Abstract
In the mammalian nose, two chemosensory systems, the trigeminal and the olfactory mediate the detection of volatile chemicals. Most odorants are able to activate the trigeminal system, and vice versa, most trigeminal agonists activate the olfactory system as well. Although these two systems constitute two separate sensory modalities, trigeminal activation modulates the neural representation of an odor. The mechanisms behind the modulation of olfactory response by trigeminal activation are still poorly understood. We addressed this question by looking at the olfactory epithelium (OE), where olfactory sensory neurons (OSNs) and trigeminal sensory fibers co-localize and where the olfactory signal is generated. Our study was conducted in a mouse model. Both sexes, males and females, were included. We characterize the trigeminal activation in response to five different odorants by measuring intracellular Ca2+ changes from primary cultures of trigeminal neurons (TGNs). We also measured responses from mice lacking TRPA1 and TRPV1 channels known to mediate some trigeminal responses. Next, we tested how trigeminal activation affects the olfactory response in the olfactory epithelium using electro-olfactogram (EOG) recordings from wild-type (WT) and TRPA1/V1-knock out (KO) mice. The trigeminal modulation of the olfactory response was determined by measuring responses to the odorant, 2-phenylethanol (PEA), an odorant with little trigeminal potency after stimulation with a trigeminal agonist. Trigeminal agonists induced a decrease in the EOG response to PEA, which depended on the level of TRPA1 and TRPV1 activation induced by the trigeminal agonist. This suggests that trigeminal activation can alter odorant responses even at the earliest stage of the olfactory sensory transduction.
SIGNIFICANCE STATEMENT Most odorants reaching the olfactory epithelium (OE) can simultaneously activate olfactory and trigeminal systems. Although these two systems constitute two separate sensory modalities, trigeminal activation can alter odor perception. Here, we analyzed the trigeminal activity induced by different odorants proposing an objective quantification of their trigeminal potency independent from human perception. We show that trigeminal activation by odorants reduces the olfactory response in the olfactory epithelium and that such modulation correlates with the trigeminal potency of the trigeminal agonist. These results show that the trigeminal system impacts the olfactory response from its earliest stage.
Introduction
Airborne chemicals are detected by olfactory sensory neurons (OSNs) in the olfactory epithelium (OE). The signal is peripherally transduced into action potentials, conveyed to the olfactory bulb (OB), and further processed and transmitted to cortical areas. Most studies of the olfactory system consider the OB as the first station of modulation of olfactory information (Schmidt and Strowbridge, 2014; Liu et al., 2015; Brunert and Rothermel, 2021). Few studies have explored modulation in the OE (Bouvet et al., 1988; Hegg et al., 2003; Daiber et al., 2013), although multiple systems could affect the olfactory signals in OSNs. The ethmoidal branch of the trigeminal nerve innervates both the OE and OB (Schaefer et al., 2002), and trigeminal-olfactory mutual modulation has been reported at the peripheral, central, and perceptual levels (Cain et al., 1980; Gudziol et al., 2001; Brand, 2006; Bensafi et al., 2007; Frasnelli et al., 2007; Lötsch et al., 2016; Tremblay and Frasnelli, 2018). fMRI studies showed cortical areas processing both nociceptive and olfactory stimuli (Bensafi et al., 2007; Lötsch et al., 2012; Pellegrino et al., 2017), while psychophysical studies demonstrated changes in trigeminal sensitivity influence the perception of odorants (Cain et al., 1980). The vast majority of odorants are also trigeminal agonists (Cometto-Muñiz and Cain, 1990; Cometto-Muñiz and Abraham, 2016). They typically activate the trigeminal system at medium to high concentrations, suggesting that when odorants enter the nasal cavity, both OSNs and trigeminal free-ending sensory fibers are activated (Doty et al., 1978; Cometto-Muñiz and Cain, 1990; W. Silver, 1992; Cometto-Muñiz and Abraham, 2016; Lötsch et al., 2016). When activated by odorants, different subsets of these trigeminal fibers will evoke specific sensations, described as pungent, tingling, stinging, burning, cooling, warming, painful, and irritating (Basbaum et al., 2009; Viana, 2011; Licon et al., 2018). Psychophysically, the trigeminal potency of odorants is described as the level of perceptual irritation they can evoke (Doty et al., 1978). Methods to determine the trigeminal potency of odors based on this definition are limited and provide only a subjective qualitative evaluation (Doty et al., 1978; Cometto-Muñiz et al., 2005). Currently, there is no quantitative parameter to complement such classification, highlighting the need to analyze trigeminal neuronal responses to odorants.
