Altering salivary protein profile can decrease aversive oromotor responding to quinine in rats

Abstract

Bitter taste is often associated with toxins, but accepting some bitter foods, such as green vegetables, can be an important part of maintaining a healthy diet. It has previously been shown that animals exposed to quinine upregulate a set of salivary proteins (SPs), and those with upregulated SPs have increased rates of feeding on a quinine diet as well as increased brief-access licking to and higher detection thresholds for quinine. These studies suggest that SPs alter orosensory feedback; however, they rely on SPs upregulated by diet exposure and cannot control for the role of learning. Here, we use taste reactivity to determine if SPs can alter bitter taste in animals with no previous bitter diet experience. First, saliva with proteins stimulated by injections of isoproterenol and pilocarpine was collected from anesthetized rats; this “donor saliva” was analyzed for protein concentration and profile. Bitter-naïve rats were implanted with oral catheters and infused with taste stimuli dissolved in saliva that contained all of the SPs from the donors, saliva that was filtered of SPs, water, or artificial saliva. Their orofacial movements were recorded and quantified. We found that presence of quinine increased movements associated with aversive stimuli, but adding SPs to the infusion was sufficient to reduce aversive oromotor responding to quinine. The effect was dependent on the total protein concentration of the saliva, as protein concentration increased aversive responses decreased. Additionally, infusions of whole saliva altered aversive responding to quinine, but not other stimuli (citric acid, NaCl, sucrose). Our work suggests that effect of these SPs is specific and the presence of SPs is sufficient to decrease aversive orosensory feedback to bitter stimuli.

1. Introduction

Variation in bitter acceptance is subject to extensive research - this work is important, because bitter sensitivity has functional health relevance. A healthy diet often contains foods that are bitter, e.g., many green vegetables. In addition to being an important source of fiber, vitamins, and nutrients [1–4], plants can also be sources of beneficial phytochemicals [5–11] some of which can be perceived as bitter. Additionally, there appears to be a relationship between plant consumption and rate of weight gain [12] and individuals that consume more plants (fruit and vegetables) have lower body mass indexes [13]. However, while there is no controversy about the benefits of fruit and vegetable consumption, and there are many interventions designed to increase intake in both pediatric and adult populations [14–17], it is difficult to increase intake of these foods, in part because of their sensory properties [16,18]. One of the most common interventions to increase intake and palatability of vegetables is to incorporate these foods into the diet through repeated exposure [18,19], which is thought to increase acceptance through a learning process, like learned safety or evaluative conditioning [20,21]. A similar process has been suggested as a mechanism for habituation to alcohol, another substance that is bitter (in addition to having other flavor and taste qualities). Like vegetables, repeated exposure to alcohol increases the palatability of the solution [22,23]. In addition to the proposed learning models for increased acceptance of foods after repeated exposure, our previous work has suggested that there is a physiological shift during repeated exposure to bitter that may contribute to increased acceptance.

We have demonstrated that bitter diet exposure drives salivary protein (SP) alteration, which we have proposed promotes increased bitter acceptance [24–27]. We have demonstrated that the induction of SPs is predictive of changes in feeding behaviors associated with or-osensory and postingestive feedback, as well as detection thresholds for quinine [25]. We have also demonstrated there is reduced chorda tympani nerve activity when quinine is applied to the tongue in saliva containing proteins (compared to saliva without proteins) [24]. Together, these data suggest that SPs increase acceptance of bitter stimuli by reducing sensitivity to them [25].

