https://orcid.org/0000-0003-3936-1885
Abstract
Outbred rats differ in their preference for the artificial sweetener sucralose. Psychophysical assessments have shown that the taste of sucralose is differentially generalized to either sucrose or a sucrose-quinine (QHCl) mixture in sucralose preferers (SP) and sucralose avoiders (SA), respectively. It remains to be determined if these differences in the psychophysical assessment of the taste of sucralose are due to an insensitivity to any bitter-like taste component of sucralose in SP or reduced sensitivity to a sweet-like component in SA that may mask any putative aversive side-taste in SP. Here, we exploited the proposed chemotopic organization of the rostral nucleus of the solitary tract (rNTS) to further parse out the root differences in the perception of the salient taste qualities of sucralose using Fos-like immunoreactivity (FLI) to approximate neural activation following intraoral delivery of sucrose, QHCl, and sucralose solutions in previously categorized SA and SP. First, we confirmed previous reports that the medial third of the NTS is primarily responsive to intraoral infusions of the bitter taste stimulus QHCl while sucrose produces a more diffuse pattern of FLI. Upon comparing the FLI generated by intraoral sucralose, we found that the pattern in SA was indistinguishable from that of QHCl while SP displayed a pattern of FLI more representative of a sucrose-QHCl mixture. We conclude that SA, relative to SP, may be less sensitive to the sucrose-like properties of sucralose and that an enhanced sensitivity to these sucrose-like qualities may mask a QHCl-like quality in SP.
Introduction
Determining the relative contributions of each component of a taste mixture in rodent models is problematic. In general, intake-driven behavioral measures conflate the stimulus quality of a compound with its hedonic properties and thus confound differences in the willingness to consume a stimulus mixture with differences in the perception of the overall salient taste qualities of the mixture. For example, many artificial sweeteners elicit a concentration-dependent change in whether a sweet- or bitter-like taste component is the salient quality (Schiffman et al. 1995) yet the degree to which either quality is currently being perceived is difficult to measure in rodent models that rely primarily on intake-driven behavioral paradigms. The ability to determine the salient taste qualities of a taste mixture is an important and potentially powerful experimental tool as it can aid in detecting even slight phenotypic differences in normal gustatory function.
Of particular relevance to this issue is the phenotypic difference in the hedonic response to the artificial sweetener sucralose observed in outbred rats (Sclafani and Clare 2004; Loney et al. 2011; Bacharach and Calu 2019). Approximately 75% of rats are largely indifferent to sucralose at lower concentrations and actively avoid it at higher concentrations (> ~0.025 mM) and thus are categorized as sucralose avoiders (SA). Conversely, the remaining 25% of rats avidly consume sucralose at these same concentrations and are referred to as sucralose preferrers (SPs; Loney et al. 2011)). In psychophysical paradigms, SA detect a bitter-like taste quality in sucralose that SP fail to detect at the concentrations tested (Loney et al. 2012a). The mechanism driving this difference in perception is currently unknown. It could be that SA perceive a bitter-like quality from sucralose that SP are taste-blind to or, conversely, SP may be more sensitive to the sweet-like quality of the stimulus that in turn masks the “bitterness” that SA respond to; a distinction that is difficult to parse in rodent models.
The first-order central gustatory relay, the rostral nucleus of the solitary tract (rNTS), displays a tight systematic organization in rodents. This anatomical organization is observed across both the rostral-caudal axis and the medial-lateral axis. The terminal fields of afferent fibers originating from the oral cavity are segregated such that projections from the IXth cranial nerve terminate more caudal and medial to those of the VIIth cranial nerve (Torvik 1956; Contreras et al. 1982; Hamilton and Norgren 1984). Furthermore, this medial-lateral differentiation is suggestive of a chemotopic pattern of organization. Specifically, quantifying Fos-like immunoreactivity (FLI) as an approximation for neural activity, numerous studies have demonstrated that oral stimulation by “bitter” stimuli, including quinine hydrochloride (QHCl), elicits an abundance of FLI in the medial portion of the rNTS, whereas oral stimulation by “sweet” stimuli, including sucrose, elicits a more diffuse pattern of activation (Harrer and Travers 1996; King et al. 1999, 2000; Travers 2002; Chan et al. 2004). While differences in the taste quality of the respective stimuli offer the most parsimonious explanation for the proposed chemotopic patterns of FLI across the medial–lateral axis of the rNTS, there are other potential explanations. One possible caveat to this assumption is that QHCl and sucrose not only differ in both their taste quality evaluation and their efficacy to activate either the IXth or VIIth cranial nerves (Danilova and Hellekant 2003), but also result in starkly different patterns of reflexive oromotor and somatic movements (Grill and Norgren 1978). As a result, different patterns of motor output could theoretically result in different patterns of sensory feedback and thus drive the observed differences in the patterns of QHCl- and sucrose-evoked FLI (Harrer and Travers 1996). In support of this possibility, the amount of FLI present in the dorsal-medial subfield of the rNTS is highly correlated with the number of gapes elicited by the stimulus (King et al. 1999, 2000).
