The Nasal Solitary Chemosensory Cell Signaling Pathway Triggers Mouse Avoidance Behavior to Inhaled Nebulized Irritants

Nasal solitary chemosensory cells mediate avoidance behavior to “bitter-tasting” irritant compounds

During the first session of the avoidance assay for control experiments, when both chambers contained nebulized saline, WT and KO mice showed no avoidance or preference, spending equal time in both chambers (Extended Data Fig. 3-1). For the avoidance behavior assay experiments, after a few brief explorations of the two chambers, WT mice learned quickly to avoid the chamber containing Den or Cyc, and settled in the saline chamber, while showing signs of discomfort, respiratory reflexes, and freeze behavior. Naive WT mice exposed to 10 mm Den or 10 mm Cyc spent statistically much more time (around 70% of the entire time) in the chamber containing nebulized saline solution (WT represented in green; one-way ANOVA followed by Tukey’s post hoc test, p < 0.0001; Fig. 3A). In contrast, KO mice missing elements of the SCC signaling cascade (TrpM5-KO and Gnat3-KO) or missing SCCs (Skn1a-KO) spent 36–46% of the time in the saline chamber, with no statistical preference for the chamber containing Den (all three KOs) or Cyc (TrpM5-KO; Fig. 3A,C, with C representing the data analysis of just the KO mice stimulated with either Den or Cyc from A; one-way ANOVA followed by Tukey’s post hoc test, p>0.99). During the experiments, mice never licked their fur and/or the chamber surfaces (Extended Data 1). The heat map in Figure 3B represents the second and third 5-min recording sessions for each mouse (n = 10), with red and blue representing preference and avoidance for the Den/Cyc, filled chamber, respectively. The absence of avoidance behavior and protective reflexes that drive the avoidance response to Den or Cyc in KO mice suggests that functional SCCs and activation of SCCs by Den or Cyc are necessary for the avoidance response.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Avoidance behavior to bitter compounds in wild-type (WT) mice and solitary chemosensory cell (SCC) pathway knock-out (KO) mice. A, Naive WT mice exposed to denatonium (Den) and cycloheximide (Cyc) exhibited avoidance responses, spending ∼70% of the time in the saline chamber. In contrast, the three KO mice showed no avoidance but instead a certain degree of interest in the irritant-containing chamber. n = 10 mice per group, 60 mice in total. Behavioral scores below 0 indicate avoidance of irritants and above 0 indicate attraction. ****p ≤ 0.0001, NS: no significance. Bars and symbols reflect mean ± SEM. B, Heat map representation of behavior score for each mouse for the data in A. Each rectangle represents the ratio of time a mouse spent in the left versus right chamber. Blue and red represent an increase in avoidance and preference, respectively, with color tones indicating intensity; gray is the baseline (behavior score = 0) representing indifference (see Extended Data Fig. 3-1 for supporting information/data). C, Representation of the data analysis of just the KO mice stimulated with either Den or Cyc from A. All KO mice exposed to either Den or Cyc show nonstatistical differences between each other in their behavioral score. NS: no significance. Bars and symbols reflect mean ± SEM.

Avoidance behaviors positively correlate with denatonium concentration

Naive WT mice responded to nebulized Den in a dose-dependent manner. When exposed to 2, 5, or 10 mm Den, WT mice exhibited a preference at 2 mm Den, a neutral behavior at 5 mm, and a marked avoidance response at 10 mm (two-way ANOVA followed by Tukey’s post hoc test; Fig. 4A). The fact that WT mice do not avoid nebulized Den at 2 mm is surprising considering that mice avoid drinking Den solution starting at 0.1 mm because of its bitter taste (Finger et al., 2005). It is difficult to make a correlation between chemicals delivered in the atomized form through the nose or dissolved in the mucus that plausibly reaches the tongue through the nasal cavity. Thus, although Den is delivered at 2 mm concentration as an aerosol into the nose, the threshold concentration that, presumably, reaches the tongue is unknown and, possibly, not enough to activate the bitter taste receptors. While we cannot imply that the taste system is not involved in the avoidance response with this experiment, we can speculate that as observed during the avoidance assay experiments mice are, probably, not getting any Den activating the taste system from either liking the mouse body hairs or direct contact with the chamber surfaces, and if any Den reaches the tongue, it would be retronasally at an unknown concentration. Another interesting observation in this experiment was that WT showed a preference instead of avoiding the chamber nebulized with 2 mm Den, suggesting that mice possibly used their sense of smell to detect Den and explored the chamber containing Den because it “smells like something” while the saline nebulized chamber had “no smell.” Coppola and colleagues reported that naive WT mice can smell Den as well as other bitterants, and mice with a bilateral bulbectomy do not learn or retain the ability to avoid Den vapor, implicating olfaction as the mode of detection (Coppola and Slotnick, 2018). We confirmed in this experiment that WT mice avoid 10 mm Den vapor in line with previous Ca-imaging experiments where SCCs detect Den starting at 10 mm (Gulbransen et al., 2008; Tizzano et al., 2010; Saunders et al., 2014), suggesting a primary role for the nasal SCCs in the avoidance behavior to Den.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Dose-dependent avoidance behavior to denatonium (Den) in naive wild-type (WT) and solitary chemosensory cell (SCC) pathway knock-out (KO) mice. A, Avoidance behavior increased in WT mice as the Den concentration increased. WT mice preferred 2 mm Den (den2) and avoided 10 mm Den (den10). TrpM5-KO, Gnat3-KO, and Skn1a-KO mice spent equal time in the irritant and saline aerosolized chambers, showing no significant difference at the three Den concentrations. n = 10 mice per group, 120 mice in total. Behavioral scores below 0 indicate avoidance of irritants and above 0 indicate attraction. ****p ≤ 0.0001, NS: no significance. Bars and symbols reflect mean ± SEM. B, Heat map representation of behavior score for each mouse for the data in A. Each rectangle represents the ratio of time the mouse spent in the left versus right chambers. Blue and red represent an increase in avoidance and preference, respectively, with color tones indicating intensity; gray is the baseline (behavior score = 0) representing indifference.

