Influence of stimulus and oral adaptation temperature on gustatory responses in central taste-sensitive neurons

Abstract

The temperature of taste stimuli can modulate gustatory processing. Perceptual data indicate that the adapted temperature of oral epithelia also influences gustation, although little is known about the neural basis of this effect. Here, we electrophysiologically recorded orosensory responses (spikes) to 25°C (cool) and 35°C (warm) solutions of sucrose (0.1 and 0.3 M), NaCl (0.004, 0.1, and 0.3 M), and water from taste-sensitive neurons in the nucleus of the solitary tract in mice under varied thermal adaptation of oral epithelia. Conditions included presentation of taste stimuli isothermal to adaptation temperatures of 25°C (constant cooling) and 35°C (constant warming), delivery of 25°C stimuli following 35°C adaptation (relative cooling), and presentation of 35°C stimuli following 25°C adaptation (relative warming). Responses to sucrose in sucrose-oriented cells (n = 15) were enhanced under the constant and relative warming conditions compared with constant cooling, where contiguous cooling across adaptation and stimulus periods induced the lowest and longest latency responses to sucrose. Yet compared with constant warming, cooling sucrose following warm adaptation (relative cooling) only marginally reduced activity to 0.1 M sucrose and did not alter responses to 0.3 M sucrose. Thus, warmth adaptation counteracted the attenuation in sucrose activity associated with stimulus cooling. Analysis of sodium-oriented (n = 25) neurons revealed adaptation to cool water, and cooling taste solutions enhanced unit firing to 0.004 M (perithreshold) NaCl, whereas warmth adaptation and stimulus warming could facilitate activity to 0.3 M NaCl. The concentration dependence of this thermal effect may reflect a dual effect of temperature on the sodium reception mechanism that drives sodium-oriented cells.

the excitation of gustatory receptor cells by taste stimuli classified as sweet by humans involves molecular mechanisms with strong sensitivity to temperature. For instance, the cellular transduction of sucrose is mediated by inward current carried by the transient receptor potential melastatin 5 (TRPM5) ion channel (Perez et al. 2002; Zhang et al. 2003), which displays increased current passage with warming from 15°C to 35°C (Talavera et al. 2005). Warming taste solutions increases peripheral nerve activity to sucrose (Breza et al. 2006; Lu et al. 2012; Yamashita et al. 1970; Yamashita and Sato 1965), an effect that is lost following genetic deletion of TRPM5 (Talavera et al. 2005). Stimulus warming accordingly enhances taste responses to sucrose in oral sensory neurons in the rostral nucleus of the solitary tract (NTS) (Wilson and Lemon 2013, 2014), which receives the central terminations of cranial nerve fibers carrying gustatory and oral somatosensory information. Furthermore, latency to respond to sucrose and the slope of the sucrose concentration-response function in NTS neurons are inversely related to stimulus temperature (Wilson and Lemon 2014), demonstrating that progressive change in stimulus temperature can systematically modify the timing of gustatory activity to sucrose in the mammalian brain and how this activity changes with concentration.

Receptor mechanisms underlying salt taste transmission also show strong temperature dependence. The transduction of sodium salts involves passage of Na⁺ ions through epithelial sodium channels (ENaCs) expressed by taste receptor cells, as evidenced by reduction in neural activity to NaCl following pharmacological blockade of ENaC by the diuretic amiloride (e.g., Heck et al. 1984) and confirmed in genetic and molecular reports (Chandrashekar et al. 2010; Shigemura et al. 2008). Electrophysiological studies of ENaCs in expression systems describe a marked sensitivity of this channel to temperature, where, for instance, decreasing temperature (i.e., cooling) can increase amiloride-sensitive Na⁺ current and channel open probability (Askwith et al. 2001; Awayda et al. 2004; Chraibi and Horisberger 2003). Furthermore, the transient receptor potential vanilloid 1 (TRPV1) ion channel, which is sensitive to capsaicin and noxious heat stimulation (Caterina et al. 1997; Tominaga et al. 1998), may also contribute to the oral transduction of sodium salts. For instance, recordings from peripheral nerves show that ENaC-independent responses to lingual application of NaCl are facilitated when solutions are heated to temperatures that approach or exceed threshold for noxious heat, an effect blocked through oral application of antagonists of TRPV1 (Lyall et al. 2004). Thermal sensitivity of receptors responsive to sodium agrees with other studies that showed that change in stimulus temperature can modify peripheral (Breza et al. 2006; Lundy and Contreras 1997; Nakamura and Kurihara 1988; Ninomiya 1996; Ninomiya et al. 1996; Sato and Yamashita 1965; Yamashita et al. 1970; Yamashita et al. 1964) and central (Wilson and Lemon 2013) gustatory activity to NaCl. These data, along with those for sucrose, support the postulate that stimulus temperature importantly guides the neural representation of sensory input in gustatory pathways.

Whereas psychophysical studies also support a role for stimulus temperature in the generation of neural and perceptual signals for taste stimuli (see Green and Frankmann 1987 and Torregrossa et al. 2012), some of this work indicates that, under certain conditions, variation in the temperature of oral epithelia can also modulate taste signals. For instance, thermal stimulation of different regions of the tongue is sufficient to elicit sensory qualities common to sweet, salty, sour, or bitter taste sensations in humans (Cruz and Green 2000; von Békésy 1964), implying that lingual temperature may, by itself, activate or gate the actions of taste receptors. Furthermore, cooling taste solutions reduces the perceived intensity of sucrose in humans (Bartoshuk et al. 1982; Calvino 1986; Green and Nachtigal 2012), although some studies reported that this effect was apparent only when both the tongue and solution were similarly cooled (to 20°C), with stimulus cooling causing only slight reductions in sucrose intensity when the tongue was thermally controlled and warmed (to 36°C) (Green and Frankmann 1987). Tongue temperature can approximate a constant, near core, “warm” value with the mouth closed (Green 1984; Green and Gelhard 1987), yet certain behaviors may decrease oral temperature. Green (1986) reported that, in humans, the temperature of the tongue is slightly reduced when the mouth is closed for 30 s compared with 60 s, which suggests that oral epithelia can lose heat to evaporative cooling when the mouth is opened. Furthermore, sipping chilled water can rapidly decrease temperature inside the human mouth by several degrees Celsius (Pangborn et al. 1970), giving rise to the potential for heat loss to cooled ingesta by oral epithelia.

The present work used in vivo neurophysiological recordings in inbred mice to test the hypothesis that variation in the adapting temperature of oral epithelia modifies central activity for gustation. To do this, we recorded gustatory responses to temperature-adjusted solutions of sucrose and NaCl from single taste-sensitive neurons in the NTS of mice under oral adaptation to physiologically warm and relatively cool temperatures. Results suggest that orosensory activity to gustatory stimuli in medullary circuits for taste is influenced in some cases by interplay between oral adaptation and taste stimulus temperature.

MATERIALS AND METHODS

Mice.

Thirty adult male and female C57BL/6J (B6) mice (stock no. 000664; The Jackson Laboratory) weighing 18–35 g on the day of recording were used. All mice were housed in a vivarium that maintained a 12:12-h light-dark cycle and an ambient temperature of ∼23°C. Food and water were available ad libitum.

Single-unit electrophysiology.