The present study sought to measure differences in the pattern of FLI elicited by “sweet” and “bitter” taste stimuli in the rNTS to probe mechanistic differences in gustatory function of outbred rats that may offer insight into the salient taste qualities of sucralose in SA and SP. Specifically, we systematically examined the pattern of FLI expression across the mediolateral axis of the rNTS in SA and SP in response to intraoral infusions of QHCl and sucrose. We then compared the patterns of FLI generated by sucralose in either phenotype to that generated by both sucrose and QHCl.
Materials and methods
Animals and housing
Naïve adult male Long-Evans rats (n = 42), classified as SA and SP (see below), were individually housed in polycarbonate cages in a temperature- and humidity-controlled room maintained on a 12:12 h light cycle. Standard rodent chow (Purina 5001) and water were available ad libitum in the animals’ home cages. All behavioral assessments involving intraoral delivery of taste stimuli were conducted in a rectangular Plexiglas chamber (305 mm2). The test chamber was equipped with a mirror, positioned at a 45° angle below the transparent floor of the chamber, which provided clear visualization of the rat’s oromotor and somatic movements during oral infusions. Animal usage and experimental protocols were approved by the Florida State University Institutional Animal Care and Use Committee.
Sucralose classification
Prior to inclusion in the present study, a large cohort of rats was given a series of four 2-bottle preference tests between deionized water (dH2O) and increasing concentrations of sucralose (0.0025, 0.025, 0.25, and 2.5 mM). It was necessary to screen a larger cohort of rats in order to obtain an adequate number of SP due to the imbalance in the phenotypic split (i.e., ~75% SA and ~25% SP). This preference curve was based on previously published sucralose classification procedures that reliably categorize rats as SP or SA (Loney et al. 2011, 2012b); Figure 1). Rats were permitted to drink freely from either bottle for 48 h per concentration. In order to control for side preferences, bottle positions were switched at the end of each 24-h period. Preference ratios were calculated by dividing the amount of sucralose consumed by the total fluid consumption during each 48-h preference test. Sucralose classification procedures lasted for a total of 8 days.
Preference for sucralose, relative to water, in a series of 48 h 2-bottle preference tests. As has been demonstrated previously, rats can be grouped into 2 distinct phenotypic categories of sucralose avoiders (SA) and sucralose preferrers (SP) based on their preference for 0.25 mM sucralose. SA were largely indifferent at the 2 lowest concentrations of sucralose (0.0025 and 0.025 mM) and actively avoided sucralose at the 2 highest concentrations (0.25 and 2.5 mM). In comparison, SP avidly consumed sucralose at these 2 highest concentrations and vastly perferred sucralose over water. Analysis revealed a significant phenotype x sucralose concentration interaction, F(1,3) = 80.11, P < 0.001. *Significant group difference (P < 0.001).
Surgical procedures
The surgical procedure was adapted from that described previously (Eckel and Ossenkopp 1995). Briefly, fixation of intraoral cannulas was achieved through the insertion of a 7-gauge needle subcutaneously, originating at the level of the scapular region. This needle was guided under the skin until it exited through the left oral mucosa just lateral to the maxillary molar. Following insertion of the needle, polyurethane tubing (PE-90; VWR) was advanced through the needle and a Teflon washer was threaded onto the tubing and held against the oral mucosa. The tubing was heat flared, cooled, and then pulled snugly against the washer and rat’s cheek. The needle was subsequently removed, leaving behind the tubing. The excess tubing was cut, and a 20-gauge adapter cap was affixed to the externalized tubing to prevent reentry into the subcutaneous space. This procedure was then repeated on the right side. Both cannulas were cleaned and flushed daily in order to prevent infection and maintain patency. All surgeries were conducted in rats maintained at a surgical level of anesthesia by inhalation of 2–4% isoflurane. All rats received intraperitoneal (i.p.) injections of gentamicin (0.2 mg; Fort Dodge) and butorphenol (0.05 mg; Henry Schien) following surgical procedures and were maintained on wet mash (powdered 5001 + dH2O) for at least 3 days following surgery. All rats continued to receive prophylactic injections of gentamicin (0.2 mg, i.p.) for 3 days following surgery. Testing occurred no fewer than 7 days following surgery.
Test stimuli
Test stimuli consisted of 10 mM sucralose, 3 mM QHCl, and 1 M sucrose. Stimulus concentrations were chosen to remain consistent with previously published studies (Harrer and Travers 1996; King et al. 1999; Loney et al. 2012a). All test solutions were prepared fresh daily in dH2O using reagent grade chemicals (QHCl and sucralose; Sigma–Aldrich) and commercially available sucrose (Publix). All stimuli were delivered at room temperature.
