A Novel Flp Reporter Mouse Shows That TRPA1 Expression Is Largely Limited to Sensory Neuron Subsets

Abstract Transient receptor potential ankyrin 1 (TRPA1) is a polymodal cation channel that is activated by electrophilic irritants, oxidative stress, cold temperature, and GPCR signaling. TRPA1 expression has been primarily identified in subsets of nociceptive sensory afferents and is considered a target for future analgesics. Nevertheless, TRPA1 has been implicated in other cell types including keratinocytes, epithelium, enterochromaffin cells, endothelium, astrocytes, and CNS neurons. Here, we developed a knock-in mouse that expresses the recombinase FlpO in TRPA1-expressing cells. We crossed the TRPA1Flp mouse with the R26ai65f mouse that expresses tdTomato in a Flp-sensitive manner. We found tdTomato expression correlated well with TRPA1 mRNA expression and sensitivity to TRPA1 agonists in subsets of TRPV1 (transient receptor potential vanilloid receptor type 1)-expressing neurons in the vagal ganglia and dorsal root ganglia (DRGs), although tdTomato expression efficiency was limited in DRG. We observed tdTomato-expressing afferent fibers centrally (in the medulla and spinal cord) and peripherally in the esophagus, gut, airways, bladder, and skin. Furthermore, chemogenetic activation of TRPA1-expressing nerves in the paw evoked flinching behavior. tdTomato expression was very limited in other cell types. We found tdTomato in subepithelial cells in the gut mucosa but not in enterochromaffin cells. tdTomato was also observed in supporting cells within the cochlea, but not in hair cells. Lastly, tdTomato was occasionally observed in neurons in the somatomotor cortex and the piriform area, but not in astrocytes or vascular endothelium. Thus, this novel mouse strain may be useful for mapping and manipulating TRPA1-expressing cells and deciphering the role of TRPA1 in physiological and pathophysiological processes.


Significance Statement
The ion channel transient receptor potential ankyrin 1 (TRPA1) is activated by oxidative stress, pollutants, cold, and downstream of GPCR activation.TRPA1 is expressed on some nociceptive sensory afferent nerves, and its activation evokes pain and nocifensive reflexes.However, there is disagreement regarding the extent of TRPA1 expression, both within specific afferent subsets and in nonafferent cells.We developed a TRPA1 Flp mouse that expresses Flp recombinase in TRPA1-expressing cells, allowing for their identification and manipulation.TRPA1 was reported in vagal and DRG subsets, and their chemogenetic activation in the skin evoked pain.TRPA1 was also reported in gut subepithelial cells, cochleal supporting cells, and some cortical neurons.We found no evidence of TRPA1 reporting in enterochromaffin cells, epithelium, endothelium, keratinocytes, Schwann cells, smooth muscle cells, or astrocytes.
Most initial screenings of TRPA1 transcript expression found little evidence in cell types other than sensory afferents (Story et al., 2003;Nagata et al., 2005;Jang et al., 2012).However, subsequent studies have demonstrated evidence of TRPA1 expression in multiple nonafferent cell types, including keratinocytes (Kwan et al., 2009), epithelium (Nassini et al., 2012), enterochromaffin cells (Cho et al., 2014;Bellono et al., 2017), cerebral artery endothelium (Earley et al., 2009), astrocytes (Shigetomi et al., 2012), Schwann cells (De Logu et al., 2017), and CNS neurons (Koch et al., 2011).In some cases, TRPA1 expression was identified using low-specificity antibodies or very low transcript numbers, and thus their determination is controversial.TRPA1 has been implicated in anxiety (de Moura et al., 2014), cognition, and memory (Lee et al., 2016a,b), and thus it is important to fully map its expression to take into account potential side effects of therapeutics developed to target TRPA1.
Here we have developed a TRPA1 Flp reporter mouse, by replacing the endogenous TRPA1 stop codon with a 2A-Flp O cassette.Flp O is a codon modified and thermostable version of Flp recombinase (Raymond and Soriano, 2007).Similar to the commonly used Cre/lox system, Flp recombines specific FRT sites in DNA (Fenno et al., 2014), thus allowing for selective manipulation of targeted genes in TRPA1-expressing cells.By crossing the TRPA1 Flp mouse with a reporter strain (R26 ai65f ) that expresses the fluorescent tdTomato in a Flp-sensitive manner, we have mapped TRPA1 expression to subsets of DRGs and vagal ganglia neurons innervating multiple peripheral organs, including the esophagus, gut, airways, bladder, and skin.Furthermore, chemogenetic activation of TRPA1-expressing fibers evokes acute nocifensive/pain behaviors.We found very limited tdTomato expression in other cell types, but most notably in subepithelial cells in the mucosa of the stomach and colon, although these did not express serotonin, a marker of enterochromaffin cells.Our TRPA1 Flp mouse may prove useful for mapping and manipulating the contribution of TRPA1-expressing cells to physiological and pathophysiological processes.