Animals were prepared for electrophysiological recording in accordance with guidelines from the University of Oklahoma Institutional Animal Care and Use Committee, which reviewed and approved our procedures, and the National Institutes of Health. Mice were acutely anesthetized using a mixture of ketamine (100 mg/kg ip) and xylazine (10 mg/kg ip). Atropine (24 μg/kg ip) was administered to reduce bronchial secretions. Once anesthetized, mice were tracheotomized to facilitate breathing during flow of solutions into the mouth and for maintenance of gas anesthesia. Mice were secured in a stereotaxic instrument with ear bars (model 930; David Kopf Instruments, Tujunga, CA), and the skull was leveled to position lambda and bregma on the same dorsal-ventral plane. The lower incisors were trimmed using rongeurs. A silk thread was passed caudal of these teeth and drawn lightly taut to deflect the mandible downward. The tongue was extended from the mouth by a small rostroventral suture. Anesthesia was maintained throughout recording sessions by 0.6–0.8% isoflurane in oxygen, which the mice freely respired through their tracheostomy tube (Wilson and Lemon 2014). Body temperature was kept at 37°C by a heating pad. A portion of the occipital bone was removed, and parts of the cerebellum were gently aspirated to allow dorsal access to the medulla. The rostral, taste-sensitive region of the NTS was targeted at 5.8–6.2 mm posterior to bregma, 1.2 mm lateral from the midline, and 500–800 μm below the surface of brain stem (Franklin and Paxinos 2008).

Trains of extracellular action potentials were recorded from taste-sensitive NTS neurons using conventional electrophysiology methods. Tungsten microelectrodes (z = 2 to 5 MΩ; FHC, Bowdoinham, ME) sampled unit electrophysiological activity, which was AC amplified (Grass P511, high-z probe), band-pass filtered (0.3–10 kHz), and monitored on an oscilloscope and loudspeaker. An electronic micropositioner (Model 2660; David Kopf Instruments, Tujunga, CA) advanced the electrode ventrally through the brain stem in 1-μm steps. Spikes generated by individual neurons were identified based on waveform consistency. For isolated thermo-gustatory neurons, spikes were digitally sampled at 25 kHz (1401 interface and Spike2 software; CED, Cambridge, UK) and time-stamped to the nearest 0.1 ms. All analyses of response data were performed offline.

Taste stimuli.

Stimulus delivery was accomplished using a custom apparatus, as described for mice (Wilson et al. 2012; Wilson and Lemon 2013, 2014). Solution flow rate was ∼0.9 ml/s at ambient temperature, although this rate could nominally increase with warming; this variance does not impact response quantification (Wilson and Lemon 2014). The taste delivery system was designed to stimulate broad regions of oral epithelia with solutions, including the rostral and caudal tongue, and the palate (Wilson et al. 2012).

All taste chemicals (Sigma, St. Louis, MO) were high in purity and dissolved in deionized water. Multiple concentrations of sucrose and NaCl were tested at 25°C and 35°C on independent trials following oral application of a thermal-adapting rinse of water adjusted to 25°C and 35°C, as described below. Moreover, neurons were also tested with 25°C and 35°C water as stimuli following thermal adaptation to assess their temperature sensitivity characteristics in the absence of taste input. Stimulus solutions were stored in airtight glass bottles that were adjusted to either 25°C or 35°C in separate recirculating water baths. As referenced in this article, 25°C (near room temperature) was considered “cool,” whereas 35°C (near physiological temperature) was “warm.” Human psychophysical studies have reported that oral stimulation with temperatures near 25°C are perceived as cool, whereas the threshold for perception of warmth with increments in the temperature of oral tissue lies near 35°C (Green 1984, 1986).

Each trial was composed of prestimulus and stimulus periods. A trial began with continuation of oral delivery of the adapting thermal water rinse, as described below, for 5 s. At this point, flow switched to the taste/thermal stimulus, which was delivered for 5 s. Switching between prestimulus rinse and stimulus solutions was accomplished using a 3-way fluid valve that was controlled by the data acquisition system. Flow switched back to the adapting rinse following completion of stimulus delivery. The oral adapting rinse continued in between trials to facilitate thermal adaptation. The intertrial interval was ∼2–3 min and allowed cells to return to prestimulus activity levels. During this interval, the stimulus delivery pathways and tubing were thoroughly rinsed with water adjusted to the temperature of the adapting rinse. Mice did not ingest solutions, as their open mouths and anesthetized states prevented swallowing.

Once isolated, the gustatory tuning profile of each neuron was characterized through initial measurement of responses, defined as described below, to a set of 25°C “prototype” taste stimuli presented in random order. Prototype stimuli included 0.5 M sucrose (disaccharide), 0.1 M sodium chloride (NaCl, sodium salt), 0.01 M citric acid (organic acid), and 0.01 M quinine-HCl (alkaloid). These stimuli fall into taste categories described as, respectively, “sweet,” “salty,” “sour,” and “bitter” by humans. The temperature of the adapting rinse during prototype testing was fixed at 25°C; the adapting rinse temperature was controlled on all trials by a fluid-to-fluid heat exchanger (cf. Wilson and Lemon 2014).

Next, cells were tested with multiple concentrations of NaCl (0.004, 0.1, and 0.3 M) and sucrose (0.1 and 0.3 M), and also water, under varied adapting rinse and stimulus temperatures. Trials for these stimuli involved four unique adaptation-stimulus temperature conditions (see Fig. 2A): adaptation to 25°C followed by delivery of a 25°C stimulus (referred to herein as constant cooling); 35°C adapting rinse, 25°C stimulus (relative cooling); 35°C adapting rinse, 35°C stimulus (constant warming); and 25°C adapting rinse, 35°C stimulus (relative warming). Constant cooling, where both thermal adaptation and taste stimulation occurred at 25°C, served as the baseline condition for relative warming, where stimulus temperature increased following adaptation to 25°C. Likewise, constant warming, at 35°C, was the baseline condition for relative cooling, where stimulus temperature decreased following 35°C adaptation. A fine thermocouple probe positioned inside the tip of the oral delivery tube continuously monitored rinse and stimulus temperatures at the moment of oral delivery. A digital thermometer (BAT-12; Physitemp Instruments, Clifton, NJ) coupled to the probe provided an analog voltage signal that covaried with temperature and was sampled (at 1 kHz) alongside neural activity by the data acquisition system. Temperatures stated above are target values, although actual mean stimulus temperature, as measured during the last 3 s of stimulus delivery, marginally deviated from target by, on average, 1.3°C under the relative warming (mean stimulus temperature = 33.7°C) and relative cooling (mean stimulus temperature = 26.3°C) conditions. This deviation was probably caused by, respectively, heat loss to the cool and heat gain from the warm adapting rinse by the fluid control valve, which in turn imparted its thermal character to the stimulus as it flowed though.

Fig. 2.

Examples of raw data. A: family of traces showing real-time measurement of solution temperature at the moment of oral delivery against time on 4 separate trials, where a cool (25°C) stimulus was delivered following cool adaptation of oral epithelia [constant cooling (C)], a warm (35°C) stimulus was presented following warm adaptation [constant warming (W)], a cool stimulus was delivered following warm adaptation [relative cooling (W→C)], and a warm stimulus was given after cool adaptation [relative warming (C→W)]. B: electrophysiological sweeps showing responses by an S-type neuron to 0.1 and 0.3 M sucrose recorded across the 4 thermal conditions. The ordering of the stack of sweeps indicated for 0.1 M sucrose also applies to 0.3 M sucrose and data for NaCl in C. Upward and downward arrowheads indicate stimulus onset and offset, respectively. C: electrophysiological sweeps showing responses by an N-type neuron to 0.004, 0.1, and 0.3 M NaCl recorded across the 4 thermal conditions.

Concentrations of NaCl covered a broad intensity range for B6 mice, from near detection threshold (0.004 M), as determined in NaCl avoidance tasks (Ishiwatari and Bachmanov 2012), to approximate indifference compared with water (0.1 M) to aversive (0.3 M) (Bachmanov et al. 1998; Ninomiya et al. 1989). Sucrose concentrations were detectable by mice (Treesukosol and Spector 2012) and appetitive (e.g., see Bachmanov et al. 2001).