Oral infusions
At study onset, rats were assigned to 1 of 4 oral-infusion groups: 10 mM sucralose (n = 8 SP and 8 SA), 1 M sucrose (5 SP and 5 SA), 3 mM QHCl (5 SP and 5 SA), or dH2O (3 SP and 3 SA). Training and test sessions were modeled after previous studies (Harrer and Travers 1996; King et al. 1999). For 3 habituation days, individual rats were placed in the test chamber and left undisturbed for 30 min. Following this adaptation period, dH2O was infused into the oral cavity via PE-90 tubing that was attached at one end to the adapter cap of the rat’s cannula. The other end of the tubing was attached to a syringe secured in an automated infusion pump (Harvard Apparatus) programmed to deliver 7 ml of dH2O over 30 min at a rate of ~0.233 ml/min. Test days immediately followed the third habituation day and the procedure emulated that of the habituation days with the exception that all rats received their assigned test stimulus. The first 1 min of stimulus delivery was videotaped with a digital camera (Sony DCR) for subsequent analyses of typical taste reactivity oromotor and somatic movements (Grill and Norgren 1978). Specifically, ingestive responses consisted of rhythmical mouth movements, tongue protrusions, and lateral tongue protrusions while aversive rejection responses consisted of gapes, paw flails, chin rubs, and passive drips. Following cessation of the test stimulus delivery, the rat was left in the testing chamber undisturbed for an additional 45 min.
Taste reactivity analyses
Digital videos of the first 1 min of the oral infusions were played back at 1/5th speed and both ingestive and aversive oromotor and somatic movements (Grill and Norgren 1978; Spector et al. 1988) were quantified by an experimenter who was blind to treatment condition. One SA in the sucrose test condition reared on its hind legs during most of the infusion. Because this occluded the mouth from view, this rat’s data were not included in the behavioral analyses.
Brain histology
Immediately following the 45-min post-infusion period on test days, rats received i.p. injections of an overdose of SomnaSol (Henry Schein, Dublin, OH) and, when unresponsive, were intracardially perfused with chilled saline followed by sodium phosphate-buffered (7.4 pH) 4% paraformaldehyde. Brains were removed and post-fixed in paraformaldehyde overnight at 4°C. Following fixation, brains were cryoprotected with 30% sucrose in 0.1 M phosphate-buffered saline (PBS). Cryoprotected hindbrains were cut in the coronal plane on a freezing sliding microtome (Thermo Scientific HM450) into 40 µm sections and divided into 2 series of alternating sections through the NTS (~14.6–11.76 mm caudal to bregma). The first series was processed for FLI and the second series was stored in 0.1 M PBS for subsequent thionin staining.
Immunohistochemical processing for FLI consisted of a series of washes with 0.1 M PBS followed by blocking with 0.5% bovine serum albumin (BSA) in a 0.3% triton/PBS solution. Following blocking, tissue sections were incubated in primary c-Fos antibody (AB-5 polyclonal rabbit antibody; Calbiochem), at a dilution of 1:10 000 in PBS/BSA overnight (20 h) at 4°C. On the following day, the tissue was washed in PBS/BSA and then incubated for 1 h in goat anti-rabbit secondary antibody (BA-1000, Vector Laboratories) at a dilution of 1:200 in PBS/BSA followed by incubation in an avidin-biotin mixture (Elite kit; Vector Laboratories) in PBS for 30 min. Tissue sections were then stained with DAB (3,3’-diaminobenzidine-HCL; Vector Laboratories) for ~5 min. The thionin staining protocol consisted of defatting with xylenes preceding staining with 0.1% thionin for 7 min. c-Fos- and thionin-stained tissue sections were mounted on gelatin-subbed slides and dehydrated through ascending ethanol solutions and cleared with xlyenes prior to coverslipping.