Tissue collection and immunohistochemistry
Mice were killed by CO 2 inhalation and transcardially perfused with ice-cold PBS.Vagal ganglia and DRGs from T1 to T5 and L3 to L5 were dissected out and fixed for 1 h in 3.7% formaldehyde (FA) at 4°C.Brainstem and spinal cord were dissected out and fixed for 4 h in 3.7% FA at 4°C.Brain, stomach, and colon were collected and fixed overnight in 3.7% FA at 4°C.Glabrous and hairy skin was removed from the hindpaw and fixed in 3.7% FA at 4°C overnight.For the cochlea, the temporal bones were collected and fixed in 4% PFA overnight at 4°C.The temporal bones were washed three times in PBS for 10 min and transferred to 10% EDTA, dissolved in H 2 O at pH 7.2, solution for overnight decalcification at 4°C.Tissue firmness was checked by pressing tissue with forceps to ensure adequate decalcification.The trachea, esophagus, and bladder were dissected out and cut longitudinally, with the ends pinned in a SYLGARD-covered dish and fixed in 3.7% FA overnight at 4°C.The muscle and the mucosal layers of the esophagus were carefully separated using fine scissors.The tissues were then unpinned and rinsed free floating three to four times at room temperature in PBS.
All tissues to be sliced using a cryostat were first cryoprotected in 20% sucrose.Stomach, colon, and skin were sectioned in 20 mm slices, ganglia were sectioned sagitally in 20 mm slices, brainstem and spinal cord were sectioned coronally in 40 mm slices, and brains were sectioned sagitally and coronally in 80 mm slices and collected onto Superfrost Plus slides.Slides were air dried at room temperature in the dark overnight.Slides were washed with PBS three times for 10 min, and tissue was permeabilized with 0.3% Triton X-100 in PBS (PBSTx) for 15 min followed by blocking for 1 h with 1% bovine serum albumin (BSA), 10% donkey serum (DS), and 0.3% PBSTx.Tissue was incubated with primary antibodies (Table 1) diluted in blocking buffer overnight at 4°C.After washing with 0.2% Tween 20 in PBS (PBST) three times for 10 min, the tissue was incubated with secondary antibodies (Table 1) in 1% BSA/5% DS in 0.2% PBST for 1 h.The tissue was washed Complete lungs with attached bronchi and trachea were dissected out and fixed in 3.7% FA overnight at 4°C with gentle agitation.Lungs were washed with ice-cold PBS three times for 30 min, followed by cryoprotection in 30% sucrose solution.The lungs were flushed with PBS three times and inflated with 2% low melting agarose solution.After the agarose solution had solidified, the lung lobes were separated and snap frozen with OCT compound for cryosection.Eighty micrometer lung slices were collected in cryoprotectant-filled wells and stored in À20°C.Using identical blocking solutions, permeabilizing solutions, and antibody solutions (see above), lung slices were washed in PBS three times for 10 min and permeabilized for 20 min.Lung slices were then blocked for 1.5 h followed by primary antibody incubation overnight at 4°C.The slices were then washed three times for 20 min and incubated with secondary antibodies for 2 h.The slices were washed again and mounted onto glass slides with VECTASHIELD Antifade Mounting Medium with DAPI.