The ordering of adaptation-stimulus temperature conditions for water, 0.1 and 0.3 M sucrose, and 0.004, 0.1, and 0.3 M NaCl was randomized for each cell. To begin, one of the two adapting rinse temperatures was randomly selected (e.g., 25°C), and water adjusted to this temperature by the heat exchanger was applied continuously to oral epithelia, except, of course, during periods of stimulus delivery. After ∼5 min of adaptation, one of the two stimulus temperatures was chosen at random (e.g., 35°C), and water, sucrose, and NaCl stimuli adjusted to this temperature were tested as a group, followed by water, sucrose, and NaCl solutions adjusted to the remaining stimulus temperature (e.g., 25°C). Trials were randomized across stimuli and concentrations within each temperature group. Following completion of this sequence, the temperature of the adapting rinse was changed to the remaining value (e.g., 35°C), and the mouth was bathed with this rinse for ∼5 min, after which randomization and testing of 25°C and 35°C stimuli was repeated, as described above.

Responses to temperature-varied solutions of 0.004, 0.1, and 0.3 M NaCl mixed with and without the competitive ENaC inhibitor amiloride (30 μM) were recorded from an additional group of neurons to explore receptor mechanisms mediating thermo-gustatory activity to NaCl. The ordering of adaptation-stimulus temperature conditions for amiloride testing was randomized, as was the ordering of stimulus concentration and amiloride treatment conditions. Trials that included amiloride began with flow of distilled water for 5 s, followed by delivery of the NaCl-amiloride mixture for 5 s; in some cases, we tested additional trials that began with flow of amiloride (30 μM) rather than water to inhibit ENaC prior to NaCl-amiloride delivery. On completion of each trial that involved amiloride, the tongue was rinsed with 20 ml of 0.1 M NaCl followed by distilled water. Trials were separated by >3 min to avoid carryover effects.

At the end of a daily recording session, a weak electrical current (100 μA/1.5 s) was passed through the recording electrode to make an electrolytic lesion of brain tissue at the recording site. For mice where multiple neurons were sampled, only one lesion was made at the location of the final cell. Anesthetized mice were then overdosed with 20% (wt/vol) urethane and perfused transcardially with isotonic saline followed by a mixture of 4% paraformaldehyde and 3% sucrose. Brains were removed and stored at least overnight in a mixture of 4% paraformaldehyde and 20% sucrose. Brain stems were cut by microtome into coronal sections (40 μm), mounted onto slides, and stained with thionin. Lesion sites were compared against an atlas of the mouse brain (Franklin and Paxinos 2008) to verify electrode placement.

Data analysis.

The strategy for data analysis was to classify cells into unit types and to explore the effect of temperature on gustatory activity to sucrose and NaCl in neurons oriented to sucrose (S-type units) or NaCl (N-type units), respectively. Hierarchical cluster analysis was applied to a matrix of correlation distances among neurons computed from their responses to prototype stimuli to identify S- and N-type cells. For the purpose of cluster analysis, taste responses to prototype stimuli were calculated as the sum of spikes that arose during the 5-s stimulus presentation period minus the sum of spikes during the 5-s prestimulus period.

Responses to each thermoconcentration of sucrose and NaCl were analyzed in detail for S- and N-type units, respectively, that showed significant activation to each input, as assessed through detection of latency to first spike on individual stimulus trials (Wilson and Lemon 2014). S- and N-type units that did not display significant latencies on all sucrose or NaCl trials, respectively, were discarded from further analysis. This measure aimed to avoid biasing statistical estimates of thermal effects on taste activity with data from cells that showed weak or null gustatory responses, although only a few recorded cells met discard criterion, as described below. What is more is that first spike latency was used to define the beginning of the thermo-gustatory response window on individual sucrose and NaCl trials so that factorial analysis of temperature effects on firing rates to these inputs, as described below, reflected spiking activity during periods of significant responding. Finally, the effects of adapting and stimulus temperature on latency were also evaluated within each cell class.

On each trial, latency was defined as the time of the first spike since stimulus onset where the firing rate became significantly greater than the prestimulus rate. Only significant increases in discharge were assessed, as taste-sensitive NTS neurons in mice invariably showed excitation when responding to sucrose or NaCl. Stimulus onset was defined as the 5-s mark of each trial, which corresponded to when solution flow switched from the adapting rinse to the stimulus. This mark preceded stimulus contact with the mouth and neural activation by a brief (<1 s) delay during rinse-stimulus switchover, although trial structure was constant across all stimulus presentations. Stimulus onset was used as the zero point for expression of spike time stamps used to detect response latency. Latency was quantified by a binless algorithm (cf. Bair and Koch 1996; Chase and Young 2007) adapted for analysis of NTS neurons (Wilson and Lemon 2014). Briefly, an iterative Poisson method estimated the probability that the firing rate at each sequential spike during taste delivery was due to lingering prestimulus drive, not taste input. When this probability became <10⁻⁶, the firing rate was considered unusually high relative to average prestimulus firing across all trials, which was indexed separately for 25°C and 35°C adaptation conditions. Thus, the time of the first peristimulus spike where this criterion was met was taken as response latency (see Fig. 3A). Latency was left undefined if criterion was not met for spikes falling within 4 s of stimulus onset.

Fig. 3.

Effects of adaptation and stimulus temperature on latency to first spike to sucrose in S-type neurons. A: rastergrams from a sample S-type cell depicting detection of latency to first spike on individual trials for 0.1 and 0.3 M sucrose across adaptation-stimulus temperature conditions. The blackened raster spike on each trial denotes latency, defined as the time of the first action potential during stimulus delivery where the firing rate of the cell became significantly greater than the prestimulus rate (see materials and methods). Latency to first spike for 0.1 M sucrose on each adaptation-stimulus temperature condition was as follows: C, 2.24 s; W, 1.59 s; W→C, 1.68 s; C→W, 1.72 s. Latencies for 0.3 M sucrose were as follows: C, 1.48 s; W, 0.99 s; W→C, 0.94 s; C→W, 1.06 s. B: median latency (±68% confidence limits) to first spike for 0.1 and 0.3 M sucrose for each adaptation-stimulus temperature condition, abbreviated as above (*P < 0.001). Confidence limits were approximated using a bootstrap-resampling procedure (see materials and methods).

Due to repeated-measures and deviation from normality, latency data were compared between temperature conditions using sign tests, with Bonferroni-adjusted α for multiple comparisons. A bootstrap resampling procedure estimated a confidence interval (CI) of median latency for each stimulus and thermal condition (cf. Wilson and Lemon 2014). To do this, latencies were resampled with replacement 1,000 times, where the n of each resample was equivalent to the number of actual latencies. On each resample, a Studentized bootstrapped-t was computed as T* = (m* − m)/s*, where m* was the median of the bootstrapped latencies, m was the median of the actual latency data, and s* was the standard deviation (SD) of the medians of 100 nested bootstrapped resamples of the current bootstrap. A 68% CI, corresponding to ∼1 SD, was computed using the distribution of T* values. This CI was given by $T_{\mathrm{0.16}}^{*}$ × S_M* + m (lower bound) and $T_{\mathrm{0.84}}^{*}$ × S_M* + m (upper bound), where $T_{\mathrm{0.16}}^{*}$ and $T_{\mathrm{0.84}}^{*}$ were the 16th and 84th percentile T* values, respectively, and S_M* was the standard deviation of the distribution of medians for the 1,000 bootstrap resamples of latency.