In contrast to WT, all the KO mice (TrpM5-KO, Gnat3-KO, and Skn1a-KO) lacked any significant differences in avoidance responses to 2, 5, and 10 mm Den (Fig. 4A). Naive KO mice did not show any statistically measurable preference behavior at all three Den concentrations (Fig. 4A). The TrpM5-KO, Gnat3-KO, and Skn1a-KO mice are missing elements or cells of the taste chemosensory pathway in several mucosae. In the MOE, MCs participate to maintaining the olfactory function and the lack of attraction to Den may be attributable to these chemosensory cells, thus their function may assist the OSNs in the detection of this compound. The heat map in Figure 4B represents the second and third 5-min recording sessions for each mouse (n = 10), with red and blue representing preference or avoidance for the Den-filled chamber, respectively. These data support the ideas that the avoidance behavior positively correlates with Den concentration and that an intact SCC signaling pathway is necessary for the avoidance response.

Improved avoidance in wild-type and preference in knock-out naive mice after repeated exposures to denatonium

The avoidance response in WT mice and the preference response in TrpM5-KO, Gnat3-KO, and Skn1a-KO mice improve significantly after repeated exposure to nebulized Den (Fig. 5). Naive WT mice exposed to 2 and 5 mm Den (time 0) showed no significant changes in the time spent in saline versus Den chamber during following exposures (times 1 and 6 h; Fig. 5A, top left). At 10 mm Den, the concentration that activates nasal SCCs, the avoidance response observed in naive WT mice was faster with significantly less time spent in the Den-filled chamber at time 6 h than at time 0 (two-way ANOVA followed by Tukey’s post hoc test; Fig. 5A, top left). The average time naive WT mice spent in the 10 mm Den chamber was 31.2%, which decreased to 25.8% and 18.8% at the second and third exposures (p = 0.0016). The Gnat3-KO and Skn1a-KO mice significantly improved their preference behavior after the first exposure to all concentrations of Den and spent more time in the Den-filled chamber (p < 0.0001; Fig. 5A, bottom), while the TrpM5-KO mice showed a significantly improved preference response only at 5 and 10 mm Den and at the 6-h time point (Fig. 5A, top right). Interestingly, at 2 mm Den only the WT but not the KO mice shows a consistent preference for Den at all three time points (0,1, and 6 h), with the KO mice learning after the first exposure to recognize Den using other clues than the SCC pathway (probably the smell of Den) and improving their preference for Den similarly to the WT mice (Extended Data Fig. 5-1). Both data in Figures 4A and 5A support the possible involvement of MCs in the preference behavior for 2–10 mm Den in naive mice. Upon repeated exposure to all three Den concentrations, other epithelial cells contribute to the preference response in the TrpM5-KO, Gnat3-KO, and Skn1a-KO mice. The absence of MCs in these KO mice suggests that MCs may be involved in the preference response to naive presentation of Den observed in WT mice, while the OSNs contribute by improving the preference behavior responses after repeated Den exposure in the KO mice and increase the avoidance response in the WT mice (Fig. 5A). The heat map (Fig. 5B) represents the second and third 5-min recording sessions for each mouse (n = 10) at each time point (0, 1, and 6 h), with red and blue colors representing preference and avoidance for the Den-filled chamber, respectively. These data show that the avoidance and preference responses in WT and KO mice positively correlate with the Den concentration and the number of subsequent exposures, indicating that mice can learn to identify Den more quickly with the assistance of other sensory systems, like taste and smell, that contribute to the behavioral responses after repeated exposure to Den.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Dose-dependent avoidance behavior after repeated exposure to denatonium (Den) in wild-type (WT) and solitary chemosensory cell (SCC) pathway knock-out (KO) mice. A, Avoidance in WT mice and preference in KO mice after the first naive exposure to Den improved after repeated exposures to all Den concentrations. WT mice showed faster avoidance responses for 10 mm but not for other concentrations of Den directly proportional to the number of repeated exposures. The KO mice also showed a significant improvement in preference behavior after repeated exposure either to Den 5 and 10 mm (TrpM5-KO) or to 2, 5, and 10 mm (Gnat3-KO and Skn1a-KO; see Extended Data Fig. 5-1 for supporting information/data). n = 10 mice per group, 120 mice in total. Behavioral scores below 0 indicate avoidance of irritants and above 0 indicate attraction. *0.01 < p ≤ 0.05, **0.001 < p ≤ 0.01, ***0.0001 < p ≤ 0.001, ****p ≤ 0.0001, NS: no significance. Bars and symbols reflect mean ± SEM. B, Heat map representation of behavior score for each mouse for the data in A. Each rectangle represents the ratio of time the mouse spent in the left versus right chambers. Blue and red represent an increase in avoidance and preference, respectively, with color tones indicating intensity; gray is the baseline (behavior score = 0), representing indifference.

