because amiloride was previously shown to significantly raise the

thresholds of both NaCl and Na-gluconate using a similar procedure (Geran & Spector

2000a; 2000b). Earlier detection threshold experiments have also reported that NaCl and

KC1 sensitivities were compromised by bilateral transaction of the CT (CTX) suggesting

that input from this nerve is necessary for normal salt detectability (Geran et al, 1999;

Kopka & Spector, 2001; Slotnick et al., 1991; Spector, Schwartz & Grill, 1990). To

further test this hypothesis NH4C1 detection threshold was also measured after severing

the CT nerves.

Methods

General Methods

All procedures were approved by the Institutional Animal Care and Use Committee

(IACUC) at the University of Florida. Ten adult male Sprague-Dawley rats were placed

on a water-restriction schedule with water bottles removed about 24h prior to the start of

training or testing on Monday and replaced after the last session Friday. During the









week, animals worked for water access in the gustometer apparatus (see Spector,

Andrews-Labinski & Letterio, 1990). Subjects were cared for exactly like subjects in

Chapter 3, except where noted. The animals were trained to press one lever immediately

after a presentation of 0.2 M NH4C1 and the other lever immediately after a distilled

water presentation. As subjects grew more proficient, more concentrations were added to

the stimulus array and the parameters of the trial structure were altered such that the

decision phase was reduced and the time out increased. Subjects moved to the next phase

of alternation or training when they reached at least a 70% overall performance criterion.

See Table 4-1 for the training schedule. Each session lasted 40 min with the rats allowed

to complete as many trials as possible during this time. In the last phase of training,

stimuli were presented in randomized blocks of 10 consisting of 5 NH4C1 concentrations

and 5 distilled water presentations. This final training phase was exactly like the first

week of testing. As performance did not change over the last week of the final training

phase (Discrimination Training III), this period was retroactively defined as the first

week of presurgical threshold testing.

Testing

Detection threshold was tested over the course of 4 weeks using a total of 8 NH4C1

concentrations ranging from .00325 to .4 M. Each Monday rats received the standard

stimulus array (i.e., 0.4, .2, .1, .05 & .025 M NH4C1). A different array was presented

Tuesday through Friday containing at least 2 concentrations from the standard array and

several lower concentrations. At least one stimulus in the test array was replaced with a

lower concentration each week until performance approached a minimum asymptote.

Rats were given a second week of testing with the lowest concentration array to increase

the number of trials for these stimuli. See Table 4-2 for test stimulus presentation









schedule. During presurgical and postsurgical amiloride testing, 100 [LM amiloride was

used as the solvent for all NH4C1 stimuli and in place of water for both stimulus

presentations and reinforcement. A water control test was performed following the final

postsurgical test to assess whether the rats were capable of responding to extraneous cues

unrelated to the chemosensory properties of the NH4C1 stimuli. This was accomplished

by filling each reservoir with distilled water and assigning half of these to the left lever

and half to the right. See Table 4-3 for phases of the experiment.

Surgery

Rats were divided into 2 groups counterbalanced for weight, overall performance and

number of trials initiated. All rats were anesthetized with a mixture of ketamine (125

mg/kg body weight) and xylazine hydrochloride (5 mg/kg) injected intramuscularly. Five

of these rats received bilateral CT nerve cauterization. This was accomplished by

retracting the external ear canal to expose the tympanic membrane. The membrane, along

with the rim of the ear canal, the malleus and the CT nerve were then cauterized. This

procedure stimulates the production of cerumen, which keeps the CT from reinnervating

taste buds in the anterior tongue for at least 118 days (Kopka & Spector, 2001). The 5

rats in the sham-transected group had each ear retracted and the tympanic membrane

punctured with microforceps. All animals received subcuteneous injections of penicillin

(30,000 units Flocillin) and an analgesic (Ketorolac, 2 mg/kg body weight) immediately

following surgery and for the next 3 days. Rats were given 6-7 days to recover from

surgery before testing resumed.

Histology

After postsurgical testing, rats were deeply anesthetized with sodium pentobarbital

(i.p.) and rapidly perfused with saline and 10% buffered formalin. The tongue and palate









were removed and stored in formalin. Staining was accomplished by placing the anterior

portion of the tongue from the tip to the intermolar eminence in distilled water for 30 min

then dipping it in 0.5% methylene blue until dark and rinsing the tissue with water to

remove excess stain. The epithelium of the anterior tongue was then pressed between 2

glass slides and examined under a light microscope. The number of intact fungiform

papillae and taste pores were counted on each tongue. Taste pores appeared as small

round dots surrounded by blue circles under the microscope (Parks & Whitehead, 1998;

St. John et al., 1995). The presence of a discernable blue dot was counted as a pore for

the purposes of this experiment. Results using this method correspond with those from a

hematoxylin and eosin stain and also correlate with the degeneration and regeneration of

taste buds concomitant with denervation and reinnervation by the CT nerve (St John et

al., 1995). Histology was performed blind to the rat's surgical treatment.