The time course of firing to individual concentrations of sucrose and NaCl was analyzed in S- and N-type neurons, respectively, that showed significant activation to each input across all four adaptation-stimulus temperature conditions. An initial statistical assessment revealed that responses to sucrose and NaCl differed across these conditions, although these differences were not the same at all concentrations (temperature condition × concentration interactions, 2-way repeated-measures ANOVA, P < 0.007). Thus, neuronal firing rate, operationally defined as spikes per 500 ms epoch post latency, to each concentration of sucrose or NaCl was analyzed independently. To begin, spike trains were divided into half-second bins. A three-way repeated-measures ANOVA using adapting temperature (2 levels: cool and warm), stimulus temperature (2 levels: cool and warm), and time bin (n levels) as factors was applied to evaluate how these parameters influenced firing rates to sucrose and NaCl. Because stimulus onset was fixed at 5 s into each trial, although latency to first spike was variable across stimulus and thermal conditions, as reported below, the number of response bins captured during significant taste-induced firing was not equivalent across stimuli. However, five bins were used for all stimuli except 0.1 M sucrose, for which only three bins were analyzed due to a marked delay (i.e., >3 s, yet the stimulus was only 5 s long) in first spike latency to this input under constant cooling in three S-type cells. Thus, analyses of temperature influence on firing rates to 0.004, 0.1, and 0.3 M NaCl and 0.3 M sucrose were based on 2.5 s of unit activity postlatency, whereas firing rates to 0.1 M sucrose were evaluated using 1.5 s of activity.

Responses to temperature-varied mixtures of NaCl and amiloride were analyzed in an additional sample of N-type cells using repeated-measures ANOVAs. These analyses used NaCl concentration, amiloride treatment condition, stimulus temperature, and adaptation temperature as factors. To simplify analysis of amiloride effects, the total number of spikes that arose during the 3.5-s period that followed first-spike latency quantified taste responses on compared trials. A longer response window than above was used because mixtures of 0.1 and 0.3 M NaCl and amiloride delivered after adaptation to cool or warm water induced an early, transient response prior to inhibition of sodium activity by amiloride.

The assumption of sphericity among differences between group means in each repeated-measures ANOVA was evaluated using Mauchly's test. When the null hypothesis of sphericity was rejected, degrees of freedom (df) for F values were corrected using the Greenhouse-Geisser procedure. The corrected df are reported herein as real numbers, where applicable. Omnibus F values were evaluated using α = 0.05. Pairwise comparisons among group means that followed significant F values were carried out using paired t-tests under Bonferroni-adjusted α.

All ANOVAs were executed in SPSS (IBM, Somers, NY), with pairwise comparisons performed in the R programming language (R Foundation for Statistical Computing, Vienna, Austria). All other analyses and the generation of data plots were carried out in MATLAB (The MathWorks, Natick, MA) using standard routines and custom code. Utilities from the statistics and bioinformatics modules for this platform were used.

RESULTS

General response characteristics.

Trains of action potentials were recorded from 49 NTS neurons in B6 mice. Thirty cells were sampled from female mice, and 19 cells were recorded from males. Each cell was tested with all stimuli and thermal conditions (28 trials) at least once, and a total of 1,372 stimulus response trials were analyzed. Recorded neurons showed low baseline spike discharge rates across all trials, although the mean prestimulus rate was higher (F_1,48 = 13.67, P = 0.001) during adaptation to warm water (1.68 Hz ± 1.87 SD) compared with cool (1.04 Hz ± 1.14 SD). Sex did not influence responses to the prototype stimuli at 25°C [nonsignificant (ns) sex × stimulus interaction, 2-way ANOVA, P = 0.7; ns main effect of sex, P = 0.7] and was not included as a factor in analyses. Cluster analysis identified 16 S-type and 26 N-type cells (Figs. 1, A and B, and 2, B and C), with units in both groups composing 86% of sampled neurons. For these groups, trials that used water as a stimulus under relative warming and relative cooling were analyzed to assess sensitivity to temperature change. The number of spikes per 5-s delivery of water during relative cooling did not differ between S- and N-type cells (independent-samples t-test, P > 0.05). On the other hand, stimulation with water during relative warming evoked a larger number of spikes during the 5-s stimulus window in S-type (4.37 net spikes ± 3.86 SD) compared with N-type (−0.65 net spikes ± 2.46 SD) neurons (independent-samples t-test corrected for unequal variances, t_22.6 = 4.7, P < 0.001). Thus, S-type neurons showed heightened sensitivity to warming from 25°C in the absence of gustatory drive. Histological analysis revealed that electrode positioning indeed targeted the NTS (Fig. 1C).

Fig. 1.

Neural groups and verification of electrode placement. A: dendrogram showing the outcome of cluster analysis applied to sort neurons into groups. Characters denoting each group refer to response profiles in B. B: means ± 1 SE responses by each neural group to 25°C solutions of 0.01 M citric acid (C), 0.1 M NaCl (N), 0.01 M quinine-HCl (Q), and 0.5 M sucrose (S). N-type, sodium-oriented cells; B-type, broadly responsive cells; H-type, acid/electrolyte-oriented cells; S-type, sucrose-oriented neurons. C: coronal section (40 μm) through mouse brain stem (left) along with neuroanatomic sketch (right) highlighting the location of the nucleus of the solitary tract (NTS) relative to general landmarks (cf. Franklin and Paxinos 2008). Arrowhead indicates an electrolytic lesion made following unit recording. SpVe, spinal vestibular nucleus; sp5, spinal trigeminal tract.

Analysis of sucrose activity in S-type cells.

Latency to first spike was estimated on each sucrose trial to quantify neuronal firing from the time point of significant activation and to determine how change in adapting temperature affected lag to respond to sucrose. Latency to respond to 0.1 and 0.3 M sucrose was detected across all adaptation-stimulus temperature conditions for all but one S-type neuron, which was excluded from further analysis. Figure 3A demonstrates measurement of latency in one of the 15 analyzed S-type cells. For this unit, contiguous warming between adaptation and stimulus epochs of taste trials (i.e., the constant warming adaptation-stimulus temperature condition) and warming taste solutions following cool adaptation (relative warming) appeared to reduce latencies for both 0.1 and 0.3 M sucrose compared with constant cooling. This observation follows prior data from NTS cells showing, under oral adaptation to ambient temperature water, that warming sucrose solutions decreased response latency, whereas cooling substantially increased latency (Wilson and Lemon 2014). Yet for the unit in Fig. 3A, cooling taste solutions following warm adaptation (relative cooling) appeared to exert no effect on latencies for either 0.1 or 0.3 M sucrose compared with constant warming.

Analysis of latency data across all S-type neurons revealed a common trend. For 0.1 and 0.3 M sucrose, contiguous warming between the adaptation and stimulus phase of taste trials (constant warming) and warming sucrose following cool adaptation (relative warming) significantly decreased latency relative to constant cooling (sign tests, P < 0.001; Fig. 3B), which induced the longest latencies among significant responses to each sucrose concentration. However, cooling sucrose solutions following warm adaptation (relative cooling) produced latencies for either concentration that were not different from those measured under constant warming (ns sign tests, P > 0.05), which decreased lag. Thus, warmth adaptation of the mouth counteracts the delay in response associated with cooling sucrose solutions.