Transient receptor potential cation channels (TRP channels), such as vanilloid 1 (TRPV1), ankyrin (TRPA1), and melastatin 8 (TRPM8), play key roles in the detection of odorants by the trigeminal system (Nilius and Owsianik, 2011; Nguyen et al., 2017). TRPV1 and TRPM8 are largely expressed on different subsets of trigeminal sensory fibers, except for a small population of TRPM8-expressing neurons that express TRPV1 as well (Hjerling-Leffler et al., 2007; Takashima et al., 2007; Huang et al., 2012; Nguyen et al., 2017). While TRPA1 is mostly co-expressed with TRPV1, a population of trigeminal sensory neurons expresses only TRPV1 (Bautista et al., 2005; Kobayashi et al., 2005; Nguyen et al., 2017; Yang et al., 2022). TRPA1 and TRPV1-expressing sensory fibers are also peptidergic and, when activated, release ATP and neuropeptides such as calcitonin gene related peptide (CGRP) into the surrounding epithelium (Holzer, 1998; Ding et al., 2000; Fabbretti et al., 2006; Shevel, 2014). The nasal mucosa, including the OE, is extensively innervated by peptidergic trigeminal fibers, where they can be detected alongside OSNs (Schaefer et al., 2002; W.L. Silver and Finger, 2009; Daiber et al., 2013). Previous work has shown a reduction of the OE response to odorants during the application of CGRP or ATP (Hegg et al., 2003; Daiber et al., 2013). We asked whether trigeminal activation modulates OSN activity in the OE. We show that activation of TRPA1 and TRPV1 channels by odorants reduces OSN responses. The stronger the trigeminal potency of an odorant, the greater the inhibition of the olfactory response. This suggests that trigeminal fibers can regulate the odorant response at its earliest stage within the OE.
Materials and Methods
Animals and ethical approval
C57BL6/J mice (purchased from The Jackson Laboratory) were used as wild-type (WT) mice. TrpA1/V1-double knock-out (KO) mice, on a C57BL6 background (TRPA1/V1-KO), were a generous gift from Diana Bautista, University of California Berkeley (Gerhold and Bautista, 2008). In these mice, exon 23 (residues 901–951), which encodes the putative pore, and part of the sixth transmembrane domain of the TRPA1 receptor, is deleted (Bautista et al., 2006), as well as the fifth and all of the sixth putative transmembrane domains and the pore-loop domain of the TRPV1 receptor (Caterina et al., 2000). All animals were bred and housed in the animal facility of the Monell Chemical Senses Center in conventional polycarbonate caging with wood chip bedding (Aspen). Animals were kept on a 12/12 h light/dark cycle (lights on at 6 A.M., lights off at 6 P.M.), and ad libitum access to food and water. For all experiments, mice were euthanized during the light phase, typically between 11 A.M. and 2 P.M. A balanced number of male and female mice were used for all EOG experiments. The EOG experiments comprised 198 mice of the WT strain, with 95 males and 103 females. Among the 155 mice of the TRPA1/V1-KO strain, 74 were females and 81 were males. For the preparation of primary trigeminal ganglia, neonate mice were selected randomly, as their sex cannot be determined at this early stage.
All experimental procedures were performed in accordance with the National Institutes of Health Guidelines for the Care and Use of Animals and approved by the Monell Chemical Senses Center Animal Care and Use Committee. Every effort was made to minimize the number of animals used and their suffering.
Primary trigeminal culture
Animals were euthanized by CO2, followed by decapitation. To ensure a sufficient number of neurons for each experiment, we used trigeminal ganglia from three to four mouse neonates (postnatal days: 3-9) for each preparation. After being surgically removed, ganglia were transferred into Ca2+ and Mg2+ free HBSS including 1% penicillin/streptomycin (PS; 100 IU, and 100 µg/ml). For the dissociation, trigeminal ganglia were finely triturated, transferred into a 15-ml tube, and incubated in 5 ml 0.05% trypsin Ca2+ and Mg2+-free HBSS-PS solution for 10 min at 37°C. Five milliliters of HBSS-PS solution was then added to stop active trypsin and centrifuged for 3 min at 300 × g. The supernatant was carefully discarded. After that, the TGNs were incubated in 5 ml 0.05% collagenase A HBSS-PS solution for 20 min at room temperature, and 5-ml HBSS-PS solution was added and centrifuged for 3 min at 300 × g. The supernatant was discarded. One-milliliter DMEM was added into the tube and triturated ∼10–20 times at moderate force with a fire-polished pipette and seeded onto No#1 15-mm round coverslips coated by ConA at 37°C overnight.