Data Analysis

The percentage of correct responses on NH4Cl trials was adjusted for false alarm

probability (see Gescheider, 1997). This was accomplished using the following equation

for corrected hit rate, or P(Hit),:


P(HUt) P(Hit) P(FA)
) 1-P(FA)


where P(Hit) was the proportion of NH4Cl trials on which the rat pressed the correct

lever and P(FA) was the proportion of water trials on which the animal pressed the wrong

lever. Hit rates were corrected for each rat at each NH4C1 concentration. The following

logistic function was then used to fit curves to the corrected hit rate values for each

animal:










1+10 "b(xc



where a = maximum asymptote of performance, b = slope, x = NH4C1 logio concentration

and c = threshold. Threshold was defined as the NH4C1 concentration at one-half the

maximum asymptote of performance. Analyses of variance (ANOVAs), paired and

independent t-tests and the normal approximation of the binomial distribution (Brown &

Hollander, 1977) were used to assess statistical significance. Alpha was set at the

conventional .05 level.

Results

Presurgical Detection Threshold

Mean detection threshold for the first 20-session threshold was .012 M NH4C1 + .001.

With the addition of 100 [LM amiloride, mean threshold decreased significantly (p < .002,

Figure 4-1). In other words, the rats performed better at near threshold NH4C1

concentrations with amiloride than prior to treatment. When NH4C1 threshold was again

measured without amiloride, sensitivity improved further still (p < .006, mean threshold =

.009 M + .001. Figure 4-2). Two-way ANOVAs (phase x concentration) of corrected hit

rates indicated main effects for both phase: (F(1,9) > 17, p < .003) and concentration:

(F(7,63) > 468, p < .001) as well as for their interaction (F(7, 63) > 4, p < .002) when

each of the 2 presurgical tests without amiloride was compared to the amiloride test.

From the first presurgical measurement to the last, mean threshold decreased by .25 logo

units + .03 (p < .001; see Figure 4-3 for individual shifts in threshold).









Postsurgical Detection Threshold

Mean NH4Cl threshold increased following CT transaction by .54 logo units + .09 to

.04 M (p < .004. Figure 4-4). This value differs significantly from that of the sham-

operated rats (p < .001, see Figure 4-5). Threshold did not change significantly with

surgery for sham-operated rats compared with the last presurgical threshold assessment

(p > .08). Nor did amiloride alter NH4Cl threshold in either the CT-transected or sham-

transected rats after surgery (p > .29 for both. Figure 4-6).

Water Control Test and Histology

No rat scored better than chance when all stimuli were replaced with distilled water

(one-tailed t-tests using the normal approximation of the binomial distribution, p > .17 for

each. Figure 4-7). Histological analysis indicated no significant difference in the number

of fungiform papillae for the 2 surgical groups (CTX: 106 + 34 SHAM: 139 + 23. p >

.11). Rats in the CT-transected group, however, had far fewer taste pores on the anterior

tongue than rats in the sham-transected group (6 + 8 and 132 + 23 respectively, p < .001).

Discussion

It is clear from these data that the rats were responding to the chemical properties of

the stimuli rather than to extraneous cues such as temperature or sound, as none of the

rats scored better than chance on the water control test (Figure 4-7). The lowest mean

detection threshold for NH4Cl (.009 M) recorded in this experiment was slightly higher

than that found previously for NaCl (.005 to.006 M) using the same 2-lever operant

procedure (Geran & Spector, 2000a; 2000b; Kopka & Spector, 2001). In contrast, the

detection threshold for KC1 (.04 M) was considerably higher than for either of the other

chloride salts tested to date (Geran et al., 1999). The detectability functions for these 3

salts will be compared in greater detail in the final chapter.









Ammonium Chloride Detectability Depends Upon an Amiloride-Insensitive Route of
Transduction

A subset of electrophysiological studies has reported that amiloride significantly

suppressed CT responding to low and mid-range concentrations of NH4Cl (Kloub et al.,

1997; Lundy & Contreras., 1997; Lundy et al., 1997), suggesting that amiloride might

also compromise NH4C1 detection threshold. Instead, threshold decreased in the presence

of the ENaC blocker. Amiloride did not seem to be the cause of the increase in NH4C1

sensitivity, however, as threshold decreased further still when the rats were tested a

second time in the absence of amiloride (Figure 4-2). In addition, amiloride failed to

decrease postsurgical threshold for either surgery group. It is unclear whether this

decrease in threshold with repeated testing represents an aspect unique to NH4C1 such as

an upregulation of ammonium-sensitive receptor elements. Alternatively, the decrease in

threshold could be due to increased performance resulting from previous experience with

low concentrations of the stimulus.