The effects of adapting temperature, stimulus temperature, and time on neuronal firing rates to sucrose were evaluated separately for each concentration using three-way ANOVA. Cooling and warming solutions of 0.1 and 0.3 M sucrose caused a change in response to these stimuli that varied with adapting temperature (adapting temperature × stimulus temperature interactions, F_1,14 > 8, P < 0.05), although these effects were constant across time for both concentrations (ns adapting temperature × stimulus temperature × time interactions, P > 0.05; ns adapting temperature × time interactions, P > 0.05; ns stimulus temperature × time interactions, P > 0.05). Contiguous warming between adaptation and stimulus epochs of taste trials (constant warming) increased firing rates to 0.1 and 0.3 M sucrose compared with the constant cooling condition (Bonferroni-adjusted pairwise comparisons, P < 0.001; Fig. 4). Thus, constant cooling suppressed whereas constant warming enhanced unit activity to sucrose. Similarly, warming taste solutions following cool adaptation (relative warming) also increased firing rates to 0.1 and 0.3 M sucrose compared with constant cooling (Bonferroni-adjusted pairwise comparisons, P ≤ 0.001). On the other hand, cooling taste solutions following warm adaptation (relative cooling) only moderately reduced unit firing to 0.1 M sucrose (Bonferroni-corrected pairwise comparison, P < 0.05; Fig. 4A) and did not affect activity to 0.3 M sucrose (Bonferroni-adjusted pairwise comparison, P = 0.39; Fig. 4B) compared with constant-warming, which facilitated activity. Thus, warmth adaptation lessens or removes the attenuation in response to sucrose normally induced by cooling. These data indicate that oral adapting temperature, in addition to stimulus temperature, can significantly influence neuronal sensitivity to sucrose in medullary taste-sensitive neurons.

Fig. 4.

Effects of adaptation and stimulus temperature on firing rate to sucrose in S-type neurons. Means ± 1 SE firing rates (spikes/500 ms) to 0.1 (A) and 0.3 M (B) sucrose during each adaptation-stimulus temperature condition (*P < 0.05; ***P < 0.001). Firing rates for 0.1 M sucrose reflect average spike discharge per half-second for 1.5 s postlatency; firing rates for 0.3 M sucrose represent mean discharge per half-second for 2.5 s postlatency. Thermal conditions were C, W, W→C, and C→W. Error bars were normalized for between-neuron variability using the method of Cousineau (2005), whereas statistical comparisons were performed on uncorrected data.

Analysis of NaCl activity in N-type cells.

A separate analysis on temperature effects on firing rate to NaCl was performed for each NaCl concentration, as described in materials and methods. Cells were included in the analysis of data for one concentration only if they showed significant response latencies to that concentration on each of the four adaptation-stimulus temperature trials. Nine of the 26 N-type neurons that were recorded met this criterion for 0.004 M NaCl, whereas 25 N-type cells significantly responded to 0.1 and 0.3 M NaCl during each adaptation-stimulus temperature condition. The analyses of firing rates to 4, 100, and 300 mM NaCl that follow are based on these respective cell subsets.

Although increasing solution concentration decreased latency to first spike to NaCl (sign tests, P < 0.001), as described previously (Breza et al. 2010; Marowitz and Halpern 1977), latencies within each concentration did not vary across adaptation-stimulus temperature conditions (Bonferroni-adjusted sign tests, P > 0.04; Fig. 5). The lack of thermal influence on latency was distinct for NaCl compared with neural activity to sucrose, where lag to response was significantly modulated by temperature (Fig. 3) (cf. Wilson and Lemon 2014).

Fig. 5.

Latency to first spike to NaCl in N-type neurons. Bars represent median latencies (±68% confidence limits) for 0.004, 0.1, and 0.3 M NaCl across adaptation-stimulus temperature conditions. Conditions were C, W, W→C, C→W. Confidence limits were approximated using a bootstrap-resampling procedure (see materials and methods).

Firing rates to 0.004 M NaCl in N-type neurons were enhanced by cooling stimulus solutions, collapsed across adapting temperature and time (main effect of stimulus temperature, F_1,8 = 8.3, P < 0.05; Fig. 6A). There was no significant interaction between stimulus temperature and adaptation temperature or time on firing rates to 0.004 M NaCl (P > 0.05). Adapting temperature affected firing rate to 0.004 M NaCl, although the effect was time dependent (adapting temperature × time interaction: F_4,32 = 17.6, P < 0.001); cool adaptation increased the firing rate to 0.004 M NaCl compared with warm adaptation during the first 1.5 s of the stimulus response (Bonferroni-adjusted pairwise comparisons, P < 0.05; Fig. 6B).

Fig. 6.

Effects of stimulus temperature, adaptation temperature, and time on firing rates to NaCl in N-type neurons. A, C, and E: bar graphs show average responses (spike/500 ms ± 1 SE) to 0.004 (A), 0.1 (C), and 0.3 M (E) NaCl at cool and warm stimulus temperatures (*P < 0.05; ***P < 0.001). Firing rates represent mean spike discharge per half-second for 2.5 s following latency to activation. Data are collapsed across time window and adaptation temperature to represent the main effect of stimulus temperature, which was significant for activity to 0.004 and 0.3 M NaCl (see results). B, D, and F: point and line graphs depict the influence of adaptation temperature and time on average responses (spikes/500 ms ± 1 SE) to 0.004 (B), 0.1 (D), and 0.3 M (F) NaCl collapsed across stimulus temperature; a significant interaction between adapting temperature and time was found for unit firing to each concentration of NaCl (see results). Error bars were normalized for between-unit variability using the method of Cousineau (2005), whereas statistical comparisons were performed on uncorrected data.

Firing rates to 0.1 M NaCl in N-type cells were unaffected by change in stimulus temperature (ns main effect P = 0.3, ns stimulus temperature × time interaction P = 0.17; Fig. 6C), although spiking to this stimulus did vary with adapting temperature and time (adapting temperature × time interaction F_1.8,43.9 = 8.1, P = 0.001). Specifically, adaptation to warmth attenuated firing rates to 0.1 M NaCl compared with cool adaptation during the first second of the response (Bonferroni-adjusted pairwise comparisons, P < 0.01; Fig. 6D). The interaction between adaptation and stimulus temperature on firing rates to 0.1 M NaCl was not significant (P > 0.05).

Finally, firing rates to 0.3 M NaCl in N-type units were moderately elevated when stimulus solutions were warmed compared with cooled, collapsed across adaptation temperature conditions and time (main effect of stimulus temperature F_1,24 = 13.1, P = 0.001; Fig. 6E). There was no significant interaction between stimulus temperature and adaptation temperature or time on firing to 0.3 M NaCl (P > 0.05). On the other hand, adaptation temperature modulated firing rates to 0.3 M NaCl in a manner that relied on time (adaptation temperature × time interaction F_2.0,48.8 = 9.3, P < 0.001). Warmth adaptation attenuated spiking to 0.3 M NaCl relative to cool adaptation during the first 0.5 s of the response but enhanced firing to this input at 2.5 s postlatency (Bonferroni-adjusted pairwise comparisons, P < 0.05; Fig. 6F).

Mixtures of NaCl and amiloride were tested on an additional sample of 10 N-type cells to evaluate the potential contribution of amiloride-sensitive channels to thermo-gustatory activity to sodium. Amiloride eliminated significant firing to 0.004 M NaCl across adaptation-stimulus temperature conditions, as latencies were not detected on these trials (Table 1). On the other hand, latency to first spike was detected on all trials for 0.1 and 0.3 M NaCl with and without amiloride (Table 1). Factorial ANOVA applied to activity to 0.1 and 0.3 M NaCl in this group of N-type neurons revealed that the effect of stimulus temperature on neuronal firing varied with amiloride treatment (stimulus temperature × amiloride treatment interaction, F_1,9 = 8.58, P = 0.02) but was independent of adaptation temperature (ns stimulus temperature × adaptation temperature interaction, P = 0.11) and concentration (ns stimulus temperature × concentration interaction, P = 0.07). Expectedly, application of amiloride substantially reduced responses to 0.1 and 0.3 M NaCl (main effect of amiloride, F_1,9 = 78.7, P < 0.001). Although activity to these stimuli was enhanced when solutions were warmed compared with cooled (Bonferroni-adjusted pairwise comparisons, P ≤ 0.001; Fig. 7A), the difference in firing between warm and cool NaCl under control conditions (22.4 spikes difference ± 3.9 SD; Fig. 7B) was reduced (dependent-samples t-test, t₉ = 3.0, P < 0.05) in the presence of amiloride (9.4 spikes difference ± 2.0 SD). Residual thermo-gustatory activity to NaCl on amiloride trials was due to an early, abbreviated response to NaCl (Fig. 7C), which was likely contributed by transient competition between Na⁺ and amiloride for receptor access, as oral adaptation to amiloride, as opposed to water, entirely abolished responses to NaCl (Fig. 7D). Assuming a peripheral locus of effect, the above results suggest that amiloride-sensitive channels wholly mediate temperature effects on gustatory activity to NaCl in N-type neurons.