The staining of the esophageal muscle and mucosa, trachea and bladder were performed in 0.5 ml tubes.The washing in filtered PBS between staining steps was performed in 15 ml tubes on the rotor at room temperature.Tissues were permeabilized in 1% Tween 20 (Sigma-Aldrich) in filtered PBS at room temperature for 6 h and washed in filtered PBS (3 Â 20 min on rotator).All tissues except trachea were then incubated in blocking solution [1% BSA in PBS (BSA/PBS)] with four drops of Avidin solution (Avidin/Biotin Blocking Kit; catalog #SP-2001, Vector Laboratories) and incubated at 4°C for 4 h.The tissues were then washed 5 Â 20 min on rotator and then incubated with primary antibodies diluted in BSA/PBS with four drops of Biotin (as necessary) for 48 h at 4°C.The tissues were then washed in filtered PBS (10 Â 30 min), incubated with secondary antibodies and/or goat biotin-XX conjugate anti-rabbit IgG (H 1 L) secondary antibody (catalog #B2770, Thermo Fisher Scientific) diluted 1:100 in BSA/PBS overnight at room temperature, then washed in PBS (10 Â 20 min), incubated with streptavidin (Alexa Fluor 548 conjugate; catalog #S11223, Thermo Fisher Scientific) diluted 1:100 in filtered PBS for 5 h at room temperature; washed in filtered PBS (3 Â 20 min); incubated in anti-fade, pH 8.6, glycerol (Sigma-Aldrich) for 24 h at room temperature; and stored at 4°C.The esophagus preparation was placed mucosal side up on a glass slide and covered with a coverslip (24 Â 50 mm).
Tracheal tissue was blocked with BSA/PBS alone at 4°C for 4 h.The trachea was then washed 5 Â 20 min on rotator and then incubated with primary antibodies diluted in BSA/PBS for 48 h at 4°C.The trachea was then washed in filtered PBS (10 Â 30 min) and incubated with secondary antibodies in BSA/PBS overnight at room temperature, then washed in filtered PBS (3 Â 20 min), incubated in anti-fade, pH 8.6, glycerol (Sigma-Aldrich) for 24 h at room temperature, and stored at 4°C.
In some cases, TRPA1 Flp/1 R26 ai65f/1 mice were administered an intraganglionic injection of the constitutively active AAV9-hSyn-GFP (;2 Â 10 13 GC/ml; catalog #50465, Addgene) into the vagal ganglia to drive GFP expression in vagal neurons.Using a procedure that has been described previously (Kim et al., 2022), mice were anesthetized with ketamine (50 mg/kg) and dexmedetomidine (0.5 mg/kg) via intraperitoneal injection.An ;2 cm incision was made over a shaved superficial portion of the masseter muscle area.The skin was retracted, and the vagus nerve was located and then separated from the common carotid artery and the anterior laryngeal nerve using a cotton tip.The virus microinjection assembly consisted of a pulled glass micropipette (tip diameter, ;20 mm) attached to a 1 ml syringe via plastic tubing.The tip of the micropipette was gently inserted into the exposed vagal ganglia and then 0.7 ml of AAV9-hSyn-GFP was injected using ;0.5 psi pressure.Three to 6 weeks later, the mice were used for analysis.
Images were taken with the Dragonfly Spinning Disk Confocal Microscope (Andor) equipped with a Zyla 4.2 PLUS sCMOS Camera (Andor; 2048 Â 2048 pixel with 6.5 mm pixel size).The pinhole size was 25 mm.We used either a 10Â UPLSAPO [0.4 numerical aperture (NA)], a 20Â UPLSAPO (0.75 NA), or a 40Â UPLSAPO (1.25 NA, silicone oil-immersion) objective was used, depending on the study.Fluorophores were excited by laser wavelengths at 405, 488, 561, or 637 nm.The z-stacked multitile images were stitched using either Fusion software or Imaris Stitcher.All 3D images were further processed using Imaris software.The identification of anatomic structures and subnuclei were based on the mouse brain map (Paxinos and Franklin, 2012).In mice, the nodose vagal ganglion and the jugular vagal ganglion are partially fused into a single structure.Nevertheless, these two ganglia can be approximately discriminated visually in sagittal slices of the vagal ganglia, as shown previously (Nassenstein et al., 2010;Kim et al., 2020a,b).Cell counts and somal diameter of neurons within the vagal ganglia and DRGs were obtained using Fiji software.

Sensory ganglia dissociation
Mice were killed by CO 2 asphyxiation followed by exsanguination.DRGs or vagal ganglia were isolated in Ca 21 -free, Mg 21 -free HBSS, incubated in HBSS containing type 1 collagenase (2 mg/ml) and dispase II (2 mg/ml), and then mechanically dissociated with fire-polished pipettes.Individual neurons were washed and resuspended in L-15 media supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, and 100 mg/ml streptomycin, then were plated onto poly-D-lysine-coated and laminincoated coverslips.As a positive ganglionic control for the RT-PCR studies (see below), we transferred 2 ml of the cell suspension into PCR tubes containing 1 ml RNase OUT.