Gilbertson and colleagues (1993), reported that vasopressin, a hormone involved in

fluid homeostasis, increases the inward Na+ current through amiloride-sensitive channels

in the taste receptor cells of hamsters. Although the amiloride-sensitive transduction

pathway of the hamster is different in significant ways from that of rats, it is conceivable

that such a mechanism could potentially have been activated by the restricted fluid access

experienced by the rats in the current experiment. One would expect such a mechanism to

have a greater effect on NaCl threshold than NH4C1 threshold, however, as this task is

more dependent on amiloride-sensitive units. Sodium chloride threshold increases

significantly in the presence of amiloride, and returns to the pre-amiloride threshold when

tested following amiloride treatment (Geran & Spector, 2000a; 2000b). This suggests that









the decrease in NH4C1 threshold observed during and after amiloride treatment was not

due to an upregulation of amiloride-sensitive receptors, although amiloride-insensitive

receptors permeable to NH4+ could have been affected. There is currently no evidence to

support such a hypothesis, however. In fact, NH4C1 vs. KC1 discrimination performance

was unchanged with amiloride treatment (Geran et al., 2002). Thus, the decrease in

NH4C1 threshold with each presurgical test is most likely the result of continued

experience with the task.

In addition to having little to no effect on the perceived intensity of NH4C1 at low

concentrations, amiloride also seems to be without effect on the taste quality of this salt.

For instance, amiloride treatment failed to compromise performance on a KC1 vs. NH4C1

task (Geran et al., 2002). In addition, taste aversions conditioned to NH4C1 resulted in the

same pattern of generalization regardless of whether the animals were treated with

amiloride at the time of conditioning (Hill et al., 1990). Thus, amiloride does not seem to

appreciably affect the perceived intensity of NH4C1 at low concentrations, or its

perceived taste quality at mid-range to high concentrations (i.e., .05 to .4 M. Geran et al.,

2002; Hill et al., 1990). It is unlikely that the drug compromises NH4C1 intensity at higher

concentrations as the electrophysiology indicates very little suppression, if any, at

concentrations above .3 M (Kloub et al., 1997; Lundy et al., 1997). Therefore, it is more

likely that amiloride affects the taste quality of NH4C1 at low concentrations, if it has any

effect at all on taste-guided performance to NH4C1. The current procedure only measures

detection, not recognition (see Gescheider, 1997). Thus, although the rats in this study

were able to discriminate NH4C1 dissolved in amiloride from amiloride alone, they may









have nevertheless been unable to recognize the stimulus as NH4C1 at near threshold

concentrations.

The Chorda Tympani Nerve is Necessary for Normal Ammonium Chloride
Detection

Bilateral transaction of the CT significantly increased NH4C1 detection threshold

(Figure 4-4), suggesting that the information carried by this nerve is necessary for normal

detection of this salt. It is also possible that the GSP and/or GL are necessary, although

insufficient, for normal NH4C1 detection. The result of this experiment, along with the

increases previously reported for both KC1 and NaCl threshold with CT transaction

(Geran et al., 1999; Kopka & Spector, 2001; Slotnick et al., 1991; Spector, Schwartz &

Grill, 1990), support the hypothesis that the CT nerve is important for normal salt

detection. This nerve also appears to be important for the recognition of salt stimuli as

CTX impairs performance on a NaCl vs. KC1 discrimination task (Kopka et al., 2000; St

John, Markison, Guagliardo et al., 1997; Spector & Grill, 1992). It is unclear whether the

role of the CT in salt recognition extends to NH4C1, as CTX does not alter unconditioned

licking to this salt in a 2-bottle preference test (Sollars & Bernstein, 1996). Of course,

factors other than taste quality, such as hedonic value or the postingestive consequences

of NH4C1 consumption, could have influenced performance on the 2-bottle test (see

Spector, 2000).

Like the CT, the GSP also appears to carry behaviorally-relevant information about

salt stimuli (Kopka et al., 2000; Roitman & Bernstein, 1999; St. John, Markison &

Spector, 1997). For instance, amiloride abolishes NaCl vs. KC1 discrimination while

CTX merely impairs it, suggesting that residual discrimination after CTX relies upon

amiloride-sensitive, sodium-selective receptors innervated by the GSP (see Kopka et al.,









2000; Roitman & Bernstein, 1999; St. John, Markison & Spector, 1997). The GL is not

thought to contain AS taste receptor cells (Doolin & Gilbertson, 1993; Kitada et al.,

1998). Thus, both the CT and GSP appear to carry information important for NaCI

recognition, and could perhaps also be necessary for normal NH4C1 recognition. It is not

clear at present whether the CT, GSP, or the combined input of the two is required for the

rat to accurately perceive the taste quality normally associated with NH4C1. If the GSP is

required for this task, one would expect CTX to impair NH4C1 vs. KC1 discrimination

performance significantly less than combined transaction of the CT and GSP.