Microscopic analysis of brain sections
Microscopic tracings of subfields and counting of FLI were conducted by an experimenter blind to both SP/SA group assignment and test stimulus condition. Analyses were performed on one section at the level of the intermediate rostral, gustatory portion of the nucleus of the solitary tract (IRgNTS, located ~12 mm posterior to bregma and ~500 µm caudal to the most caudal portion of the dorsal cochlear nucleus; cDCN). The IRgNTS was chosen based on the described terminal fields of the gustatory nerves (Hamilton and Norgren 1984; King et al. 1999; Chan et al. 2004) as the IRgNTS receives afferent fibers from both the glossopharyngeal (IXth) and chorda tympani (VIIth) nerves and is the area in which previously published differences in the pattern of FLI between sucrose and QHCl was most pronounced (Harrer and Travers 1996; King et al. 1999). Images of the IRgNTS captured at 10× were taken with a light microscope equipped with a digital video camera (Olympus). The nucleus was divided into 6 subfields as described previously (King et al. 1999, 2008; Chan et al. 2004); adjacent thionin-stained sections were utilized to distinguish the anatomical landmarks constituting the nuclear border where necessary. Briefly, by drawing directly onto the digital image of the IRgNTS, the nucleus was divided into 3 equal sections by constructing 2 lines perpendicular to the horizontal, long axis of the nucleus. Next, at each of these intersecting points, the dorsal-ventral axis of the nucleus was measured, divided in half, and then the horizontal axis redrawn such that it connected each of these halfway points (see Figure 2). This resulted in 6 subfields of the IRgNTS: dorsal medial (DM), ventral medial (VM), dorsal central (DC), ventral central (VC), dorsal lateral (DL), and ventral lateral (VL). Such an approach served to parse out each subfield in such a way that the dorsal and ventral areas were roughly equal in area. Clearly labeled FLI-positive neurons (i.e., those cells containing dark, punctate nuclear staining) were then counted manually in each of the 6 subfields by experimenters blinded to experimental conditions.
Representative sections containing FLI in coronal brain sections from SA and SP following intraoral delivery of dH2O (A, B), QHCl (C, D), sucrose (E, F), and sucralose (G, H). The amount of FLI present in each section was counted and analyzed as a function of subfield and sucralose preference profile.
Data analysis
In order to determine the degree to which the taste stimuli (QHCl, sucrose, and sucralose) produced more FLI than dH2O, and which subfields of the NTS displayed increased taste-specific FLI, a 2-factor ANOVA was conducted on the raw FLI counts contained across each area of the NTS in response to each of the 4 intraoral stimuli. As we were specifically interested in the taste-specific induction of FLI, the average Fos count from each NTS subfield generated by dH2O was subtracted from the average count in that same subfield generated by intraoral sucralose (Stratford and Finger 2011; Stratford and Thompson 2016), QHCl, and sucrose. Analyses of both the relative number and proportion of FLI within the 6 subfields of the IRgNTS were accomplished through 2- and 3-factor mixed-design ANOVAs, where appropriate. Sucralose preference profile and stimulus condition served as between-subject factors and IRgNTS subfield served as the within-subject factor. Statistically significant main and interactive effects (P < 0.05) were further explored with Bonferroni-corrected t-tests.
Independent-samples t-tests were used to analyze differences between SA and SP in the total amount of FLI present in the NTS as well as the number of ingestive and aversive taste reactivity responses elicited by each of the 3 taste stimuli. Ingestive responses included the sum of tongue protrusions and lateral tongue protrusions (observed almost exclusively in response to appetitive stimuli) as well as the sum of rhythmic mouth movements (RMMs) and paw licks. The aversive responses included the sum of gapes, paw flails, head shakes, chin rubs, and passive drips.
Results
Oral stimulation induces FLI in a taste-specific manner within the IRgNTS
Each of the taste stimuli elicited more robust and comparatively darker stained FLI relative to that elicited by oral stimulation with dH2O, and none of the taste stimuli, nor dH2O, resulted in substantial FLI in the lateral region areas of the IRgNTS (Figure 3). Statistical analysis confirmed these observations. The presence of FLI in the IRgNTS was influenced by the main effect of stimulus condition (F(3,38) = 18.21, P < 0.001). Post hoc analyses revealed that oral infusions of all 3 taste stimuli (QHCl, sucrose, and sucralose) produced significantly greater amounts of FLI relative to oral infusions of dH2O in all rats (Ps < 0.001; data not shown). An interaction between stimulus condition and NTS subfield (F(15,170) = 10.89, P < 0.001) revealed that each of the taste stimuli elicited more FLI than dH2O in the medial and central subfields (DM, VM, DC, and VC) of the IRgNTS (Ps < 0.05; data not shown). This taste-specific difference in the induction of FLI was restricted to these subfields as none of the taste stimuli elicited more FLI than dH2O within the lateral subfields (DL and VL; Ps > 0.52; data not shown.
Schematic depicting the level of the rNTS and approximate subfields analyzed in the present set of experiments. A coronal plate (adapted from Paxinos and Watson 1998) depicting the approximate level of the IRgNTS (~12.3 mm posterior to bregma) at which the rNTS was subjected to subfield analyses of the pattern of FLI in response to intraoral 3 mM QHCl, 1 M sucrose, and 10 mM sucralose.
Next, we examined the total amount of taste-specific FLI generated by the 3 taste stimuli (QHCl, sucrose, and sucralose; Figure 4) as a function of sucralose preference profile. This 2-factor ANOVA revealed a significant main effect of taste stimulus (F(2,30) = 5.57, P < 0.01) driven by the fact that sucralose produced more FLI within the IRgNTS relative to both QHCl and sucrose (Ps < 0.05). Importantly, there were no main or interactive effects of sucralose preference profile, indicating that SA and SP did not differ in the total amount of FLI generated by any taste stimulus.