Neurons were incubated at 37°C in antibiotic-free L-15 media supplemented with 10% FBS and used within 24 h.

Single-neuron RT-PCR
Coverslips with dissociated vagal sensory neurons were superfused with Krebs bicarbonate solution (KBS; in mM: 118 NaCl, 5.4 KCl, 1 NaH 2 PO 4 , 1.2 MgSO 4 , 1.9 CaCl 2 , 25 NaHCO 3 , and 11 dextrose, gassed with 95% O 2 -5% CO 2 ).tdTomato-expressing neurons were identified by fluorescence (554 nm excitation, 581 nm emission) and a BX51WI Microscope (Olympus).Individual neurons were collected with a borosilicate glass pipette by applying a gentle negative pressure.Each single neuron was stored in a PCR tube containing 1 ml RNase OUT and stored at À80°C.The KBS surrounding the vagal neurons was sampled (;2 ml) as a negative control for the RT-PCR.
First-strand cDNA was synthesized with SuperScript III First-Strand Synthesis System for RT-PCR by following the manufacturer instructions.Primers and dNTP (deoxynucleotide triphosphate) mix were added into each neuron-containing PCR tube.Samples were incubated at 75°C for 10 min and then placed on ice for at least 1 min.cDNA synthesis mix containing 10Â reverse transcriptase (RT) buffer, MgCl 2 , DTT, and SuperScript III RT was added into each sample.Samples were incubated at 50°C for 50 min, followed by 85°C for 5 min.cDNA was stored at À20°C until used for PCR amplification.cDNA was amplified with HotStarTaq DNA polymerase for 50 cycles of denaturation at 94°C for 30 s, followed by annealing at 60°C for 30 s and extension at 72°C for 1 min.Customized intron-spanning primers for mouse TRPA1, TRPV1, P2X 2 , preprotachykinin A (PPT-A), and vesicular glutamate transporter 2 (VGluT2; Table 2) were used as previously reported (Nassenstein et al., 2008(Nassenstein et al., , 2010)).Products were visualized in 1.5% agarose gel with GelRed, with a 100 bp marker.VGluT2 was used as a positive control.

FURA-2AM Ca 21 imaging and analysis
Neuron-coated coverslips were incubated with 4 mM fura-2 AM for 30-60 min at 37°C.Coverslips were loaded into a chamber on an inverted microscope (model Eclipse Ti, Nikon), fitted with a CFI Super Fluor 10Â objective and a CoolSnap HQ2 camera (Photometrics), and perfused with heated (33-34°C) HEPES buffer (154 mM NaCl, 1.2 mM KCl, 1.2 mM MgCl 2 , 2.5 mM CaCl 2 , 5.6 mM D-glucose).Cells were subjected to fluorescent light delivered by a 300 W (5000 lumens) PE300BFA Xenon Parabolic Bulb housed in a Lambda LS (Sutter Instruments), with appropriate excitation and emission filters.Bright-field and fluorescence images (535 nm excitation, 610 nm emission) were collected to facilitate subsequent analysis and the determination of tdTomato expression, respectively.Changes in [Ca 21 ] i were monitored every 6 s using sequential excitation at 340 and 380 nm (510 nm emission) and evaluated ratiometrically using the 340:380 ratio (R).All drugs were diluted in HEPES buffer.AITC (100 mM) was used to determine the functional expression of TRPA1 (Jordt et al., 2004), and capsaicin (1 mM) was used to determine the functional expression of TRPV1 (Caterina et al., 1997).Neurons were further characterized by response to KCl (75 mM) before ionomycin (5 mM), which evoked a maximal Ca 21 response.Region of interest (ROI) analysis for individual neurons was performed using Nikon Elements.ROIs with an unstable, high, or noisy baseline were eliminated from analysis.Neurons that failed to exhibit an increase in [Ca 21 ] i to AITC, capsaicin, or KCl challenges (.30% the ionomycin maximal response) were eliminated.An individual neuron was considered to be sensitive to a given agent if R agent .(R bl 1 2 * SD bl ) 1 0.03; where R agent is the average 340:380 ratio during treatment, R bl is the average 340:380 ratio before treatment, and SD bl is the SD of R bl .Neurons were grouped by tdTomato expression and sensitivity to AITC and capsaicin.In some cases, the "d " evoked response to a given treatment was calculated as the R agent -R bl .