Table 1.

Median latencies to 1st spike for all thermoconcentration solutions of NaCl mixed with (A+) and without (A−) amiloride

	Cool Adapting Temperature				Warm Adapting Temperature
	Cool stimulus		Warm stimulus		Cool stimulus		Warm stimulus
NaCl, M	A+	A−	A+	A−	A+	A−	A+	A−
0.004	Ø	1.39	Ø	1.16 (9)	Ø	1.24 (7)	Ø	1.04 (6)
0.1	0.79	0.66	0.78	0.67	0.67	0.64	0.68	0.69
0.3	0.58	0.46	0.57	0.51	0.48	0.40	0.48	0.44

Data are based on 10 N-type neurons. Parenthetical values give the n of cells that showed significant latency to 1st spike and were used to compute the median; otherwise, n = 10. Ø, Zero cells showed significant latency to respond to NaCl.

Fig. 7.

The influence of amiloride on thermo-gustatory activity to NaCl in an additional sample of N-type neurons (n = 10). A: average firing (spikes/3.5 s ± 1 SE) to cool (25°C) and warm (35°C) NaCl mixed with (A+) and without (A−) 30 μM amiloride (***P ≤ 0.001). Data are collapsed across concentration (0.1 and 0.3 M) and adaptation temperature to represent the interaction between amiloride condition and stimulus temperature on the magnitude of the NaCl response (see results). B: mean difference (± 1 SE) in firing between cool and warm NaCl in the A+ and A− conditions in A (*P < 0.05). Error bars in A and B were normalized for between-neuron variability using the method of Cousineau (2005), whereas statistical comparisons were performed on uncorrected data. C: mean time course of firing to 0.004, 0.1, and 0.3 M NaCl across cells during the A+ (dashed line) and A− (solid line) conditions. Grayed areas surrounding each line denote ± 1 SE. D: raw electrophysiological traces showing activity by one N-class neuron to cool taste stimuli following oral adaptation to cool water (left), to 0.1 and 0.3 M NaCl (cool) following adaptation to a cool amiloride solution (middle), and the response of this cell to 0.3 M NaCl mixed with and following oral adaptation to amiloride (right). Upward and downward arrowheads indicate stimulus onset and offset, respectively.

DISCUSSION

An influence of adapting temperature on central gustatory activity to sucrose.

Building on prior data that showed stimulus temperature can influence neural activity to sucrose (Breza et al. 2006; Lu et al. 2012; Talavera et al. 2005; Wilson and Lemon 2013, 2014; Yamashita et al. 1970; Yamashita and Sato 1965), the present work describes a novel effect of variation in oral adaptation temperature on neural sensitivity to this stimulus. Different adapting rinse temperatures were tested in some of the prior electrophysiological work on thermal influence on sucrose activity, although sucrose solutions were also adjusted to and tested at rinse temperature (Breza et al. 2006; Lundy and Contreras 1999; Talavera et al. 2005; Yamashita and Sato 1965). Thus, separating the effects of adapting from stimulus temperature was precluded in these studies, as both were presumably equal. By testing adapting rinse and sucrose solutions at multiple isothermal and anisothermal temperatures, here we show that change in adapting temperature can influence neural activity for sucrose taste. Cooling sucrose solutions to 25°C substantially decreased firing rates to 0.1 and 0.3 M sucrose in the NTS compared with warming, but only when the mouth was cool-adapted to 25°C (Fig. 4). Following warm adaptation to 35°C, firing rates to both cooled and warmed solutions of sucrose were relatively enhanced and only moderately different (0.1 M) or indistinguishable (0.3 M) from one another. Thus, warm adaptation largely counteracted the attenuation in magnitude and also onset (Fig. 3) of gustatory activity to sucrose imposed by cooling the stimulus solution.

Notwithstanding periods of taste delivery, the thermal adapting rinse was continuously applied to the mouth during data acquisition trials and intertrial periods, which were ∼2–3 min for both cool and warm adaptation conditions. The surface temperature of the tongue in rodents can reach a steady reduced or elevated value after <1 min of adaptation to chilled (e.g., 10°C) or heated (e.g., 40°C) water, respectively (Yamashita and Sato 1965). Thus, the present method of continuous thermal adaptation might have effectively changed the temperature of the oral epithelial surface. It is important to note that temperature measurements in the present work indexed solution temperature at the moment of oral delivery and cannot be interpreted as direct measurements of oral epithelia temperature. Nonetheless, a change in the thermal environment of taste receptor cells agrees with a potential mechanistic explanation of the present effects involving warmth potentiation of TRPM5-dependent receptor pathways for sucrose (Talavera et al. 2005). It is possible that these pathways sensitized during extended stimulation with 35°C under warm adaptation, increasing the level of membrane depolarization in taste cells. This postulate agrees with our observation that prestimulus spiking activity in NTS neurons was higher under warm adaptation as opposed to cool. Heightened taste cell depolarization could in turn have led to enhanced firing to sucrose in downstream S-type neurons in the brain stem, even when sucrose was cooled to near-ambient temperature. Analogous sensitization effects of warmth adaptation were reported in psychophysical studies on sucrose taste perception. For instance, although cooling can reduce the intensity of sucrose perception (Bartoshuk et al. 1982; Calvino 1986; Green and Nachtigal 2012), only “slight and inconsistent” decrements in the sweetness of cooled (20°C) sucrose were reported in humans when the tongue was thermally controlled and warm-adapted to 36°C (Green and Frankmann 1987). These data, along with the present findings, indicate that adaptation temperature importantly influences gustatory responses to sucrose and that warming the mouth can, in some cases, facilitate activity to sucrose independently of stimulus temperature.

For humans and also mice, the temperature of oral epithelia would be expected to hold near-constant value when the mouth is closed (cf. Green 1984; Green and Gelhard 1987), and the 35°C adapting rinse used presently intended to target a normal closed-mouth temperature approximated for mammals (cf. Green 1986). On the other hand, the 25°C adapting rinse was at a temperature lower than closed-mouth, although oral epithelia may at times display decreased temperatures upon opening of the mouth and during the consumption of cooled ingesta (cf. Green 1986; Pangborn et al. 1970), as discussed above. What is more is that dipping the tongue into a 37°C solution of sucrose following adaptation to sucrose at 21°C counteracts the attenuation in sweetness associated with the cooled solution (Green and Nachtigal 2012). Such “rewarming” of the tongue was interpreted to explain in part the phenomenon where tasting sweet stimuli with the tongue outside the mouth, conducive to evaporative cooling of lingual epithelia, produces a reduced sensation of sweetness that quickly grows in magnitude once the tongue is retracted into the mouth, which would rewarm the tongue (Green and Nachtigal 2012). Thus, the temperature of oral epithelia can seemingly vary with ingestive behaviors and shape neural and perceptual responses to sucrose taste.