The total amount of FLI elicited across the IRgNTS following intraoral infusions of QHCl, sucrose, or sucralose as a function of sucralose preference profile. There were no differences between SA and SP in the total amount of FLI elicited by any stimulus.
QHCl and sucrose elicit a different pattern of FLI within subregions of the IRgNTS
In order to probe the pattern of FLI generated across the mediolateral axis of the IRgNTS as a function of sucralose preference profile we conducted a 3-factor mixed-design ANOVA comparing QHCl- and sucrose-elicited FLI in each IRgNTS subregion. This analysis revealed no main or interactive effects of sucralose preference profile (Ps = 0.67–0.79) as a function of taste stimuli (i.e. QHCl or sucrose). As such, the data were collapsed across the sucralose preference profile and analyzed by a 2-factor mixed-design ANOVA. This analysis revealed a main effect of IRgNTS subfield (F(5,90) = 45.40, P < 0.001) and a significant interaction between taste stimulus and subfield (F(5,90) = 12.95, P < 0.001) (Figure 5). Post-hoc analyses of the stimulus by subfield interaction revealed that QHCl elicited significantly more FLI, relative to sucrose, in the DM subfield, whereas sucrose elicited more FLI, relative to QHCl, in the DC subfield (Ps < 0.01; Figure 5). There was a trend for QHCl to elicit more FLI in the VM subfield, but this effect did not survive Bonferroni correction. Altogether, ~62% of the FLI elicited by QHCl was observed in the medial third of the IRgNTS (DM and VM) while the FLI elicited by sucrose was more evenly distributed across the medial and central subfields (43% and 46%, respectively).
QHCl and sucrose differed in the average amount of FLI present in each dorsal medial and dorsal central subfields of the NTS. Analysis of the average FLI located in each of the 6 subfields of the IRgNTS revealed that QHCl resulted in significantly larger amounts of FLI in the DM subfield while sucrose resulted in significantly larger amounts of FLI in the DC subfield, relative to each other. There were no differences in pattern of FLI generated by either QHCl or sucrose in SA and SP, as such, data are presented as collapsed across sucralose preference profile. *QHCL different from sucrose, P < 0.05.
Sucralose elicits a pattern of FLI within the IRgNTS that differs in SA and SP
A 2-factor mixed-design ANOVA revealed a significant interaction between sucralose preference profile and IRgNTS subfield (F(5,70) = 4.03, P < 0.01). Within the DC subfield, which is maximally responsive to sucrose, oral infusion of sucralose elicited significantly more FLI in SP, relative to SA (P < 0.05; Figure 6). There was no difference across SA and SP in the amount of FLI generated in the DM subfield or any of the other IRgNTS subfields (see Figure 4). We did observe, however, that ~62% of the FLI elicited by sucralose in SA was concentrated in the medial third of the IRgNTS, a pattern identical to that observed following oral infusion of QHCl. In comparison, sucralose elicited a more evenly distributed pattern of FLI in SP, although, unlike that observed following oral infusion of sucrose, ~50% of the FLI was observed in the medial third of the IRgNTS, with the DM subfield containing the largest quantity of FLI.
The pattern of FLI generated by intraoral sucralose differed in SA and SP. These differences in the pattern of FLI elicited by sucralose between SA and SP mirror the differences in the pattern of FLI elicited by QHCl and sucrose. Analyses of the pattern of FLI generated by sucralose across all 6 subfields of the IRgNTS in SA and SP revealed that SP display significantly greater amounts of FLI in the DC subfield, relative to SA, which is the same subfield in which sucrose generated significantly more FLI than QHCl. *SP greater than SA, P < 0.05.
To further characterize the differential pattern of sucralose-induced FLI in SP and SA, we examined the proportion of FLI induced by sucralose across IRgNTS subregions in SA and SP with that induced by oral infusion of sucrose and QHCl. This analysis revealed a significant 2-way interaction between stimulus and subfield (F(15,160) = 7.36, P < 0.0001). Post-hoc analyses comparing the proportion of FLI in each of the subfields of the IRgNTS generated by sucralose to that generated by either QHCl or sucralose in SA (Figure 7A) and SP (Figure 7B) were conducted separately. These analyses revealed significant differences in the proportion of FLI generated by sucralose in SA compared to sucrose in the DM and DC subfields (Ps < 0.05) with no statistical differences between sucralose and QHCl. In SP, there was a significant difference in the proportion of FLI generated by sucralose in only the DC subfield compared to QHCl (P < 0.05) with no differences between sucralose and sucrose.