In contrast, input from the GL does not appear to be necessary for the perception of

salt taste quality (Markison et al., 1995; Spector & Grill, 1992; Chapter 3). Instead, the

GL appears to be more important for the stimulation of unconditioned aversive gustatory

reflexes, like gaping upon contact with quinine, than for the perception of taste quality

(St. John & Spector, 1998; Travers et al., 1987). For instance, combined transaction of

the CT and GSP nerves drops performance on a KC1 vs. NH4C1 discrimination task to

chance while GL transaction is without effect (Geran et al., 2002; Chapter 3).

Conclusions

Like KC1 detection, normal NH4C1 detectability appears to depend upon amiloride-

insensitive receptors innervated by the CT nerve. It is unclear whether the same

population of receptors is responsible for both NH4C1 and KC1 detection. The fact that

KC1 and NH4C1 are easily discriminated by the rat (Geran et al., 2002) suggests that

differences in activation of the NTS exist for the 2 salts but have not yet been found.

The increase in detection threshold for NH4C1 with CT transaction lends further

support to the hypothesis that this nerve is necessary for the normal detection of salt

stimuli. This manipulation has also impaired both NaCl and KC1 detectability in previous






69


tests (Geran et al., 1999; Kopka & Spector, 2001; Slotnick et al., 1991; Spector, Schwartz

& Grill, 1990). Finally, the current experiment was also useful in that it provides a

detection threshold for NH4C1 that can be used to determine testing concentrations for

future psychophysical studies.













Table 4-1. Training Schedule
Time Limited Stimulus
S Phase Out Holda Stimuli Presentation
Days (s) (s) Schedule
0.2 M NH4CI
9 Shaping I none 180 0.2 M NH4 constant
or DH20
4 Shaping II b none 180 0.2 M NH4stant
or DH20
alternated
0.2 M NH4CI after "x"
7-14 Alternation 10 15 0 NH ar
and DH20 correct
responses
2 Discrimination 10 10 0.2 M NH4CI semi-random d
Training I and DH20
S Discrimination 0.2 M NH4CI semi-random
3 20 10 semi-random
Training II and DH20
0.025, 0.05,
Discrimination 0.1, 0.2, 0.4
8-15 e 20 5 semi-random
Training III e M NH4CI and
S__DH20

a Limited hold refers to the amount of time the rat is given to make a response.
b During Shaping II the rat is trained on the opposite stimulus and lever as in Shaping I.
SA stimulus is presented repeatedly until a certain number of correct responses are
made. This required number of responses, known as the alternation criterion, decreases
with each session. Eight for the first session, six for the second, four for the third, three
for the fourth, and two for the fifth. It is not necessary that the correct responses be
consecutive.
d Stimuli were presented in randomized blocks.
e The number of days shown here includes the first week of presurgical threshold testing.












Table 4-2. Test Stimulus Presentation Schedule


NH4CI Concentrations (M) in Test Array


Table 4-3. Experiment Schedule


Phase # of Sessions
(or Days)
Training 35
Presurgical Threshold Testing 1 20
Presurgical Amiloride Testing 20
Presurgical Threshold Testing 2 20
Surgery 2
Recovery 6-7
Postsurgical Threshold Testing 20
Postsurgical Amiloride Testing 20
Water Control Test 1


.0065


.00325


1.4 .2 .1 .05 .025 .013
Week 1 &
Monday:
Standard
Array
Week 2
(ue to Fri)
Week 3
(ue to Fri)
Week 4
(ue to Fri)













100


so


I 60


w 40 /

0 /
O 20- /




0.01 0.1
NH4CI CONCENTRATION (M)


Figure 4-1. Effect of amiloride on NH4C1 detection. Mean performance with and without
amiloride (100 ptM). Threshold decreased significantly (i.e. sensitivity increased) with
amiloride (p < .002). Amiloride performance is represented with open symbols.













100


LU
80


I 60
LJ

w 40

0
O 20 /




0.01 0.1

NH4CI CONCENTRATION (M)

Figure 4-2. NH4C1 threshold decreased again following amiloride treatment. Threshold
decreased significantly following amiloride testing (p < .006). Open symbols represent
amiloride performance.







74






1.0
CTX SHAM
0 0.8
oU
S0.6

0.4
Z ,

UJ WU
QC9 0.0

Eiy -0.2
S-0.4
1 4 5 6 9 M 2 3 7 8 10 M
RAT NUMBER

Figure 4-3. Individual shifts in presurgical threshold. Shifts for each animal followed by
the mean for each surgical group. All values are before surgery. Each animal showed a
decrease in threshold between the first and second measurement ofNH4C1 detectability in
the absence of amiloride. Animals on the left underwent transaction of the chorda
tympani nerve (CT) after these shifts in threshold were measured while animals on the
right received sham surgery.