Warmth adaptation and stimulus warming decreased latency to fire to sucrose in S-type NTS neurons (Fig. 3), similar to previous findings (Wilson and Lemon 2014). As discussed above, a physiological mechanism supporting this effect may involve an influence of temperature on the TRPM5 component of taste receptors for sucrose. Current flow through TRPM5 at positive voltage appears to increase with warming from ∼25°C and lower temperatures to 35°C (Talavera et al. 2005). This suggests that a depolarizing step in the membrane potential of taste receptor cells harboring TRPM5 would result in greater inward current flow at warm temperatures compared with cool, potentially leading to more rapid cellular activation and signal transmission to follower neurons. Yet the contribution of TRPM5 to temperature effects on the timing of sucrose activity awaits investigation. It is noteworthy that the lack of thermal influence on latency to fire to NaCl in N-type units may reflect the disparate receptor process involved with transduction of this stimulus.

Although TRPM5 emerges as a likely and parsimonious molecular candidate for mediating thermal influence on gustatory activity to sucrose, this postulate is not without caveat. In addition to its role in sucrose taste, the warmth-sensitive (Talavera et al. 2005) TRPM5 ion channel also contributes critically to the transduction of select bitter-tasting stimuli, including the bitter prototype quinine (Damak et al. 2006; Zhang et al. 2003). This reliance of quinine taste on heat-sensitive TRPM5 predicts that temperature should impact gustatory responses to quinine, which is along the vein of the marked influence of temperature on activity to sucrose if we assume that thermal action on this input arrives through TRPM5 (e.g., see Lu et al. 2012; Talavera et al. 2005; and Wilson and Lemon 2014). However, whereas change in stimulus temperature from cool (22°C) to warm (37°C) values can induce supralinear change in NTS unit activity to oral delivery of sucrose, such an effect is not found with quinine, a stimulus that elicits taste responses with comparably low sensitivity to temperature (Wilson and Lemon 2013). The nonuniform effect of temperature across gustatory responses to sucrose and other stimuli whose transduction involves TRPM5 suggests that other mechanisms may contribute to thermal modulation of sucrose activity. One possibility is that temperature may exert unique influence on the receptor proteins involved with sucrose taste transduction (cf. Green and Nachtigal 2012; Wilson and Lemon 2013), although this remains to be experimentally studied. Other potential mechanisms could include activation of central gustatory sensitization and habituation processes (cf. Di Lorenzo and Lemon 2000) that are engaged through extended oral stimulation with warm and cool adapting temperatures. Moreover, the NTS receives projections from the trigeminal (V) nerve (e.g., Blomquist and Antem 1965; Contreras et al. 1982; Corson et al. 2012; Hamilton and Norgren 1984; Marfurt and Rajchert 1991; Whitehead and Frank 1983), which conveys oral pain, touch, and temperature sensation. NTS neurons implicated for contributing to gustatory sensation have been shown to receive synaptic input from mandibular processes of the trigeminal nerve (Boucher et al. 2003; Braud et al. 2012; Felizardo et al. 2009), suggesting that these neurons may also combine information about gustation and oral somatosensation originating along anatomically and presumably functionally distinct afferent routes. It is possible that thermal stimulation of trigeminal pathways and the convergence of V input onto NTS neurons may have contributed in part to the present results, although this was not directly tested. The present data show that adaptation temperature can markedly influence gustatory activity to sucrose in the brain stem, although additional studies are needed to delineate mechanisms that support the observed effects.

Temperature and central gustatory activity to sodium.

The influence of adaptation and stimulus temperature on responses to NaCl in the present sample of N-type NTS neurons progressively varied with stimulus concentration. Firing rates to 0.004 M (perithreshold) NaCl were enhanced by stimulus cooling and also cool adaptation of oral epithelia, which markedly raised activity to 4 mM NaCl during an extended period of the taste response (Fig. 6). Increasing NaCl concentration to 0.1 M induced responses that were relatively less susceptible to change with adaptation temperature and unaffected by cooling or warming of stimulus solutions. On the other hand, responses to 0.3 M NaCl were moderately enhanced by stimulus warming, with warmth adaptation also facilitating activity to this input during a late window of the taste response; across concentrations, this was the only instance where warmth adaptation enhanced unit firing to NaCl. Thus, warming could enhance unit firing to concentrated NaCl, whereas cooling facilitated neural responses to low-intensity solutions of NaCl. Considering perithreshold NaCl, the increased firing to this input under stimulus cooling and cool adaptation may suggest gustatory adaptation, or habituation, to weak intensities of NaCl decreases with decreasing temperature (cf. Fig. 6, A and B). This contrasts with the present and prior data on thermo-gustatory activity to sucrose, where firing to low (0.1 M; Fig. 4A) and perithreshold (0.05 M) (Wilson and Lemon 2014) intensities of sucrose is increased by warming, not cooling, possibly reflecting a decrease in the rate of gustatory adaptation to sucrose with rising temperature (cf. Green and Nachtigal 2012).

The gustatory signal for NaCl carried by “sodium-best” NTS neurons arises from taste receptor processes sensitive to blockade by the ENaC antagonist amiloride (Boughter et al. 1999; Scott and Giza 1990; Smith et al. 1996; St. John and Smith 2000). In confirmation, our experiments showed that thermo-gustatory activity to 0.004, 0.1, and 0.3 M NaCl in N-class cells was inhibited or blocked when stimuli were mixed with amiloride and abolished when such mixtures were tested following adaptation of oral epithelia to amiloride (Fig. 7). Thus, assuming a peripheral locus of effect, enhancement, or suppression of responses to NaCl in N-class units due to cooling and warming, as reported here, would intuitively involve an influence of temperature on amiloride-sensitive channels. Along this line, Nakamura and Kurihara (1988) used whole nerve recordings from the chorda tympani (CT) nerve, which supplies taste and oral somatic sensation to the anterior tongue, to reveal that peripheral gustatory activity to NaCl is composed of two amiloride-sensitive components, each with a dependence on temperature and Na⁺ concentration that parallels the present effects of these parameters on central activity to NaCl. In their study, amiloride-sensitive CT responses to a reduced concentration (0.01 M) of NaCl peaked when solutions were cooled to ∼10°C. On the other hand, amiloride-sensitive nerve activity to an elevated NaCl concentration (0.3 M) peaked at ∼30°C. Amiloride-sensitive activity to an intermediate concentration (0.1 M) of NaCl showed, accordingly, two approximately equal peaks, one at ∼10°C and another at 30°C. This result was interpreted as evidence for two amiloride-blockable taste receptor sites for NaCl (Nakamura and Kurihara 1988), one tuned to low concentrations of Na⁺ when cooled, and another maximally stimulated by high concentrations of Na⁺ when warmed. A similar “double effect” of temperature on amiloride-sensitive receptor processes for sodium was described in data showing that amiloride-sensitive fibers of the CT nerve were divisible by whether they displayed heightened activation to NaCl tested at room temperature or cooled to ∼12°C (Ninomiya 1996).

Biophysical studies have revealed a marked influence of temperature on amiloride-sensitive Na⁺ current mediated by ENaC. At physiological temperatures (e.g., 35°C), this current shows a rapid increase to “peak” amplitude that is followed by a fast relaxation to a decaying, tonic value (Chraibi and Horisberger 2002). The reduced tonic current is thought to reflect a form of recurrent inhibition of Na⁺ current flow through ENaC induced by multiple potential mechanisms (cf. Chraibi and Horisberger 2002; Gilbertson and Zhang 1998). Recurrent inhibition of Na⁺ current also arises in taste receptor cells that show amiloride-sensitive responses to Na⁺ salts (Gilbertson and Zhang 1998). When ENaC is cooled to ∼25°C, amiloride-sensitive Na+ current retains its rapid onset-to-peak characteristic, although the rate of decline after the peak slows, resulting in heightened tonic current compared with warm conditions (Chraibi and Horisberger 2002). Further reductions in temperature (e.g., 12°C) can largely remove the self-inhibitory effect of Na⁺ on ENaC, rendering similarities, or indifference, between the peak and tonic components of amiloride-sensitive responses (Chraibi and Horisberger 2002, 2003). Thus, cooling ENaC increases tonic current during stimulation with Na⁺, whereas warming tends to decrease this current.