Mice were killed by CO 2 inhalation and exsanguination, and the vagal ganglia and T2, T3, or L5 DRGs were dissected out and pinned in a chamber superfused with KBS (4 ml/min at 37°C).The ganglion was visualized using a 25Â water-immersion objective and the Olympus FVMPE-RS multiphoton system.The tissue was excited with 960nm (InSight X3 IR laser), and the emitted light was split into two paths using a beam splitter, which was then detected with specific filter cubes and distinct photomultiplier tubes (PMT) containing gallium arsenide phosphide (GaAsP): an FVG filter cube and 488 PMTs for GCaMP6s, and an FGR filter cube and 555 PMTs for tdTomato.A z-stack (100 mm) of the ganglia was visualized using Fluoview software (Olympus), which allowed for automated increases in laser percentage and PMT voltage depending on the depth of the zplane into the ganglia.Live images were recorded every 12 s.
For the vagal recordings of TRPA1 Flp/1 R26 ai65f/1 mice with intraganglionic injection of AAV9-CAG-GCaMP6s, we determined the GCaMP6s responses to AITC (100 mM, 300 s) and then KCl (50 mM, 300 s) added to the ganglionic compartment.For the DRG recordings of Pirt Cre/1 R26 ai96/1 mice, we determined the GCaMP6s responses to AITC (100 mM, 300 s), followed by capsaicin (1 mM, 300 s) and then KCl (50 mM, 300 s) added to the ganglionic compartment.Fluoview-derived images were analyzed offline using Fiji software (ImageJ).Each ganglion was imaged up to a depth of 100 mm, in ten 10-mm-thick z-planes.We analyzed either odd-numbered or even-numbered z-planes.This limited the possibility that a given mouse neuron (which are ;20 mm in diameter) was counted twice.Each analyzed z-plane was used to mark the ROIs of tdT 1 neurons and neurons that responded to any of the treatments.A neuron was considered to be sensitive to a given treatment if its maximal GCaMP6s response during treatment was double the mean baseline GCaMP6s fluorescence of all identified neurons within that ganglion.A neuron was only included in the analysis if it responded to one of the treatments.For each neuron, the mean baseline GCaMP6s fluorescence was subtracted from the GCaMP6s fluorescence at each time point for dataset comparisons (F 1 -F 0 ).

Nocifensive/pain behavioral assessment
The nocifensive behavior of mice following intraplantar injection of the hindpaw with vehicle or CNO dihydrochloride (catalog #6329, Tocris Bioscience) was determined.In some experiments, TRPA1 Flp/Flp mice first received an intraplantar injection (10 ml) into the right hindpaw with rAAV2-hSyn-fDIO-hM3Dq [;1 Â 10 13 genome copies (GC)/ml; catalog #154868, Addgene] or (as a control) rAAV2-Ef1a-fDIO-mCherry (;1 Â 10 13 GC/ml; catalog #114471, Addgene) 4 weeks before the behavioral assessment to drive the Flp-dependent expression of the "activating" DREADD hM3Dq receptor (or mCherry as control) in TRPA1-expressing paw afferents.CNO working solution was made fresh on the day of the experiment in sterile PBS (1 or 10 mg).Mice received an intraplantar injection (20 ml) of CNO (1 or 10 mg) or vehicle (sterile PBS) into the right hindpaw and then were placed immediately into a clear cage monitored by a camera placed underneath.The duration (in seconds) that the animal spent flinching the injected hindpaw over a 15 min period was recorded by the camera.The recordings were then scored by an observer blinded to treatment allocation and mouse strain.For each experiment, the response to vehicle and CNO was recorded in the same set of mice, with the treatments separated by 1 week (paired experiments).Thus, the following studies were performed: (1) TRPV1-hM3Dq mice (n ¼ 5: two females, three males) given intraplantar injection of vehicle and 1 mg of CNO (paired experiments); (2) TRPA1 Flp/Flp mice 4 weeks after intraplantar injection with rAAV2-Ef1a-fDIO-mCherry (n ¼ 3: one female, two males) given intraplantar injection of vehicle and 1 mg of CNO (paired experiments); and (3) TRPA1 Flp/Flp mice 4 weeks after intraplantar injection with rAAV2-hSyn-fDIO-hM3Dq (n ¼ 6: four females, two males) given intraplantar injection of vehicle and 1 mg of CNO and 10 mg of CNO (paired experiments).