Sensory input from the olfactory but not the taste system contributes to the behavior responses

WT mice can taste and avoid Den at 0.1–1 mm concentrations (Finger et al., 2005), while nasal SCCs respond to 10 mm or higher concentrations of Den (Gulbransen et al., 2008; Tizzano et al., 2010; Saunders et al., 2014). Thus, a mouse should be able to detect the bad “bitter” taste of 2 mm Den. However, in our avoidance assay experiments, WT mice did not avoid but showed a preference behavior for 2 mm Den (Fig. 4). Moreover, mice during the avoidance assays do not perform any licking of their fur or of the chamber surfaces (Extended Data 1). Altogether, the evidence suggests that the taste system is probably not involved in the behavioral response to vaporized Den at this concentration. In TrpM5-KO, Gnat3-KO, or Skn1a-KO mice, that lack both the SCC and taste pathways, the avoidance response to 2–10 mm Den was missing, implying that functional SCCs and activation of SCCs by Den are necessary for the avoidance behavior, but not providing conclusive evidence that would exclude the participation of the taste system in the avoidance behavior. As discussed above, the KO mice showed a preference behavior for the chamber nebulized with Den, probably based on input from a sensory system other than SCCs and taste.

To determine the role of the taste system in the behavioral responses to Den, we used the bitter-ageusic transgenic P2X2/3-double-KO mouse, which is missing a functional taste system. The P2X2/3-double-KO mouse avoids 10 mm Den similarly to the control WT (Fig. 6A). When naive or exposed repeatedly to 10 mm Den, both P2X2/3-double-KO and WT mice showed no statistical differences in their avoidance behavior. Like in previous experiments (Den 10 mm; Fig. 5A, top left), both P2X2/3-double-KO and wild-type mice seem to perform better in subsequent time point exposures (1 and 6 h; Fig. 6A). The heat map in Figure 6B represents recording sessions for each mouse (n = 10), with red and blue representing preference or avoidance for the Den-filled chamber, respectively. These data support the idea that the taste system is not involved in the avoidance behavioral response to Den and that the SCC signaling pathway is mainly responsible for the avoidance response.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Sensory input from the taste system does not contribute to the behavior responses. A, Bitter-ageusic transgenic P2X2/3-double-KO mice still avoid 10 mm Den, implying that SCCs but not the taste system are involved in the avoidance behavior response to nebulized irritants. Both P2X2/3-double-KO and wild-type mice respond similarly to Den at all time points. n = 10 mice per group, 20 mice in total. Behavioral scores below 0 indicate avoidance of irritants and above 0 indicate attraction. NS: no significance. Bars and symbols reflect mean ± SEM. B, Heat map representation of behavior score for each mouse for the data in A. Each rectangle represents the ratio of time mice spent in the left versus right chambers. Blue and red represent an increase in avoidance and preference, respectively, with color tones indicating intensity; gray is the baseline (behavior score = 0) representing indifference.