In our previous work, SP profiles were altered primarily though extended exposure to a bitter diet. However, to isolate the effect of the protein from the learning models described above, one must control for bitter experience. To control for bitter experience, we and others have used injections of isoproterenol, a beta-adrenergic agonist that stimulates SP production, [24,28]. Unfortunately, the necessary doses of the drug cause cardiotoxicity in addition to altering SP profile, which limits the utility of the treatment. To separate the possible contribution of experience from that of SPs, and to avoid the potential side effects of isoproterenol injections in awake rats, we have developed a “donor saliva” method was suggested by a reviewer and we were instructed to make that change during the proofs process. Saliva is collected from one group of rats and delivered to a group of rats that are naïve to the induction process, which allows us to eliminate the possible effects of bitter diet experience on orosensory responding. We paired our donor saliva method with the taste reactivity paradigm, which quantifies the reflexive and taste-guided oral motor responses to taste stimuli after an animal is given an oral infusion of a stimulus [29–34]. For example, rats infused with a sweet stimulus display stereotypical ingestive responses like rhythmic mouth movements and tongue protrusions, while rats given a bitter stimulus display rejection responses like gapes, paw flails, and chin rubs. Pairing donor saliva with taste reactivity allows us to observe effects of SPs on orosensory responding, while controlling for the animals’ diet experience. Additionally, stimuli are delivered regardless of the animals’ motivation to approach them, so rats can be tested while food- and water-replete. Using the taste reactivity paradigm, we conducted three experiments to clarify the effect of SPs on orosensory responses to taste stimuli. First, to determine if SPs alter oromotor responding to quinine, we measured oromotor responses while infusing quinine into the oral cavity in 1 of 4 diluents: whole saliva (donor saliva, containing proteins), filtered saliva (donor saliva with proteins removed), artificial saliva, or water. Second, to determine if there was a concentration-dependent effect of the proteins, we infused quinine in a range of SP concentrations. Finally, to determine if SPs alter oromotor responding to other taste qualities, we infused citric acid, NaCl, quinine and sucrose in either whole saliva or artificial saliva.

2. Materials and methods

2.1. Subjects

Infusion subjects were 28 adult male Long Evans rats (Charles River Breeding Laboratory, Raleigh, NC), weighing 150–250 g at study onset. The colony room was maintained at 20 ± 2 °C with a 12:12 h light/dark cycle. All training and testing was performed during the lights on phase. All animal procedures were approved by the University at Buffalo Animal Care and Use Committee.

2.2. Donor saliva collection and preparation

Donor saliva was collected and prepared similarly to our previously described protocol [26]. Animals were anesthetized with 2–4% isoflurane, and injected subcutaneously with isoproterenol (0.2 mg/ml) to stimulate protein production and pilocarpine (2 mg/ml) to stimulate salivary flow [35]. Saliva was aspirated from around the tongue and ventral cheek using a 200 μl pipette. Rats (n = 20) produced 1–3 ml of saliva per collection, with each sample ranging from 10–15 mg/ml of protein. This concentration is much higher than what we have recorded in the absence of the drug (3 mg/ml) [26]. Saliva samples were combined into a single homogenous sample, referred to as donor saliva. The donor saliva was then diluted with artificial saliva (0.015 M NaCl, 0.022 M KCl, 0.003 M CaCl², and 0.0006 M MgCl²; pH 5.8 ± 0.2) to 3 mg/ml total protein content. Artificial saliva was used instead of water to maintain the ionic composition of the saliva.

Donor saliva was used for experiments either as whole saliva or filtered saliva. While whole saliva was not prepared further, filtered saliva was obtained by filtering a subset of the donor saliva with a 30 μm Microsep Advance Centrifugal Device (Pall Corporation) and centrifugation (12,000 × g for 15 min) to remove the proteins. We confirmed that filtration removed all proteins from the donor saliva with the BCA protein assay method (Pierce Protein Biology Products), which showed that our filtered saliva sample was below the measureable range (<20 μg/ml total protein), and gel electrophoresis (see Fig. 1), as previously described [26].

Fig. 1. Comparison of whole and filtered saliva.

Molecular weight markers (MW) and samples from donor saliva either unfiltered (W, whole saliva) or filtered of all proteins (F). Infusions for all experiments used the same whole and filtered saliva stocks, which were derived from a single, homogenized source (donor saliva).

2.3. Intraoral catheterization

Intraoral cannulae were implanted bilaterally [29, 36]. Surgical procedures were conducted using sterile technique while rats were anesthetized via 2–4% isoflurane. A 15 G needle was inserted under the skin at the dorsal surface of the neck near the shoulder blades, and run subcutaneously until it emerged inside the oral cavity, rostral to the first maxillary molar. A length of PE-90 tubing (Intramedic) was then threaded through the barrel of the needle until it emerged in the oral cavity, after which the needle was withdrawn while the tubing remained on the needle track. The tubing externalized at the shoulder blade was threaded through a small length (~3 cm) of Silastic™ tubing (Dow Corning), and the ends of tubing were threaded through Teflon washers and heat flared to secure on both the external and oral sides. Following surgery and for 3 days post-operatively, rats were injected with carprofen (5 mg/kg, analgesic) and enrofloxacin (5 mg/kg, antibiotic). To minimize the risk of clogging or infection, the cannulae were flushed daily with saline. Rats were allowed 3 to 5 days for recovery before habituation and training.