Ca2+ imaging
For our experiments, we used five different stimuli: 2-phenylethanol (PEA), pentyl acetate (PA), cinnamaldehyde (CNA), allyl-isothiocyanate (AITC), menthol (MNT), and capsaicin (CAP). All were purchased from Sigma-Aldrich (CAS numbers: AITC, 57-06-7; CAP, 404-86-4; CNA, 14371-10-9; MNT: 89-78-1; PA: 628-63-7; PEA: 60-12-8); 1 m (PA, AITC, CNA, CAP, and MNT) and 3 m (PEA) stock solutions were first prepared in a methanol solution and then further diluted in Ringer solution to reach the final concentrations tested. The final methanol concentration in each solution was never >1%. Cellular responses to these odorants were measured using a ratiometric Ca2+ imaging technique as previously described (Gomez et al., 2005). The cells were loaded with 5 μm acetoxymethyl-ester of fura-2 (fura-2 AM) and 80 μg/ml pluronic F127 (Invitrogen, Eugene, Oregon) for at least 30 min at room temperature, settled in a recording chamber, and superfused with Ringer's solution or Ringer's solution containing different chemical compounds via a valve controller (VC-8, Warner) and perfusion pump (Perimax 12, SPETEC). Stimulation and washout duration was 20–30 s and 10 min at 3 ml/min perfusion rate, respectively, which depends on the chemical characteristics of the applied compound. There was a 10-s delay between solenoid valve activation and the arrival of stimulus compounds at the TGNs. Ca2+ imaging recordings were obtained using a Zeiss microscope equipped with a MicroMax RS camera (Roper Scientific Inc.) and a λ 10–2 optical control system (Sutter Instrument Co). Excitation from a monochromator was set at 340 and 380 nm with a 510-nm emission filter and the cellular fluorescence was imaged with a 10× objective (Zeiss). Images were digitized and analyzed using MetaFluor software (Molecular Devices). Among the multiple types of cells in the dissociated tissue preparation, TGNs were recognized based on morphology and positive response to 30 mm KCl. To compare trigeminal potency across stimuli, we chose concentrations based on EC50 determined in previous literature (Bandell et al., 2004; Jordt et al., 2004; Bautista et al., 2007; Elokely et al., 2016; Lieder et al., 2020; L. Xu et al., 2020). For PA and PEA, which were not previously characterized, we used concentrations of similar magnitude to their olfactory EC50. The change in fluorescence ratio (F340/F380) was calculated for regions of interest (ROIs) drawn manually around these cells. Response magnitudes were measured as the difference between the peak magnitudes (Fpeak) during the response window (90 s following presentation of stimulus) minus the mean baseline fluorescence (F0) and then divided by the mean baseline fluorescence [(Fpeak – F0)/F0]. Based on previous literature, we considered increases of intracellular Ca2+ >3% from the baseline level of fluorescence as responses (Bryant and Kraus, 2018; J. Xu et al., 2019).
Electro-olfactogram (EOG)
Twelve- to 24-week-old mice were euthanized by intraperitoneal injection of urethane (8 mg/g of body weight, ethyl carbamate, Sigma-Aldrich) followed by decapitation. We removed the skin and lower jaw and split the skull and nasal bone along the interfrontal and internasal sutures. The olfactory endoturbinates were then exposed by removing the nasal septum.
We used the electro-olfactogram (EOG) setup and procedure similar to one previously described previously (Cygnar et al., 2010). The half-head was mounted in an interface chamber with the sensory surface in constant contact with a stream of deodorized, humidified air at a flow rate of 3 l/min. For each odorant (same as above), we prepared 5 m stock solution in DMSO (Sigma-Aldrich), with the exception of MNT stock solution, which, because of its low solubility, was diluted to a 1 m stock. For dose–response experiments we used 10−1 to 10−7 serial dilutions of the stock solutions into water. Solutions were stored in glass vials with silicone stoppers and left to equilibrate with the air headspace for at least 30 min before the experiments. As a pure irritant stimulus, we used a CO2/air mixture (50% v/v), which was prepared using a gas proportioner multitube flowmeter (Cole-Palmer), and stored in a sealed glass flask sealed during the experiment. For stimulation, odorants or CO2/air mixture were injected into the air stream with pressure pulses (100 ms, 10 psi) using a pneumatic picospritzer system (Parker Hannifin). Stimuli were presented at 1 min intervals to allow the recovery of the epithelium.
Surface potentials from the endoturbinates 2 and 2b were recorded using two recording electrodes filled with 0.05% agarose melted in Ringer's solution (mm): 140 NaCl, 5 KCl, 2 CaCl2, 1 MgCl2, and 10 HEPES (7.4 pH). Recording pipettes were pulled from borosilicate glass capillaries (outer diameter: 1.5 mm, inner diameter: 0.87 mm) to a tip aperture of 20–25 μm using a Flaming–Brown puller (Sutter Instruments). The recording electrodes and a ground electrode were connected to two Warner DP-301 amplifiers. The 1 kHz low-pass signal was digitized (CED Micro 1401 mkII digitizer) and processed by a PC. Signal acquisition software (Cambridge Electronic Design) was used to acquire the data at a sampling rate of 2 kHz.