Comparisons of the proportion of FLI expressed across the 6 IRgNTS subfields revealed that SA displayed a pattern of FLI that was nearly identical to that elicited by QHCl, whereas SP displayed a pattern of FLI that was more reminiscent of a QHCl/sucrose mixture. (A) In SA, the proportion of FLI elicited by sucralose differed from that elicited by sucrose in the DM and DC subfields, but was similar to that elicited by QHCl in all subfields of the IRgNTS. (B) In SP, the proportion of FLI elicited by sucralose differed from that elicited by QHCl in the DC subfield, but was similar to that elicited by sucrose in all subfields of the IRgNTS. *Significant stimulus difference, P < 0.05.
Taste reactivity responses
Ingestive and aversive oromotor responses to oral infusions of QHCl, sucrose, and sucralose are presented in Table 1. Both the total ingestive and aversive responses elicited by quinine and the ingestive responses elicited by sucrose (no aversive responses were observed) were similar in SA and SP (t(7–8) = –0.03 to 1.38; Ps = 0.21–0.98). In comparison, the ingestive responses elicited by intraoral infusion of sucralose were greater in SP, relative to SA (tongue protrusions, t(14) = 2.60, P < 0.05; RMMs, t(14) = 3.81, P < 0.01; Table 1) and they maintained a high rate of ingestive responses throughout the first min of the oral infusion. In contrast, SA displayed some ingestive responses upon stimulus delivery but rapidly switched to trains of aversive responses within the first 30 s of stimulus delivery. The aversive responses elicited by sucralose also differed in SA and SP (t(14) = 14.07, P < 0.001). Oral sucralose stimulation elicited gapes in every SA within the initial 1 min of stimulus delivery, whereas no aversive mouth movements of any kind were observed in SP. SP were observed to passively drip sucralose further into the 30-min infusion, well after the end of the analysis period, a response that was not observed in response to sucrose infusions.
Ingestive and aversive taste reactivity measures in response to 1-min intraoral infusions of QHCl, sucrose, and sucralose in sucralose avoiders (SA) and sucralose preferrers (SP)
| Stimulus | RMMs and paw licks | Tongue protrusions | Aversive responses (gape, chin rub, paw flail) | |||
|---|---|---|---|---|---|---|
| SA | SP | SA | SP | SA | SP | |
| QHCl | 94 ± 22.0 | 152.0 ± 36.2 | 1.6 ± 1.1 | 7.4 ± 4.5 | 31.4 ± 4.0 | 31.2 ± 6.1 |
| Sucrose | 180.8 ± 11.5 | 209.2 ± 50.3 | 14.3 ± 2.7 | 17.2 ± 6.1 | 0 | 0 |
| Sucralose | 105.9 ± 18.0* | 213.0 ± 20.6 | 2.9 ± 1.6* | 11.9 ± 3.0 | 25.5 ± 1.7* | 0 |
| Stimulus | RMMs and paw licks | Tongue protrusions | Aversive responses (gape, chin rub, paw flail) | |||
|---|---|---|---|---|---|---|
| SA | SP | SA | SP | SA | SP | |
| QHCl | 94 ± 22.0 | 152.0 ± 36.2 | 1.6 ± 1.1 | 7.4 ± 4.5 | 31.4 ± 4.0 | 31.2 ± 6.1 |
| Sucrose | 180.8 ± 11.5 | 209.2 ± 50.3 | 14.3 ± 2.7 | 17.2 ± 6.1 | 0 | 0 |
| Sucralose | 105.9 ± 18.0* | 213.0 ± 20.6 | 2.9 ± 1.6* | 11.9 ± 3.0 | 25.5 ± 1.7* | 0 |
*Different than SP, P < 0.05.
Ingestive and aversive taste reactivity measures in response to 1-min intraoral infusions of QHCl, sucrose, and sucralose in sucralose avoiders (SA) and sucralose preferrers (SP)
| Stimulus | RMMs and paw licks | Tongue protrusions | Aversive responses (gape, chin rub, paw flail) | |||
|---|---|---|---|---|---|---|
| SA | SP | SA | SP | SA | SP | |
| QHCl | 94 ± 22.0 | 152.0 ± 36.2 | 1.6 ± 1.1 | 7.4 ± 4.5 | 31.4 ± 4.0 | 31.2 ± 6.1 |
| Sucrose | 180.8 ± 11.5 | 209.2 ± 50.3 | 14.3 ± 2.7 | 17.2 ± 6.1 | 0 | 0 |
| Sucralose | 105.9 ± 18.0* | 213.0 ± 20.6 | 2.9 ± 1.6* | 11.9 ± 3.0 | 25.5 ± 1.7* | 0 |
| Stimulus | RMMs and paw licks | Tongue protrusions | Aversive responses (gape, chin rub, paw flail) | |||
|---|---|---|---|---|---|---|
| SA | SP | SA | SP | SA | SP | |
| QHCl | 94 ± 22.0 | 152.0 ± 36.2 | 1.6 ± 1.1 | 7.4 ± 4.5 | 31.4 ± 4.0 | 31.2 ± 6.1 |
| Sucrose | 180.8 ± 11.5 | 209.2 ± 50.3 | 14.3 ± 2.7 | 17.2 ± 6.1 | 0 | 0 |
| Sucralose | 105.9 ± 18.0* | 213.0 ± 20.6 | 2.9 ± 1.6* | 11.9 ± 3.0 | 25.5 ± 1.7* | 0 |
*Different than SP, P < 0.05.