2.4. Infusion

Animals were trained and tested in a clear cylindrical infusion chamber with a Plexiglas floor, in a paradigm which has been previously used to accurately monitor taste reactivity in rats [30,33,34,37]. The chamber is held by an open frame and a camera placed below the chamber is aimed upwards, directly at the floor of the chamber. Animals were habituated to the testing paradigm for 5 days. On the first day, rats were placed in the chamber for 30 min. On the second day, rats sat in the chamber for 30 min, after which a single-syringe pump (Harvard Apparatus) was run for 100 s. During the session, the animals were not attached to any tubing; the purpose of the session was to expose them to the noise of the pump. For the next 3 days, rats were placed in the chamber for 30 min, after which they were attached to the infusion line and given an infusion of DI water (0.3 ml/min). On these days, the rat had 1 of its 2 cannulae connected via PE tubing to syringe on the infusion pump through a swivel centered in the lid of the behavioral chamber. On testing days, a total of 0.5 ml of solution was infused, at a rate of 0.3 ml/min (infusion length of 100 s). Quinine, citric acid, and NaCl (Sigma Aldridch) or sucrose (Walmart) were mixed using a vortex into their respective solvents for 30 s, immediately prior to infusion.

Experiment 1: Do SPs alter oromotor responding to quinine?

Our first group of animals received DI water, artificial saliva, filtered saliva, whole saliva, and each of the previous solutions with 1 mM quinine added, the highest concentration we have previously used in behavioral testing. Each animal (n = 6) was exposed to each condition, presented in random order. The comparison between filtered saliva and whole saliva was used to identify the importance of SPs in altering taste reactivity, as whole saliva and filtered saliva should have matching ionic profiles. We used artificial saliva as a control because it has previously been used in behavioral and physiological studies as a facsimile of saliva [38].

Experiment 2: Do SPs have a concentration-dependent effect on oromotor responding to quinine?

Because we did not see any differences between artificial saliva and filtered saliva in our first experiment, we chose to focus only on whole saliva and artificial saliva with our second and third groups. Using artificial saliva in place of filtered saliva allowed us to reduce the amount of donor saliva needed, as it is a limited resource. In our second group, rats (n = 6) were infused with 1 mM quinine dissolved in artificial saliva or whole saliva, at several protein concentrations (3, 1.5, 0.75, and 0.375 mg/ml) in random order, to determine if the suppression effect was concentration dependent.

Experiment 3: Does the presence of SPs alter oromotor responding to diverse stimuli?

To determine SPs can affect responding to diverse taste stimuli, a third group of rats (n = 8) were infused with quinine (0.1 mM, 0.3 mM, 1 mM), NaCl (300 mM, 500 mM), or sucrose (30 mM, 100 mM) in artificial saliva or whole saliva, in random order. These same rats were also infused with citric acid (10 mM, 20 mM), which was delivered in artificial saliva, whole saliva, or filtered saliva. We chose to use filtered saliva with 10 mM concentration of citric acid because we thought there may be differences in the buffering capacity of artificial saliva and donor saliva. We chose multiple concentrations of each stimulus using brief-access licking curves from the literature [39]; this was done to ensure that we captured both high and low responding, and to help assure that any negative results were not because a given stimulus concentration was not strong enough to elicit sufficient oromotor responding, or because the stimulus concentration was too strong to be affected by our SPs.

2.5. Scoring and statistical analysis

Videos were recorded with a GoPro Hero Session 4 and uploaded daily to a Dell Optiplex 7020 (Windows 7 OS), for analysis. For each video, 30 s of in-view behavior (the first 30 s following the first TR behavior) was scored as defined by Spector, Breslin, and Grill [34]. Rearing was noted, but not considered in-view behavior, and was not included in the analysis. Ingestive responses included: rhythmic mouth movements, tongue protrusions, lateral tongue protrusions, and paw licks. Aversive responses included: gapes, chin rubs, forelimb flails, face washes, and head shakes. Passive drip was also scored, but considered a neutral response. Videos were scored at 0.2x speed using Wondershare Filmora 8 video editing software, by 2 independent, blind scorers. Scores for each video were then compared/compiled by a third blind scorer, who re-scored the video if any measure differed by 10%. Aversive movements and ingestive movements were totaled by summing the number of each behavior listed above.