Data analysis
Recordings were analyzed using Origin (8.5v, OriginLab Corp.).
The EOG recording baselines were determined as the potential values before each odor stimulus and were subtracted from the trace. Net amplitudes of odorant-induced signals were grouped according to the mice genotypes and stimulus, and averages for each group are presented with the SEM.
All statistical analysis was conducted using Jamovi (version 1.6, the Jamovi project 2021; https://www.jamovi.org). Indicated significance levels were calculated using Welch's unpaired t test, one-way ANOVA (Welch's) with Tukey's post hoc test, and Kruskal–Wallis nonparametric one-way ANOVA with Dwass–Steel–Critchlow–Fligner post hoc tests as indicated; *p < 0.05, **p < 0.01, or ***p < 0.001.
Results
Trigeminal responses to odorants
Trigeminal responses to odorants were assessed using Ca2+ imaging (Fig. 1A,B). We imaged TGNs responses to five odorants: 2-phenylethanol (PEA), pentyl acetate (PA), cinnamaldehyde (CNA), allyl-isothiocyanate (AITC) and menthol (MNT). All odorants tested except PEA are previously characterized TRPA1 or TRPM8 agonists (Peier et al., 2002; Bandell et al., 2004; Bautista et al., 2005; Richards et al., 2010). In addition, we used capsaicin (CAP) to evaluate the population of TRPV1-expressing neurons (Caterina et al., 1997; W.L. Silver et al., 2006). The population sizes of TGNs responding to different odorants in each strain were calculated as the average percentage of neurons responding to a certain stimulus in each primary culture preparation. Imaging was conducted on a total of nine preparations for the WT (mean number TGNs responding to KCl/preparation ± SD: 233.7 ± 181.8) and five preparations for the TRPA1/V1-KO (380.8 ± 221.6) for a total number of 1386 TGNs from WT and 1683 TGNs from TRPA1/V1-KO mice.
TGNs responding to odorants. A, B, Example of Ca2+ transients evoked by PEA (10 mm), PA (5 mm), AITC (50 μm), MNT (50 μm), CNA (400 μm), CAP (100 nm), and KCl in different cells in WT and TRPA1/V1-KO. C, Percentage of TGNs responding to odorants across different preparations (black dots) of the primary culture in WT and TRPA1/V1-KO. D, Rate of TGNs responding to each stimulus in WT (gray) and TRPA1/V1-KO (KO, red). Significance was calculated using Kruskal–Wallis nonparametric one-way ANOVA with Dwass–Steel–Critchlow–Fligner post hoc test. WT versus KO, *p < 0.05, **p < 0.01.
In WT mice, 35.6 ± 4.8% TGNs responded to any odorant, while in TRPA1/V1-KO only 7.5 ± 2.3% TGNs responded to the chemosensory stimuli we tested (Fig. 1C), showing a significant reduction of chemosensory TGNs in the TRPA1/V1-KO mice (p = 0.004, Kruskal–Wallis nonparametric one-way ANOVA, with Dwass–Steel–Critchlow–Fligner post hoc test). In WT (Fig. 1D), CAP activated the largest population of neurons (25.9 ± 4.9%), followed by PA (14.3 ± 2.1%), PEA (12.1 ± 2.1%), AITC (11.2 ± 2.8%), MNT (9.4 ± 1.5%) and CNA (6.35 ± 1.8%). This estimation of the population of neurons activated by CAP, a TRPV1 agonist, and AITC, a TRPA1 agonist, aligns with a previous characterization of TRPA1 and TRPV1-expression in the trigeminal ganglion of WT mice (Huang et al., 2012).
TGN primary cultures from TRPA1/V1-KO mice showed a significantly lower percentage of neurons responding to CAP (0.17 ± 0.17%, p = 0.002, Kruskal–Wallis nonparametric one-way ANOVA, with Dwass–Steel–Critchlow–Fligner post hoc test), CNA (0.21 ± 0.13%, p = 0.003), AITC (0.32 ± 0.2%, p = 0.003), PA (6.42 ± 1.9%, p = 0.014) and PEA (2.03 ± 1.1%, p = 0.004) being significantly reduced in comparison to WT (Fig. 1D). No significant differences were observed in the number of TGNs responding to MNT (5.7 ± 1.01%; Fig. 1D). These results suggest that the activation of TRPA1 and TRPV1 channels is the main mechanism necessary to evoke Ca2+ increases in response to AITC, CNA, and CAP. Furthermore, our results indicate that PEA and PA detection by the trigeminal system also relies partially on TRPA1 and TRPV1 channels.