Cooling-induced enhancement of responses to low concentrations (≤0.01 M) of NaCl in amiloride-sensitive peripheral (Nakamura and Kurihara 1988) and central (Fig. 6) taste pathways agrees with the known effects of cooling on ENaC, where decrements in temperature can enhance the amplitude of tonic current and Na⁺ passage through this channel (Askwith et al. 2001; Awayda et al. 2004; Chraibi and Horisberger 2002, 2003). On the other hand, warmth enhancement of responses to elevated concentrations (0.3 M) of NaCl in amiloride-sensitive circuits, as shown previously (Nakamura and Kurihara 1988) and presently (Fig. 6), seems at odds with ENaC, which displays reduced amiloride-blockable Na⁺ currents with warming above room temperature, as described above. Yet some reports have shown that, although cooling can indeed markedly raise the open probability of ENaC, warming can increase the conductance of this channel, which is indicative of a dual effect of temperature (Chraibi and Horisberger 2003). Although it is cumbersome to directly compare the influence of temperature on ENaC current in vitro with the in vivo performance of gustatory-sensitive neurons implicated to receive sodium taste input mediated by this channel, the available biophysical data suggest that temperature can certainly exert complex effects on ENaC function. Such effects may contribute to the present and previously observed (Nakamura and Kurihara 1988; Ninomiya 1996) dual action of temperature on gustatory activity to NaCl in amiloride-sensitive taste pathways.

It is important to consider that the amiloride-insensitive TPRV1 ion channel is also implicated for sodium transduction and responses in taste-sensitive nerves and contributes potentiated activity to high concentrations of Na⁺ paired with heat (Lyall et al. 2004). Such sensitivity might render TRPV1 as a candidate receptor contributing to warmth facilitation of gustatory activity to concentrated NaCl, as in Fig. 6. However, the present (Fig. 7) and prior data (Boughter et al. 1999; Scott and Giza 1990; Smith et al. 1996; St. John and Smith 2000) suggest that Na⁺ input to N-class neurons arrives exclusively through ENaC.

A noteworthy study by Yamashita et al. (1970) evaluated how abrupt change in tastant temperature during gustatory stimulation would impact the continuation of taste responses recorded from single CT nerve fibers. In this work, a sudden drop in stimulus temperature from 30°C to 10°C augmented ongoing unit activity to a reduced concentration (0.03 M) of NaCl in select fibers, whereas a sudden increase in stimulus temperature from 10°C to 30°C inhibited responding to this stimulus. Also revealed were a small number of fibers where abrupt warming potentiated ongoing activity to elevated concentrations of NaCl, including 0.3 M, as used presently. There are similarities between the results of this study and the present, although important procedural differences exist, including different adaptation conditions. Units were adapted to temperature-adjusted taste stimuli in Yamashita et al. (1970), which presumably promoted gustatory adaptation, as opposed to temperature-controlled water in the present effort. Along this line, Yamashita et al. (1970) showed that ongoing activity by select CT fibers to 0.1 M NaCl could be inhibited by abrupt warming of the stimulus, although the present data revealed that warming stimulus solutions did not affect, across adapting temperatures, responses to 0.1 M NaCl in N-type neurons (Fig. 6C). Other single-fiber studies on temperature-taste interactions in the CT nerve that used thermal, rather than taste, adaptation showed no effect of temperature on responses to 0.1 M NaCl in sodium-best units (Breza et al. 2006; Lundy and Contreras 1999). Although it is important to consider that Yamashita et al. (1970) classified units based on thermal rather than gustatory tuning, the above results taken together suggest that the level of thermosensory and also chemosensory adaptation of sodium receptors may guide how temperature impacts gustatory activity to NaCl.

Final considerations.

There are limited extant data on the frequency by which thermal sensitivity arises in taste-sensitive neurons in the mammalian NTS, with sometimes inconsistent findings reported across these studies. An electrophysiological description of the lamb NTS revealed that ∼30% of sensory neurons in this nucleus are sensitive to oral and epiglottal stimulation with temperature (Sweazey and Bradley 1989). Further work on rats showed that ∼33% of taste-sensitive NTS neurons were sensitive to change in oral temperature, with cooling, but not warming, inducing excitation in firing in these cells (Travers and Norgren 1995). On the other hand, other electrophysiological investigations discovered that a marked majority of gustatory-sensitive neurons in rodent NTS can respond with excitation to change in oral temperature, including cooling and also warming of oral epithelia (Ogawa et al. 1988; Wilson and Lemon 2013). Differences in findings across these studies may relate to methodological discrepancies. Along this line, the present and recent (Wilson and Lemon 2013, 2014) data have further revealed an extended receptive range of taste-sensitive NTS neurons to particular combinations of gustatory stimulation and temperature. Temperature-induced change in firing to gustatory input in NTS neurons can markedly surpass responses by these units to oral thermal stimulation alone (Wilson and Lemon 2013, 2014). Thus, thermal sensitivity by taste-sensitive neurons may be effectively indexed, in some cases, only through tests that include pairings of temperature input with taste stimuli.

It is also important to consider that functional interactions between temperature and gustation partly show species dependence. Although the present and other data from rodents, as described above, indicate that temperature modulates taste responses to sugars and salts, a recent electrophysiological study in Manduca hornworms revealed that peripheral taste responses to K⁺ salts and select sugars, including sucrose, showed no change with warming and cooling of receptors, although sensitivity to an aversive stimulus was found to rely on temperature (Afroz et al. 2013). Considering other arthropods, Drosophila display neural and behavioral sensitivity to warmth mediated by a taste receptor protein (Ni et al. 2013) expressed in gustatory-sensitive neurons implicated in avoidance function (Thorne and Amrein 2008). Such cells may combine thermal and aversive gustatory input (Montell 2013). Temperature can modulate taste responses to aversive stimuli, such as acids and bitters, in mammalian neurons (Breza et al. 2006; Lundy and Contreras 1999; Yamashita and Sato 1965), although, as described above, the effect of temperature on these responses can be markedly less pronounced compared with its influence on appetitive taste signals (Wilson and Lemon 2013). Different patterns of thermo-gustatory interaction across species may have in part an ecological basis (Afroz et al. 2013).

In closing, the present data continue to build on the literature indicating that temperature is an underlying and systematic parameter of the neural processing of taste in mammals. Given its omnipresent nature during taste experience and marked ability to modulate gustatory signals, as shown here and in other studies referenced above, temperature should be considered in equal light to stimulus concentration and quality as a modifier of gustatory activity. Future studies that test extended concentration series of tastants across several adapting and stimulus temperatures may precisely describe how change in the balance between these temperatures modulates taste activity and perception. It is conceivable in humans that this balance may dynamically and rapidly shift during the consumption of flavorful foods and drinks served warmed and chilled.

GRANTS

This work was supported by National Institutes of Health Grant DC-011579 to C. H. Lemon.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.L. performed experiments; J.L. and C.H.L. analyzed data; J.L. and C.H.L. interpreted results of experiments; J.L. and C.H.L. prepared figures; J.L. drafted manuscript; J.L. and C.H.L. edited and revised manuscript; J.L. and C.H.L. approved final version of manuscript; C.H.L. conception and design of research.