Discussion
Previous work from our group (Loney et al. 2012a) has demonstrated that orosensory differences are sufficient to explain the phenotypic differentiation in the preference for sucralose observed in outbred rats. That previous experiment demonstrated that SA and SP differ in the perception of a putative bitter-like taste quality from concentrated sucralose yet the degree to which that difference in perception was based on the inability of SP to detect an aversive taste quality or the inability of the sweet-like qualities of sucralose in SA to mask it remained undetermined. These psychophysical tests, while powerful in their ability to isolate the influences of taste quality perception from hedonic and postingestive consequences, are limited in their ability to parse out the direction of this phenomenon. The present study extends more recent behavioral findings (Torregrossa et al. 2015) suggesting that SP are, in fact, likely capable of detecting this bitter-like quality, and is the first to propose that sucralose elicits weaker sweet-like responses in SA, relative to SP. Here we show that oral stimulation with sucralose produces reliable differences in the pattern of neural activation observed in the IRgNTS, as assessed through both the counts and the proportional patterns of FLI. The differential patterns of FLI in SA and SP are remarkably similar to patterns produced by the prototypical bitter- and sweet-tasting stimuli, QHCl and sucrose, an outcome consistent with the commonly reported bittersweet taste profile of artificial sweeteners in human subjects (Schiffman et al. 1995).
Oral stimulation with taste stimuli (QHCl, sucrose, or sucralose) elicits a significant increase in FLI in the medial and central thirds of the IRgNTS over stimulation with water (Figures 3 and 4) demonstrating that examination of FLI in the rostral NTS is an effective measure of taste-specific responses. Here, the lateral third of the nucleus was minimally responsive to oral stimulation with any stimulus used in the present experiment, consistent with the proposed sparse termination in this area of afferent projections from taste buds in the oral cavity (Whitehead 1988a, 1988b). Moreover, IXth nerve transection has minimal impact on what FLI is stimulated in the lateral portion of the rostral NTS while severely compromising QHCl elicited FLI in the medial and central thirds (King et al. 1999, 2000) further demonstrating that this area appears to be minimally involved in the processing of taste-specific information.
Consistent with previous reports (Harrer and Travers 1996; Travers 2002; Chan et al. 2004), the pattern of FLI across the medial-lateral axis of the IRgNTS differed as a function of the putative taste qualities of QHCl and sucrose (Figure 5). Specifically, QHCl elicited FLI was highly concentrated in the medial third, relative to sucrose, particularly within the DM subfield, whereas sucrose resulted in a more diffuse pattern of FLI. Of note, while the pattern of FLI from sucrose was more diffuse across the medial and central thirds of the IRgNTS, it did result in a significant increase in FLI in the DC subfield relative to QHCl. As stated previously, QHCl-stimulated FLI in the medial portion of the NTS is dependent on input from the IXth nerve as transection of this nerve abolishes this segregated FLI pattern and practically prevents all taste-specific FLI resulting from oral QHCl stimulation (King et al. 1999, 2000). These differential patterns of FLI are strongly associated with the receptive fields of the oral cavity as primary afferents from the VIIth nerve terminate rostrally and laterally to those originating from the IXth nerve (Hamilton and Norgren 1984). Furthermore, relative to each other, the VIIth nerve is more responsive to stimulation with “sweet” compounds and the IXth nerve is more responsive to stimulation with “bitter” compounds (Danilova and Hellekant 2003). As such, these patterns of FLI are indicative of a chemotopic organization of the rostral NTS resultant from differential innervation by the VIIth and IXth cranial nerves.