1.0
CTX SHAM
0.8

0.6

0.4

0.2

0.0

-0.2

-0.4


1 4 5 6 9 M 2
RAT NUMBER


3 7 8 10 M


Figure 4-4. Individual shifts in performance with surgery followed by the mean for each
group. Animals that received bilateral chorda tympani transaction (left) showed a
significant increase in threshold (p < .004), while sham-operated rats (right) did not.












100 CTX SHAM

< 80 80
I-
I 60
o

0 40

0 20
0 PRE-CTX PRE-SHAM
O POST-CTX 0 POST-SHAM
0 .
0.01 0.1 0.01 0.1
NH4CI CONCENTRATION (M)



Figure 4-5. NH4C1 detectability functions pre- and post-surgery. Mean performance for
each surgical group. Rats in the chorda tympani transaction group (left) had significantly
higher thresholds (p < .004) following surgery while rats in the sham-operated group
(right) did not. Presurgical results are from the second determination of threshold in the
absence of amiloride.












100 CTX SHAM

< 80

I 60 /
o

S40/

0 20 -
0 POST-CTX 0 POST-SHAM
0 CTX+AMIL 0 SHAM+ AMIL
0-
0.01 0.1 0.01 0.1
NH4CI CONCENTRATION (M)


Figure 4-6. Postsurgical NH4C1 detectability functions with and without amiloride (100
pM). Neither rats in the chorda tympani transected group (left) nor rats in the sham-
operated group (right) showed any significant change in NH4Cl threshold with amiloride.







78



1.0
CTX SHAM

LU
C 0.8
z


O 0.6
LL
LU
0-
j0.4
_j

LU
> 0.2



0.0
1 4 5 6 9 2 3 7 8 10

RAT NUMBER




Figure 4-7. Water control test. Individual overall performance when all stimuli were
replaced with distilled water. Neither rats in the chorda tympani transected group (left),
nor rats in the sham-operated group (right) performed significantly better than chance.














CHAPTER 5
GENERAL DISCUSSION

Discrepancies Between the Electrophysiology and the Behavior In Regard to NH4CI

Regardless of whether NH4C1 vs. KC1 discrimination is due to different receptor

subtypes or different temporal properties within the same receptor subtype, significant

differences in activity at the level of the CNS should be noticeable. At present, no such

differences have been observed, although this might be due to the paucity of experiments

that include both NH4C1 and KC1 as stimuli. The overall patterns of activity are very

similar for NH4Cl and KC1 in the NTS (Erickson, 1963; Nakamura & Norgren, 1993).

This similarity in neural responding in conjunction with similar generalization patterns

for the two salts in conditioned aversion and avoidance tasks has historically been used to

support both the across-fiber pattern theory and the best-stimulus method of classification

(see Erickson, 1963; Smith et al., 2000). Although these salts have been shown to taste

similar to the rat (Erickson, 1963; Hill et al., 1990; Morrison, 1967), they are also easily

discriminated suggesting that coding theorists must also provide an explanation for this

phenomenon.

In addition to the small number of CNS experiments using both KC1 and NH4C1, it is

also possible that differences between the NTS electrophysiology and the behavior exist

because the gustatory system was altered over the course of behavioral testing. Perhaps

prolonged experience with NH4C1 amplified signals from afferents that synapsed with

cells containing ammonium-sensitive receptors. Such a mechanism would not be

observed in electrophysiological tests in which the rat's first experience with NH4C1 was









on the test day. Furthermore, the subjects in these preparations are usually deeply

anesthetized, allowing for little influence from higher brain regions on gustatory activity,

unlike in an awake animal. Along these lines, stimulation of the central amygdala,

thought to be important for attention and memory, has been shown to affect responses in

the NTS and PBN (Li et al., 2002; Lundy & Norgren, 2001).

Another difference between the electrophysiological literature and the majority of the

behavioral literature involves the depletion state at the time of testing. In order to

motivate animals to perform behavioral tasks, researchers must often restrict food or

water access, or deplete subjects of some other commodity like sodium. These methods

are not necessary for electrophysiological recordings. The few electrophysiological

experiments that have been reported using depleted animals have suggested that such a

state can affect neural responses in some areas of the gustatory system. For example,

sodium depletion, or hormones associated with depletion, affect both NTS and CT

responses to NaCl (Contreras & Frank, 1979; Herness, 1992; Lundy, 1998; McCaughey

& Scott, 2000; Tamura & Norgren, 1997). Likewise, vasopressin, a hormone important

for regulating fluid homeostasis, can also significantly increase inward Na+ and proton

currents in hamster TRCs (see Gilbertson et al., 1993).