For each odorant, we then obtained dose–response curves in the responsive subset of TGNs (Fig. 2). Maximal response amplitudes (max ΔF/F) to AITC, CNA, and PA were significantly reduced in TRPA1/V1-KO (Welch's unpaired t test, AITC: df = 19, p < 0.001; CNA: df = 52, p < 0.001; PA: df = 183, p = 0.006), while no changes were observed for MNT and PEA (Welch's unpaired t test, PEA: df = 66.6, p = 0.06; MNT: df = 134, p = 0.313). In WT, MNT had the lowest EC50 (13.86 ± 2.70 μ, n = 42), followed by AITC (61.63 ± 13.28 μ, n = 20), CNA (0.264 ± 0.08 mm, n = 53), PA (2.64 ± 0.55 mm, n = 70), and PEA (8.53 ± 1.66 mm, n = 35). In TRPA1/V1-KO, EC50 of MNT (10.14 ± 1.85 μm, n = 94) and PA (1.67 ± 0.87 mm, n = 128) showed no significant changes (Welch's unpaired t test, MNT: df = 80, p = 0.26; PA: df = 192, p = 0.35). TGNs of TRPA1-V1-KO responding to PEA showed decreased sensitivity to the odorant (77.77 ± 19.3 mm, n = 46, Welch's unpaired t test, p < 0.001, F = 45), while AITC and CNA did not evoke any response in TRPA1/V1-KO. Overall, we observed changes in either EC50 or (max ΔF/F), for all odorants except MNT, suggesting no involvement of TRPA1 or TRPV1 in its detection by the trigeminal system.
Calcium responses of TGNs to odorants. Dose–response curves from TGNs in WT (black) and TRPA1/V1-KO (KO, red) for the odorants (A) AITC, (B) CNA, (C) PEA, (D) PA, and (E) MNT.
Physiologic classification of trigeminal activation by odorants
To quantify the activity evoked by odorants in single cells expressing a given receptor we used an “activity index,” which was defined as (−log(EC50 (M)) × max ΔF/F) (del Mármol et al., 2021). We used a similar approach to quantify the overall activity induced by each agonist across the population of TGNs (Fig. 3A). We multiplied the activity index of each odorant in each mouse strain by the percentage of TGNs they activated, generating a “weighted” activity index (WAI; Fig. 3B). AITC and CNA activity indexes in TRPA1/V1-KOs are zero, because of the lack of responses to these odorants in the absence of TRPA1 and TRPV1 channels. Finally, to summarize the quantitative trigeminal properties of an odorant, we subtracted WAIKO from WAIWT to obtain a single score (TRPA1/V1-score) for each odorant (Fig. 3C). AITC had the highest TRPA1/V1-score, followed by CNA, MNT, PEA and PA.
Physiologic classification of trigeminal activation by odorants. A, Activity index of all odorants in WT (gray), and TRPA1/V1-KO (red). B, Weighted activity index of all odorants in WT (gray), and TRPA1/V1-KO (KO, red). C, TRPA1/V1-score, score values are associated with the level of activation of the TRPA1/V1-expressing fibers. The higher is the score value, higher is the level of activation.
Odorant responses in TRPA1/V1-KO mice
Using the EOG technique, we assessed whether the lack of expression of TRPA1 and TRPV1 receptors in the OE could affect the response to odorants. We determined dose–response relations in the OE for all the odorants previously tested on TGNs (except CAP, which is not an odorant). Peak response amplitudes for each concentration were fitted with a Hill equation to obtain a dose–response curve. CNA, MNT, and PEA showed no significant differences among dose–response curves obtained in WT and TRPA1/V1-KO (Fig. 4). The absence of TRPA1 and TRPV1 channels in the KO mice resulted in a leftward shift of the dose–response curve of AITC and PA, both characterized by a significant reduction of Ks (AITC: F(1,7.27) = 4.73, p = 0.047; PA: F(1,7.28), p = 0.19; Welch's one-way ANOVA, Tukey's post hoc test; Fig. 4L) and therefore sensitization to odorants. Maximal response amplitudes (Vmax) were unchanged in WT and TRPA1/V1-KO across all odorants (Fig. 4K).
EOG responses to different odorants in WT and TRPA1/V1-KO. A, C, E, G, I, Examples of EOG responses in WT (black) and TRPA1/V1-KO (KO, red) to two different concentrations (10−2 M and 10−4 M) of odorants. B, D, F, H, J, Dose–response curves obtained in WT (black) and TRPA1/V1-KO (KO, red) for AITC (B, WT n = 8, KO n = 8), CNA (D, WT n = 15, KO n = 8), PEA (F, WT n = 15, KO n = 17), PA (H, WT n = 8, KO n = 13), MNT (J, WT n = 8, KO n = 8). K, L, Maximal response amplitude (Vmax) and Ks obtained for all odorants in WT (black) and TRPA1/V1-KO (KO, red). Significance was calculated using a Welch's one-way ANOVA with Tukey's post hoc test. WT versus TRPA1/V1-KO, *p < 0.05.