Anderson-Darling tests for normality were conducted on the data. Two datasets were found to have left-skewed distributions (quinine aversive responses, experiment 1, and NaCl aversive responses, experiment 3). These data were subjected to a square-root transformation, after which they conformed to normal distribution (p’s > 0.12).

In Experiment 1, we used repeated-measures ANOVAs to compare orofacial responses (ingestive, aversive, passive drip). In each ANOVA, stimulus (with or without quinine) and solvent (whole saliva, filtered saliva, artificial saliva, and water) were within-subject factors. Following the ANOVAs, planned comparisons were done using Bonferroni-corrected t-tests. Comparisons tested for the effect of quinine in each solvent, and follow up ANOVAs were run to compare between solvents. In the second experiment, protein concentration was used as a within-subject factor in a repeated measures ANOVA. We also used a simple linear regression to further describe the relationship between protein concentration and number of aversive movements. In the third experiment, we used repeated-measures ANOVAs to compare orofacial responses (ingestive, aversive, passive drip) for each taste stimulus (quinine, citric acid, salt, sucrose). In each ANOVA stimulus concentration (2 concentrations for citric acid, NaCl, and sucrose, 3 concentrations for quinine) and solvent (whole saliva and artificial saliva for quinine, NaCl, and sucrose, whole saliva, filtered saliva, and artificial saliva for citric acid) were within-subject factors. Following the ANOVAs, planned comparisons were done using Bonferroni-corrected t-tests: for quinine and NaCl, comparisons were made to test for the effect of stimulus concentration. For citric acid, comparisons were made to test for the effect of solvent type. P-values are reported after Bonferroni correction. Statistical comparisons were conducted in Systat 12.

3. Results

3.1. Experiment 1: are SPs sufficient to alter oromotor responding to quinine?

There was a significant interaction between the presence/absence of quinine and the solvent type on aversive oromotor responses (F (3,40) = 9.18, p = 0.005), as well as a significant main effect of quinine addition (F(1,40) = 39.10, p < 0.001, Fig. 2A). Post-hoc testing compared responses with and without quinine for each solvent; they revealed that rats increased aversive oromotor movements when quinine was added to filtered saliva (t = 5.01, p = 0.004), artificial saliva (t = 4.35, p = 0.007), and water (t = 4.59, p = 0.006), but did not increase when quinine was added to whole saliva (t = 0.19, p = 0.858). Responding to whole saliva, filtered saliva, artificial saliva, and water was not different at baseline (without quinine, F (3,20) = 0.62, p = 0.610) and aversive movements for quinine in filtered saliva, artificial saliva, or water were not different from each other (F(2,15) = 3.08, p = 0.076).

Fig. 2. SPs suppress negative oromotor responding to quinine.

Data are average (n = 6,± SEM) number of movements (A, B) or seconds spent (C) in response to 30 s infusions of artificial saliva, filtered saliva, whole saliva, and water, with quinine (white bars) or without quinine (black bars). Movements are categorized as aversive (A, gapes, head shakes, chin rubs, forelimb flails, and face washes), ingestive (B, rhythmic mouth movements, tongue protrusions, lateral tongue protrusions, and paw licks), or time spent in passive drip (C). Effects on responding for each valence (aversive, ingestive, passive drip) were assessed by ANOVA comparing infusions with and without quinine, across solvent. Post-hoc analyses (Bonferroni corrected paired t-tests) were performed on aversive responses (A), showing that while addition of quinine to artificial saliva, filtered saliva, and water resulted in increased aversive responding, addition of quinine to whole saliva had no effect on aversive oromotor movements. No effects were found on ingestive movements (B) or time spent in passive drip (C). * indicates significant differences (p < 0.05).