A paper by Lundy & Contreras (1999) is one of the few studies to claim that KC1 and

NH4C1 could potentially produce discriminable perceptions. The central thesis of this

paper is that NaC1, KC1 and NH4C1 could perhaps be discriminated by the rat due to

across fiber patterns shaped predominantly by the activity of 2 classes of neurons in the

geniculate ganglion. Sodium chloride (NaC1)-specialists appear to synapse with taste

cells containing amiloride-sensitive, apically-located ENaCs that are highly selective for









Na but also permit passage of small amounts of K and to a lesser extent NH4+.

Hydrochloric acid (HC1)-generalists, on the other hand, respond well to all 3 salts, but

show a more robust response to NH4C1 than to the others. The remaining types of

ganglion cells respond equally well to all 3 salts. This hypothesis is plausible, but given

that the difference in NaCl-specialist activation with KC1 and NH4C1 stimulation is very

slight, does not sufficiently address how high concentrations of KC1 would be

distinguished either from low concentrations of NH4C1 or from other nonsodium chloride

salts and acids which produce moderate activation of HCl-generalist cells. Additionally,

the finding that NH4C1 and KC1 are discriminable even in the presence of amiloride

(Geran et al., 2002), when responses from NaCl-specialists should be suppressed,

suggests that we must look elsewhere for the underlying cause of this discrimination.

As for the discrepancy among electrophysiological tests as to whether amiloride

impairs nonsodium salt responsiveness of the CT nerve, this could be due to important

differences in the methods employed. For instance, most of the experiments reporting

inhibition with amiloride deposited a salt solution on the tongue immediately followed by

either a mixture of salt and amiloride (Lundy & Contreras, 1997; Lundy et al., 1997), or

amiloride alone (Minear et al., 1996). Thus, the nerve was adapted to the salt prior to

amiloride treatment. When the reverse was attempted, Minear and colleagues (1996)

noted that adapting the tongue to amiloride prior to salt treatment had no effect on KCl

responding. This has also been shown to hold true for NH4C1 responsiveness (Formaker

& Hill, 1988; Hill & Bour, 1985; Hill et al., 1982). Also without effect were experiments

in which the tongue was adapted to amiloride prior to the presentation of salt dissolved in

amiloride (Brand et al., 1985). This method more closely resembles that of the current









detection threshold procedure. Rats ingest amiloride when they are reinforced for

pressing the correct lever. This serves to adapt the tongue to amiloride prior to a stimulus

trial consisting of salt dissolved in amiloride or amiloride alone. Differences among

electrophysiological procedures also exist for duration of stimulus delivery, stimulus

concentration, time constant, and whether the integrated response or the number of action

potentials is used as the dependent variable, making comparisons difficult.

Implications for Chloride Salt Detectability

When the parameters from salt detectability functions measured to date were

compared, we found that KC1 was more similar to NaCl dissolved in amiloride than to

NaCl dissolved in distilled water both in slope and threshold (Figure 5-1). Although,

these functions are based on data from separate groups of subjects, this finding is

interesting because NaCl + amiloride and KC1 also share a similar taste quality (Hill et

al., 1990; Kopka et al., 2000; Spector et al., 1996) and route oftransduction at higher

concentrations (Ye et al., 1994). Taken together, these results suggest that NaCl +

amiloride and KC1 are indiscriminable to the rat at a variety of concentrations due to

activation of a common pathway. One caveat to this interpretation is that although the

slope and threshold for KC1 and NaCl + amiloride are similar, the asymptotic

performance is significantly different. This is most likely due to the fact that rats in the

amiloride experiment were trained on NaCl dissolved in distilled water, and therefore

might have performed less well at the higher concentrations due to a change in perceived

intensity or quality with amiloride at suprathreshold concentrations. Rats in the KCl

experiment, however, were trained on KC1 and therefore, presumably did not experience

any changes in intensity or quality with testing. It would be interesting to test this









hypothesis by measuring NaC1 + amiloride detectability in animals trained on a KC1 vs.

water task.

Given the results of the NaCl + amiloride experiment, it was hypothesized that the

detectability function for NH4C1 might also fit this pattern and support data from

electrophysiological and taste quality experiments. Ammonium chloride shares a similar

taste quality (Erickson, 1963; Hill et al., 1990; Morrison, 1967) and perhaps at least one

transduction mechanism with KC1 and NaCl + amiloride (DeSimone et al., 2001), but

produces a detectability function that is different from that of KC1 and instead very

similar to that of NaCl (Figure 5-2), a stimulus that is easily discriminated from

nonsodium salts at higher concentrations (Geran et al., 2002; Hill et al., 1990; Spector &

Grill, 1992). In addition to evidence that NH4C1 and KC1 activate the same transduction

pathway, or pathways (DeSimone et al., 2001), there is also evidence that NH4C1 might

utilize a separate AS pathway at lower concentrations (Kloub et al., 1997). If NH4C

transduction does depend on an AS route, this pathway cannot be blocked sufficiently by

micromolar doses of amiloride to impair detectability. Larger doses of amiloride have not

been tested as they might lead to nonspecific effects (see DeSimone & Ferrell, 1985).