To test whether the olfactory system contributes to behavioral responses preferences, we pharmacologically ablated the olfactory epithelium (OE) and microvillous cells (MVCs; SCC-like cells located superficially on the OE) with a methimazole injection (MMI) that causes the detachment of several cell types in the OE from the basal lamina and consequent loss of the olfactory function (Fig. 7). Two days post-MMI, TrpM5-GFP-positive SCCs and the respiratory epithelium (RE) are still preserved (Fig. 7B’) and look anatomically no different from in untreated mice (Fig. 7A’). MMI treatment caused detachment of the OE from the basal lamina, and only a few pieces of dead OE with few decaying GFP-positive MVCs can be seen around the nasal turbinates (Fig. 7B’’). For the avoidance behavior assay, WT and TrpM5-KO, Gnat3-KO, and Skn1a-KO mice received MMI and, after 2 d, were tested for behavioral responses to 10 mm Den. WT mice avoided Den at 0-, 1-, and 6-h time points (two-way ANOVA followed by Tukey’s post hoc test; Fig. 8A), showing that the RE is still intact and the SCC pathway is fully functional. Interestingly, the significant improvement in the WT avoidance response to 10 mm Den observed at the 6-h time point (Fig. 5A) was missing in the WT+MMI group (Extended Data Fig. 8-1A, red arrow), showing that WT mice use the sense of smell to recognize Den and learn to avoid it more quickly after multiple exposures, spending less time in the Den-filled chamber. TrpM5-KO, Gnat3-KO, and Skn1a-KO mice with MMI lost the preference behavior to 10 mm Den (two-way ANOVA followed by Tukey’s post hoc test; Fig. 8A) or the quicker preference responses (Extended Data Fig. 8-1B–D, red arrows) that the KO mice showed before MMI. This confirms that the KO mice can smell Den and probably are attracted to the smell of the Den-filled chamber in the absence of the sensory discomfort mediated by SCCs, which are nonfunctional in these KO mice. Moreover, local hyperalgesic responses to Den were not observed in either WT nor KO MMI treated mice, as underlined by the absence of any increased avoidance or preference in MMI mice (Extended Data Fig. 8-1, red arrows). We believe that the nasal epithelium injuries caused by methimazole ablation of the OE are surely triggering inflammatory reactions, but this seems not to affect the behavior assay responses or cause hyper-responsiveness of the nasal epithelium. The heat map in Figure 8B represents the second and third 5-min recording sessions for each mouse (n = 10), with red and blue representing preference or avoidance for the Den-filled chamber, respectively.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

Ablation of the olfactory sensory epithelium after methimazole injection (MMI). Top, MMI experiment timeline and schematic representation for the sectioning orientation of the TrpM5-GFP mice with MMI and saline injection. Mice were injected at day 0 where the olfactory epithelium is still intact. Two days after MMI, mice were used for the avoidance behavior experiments. TrpM5-GFP mice injected with MMI were euthanized after 2 d, and tissue was collected for immunohistochemical examination to show the ablation of the olfactory epithelium and consecutive smell loss. A, Saline-injected mice showed normal respiratory (RE) and main olfactory epithelium (MOE). A’, A’’ are magnifications of the red insets in A showing intact RE and the presence of TrpM5-GFP-positive solitary chemosensory cells (SCCs) in the RE and green fluorescent protein (GFP)-positive microvillous cells (MCs) in the MOE. B, After 2 d, MMI mice showed a normal RE (B’) with healthy GFP-positive SCCs, but the MOE (B’’) was detached from the basal lamina and is nonfunctional, resulting in the loss of the sense of smell. In the MOE, GFP-positive MCs are still visible, crumpled at the back of the turbinate, but this tissue is not alive and will disappear in a few days.

• Download figure
• Open in new tab
• Download powerpoint

Figure 8.

Sensory input from the olfactory system contributes to the behavior responses. Mice treated with methimazole lose the sense of smell and lack the improvement of avoidance or preference behavioral responses present in untreated control mice. A, After methimazole injection, the significant improvement in the avoidance response to 10 mm Den (Fig. 5A) shown by the knock-out (KO) mice was missing in WT mice at all time points (see Extended Data Fig. 8-1 for supporting information/data). TrpM5-KO, Gnat3-KO, or Skn1a-KO mice injected with methimazole did not show preference behavior to 10 mm Den. n = 10 mice per group, 40 mice in total. Behavioral scores below 0 indicate avoidance of the irritant and above 0 indicate attraction. ****p ≤ 0.0001, NS: no significance. Bars and symbols reflect mean ± SEM. B, Heat map representation of behavior score for each mouse for the data in A. Each rectangle represents the ratio of time mice spent in the left versus right chambers. Blue and red represent an increase in avoidance and preference, respectively, with color tones indicating intensity; gray is the baseline (behavior score = 0) representing indifference.