Ingestive movements were not affected by the addition of quinine to solution, or what type of solvent was used, and there was no significant interaction (presence of quinine: (F(1,40) = 1.80, p = 0.216, solvent: F (3,40) = 0.24, p = 0.866, presence of quinine × solvent: F (3,40) = 0.34, p = 0.795, Fig. 2B). Additionally, there were no significant effects found on passive drip (presence of quinine: (F (1,40) = 0.87, p = 0.379, solvent: F(3,40) = 1.11, p = 0.365, presence of quinine × solvent: F(3,40) = 0.84, p = 0.481, Fig. 2C).

3.2. Experiment 2: do SPs have a concentration-dependent effect on oromotor responding to quinine?

There was a significant effect of protein concentration on aversive responses to quinine infusion (F(4,25) = 8.06, p < 0.001, Fig. 3A). We conducted a regression analysis, which revealed a significant negative relationship between protein concentration of saliva and aversive oromotor movements (adjusted r² = 0.52, p < 0.001). We confirmed that infusion of quinine in whole saliva suppresses aversive oromotor movements, and also found that as the protein concentration of the saliva is decreased, the number of aversive movements increases. There were no effects of protein concentration on ingestive movements (F (4,25) = 0.54, p = 0.710, Fig. 3B) or time spent in passive drip (F (4,25) = 0.25, p = 0.905, Fig. 3C).

Fig. 3. SPs suppress aversive responding in a concentration-dependent manner.

Data are average (n = 6,± SEM) number of aversive/ingestive movements (A, B) or seconds spent in passive drip (C) in response to 30 s infusions 1 mM quinine in artificial saliva or whole saliva at 0.375, 0.75, 1.5 or 3 mg/ml total protein concentration. Effects on responding for each valence (aversive, ingestive, passive drip) were assessed by ANOVA comparing infusions across protein concentration. (A) As there was an effect of protein concentration on the number of aversive movements, a post-hoc analysis (regression analysis comparing protein concentration to number of aversive movements) was performed, showing that as protein concentration decreased, the number of aversive movements to quinine infusion increased. (B, C) No effects were found on ingestive movements or time spent in passive drip.

3.3. Experiment 3: does the presence of SPs alter oromotor responding to diverse stimuli?

3.3.1. Quinine

There was a significant interaction between quinine concentration and solvent (F(2,42) = 5.10, p = 0.015, Fig. 4A) such that whole saliva reduced negative responding to quinine (main effect: F(1,42) = 14.04, p = 0.003, Fig. 4A) and quinine concentration increased aversive responding (main effect: F(2,42) = 16.54, p < 0.001), but whole saliva was more effective at suppressing responding as the concentration of quinine increased. Post-hoc testing showed that the effect of whole saliva was specific to 1 mM quinine (t = 4.80, p = 0.001), and while aversive movements were not different at 0.3 mM quinine, they trended towards a decrease in responding (t = 2.33, p = 0.067). No differences were found at 0.1 mM quinine (t = 0.24, p = 0.823), which had very few aversive responses. Increasing the concentration of quinine decreased the number of ingestive movements (F(2,42) = 19.32, p < 0.001, Fig. 4B), however, there was no effect of saliva type on ingestive movements during quinine infusion (F(1,42) = 0.05, p = 0.832), and no interaction between concentration and saliva type (F(2,42) = 1.09, p = 0.356). Finally, there were no differences in passive drip time (saliva type: F(1,42) = 0.77, p = 0.399, concentration: F (2,22) = 3.120, p = 0.06, saliva type × concentration: F(2,42) = 0.08, p = 0.922, Fig. 4C).

Fig. 4. increased quinine concentration increases aversive and decreases ingestive oromotor responding.

Data are average (n = 8, ± SEM) number of movements (A, B) or seconds spent (C) in response to 30 s infusions of 0.1, 0.3, or 1 mM quinine in artificial saliva (black bars) or whole saliva (gray bars). Effects on responding for each valence (aversive, ingestive, passive drip) were assessed by ANOVA comparing infusions in artificial vs. whole saliva, across concentration. (A) There was an interaction between the concentration of quinine infused and solvent for aversive movements, and post-hoc analyses (Bonferroni-corrected paired t-tests) showed that whole saliva decreased aversive responding only at the 1 mM concentration. (B) There was an effect of concentration for ingestive movements, and no effects were found on time spent in passive drip (C). * indicates significant differences (p < 0.05)