While neither the total FLI nor the pattern of FLI elicited by QHCl and sucrose differed in SA and SP, the pattern of FLI elicited by sucralose did differ in a manner consistent with their behavioral phenotype and predicted by the differential patterns of FLI elicited by QHCl and sucrose (Figures 6 and 7). SA displayed a high proportion of FLI in the DM subfield with a majority confined to the medial third of the nucleus in a pattern that was statistically indistinguishable from that of QHCl. In comparison, SP displayed nearly identical amounts of FLI in the DM subfield yet significantly more FLI in the DC subfield, similar to FLI elicited by sucrose. As such, the pattern of FLI generated by sucralose in SA appeared exclusively QHCl-like while that in SP was more representative of a QHCl–sucrose mixed pattern (Figure 7). These results strongly suggest that the root cause of the sensory differences in the taste of sucralose between SA and SP is an increased sucrose-like signal generated by sucralose in SP that appears to mask the presence of a QHCl-like side taste. Consistent with this interpretation are previous data that demonstrate that SP do not treat sucralose as a unitary sweet-like compound but rather are self-regulating their intake of the sweetener in a manner suggestive of the presence of an aversive taste quality (Torregrossa et al. 2015). If sucralose was perceived as a unitary sweet-like stimulus then the microstructural make-up of sucralose drinking should mimic that of sham-fed sucrose, a preparation where the caloric postingestive effects are negligible. While SP preferred sucralose across all concentrations tested in that study, and their preference increased as a function of increasing concentration, the microstructural elements of drinking bouts in SP revealed no change in bout size and a decrease in the rate of licking as concentration was increased (Torregrossa et al. 2015). Conversely, sham-fed sucrose results in robust, monotonic increases in bout size and rate of licking as concentration is increased (Smith 2000).
As SA and SP, in response to sucralose, did not differ in the number of FLI in the QHCl-associated DM subfield yet did differ in the sucrose-associated DC subfield, it seems likely that this proposed increased sweet-like signal generated in SP, relative to SA, is sufficient to drive the increased hedonic responding observed to concentrated sucralose in SP whereas the presence of a bitter-like signal is sufficient to influence SP’s intake such that sucralose is not treated as a unitary sweet-like compound. These findings are consistent with previous reports (Sclafani and Clare 2004; Loney et al. 2012b; Bacharach and Calu 2019) demonstrating that the phenotypic split in response to sucralose generalizes to the preference for saccharin solutions, although these differences are less pronounced than the clear dichotomy observed in response to sucralose. Furthermore, the decreased acceptance of concentrated saccharin solutions observed in SA, relative to SP, is a stable trait that is expressed prior to classification with sucralose suggesting an innate mechanism as opposed to a learned avoidance implemented by prior sucralose experience (Bacharach and Calu 2019). It follows then that whatever neural mechanisms contribute to the sucralose phenotype likely overlap with those responsible for behavioral responding for sucrose. It is important to note that we did not observe any behavioral or immunohistochemical differences between SA and SP in the response to intraoral sucrose at the concentration tested. As such, future studies concerned with elucidating the precise mechanism driving differences in the acceptance of sucralose should likely focus on the receptors and intracellular signaling mechanisms for sweet-taste transduction.
The taste reactivity measures were largely consistent with the FLI counts within the NTS. There were no statistical differences in SA and SP in either the ingestive or aversive oromotor responses (Table 1) following intraoral delivery of QHCl or sucrose just as there were no differences in either the total FLI or pattern of FLI across the IRgNTS subfields. In contrast, SA and SP differed in both the ingestive and aversive oromotor responses following intraoral sucralose with SA displaying significantly fewer ingestive responses and significantly more aversive responses. While no aversive oromotor responses were seen in SP in response to sucralose within the 1 min analysis period, SP were observed to passively drip sucralose much later into the 30 min infusion period suggesting that intraoral sucralose was not producing a pure sucrose-like quality. This further supports the FLI pattern analyses and previous microstructural analyses (Torregrossa et al. 2015) which indicate that, at least at this concentration of sucralose, SP display a pattern of FLI that is indicative of a mixed QHCl–sucrose response while the pattern in SA was indistinguishable from that of QHCl. Additionally, SP did not differ from SA in the amount of FLI present in the DM subfield, despite never producing gapes or other typically aversive oromotor responses. As such, it would appear that the FLI response within the DM subfield of the IRgNTS is primarily chemoresponsive as opposed to being impacted by differential oromotor movements.
In summary, these data further support findings that the phenotypic differences in the hedonic response to sucralose observed across outbred rats have a sensory origin. In addition, these data suggest a potential mechanism for these differences in that SA appear to be relatively insensitive to any sucrose-like taste component of sucralose, instead generating a pattern of NTS activation in response to oral sucralose indistinguishable from that of QHCl. In contrast, the activation of FLI in the NTS observed in response to sucralose in SP, appeared to be more in-line with a sucrose-QHCl mixture, consistent with the mixed “bitter-sweet” taste quality often associated with artificial sweeteners. Furthermore, these differences are observed at the level of the NTS, the first central relay for neural information conveying taste stimuli from the oral cavity. As such, these phenotypic differences may be indicative of differences in the physiology of peripheral receptors responsible for processing the sweet-like qualities of taste signals
Acknowledgments
The authors would like to thank Camille and Michael King for their expert technical assistance in preparing the histology analyses.
Funding
U.S. Department of Health and Human Services, National Institutes of Health, National Institute on Deafness and Other Communication Disorders (T32 DC-000044), State University of New York, University at Buffalo, Start-up Funds.
Conflict of interest
The authors have no potential conflicts of interest to declare.