Sodium chloride detectability appears to depend heavily on CT input. Precisely how

heavily depends on the procedure, with values ranging from 1-2 logo units for the shift

following transaction (Kopka & Spector, 2001; Slotnick et al., 1991; Spector, Schwartz &

Grill, 1990). Kopka & Spector (2001) reported that the shift in threshold with CTX was

not further impaired with amiloride treatment, suggesting that AS taste receptor cells

innervated by the GSP are not important for normal NaCl detectability in the rat.

Performance to concentrations greater than the mean postsurgical threshold of 0.1 M









NaC1 are presumably due to combined activation of AI cells in the GL and GSP (Kopka

& Spector, 2001). Potassium chloride detectability also appears to be largely dependent

upon input from the CT. Chorda tympani transaction increases KC1 threshold by about

0.6 logo units (Geran et al., 1999) to approximately 0.1 M KC1. This shift is similar to

that observed for NH4C1, but much less than that reported for NaCl after CTX (Kopka &

Spector, 2001; Slotnick et al., 1991; Spector, Schwartz & Grill, 1990). The KC1 threshold

for intact animals, however, is markedly (- 0.5 to 0.7 logo unit) higher than that of either

NaCl or NH4C1. Thus, KC1 and NaCl result in approximately the same threshold

following CTX while the threshold for NH4C1 is slightly lower (Table 5-1). This suggests

that the remaining taste receptor cells innervated by the GL and/or GSP are more

sensitive to NH4C1 than to KC1 or NaC1. This finding is supported by the

electrophysiology literature in that each of the 3 main gustatory nerves responds more

robustly to mid-range and high concentrations of NH4C1 than to similar concentrations of

KC1 or NaCl (see Frank et al., 1983; Kitada et al., 1998; Sollars & Hill, 1998). It is not

apparent, however, if this comparatively large response to ammonium is also observed at

lower, near-threshold concentrations.

It is not known what effect, if any, the anion might have on salt detectability. To date,

detectability has been measured for only 1 nonchloride salt (NaG) in the rat and it was

found to be very similar to NaCl (Geran & Spector, 2000b). It is unclear whether other

nonchloride salts like potassium gluconate or ammonium hippurate would also produce

detectability functions that mimic those of their halogenated counterparts. Likewise, the

detection thresholds of divalent salts like calcium and magnesium chloride have not been

tested. In the future, it might be worthwhile to expand the number of stimuli tested and









nerve transactions performed, including GSP and GL neurotomies, to better assess

processes necessary for normal salt detectability in the rat.

Support for the Hypothesis That The Seventh Cranial Nerve Is More Important For
Taste Recognition and Discrimination Than The Glossopharyngeal Nerve

A variety of nerve transaction experiments have suggested that the gustatory branches

of the 7th cranial nerve appear to be involved in taste recognition and discrimination,

while the 9th cranial nerve, or GL, is more important for the expression of oromotor

reflexes to aversive stimuli (see St. John & Spector, 1998). The discrimination

experiments have involved tastants from each of the 4 main taste categories. These

discrimination have included sucrose vs. maltose, citric acid vs. quinine, KC1 vs.

quinine, NaCl vs. KC1 (St. John, Markison, Guagliardo et al., 1997; St. John & Spector,

1997; St. John & Spector, 1998; Spector & Grill, 1992; Spector et al., 1997) and now

KC1 vs. NH4C1 and NaCl vs. NH4C1. Furthermore, although the GL is the only gustatory

nerve that contains fibers highly responsive to quinine (Frank, 1991) and innervates

approximately 60% of the taste buds in the oral cavity (Miller, 1995), GL transaction has

little to no effect on quinine performance in a variety of tasks. These include

concentration-dependent avoidance in a 2-bottle preference test and detection threshold

(Akaike et al., 1965; St. John & Spector, 1996). Additionally, in their seminal paper, St.

John & Spector (1998) reported that GL transaction likewise had no effect on KC1 vs.

quinine discrimination, while combined transaction of the CT and GSP nerves did. This

was surprising given that the GL contains fibers that are differentially responsive to these

2 stimuli while neither the CT nor the GSP are thought to contain a substantial number of

quinine-sensitive units (Frank et al., 1983; Sollars & Hill, 1998), although a single fiber

analysis of the GSP has not yet been performed. These results were interpreted as









evidence that input from the CT could potentially be important for the perception of a

wide variety of taste qualities, not just the salt and acid stimuli to which it is most

responsive.