Repeated exposure of the OE to irritants reduces the EOG response to odorants
To establish whether trigeminal activation by odorants can modulate the olfactory response, we performed EOG recordings, in which we alternated brief stimulations of the OE with PEA (0.1 m), the odorant with the lowest trigeminal potency, followed by exposure of the OE to a trigeminal agonist to activate the trigeminal sensory fibers. We applied three pulses of PEA (100 ms) to the OE alternated with pulses of a given trigeminal agonist (100 ms), followed by three more PEA pulses (Fig. 5A). In between each stimulus, we allowed 1 min for the OSNs to recover from the previous stimulation and to avoid olfactory adaptation. To determine how the response to PEA would change during the recording session, in the absence of any trigeminal stimulus, we alternated PEA and a pulse of nonodorized air. EOG responses to PEA were normalized by dividing each EOG peak amplitudes (V) by the amplitude of the first PEA response (V0). The means of the normalized EOG responses were then compared (Fig. 5D,E). In the control, we observed a decrease of the PEA response amplitude during the experiment reaching ∼22% in the last PEA stimulus (Fig. 5B,D). This decline was observed in both WT and TRPA1/V1-KO. We then repeated the same experiment using CO2 (50% v/v, 100-ms pulse), a potent TRPA1 agonist (Fig. 5C,E). In WT, CO2 induced a reduction of the PEA response of ∼53% (Fig. 5E, black). Such stark reduction of the EOG response was not observed in TRPA1/V1-KO mice (Fig. 5E, red), in which the responses to PEA were no different from the control.
Repeated exposure of the OE to CO2 reduces the EOG response to PEA. A, Sequence representing the experimental protocol. B, C, Example of EOG responses to PEA in WT when OE was exposed to air or CO2. D, E, Mean EOG responses to PEA (V) normalized to the amplitude of the first PEA response (V0) (Norm. EOG response (V/V0)) in WT (black) and TRPA1/V1-KO (KO, red) after exposure to air (D, WT n = 19; KO n = 7) or to CO2 (E, WT n = 14; KO n = 12). Significance was calculated using Kruskal–Wallis nonparametric one-way ANOVA with Dwass–Steel–Critchlow–Fligner post hoc test. WT versus KO, *p < 0.05.
We then addressed whether odorants, which are also trigeminal agonists, could modulate the OE activity. Replacing the intermittent exposure of the OE to CO2 with AITC (0.1 m), which has the highest TRPA1/V1-score, induced a progressive reduction of the PEA response in WT, which was abolished in TRPA1/V1-KO mice (mean V3''/V0 WT: 0.36 ± 0.087, n = 11; KO: 0.67 ± 0.076, n = 10, p = 0.024, Kruskal–Wallis nonparametric one-way ANOVA, with Dwass–Steel–Critchlow–Fligner post hoc test; Fig. 6A).
Repeated exposure of the OE to irritants reduces the EOG response to odorants. A–F, Mean EOG responses to PEA (V) normalized to the amplitude of the first PEA response (V0) (Norm. EOG response (V/V0) after exposure of the OE to 0.1 m AITC (A), CNA (B), PEA (C), PA (D), MNT (E) and 1 mm AITC (F). Comparison of mean normalized EOG responses to PEA in WT (G) and TRPA1/V1-KO (H) after the exposure of the OE to 0.1 m AITC (triangle, dark orange lines), 1 mm AITC (circles, light orange lines), and Air (squares, gray lines). Significance was calculated using Kruskal–Wallis nonparametric one-way ANOVA with Dwass–Steel–Critchlow–Fligner post hoc test. WT versus KO, *p < 0.05, **p < 0.01.
CNA (0.1 m) and PEA (0.1 m) did not induce a difference among EOG responses in WT and TRPA1/V1-KO, which declined to the same rate by the end of the experimental protocol (Fig. 6B,C). In WT, we observed a small and temporary enhancement of the olfactory response after three CNA stimuli (WT: V3/V0: 0.81 ± 0.04, n = 23; KO: 0.75 ± 0.05, n = 22, p = 0.017). Similarly, in WT, we observed the same enhancement of the relative response amplitude (V/V0) of the stimuli 1 and 3', when PEA was used as the trigeminal agonist, but the EOG responses in response to the olfactory stimuli thereafter (2 and 3'') were not significantly different from TRPA1/V1-KO (mean V1/V0 WT: 1.04 ± 0.0481, n = 20; KO: 0.89 ± 0.03, n = 13; p = 0.011. mean V3'/V0 WT: 0.81 ± 0.06, n = 20; KO: 0.64 ± 0.06, n = 13, p = 0.031). The stimulation of the OE by PA (0.1 m) and MNT (0.02 m) induced a more robust and sustained decay of the response to PEA in TRPA1/V1-KO (Fig. 6D,E), which persisted until the end of the recordings (PA mean V3''/V0 WT: 0.87 ± 0.07, n = 18; KO: 0.64 ± 0.07, n = 13, p = 0.038; MNT mean V3''/V0 WT: 0.66 ± 0.07, n = 16; KO: 0.48 ± 0.05, n = 13, p = 0.02).