3.3.2. Citric acid

We infused citric acid in either whole saliva, filtered saliva or artificial saliva. There was no effect of the saliva type or concentration on citric acid-driven aversive movements (saliva type: F(2,35) = 0.43, p = 0.524, concentration: F(1,35) = 0.21, p = 0.656, interaction: F (1,35) = 0.06, p = 0.804, Fig. 5A). However, we did see effects on ingestive movements (Fig. 5B), which increased with concentration (F (1,35) = 10.43, p = 0.007). This effect appears to be driven largely by rhythmic mouth movements (F(1,35) = 8.53, p = 0.013). The type of saliva also changed the number of ingestive movements during citric acid infusion (F(2,35) = 5.34, p = 0.039), and there was no interaction between stimulus concentration and saliva type (F(1,35) = 0.20, p = 0.661). Although it appears that whole and filtered saliva suppress the number of ingestive movements compared to artificial saliva, post-hoc tests failed to reach significance (t = 0.77–2.21, p’s >0.069). Lastly, there was an effect of saliva type on time spent in passive drip (F (2,35) = 4.65, p = 0.052, Fig. 5C), as well as an effect of concentration (F(1,35) = 5.99, p = 0.031), but no interaction (F(1,35) = 1.12, p = 0.309). It is likely that animals spent less time in passive drip because they spent more time doing ingestive movements.

Fig. 5. SP content affects responding to citric acid and NaCl, but not sucrose.

Data are average (± SEM) measures of taste reactivity categorized as (A, D, G) aversive, (B, E, H) ingestive, or (C, F, I) time spent in passive drip. Rats were given 30 s infusions of (A-C) 10 or 20 mM citric acid, (D-F) 300 or 500 mM NaCl, or (G-I) 30 or 100 mM sucrose in artificial saliva (black bars), whole saliva (light gray bars), or filtered saliva (A-C only, dark gray bars). Effects on responding for each stimulus (citric acid, NaCl, sucrose) and valence (aversive, ingestive, passive drip) were assessed by ANOVA comparing solvent across concentration. A-CWhile no effects were found in aversive movements, number of ingestive movements increased across concentration and were altered by solvent. Post-hoc analyses (Bonferroni-corrected paired t-tests) failed to reach significance at either concentration, though they trended towards a decrease in ingestive movements for whole saliva infusions. Ingestive taste reactivity during filtered saliva infusions was not significantly different from either artificial or whole saliva, though the data suggest that filtered saliva infusions are more like those of whole saliva than artificial saliva. Additionally, there was a solvent and a concentration effect found on passive drip. d-FNo differences were found in aversive or ingestive movements during NaCl infusions. However, there was an effect of solvent on time spent in passive drip, though post-hoc tests failed to reach significance. G-IFor sucrose infusions, no effects of either saliva type or concentration were found at any valence.

3.3.3. NaCl

There was no effect of saliva type on aversive oromotor movements during NaCl infusions (F(1,28) = 1.26, p = 0.270, Fig. 5D), no effect of concentration (F(1,28) = 3.56, p = 0.069), and no interaction between saliva type and concentration (F(1,28) = 0.05, p = 0.836). There was also no effect of saliva type or concentration on ingestive oromotor movements (saliva type: F(1,28) = 3.77, p = 0.074, concentration: F (1,28) = 0.46, p = 0.510, interaction F(1,28) = 0.01, p = 0.936, Fig. 5E). There was an effect of saliva type on time spent in passive drip (F(1,28) = 5.78, p = 0.031, Fig. 5F), however post-hoc tests failed to reach significance at either concentration (300mM: t = 1.85, p = 0.106, 500mM: t = 0.84, p = 0.431). Additionally, there was no effect of NaCl concentration on time spent in passive drip (F (1,28) = 4.02, p = 0.065), and no interaction between saliva type and concentration (F(1,28) = 1.17, p = 0.298).

3.3.4. Sucrose

For sucrose infusions, there were no effects of either saliva type or concentration on ingestive movements (saliva type: F(1,28) = 0.01, p = 0.937, concentration: F(1,28) = 1.25, p = 0.285, interaction: F (1,28) = 0.01, p = 0.924, Fig. 5G), aversive movements (saliva type: F (1,28) = 0.08, p = 0.777, concentration: F(1,28) = 0.48, p = 0.499, interaction: F(1,28) = 0.31, p = 0.590, Fig. 5H), or passive drip (saliva type: F(1,28) = 1.42, p = 0.254, concentration: F(1,28) = 0.00, p = 0.945, interaction: F(1,28) = 0.01, p = 0.945, Fig. 5I).