This apparent dichotomy between the functional roles of the facial and GL nerves is

more pronounced in the catfish (Finger & Morita, 1987). For these animals, the facial

nerve is necessary for locating food while the GL is necessary for initiation of the

appropriate oromotor response once a food stimulus has been located (see Caprio et al.,

1993). In rodents, the role of the GL is not quite so obvious. Deficits in the

sensory/discriminative and hedonic domains of taste function (see Spector, 2000),

although minimal, can be produced with GL transaction in addition to changes in

oromotor reflexes. For instance, GL transaction has produced a modest impairment in the

quinine avoidance of naive rats (Markison et al., 1999). Combined transaction of the CT

and GL increases the detection threshold for quinine and compromises pre-trained

quinine avoidance in a brief-access test, although GL or CT transaction alone is without

effect on these tasks (St. John et al., 1994; St. John & Spector, 1996). This suggests that

perhaps the GL and CT carry information that is redundant for the maintenance of these

tasks in pre-trained rats. Deficits in aversive oromotor reflexes, on the other hand, are

quite pronounced with GL transaction (Travers et al., 1987), suggesting that the GL is

more important for unconditioned reflexes to aversive stimuli (see Eylam et al, 2000),

than for taste perception or hedonic responses.

While it is easy to appreciate the utility of fibers narrowly tuned for "bitter"

compounds or salts and acids in performing such a function, it is less obvious why this

nerve would contain fibers that are narrowly-tuned for sugars (Frank, 1991). It has been









hypothesized (St. John & Spector, 1998) that although the GL is not sufficient to

maintain a KC1 vs. quinine discrimination for rats trained while intact, this input may be

sufficient for rats to learn this or other discrimination over time based on cues mediated

by the GL. A second hypothesis concerning the role of the GL, is that this nerve could

provide information important for determining suprathreshold stimulus intensity. Humans

report an increase in perceived intensity of quinine following bilateral anesthesia of the

CT nerve that appears to be due to a release of inhibition (see Catalanotto et al., 1993). It

is possible that such a mechanism also exists in the rat, but has not been apparent due to

obstacles inherent in measuring suprathreshold intensity in animal subjects. Clearly more

research is necessary to determine the role of the GL in the taste-guided behavior of

rodents.

Conclusions

Overall, this series of experiments supports the possibility of a labeled line for sodium

taste quality in the rat, as the AS transduction pathway appears to be both necessary and

sufficient for the recognition and detection of sodium salts (see Bernstein & Hennessy,

1987; Geran & Spector, 2000b). The fact that amiloride does not seem to affect the taste

quality or intensity of KC1 or NH4C1 lends credence to the hypothesis that this AS

pathway is appreciably activated only by sodium and lithium ions. In contrast, the

perceived intensities of all three salts tested to date (i. e., NaC1, KC1 and NH4C1) have

been compromised by CT transaction, suggesting that this nerve might be highly

sensitive to low concentrations of both sodium and nonsodium chloride salts and/or

provide input to regions of the gustatory CNS important for salt detection. It would be

interesting to test this hypothesis using a wider array of salt stimuli and nerve

transactions.









Also of interest is the finding that KC1 and NH4C1, stimuli that produce similar

patterns of activity in the NST (Erickson, 1963), are discriminated by the rat. This result

raises questions about the effect of training on the tuning characteristics of the NST, as

well as the possibility that across-fiber pattern coding, as currently applied, might not

always be the best method for taste quality classification. Furthermore, this finding

underscores the importance of using several different behavioral tasks to describe the

perception associated with a particular taste stimulus and suggests that perhaps small

differences at the neural level might produce significant differences at the behavioral

level. Perhaps focusing on neural and behavioral differences within a prototypical taste

quality, such as "saltiness," might enable researchers to make interesting predictions

about the coding of a particular class of taste stimuli or even taste coding in general.














Table 5-1 Mean Detectio s


Stimulus Mean Threshold Shift in Mean Threshold
for Intact Rats Threshold with for CTX Rats
CTX
NaCI a ~ 0.006 M 1.0 -1.2 loglo units ~ 0.1 M
NaCI + 100 ~ 0.04 M 0.4 loglo units ~ 0.1 M
|pM amiloride a
KCI b ~ 0.03 M 0.6 logo units 0.1 M
NH4CI ~ 0.009 0.01 M 0.5 loglo units ~ 0.04 M
a T 1 C Tr 1 0 Cl ,> 1


v values trom KopKa & spector, zuu i.
b Values from Geran et al., 1999.