We then repeated the previous experiment exposing the OE to a lower concentration of AITC (1 mm) to test whether the trigeminal modulation of the olfactory response is concentration dependent. Exposing the OE to 1 mm AITC still induced a significant reduction of the odor response in WT in comparison to TRPA1/V1-KO (mean V3''/V0 WT: 0.63 ± 0.06, n = 23; KO: 0.96 ± 0.15, n = 11, p < 0.01; Fig. 6F), but significantly smaller in comparison to the one elicited by 0.1 m AITC in the same strain (p = 0.034), and not significantly different to the control with air (Fig. 6G). In TRPA1/V1-KO mice, neither concentration of AITC altered the PEA response relative to the control (Fig. 6F).
TRPA1/V1-score correlates with the level of trigeminal modulation of OE response to odor in WT
We next determined whether the ability of an odorant to activate the trigeminal sensory fibers correlates with the reduction of the olfactory response induced by the same odorant in the OE (Fig. 7). For both strains, we plotted the reduction of the PEA response (%) induced by the odorant against its TRPA1/V1-score. All data points were fitted with a linear function (WT: intercept = 16.44 ± 4.32; slope = 1.17 ± 0.25; Fig. 7B, KO: intercept = 45.84 ± 5.55; slope = −0.26 ± 0.33; Fig. 7A). In WT, the reduction of the olfactory response correlates with the TRPA1/V1-score of the odorant (R = 0.93, p = 0.019), while changes in odor responses in TRPA1/V1-KO are not linked to the trigeminal properties of the odor (R = 0.41, p = 0.49).
TRPA1/V1-score correlates with the level of trigeminal modulation of OE response to odor in WT. Correlation TRPA1/V1-score and reduction EOG response (%) in WT (A) and TRPA1/V1-KO (B).
Discussion
In this work, we quantified the trigeminal activity induced by odorants and how it modulates the olfactory response generated in the OE. Until now, trigeminal potency has been described using only psychophysical approaches (Doty, 1975; Doty et al., 1978; Cometto-Muñiz and Cain, 1990; Frasnelli and Hummel, 2007; Cometto-Muñiz and Abraham, 2016). With such methods, it is hard to separate the two sensory modalities evoked by the odorant, and they provide only a subjective evaluation of the perception evoked. Assessments of trigeminal potency of odors by patients with olfactory loss eliminates the olfactory interference from the measurements, but acquired anosmia is associated with an alteration of trigeminal perception as well (Gudziol et al., 2001). While more objective methods to measure trigeminal responses to odorants from patients like the recording of the negative mucosal potential and functional magnetic resonance are less suitable for screenings on a large scale (Kratskin et al., 2000; Bensafi et al., 2012; Pellegrino et al., 2017). Based on the responses to odorants obtained in TGNs, we developed the TRPA1/V1-score, a physiological classification of the trigeminal potency of odorants. For each odorant, this score incorporates its activity index, the size of the trigeminal population activated by it, and if it activates TRPA1 and/or TRPV1 channels. Although this score does not provide a further distinction among different chemosensory TRP channels, it is the first quantitative measure of trigeminal potency without any olfactory interference and independently from human perception. Previously, a few works have suggested the possibility of trigeminal/olfactory interaction at the periphery. Tracing of the trigeminal innervation of the nasal cavity showed previously that peptidergic sensory fibers from the ethmoidal branch of the trigeminal nerve innervate the OE and OB (Finger and Böttger, 1993; Schaefer et al., 2002).
Previous work from Hegg et al., and Daiber et al., showed that ATP and the neuropeptide CGRP can both modulate OSN responses to odorants (Hegg et al., 2003; Daiber et al., 2013). Both compounds are released on stimulation by trigeminal sensory fibers, which express TRPA1 and TRPV1 channels. Our work directly builds on Daiber's and Hegg's findings (Hegg et al., 2003; Daiber et al., 2013), addressing whether the exposure of the OE to odorants with different trigeminal potencies could modulate the olfactory response and whether different levels of trigeminal activation would affect the olfactory response differently. Our results suggest that TRPA1/V1-agonists induce a graded modulation of the olfactory response to PEA, which correlates with the level of trigeminal activation they induce. This correlation is drastically reduced in the absence of TRPA1 and TRPV1 expression, suggesting that the TRPA1/V1-score could constitute an indicator of trigeminal potency.