4. Discussion

We have previously shown that exposure to quinine or tannic acid containing diets upregulates a set of SPs, and animals with these SPs upregulated show increases in measures of acceptance and palatability, including increased consumption of bitter diets, rates of feeding, and brief-access licking to quinine [24–27]. Rats also have increased detection thresholds to quinine, suggesting that the increases in palatability may be driven by a decrease in sensitivity to the bitter stimulus [25]. Our previous findings, however, are limited in that they do not adequately separate SP upregulation from bitter diet experience, or rely on drugs that induce malaise. The taste reactivity procedure is ideal to assess palatability of quinine in saliva solutions, as it focuses on consummatory response to stimuli and allows us to control the animals’ diet experience independent of the taste stimuli tested. Additionally, since infusion volumes are low and infusion times are short, responses can reasonably be assumed to involve little post-oral feedback. The infusions result in taste reactivity i.e., stereotypical affective behavioral responses, which are unconditioned [34,40,41]. Infusions of preferred stimuli like sucrose result in ingestive oromotor movements that en-courage ingestion, like tongue protrusions, while infusions of avoided stimuli like quinine result in aversive oromotor movements that facilitate rejection, like gaping [42]. With our donor saliva paradigm, saliva affected taste reactivity in response to infusions of quinine and citric acid, each in a different valence (aversive for quinine, ingestive for citric acid), and each likely due to a different saliva constituent (proteins for quinine, ions for citric acid).

While addition of 1 mM quinine to water, artificial saliva, and filtered saliva predictably increased aversive oromotor responding, there was no change in responding when quinine was added to the whole saliva (containing SPs). This suggests that the proteins alone, in the absence of other experience, learning, or post-oral feedback, are capable of decreasing the perceived intensity of the quinine. Additionally, ingestive oromotor responses were not altered by addition of whole saliva to the infusions, which indicates that the SP effect is not increasing the positive valence of quinine, only decreasing the negative valence, and is supported by our previous finding that rats are less sensitive to bitter after upregulation. In the second experiment, we saw a concentration-dependent change in aversive responding to quinine. As SP concentration decreased, we saw an increase in aversive responses. Furthermore, movements to sucrose and NaCl were not significantly altered with the protein content of the saliva, suggesting that the effect is not a broad response to all taste stimuli. Though our previous experiments have used both TA and quinine as models for bitter taste stimuli, here we chose to focus solely on quinine, as TA has an additional astringent component [43–46], which may interact with SPs independently from their effect on bitter taste [47–51]. Together with our previous reports, these findings strongly suggest that SPs are capable of altering bitter taste.

In contrast to infusions of quinine, where whole saliva decreased aversive responding, whole saliva decreased ingestive oromotor movements in response to citric acid infusions. We chose to look at 2 concentrations of citric acid - a higher concentration that was more avoided in brief-access licking (20 mM citric acid) and a lower concentration that was less avoided (10 mM citric acid). The data for both concentrations show that whole saliva appears to attenuate ingestive oromotor movements. Based on previous studies demonstrating a buffering effect of saliva on acids, we did not want to assume that the buffering effect would be identical between filtered saliva and artificial saliva, therefore we infused citric acid in filtered saliva, artificial saliva and whole saliva. Filtered saliva appeared to result in similar responding to whole saliva but differences between the donor saliva and artificial saliva failed to reach significance. Although there is more work that needs to be done to identify the part of the saliva that is altering citric acid responding, our finding suggests that the ion content of saliva or the small peptides that may be present, rather than the SPs, are altering citric acid responding. Our experiment was not designed to determine if ion profile was important to taste reactivity, and we are unsure how the ion profile of filtered saliva/whole saliva is different from that of artificial saliva.