This modulatory mechanism likely originates from trigeminal sensory fibers rather than other cell types in the OE. Single-cell RNA-seq obtained from the OE shows a lack of expression of TRPA1 in non-neuronal cells (Tsukahara et al., 2021). Low levels of expression of TRPV1 have been detected in Trpm5+/Chat+ microvillar cells, but the involvement of these cell types in the modulation seems unlikely since AITC and CO2 are both TRPA1 and not TRPV1 agonists.
A subpopulation of trigeminal TRPA1/V1-expressing fibers are peptidergic free nerve endings, which, when stimulated, can release neuropeptides such as CGRP or neuromodulators like ATP. The activation of this population of sensory neurons by odorants might induce the release of different amounts of ATP and CGRP, as measured by the score TRPA1/V1-score. Previous studies have shown that ATP and CGRP can reduce the olfactory response in the OE (Hegg et al., 2003; Daiber et al., 2013), possibly driving Ca2+ in the dendritic and soma compartments. The increase of intracellular Ca2+ could then activate Ca2+-activated K+ currents (Kawai, 2002) and, consequently, decrease OSN responses evoked by the following stimulus. OSNs express purinergic receptors P2X4 and P2Y2 (J. Xu et al., 2016; Tsukahara et al., 2021). The release of ATP into the extracellular space could open P2X4 expressed on the membrane of OSNs and drive an intracellular Ca2+ increase (Stokes et al., 2017). Activation of P2Y2 receptors in OSNs would initiate the PLC-mediated Ca2+ signaling cascade, which leads to the release of Ca2+ from intracellular stores. Purinergically-induced intracellular Ca2+ increase in the OSNs might therefore contain two phases, an early one, driven by P2X4, and a delayed one, mediated by P2Y2. ATP in the intracellular space is quickly degraded, therefore, combining both P2X and P2Y receptors might be crucial to provide a more sustained Ca2+ increase able to affect the odor response. The neuropeptide CGRP, which is also released by trigeminal peptidergic fibers, was also shown to modulate OSN responses to odorants (Daiber et al., 2013). The activation of the CGRP receptor leads to the activation of adenylate cyclase followed by an increase of cAMP (Russell, 2011) and consequentially to the rise of intracellular levels of Ca2+. In the OSNs, CGRP has been shown to induce increases in cAMP (Daiber et al., 2013), which could contribute to driving intracellular Ca2+ increase and affect their response to odorants.
Taken together, our study shows that odorants can simultaneously activate both the olfactory and trigeminal systems in the OE and that TRPA1/V1-expressing trigeminal fibers can modulate the OSN response to odors. Such modulation is a graded reduction of the olfactory signal, which correlates with the odorant's TRPA1/V1-score. While the TRPA1/V1-score we introduced demonstrates the potential to predict the influence of prior trigeminal agonist exposure on OSN activity, it is important to acknowledge that our current results are based on a limited set of odorants. To comprehensively assess the extent of trigeminally active odorants' impact on OSN responses, further studies using a wider range of odorants will be necessary. Moreover, since in our experiments, we used a constitutional knock-out mouse for TRPA1 and TRPV1 channels, we cannot exclude the possibility of changes in the expression of other chemosensory TRP channels, such as TRPM8. Notably, the role of TRPM8-expressing sensory neurons in the periphery is complex and likely to play multiple roles (Peier et al., 2002; Proudfoot et al., 2006; Bautista et al., 2007; Takashima et al., 2007). Previous studies have shown a subpopulation of TRPM8 neurons expressing CGRP and TRPV1 (Takashima et al., 2007; Yang et al., 2022), raising the possibility that the exposure to MNT could induce the modulation of the olfactory response by activating this specific population of sensory neuron in WT and TRPA1/V1-KO. Such a hypothesis would be supported by the changes in the MNT dose–response curve we observed in TRPA1/V1-KO primary TGNs as well as by the increased reduction of the olfactory response induced by MNT in TRPA1/V1-KO.
Overall, the mechanism we describe supports and complements the previous findings of a peripheral modulation of the olfactory signal by the trigeminal system and underscores the necessity of taking into account the trigeminal potency of an odorant when analyzing olfactory sensory processing. Furthermore, the role of the trigeminal activation might be particularly relevant when considering odor mixtures encoding, with more than one component able to simultaneously activate the trigeminal system.
Footnotes
This work was supported by National Institutes of Health National Institute on Deafness and Other Communication Disorders Grants R21DC018358 (to F.G.); R01DC016598, R03DC012413, and R01DE028979 (to M.T.), and 1R01DC016647 (to J.R.). We thank Dr. Kevin Bolding and Dr. Joel Mainland for discussing the manuscript and Minliang Zhou for the help with the animal colonies.
The authors declare no competing financial interests.
- Correspondence should be addressed to Federica Genovese at fgenovese{at}monell.org