Additionally, the citric acid data must be interpreted with the caveat that ingestive movements increased as citric acid concentration increased, which is surprising, as rats lick less to 20 mM citric acid than 10 mM in a brief-access paradigm, suggesting that 20 mM is the less “positive” stimulus. Additionally, citric acid responses were higher than sucrose responses, which may also seem surprising. There have been previous reports that taste reactivity to sour stimulus infusion is inconsistent and difficult to characterize, is not reflective of sour stimulus consumption, and includes both aversive and ingestive movements [29,42,52]. It has also been suggested that the most common response to citric acid is low amplitude mouth movement, which we would have scored as rhythmic mouth movements, an ingestive behavior. It has been shown that rat pups express these movements as often to citric acid as they do to sodium saccharin [52]. We, and others, also recorded these rhythmic mouth movements in response to quinine [29,30,34]. Although it is not uncommon to include rhythmic mouth movements as “ingestive behaviors” as they appear to promote ingestion of a stimulus [53], some authors do not include rhythmic mouth movements as “ingestive” as it does not necessarily reflect hedonically positive responding [54]. Further work is necessary to explore the role of saliva in altering citric acid responding.

A limitation of this paradigm is that “donor saliva” does not precisely recapitulate the profile of diet-upregulated saliva. Though it is possible to collect saliva from awake and behaving rats after diet manipulations have altered SP content ([26,27], a technique we have used successfully in the past), the incredibly small volumes collected from each rat (~60 μl per collection) make the use of unstimulated saliva for infusions impractical in the large scale. Therefore, we use drug stimulated saliva. The stimulated saliva, like that of our diet exposure, results in protein upregulated saliva; however, injections of isoproterenol result in an upregulation of many SPs. The total protein content is approximately 3 times higher in the donor saliva, than in saliva we collect in unstimulated animals. To address this, donor saliva is diluted with artificial saliva to physiologically relevant concentrations, meaning that both whole and filtered saliva contain artificial saliva. While the dilution of our saliva sample gives us confidence that the SPs are able to change taste-driven responses in biologically relevant concentrations, it means that we can make no conclusions about the natural ion profile of saliva. Additionally, the specific profile of proteins the animals are being treated with is not identical to what they would have after diet-exposure to quinine; for example, IPR upregulates SPs that are detected at 25 kDa [24] which we do not see upregulated by quinine diet exposure [26]. With these caveats in mind, direct comparisons between these data and diet-induced effects should be made cautiously.

4.1. Conclusions and future work

These experiments suggest that SPs are sufficient to alter the aversive orosensory response to bitter stimuli, and together with our previous work, strongly suggest that SPs can alter taste. SPs may function in addition to the learning process that takes place during repeated bitter exposure, representing a physiological shift that enables animals to perceive bitter foods as less bitter.

Future work should focus on the mechanism of the physiological shift. Although evidence is mounting that the increased acceptance of bitter foods after repeated exposure is in part due to changes in taste responding after changes in SP expression, we do not know how the upregulated SPs alter bitter signaling. There are several hypotheses that have been proposed in the literature. The first hypothesis that has been proposed is that SPs form a mucosal pellicle to reduce access of the stimulus to the receptor [55]. In support of this hypothesis, a number of the proteins that correlate with taste alterations are involved in pellicle formation [26,55]. Inconsistent with this hypothesis is that SP expression alters bitter but not sweet taste [25,26] suggesting that the pellicle would have to be permeable to some but not all stimuli. Furthermore, saliva from donor animals washed across the tongue decreases taste nerve signaling [26] and orosensory responding, which suggests that pellicle formation is not necessary for SPs to exert their action. The second hypothesis is that SPs bind to the stimulus and keep it from interacting with the receptor. In support of this hypothesis, there is evidence that SPs form a complex with a variety of stimuli [56–58]. However, the ability to bind to the stimulus is not sufficient evidence that it is mediating the effect. The role of most of the ~1000 proteins produced in saliva is to bind. SPs bind to each other to form pellicles that protect the tissue and teeth. They also bind to bacteria, the tooth surface, tissue surfaces, and so on. Due to the “sticky” nature of these proteins, binding assays alone have historically revealed little about the nature of these protein interactions. The third hypothesis postulates that SPs bind to the receptor and alter its responsiveness to stimuli. There is no work to our knowledge that has systematically tested this hypothesis, and it would be an important step to untangling the relationship between SPs and taste function. Ultimately, these proteins could be targeted to assist humans in maintaining a healthy diet by increasing the attractiveness of healthy foods.