Statistical analysis
All data were analyzed in GraphPad Prism 10.The somal areas of various vagal and DRG afferent subpopulations were compared using ANOVA with Tukey's multiple-comparisons test.In the fura-2 AM studies, the mean d responses to a given irritant were compared between tdT 1 neurons and tdT -neurons using an unpaired t test.The time spent flinching after intraplantar injection of CNO was compared against vehicle responses in paired experiments: either t test for single CNO doses or ANOVA with Dunnett's multiple-comparisons test for multiple CNO dose studies.In all cases, a p value ,0.05 was considered significant.
Last, we crossed the TRPA1 Flp mouse with the Flp-sensitive R26 FLTG reporter mice to produce TRPA1 Flp/1 R26 FLTG/1 mice.Serial sections of the t-DRGs from these mice (six slices from three mice) demonstrated that 4.4% (61 of 1385) of neurons expressed tdTomato (data not shown), which is similar to data from TRPA1 Flp/1 R26 ai65f/1 mice.This indicates that the lack of efficiency in the tdT expression in DRG neurons is not dependent on the specific Flp-sensitive reporter strain.

tdTomato expression in the CNS of TRPA1 reporter mice
We investigated the expression of tdTomato in the CNS of TRPA1 Flp/1 R26 ai65f/1 mice (n ¼ 5 mice).In the midbrain and higher brain levels, tdTomato expression was extremely limited, except for minor populations of neurons within the second/third layers of the somatomotor cortex and the piriform area (Fig. 7A,B).In the brainstem, tdTomato expression was limited to nerve fibers within the nucleus tractus solitarius (nTS) and the trigeminal complex (Sp5; Fig. 7A,C), as well as the paratrigeminal complex (Pa5; data not shown).These medullary areas receive dense innervation from vagal and trigeminal nociceptive afferents (Kim et al., 2020a).tdTomato expression in the nTS was restricted to the dorsal and medial subnuclei (e.g., SolC, SolG, SolDL).In the thoracic spinal cord, tdTomato expression was observed in nerve fibers in the dorsal horn (Fig. 7D).Primarily, tdTomato 1 fibers were concentrated in the superficial laminae innervated by CGRP 1 DRG afferents (Lamina I and, to some extent, Lamina II o ) and by IB4 1 DRG afferents (Lamina II i ; Fig. 7E), although sporadic fibers were noted in deeper areas.We found no obvious tdTomato expression in astrocytes or glial cells in any CNS region.] i responses of tdT 1 (red; n ¼ 144) and tdT -(black; n ¼ 66) neurons within intact vagal ganglia to AITC (100 mM) and KCl (50 mM).G, Percentage of tdT 1 and tdT -neurons within intact vagal ganglia that were activated by AITC.

tdTomato expression in peripheral tissues of TRPA1 reporter mice
We investigated the expression of tdTomato in peripheral organs of TRPA1 Flp/1 R26 ai65f/1 mice and control R26 ai65f/ai65f mice.We noted consistent Flp-independent tdTomato expression in many LE-lectin 1 blood vessels within the esophagus (both muscle and mucosal layers), trachea, stomach, colon, and bladder of R26 ai65f/ai65f mice (Figs.8A,B, 9A,B; n ¼ 4 mice).Immunohistochemical staining with an anti-DsRed antibody confirmed the presence of tdTomato in these control mice (data not shown).tdT 1 blood vessels were only occasionally observed in the   red), the nonpeptidergic afferent marker IB4 (blue), and the peptidergic afferent marker CGRP (green).Note that IB4 also labels some CNS blood vessels in mice (Laitinen, 1987).tdT 1 fibers were found within the CGRP 1 lamina (yellow arrow) and within the deeper IB4 1 lamina (pink arrow).Scale bars: A, B, 2 mm; C, 1 mm; D, 500 mm; E, 100 mm.skin and were absent from the lungs and cochlea (data not shown).In the esophagus of TRPA1 Flp/1 R26 ai65f/1 mice (n ¼ 6), many tdT 1 nerve fibers were found in the mucosal and muscle layers (data not shown) along with tdT 1 blood vessels.No tdT 1 neuronal cell bodies were found.To determine the origin of these tdT 1 fibers, we injected the vagal ganglia of TRPA1 Flp/1 R26 ai65f/1 mice with AAV9-hSyn-GFP to label the vagal afferents with GFP.There was consistent  overlap of tdTomato, GFP, and CGRP staining, indicating that most tdT 1 fibers were vagal CGRP 1 fibers (Fig. 8C).
In addition to the tdT 1 blood vessels that were also observed in control R26 ai65f/ai65f mice (Fig. 9A,B), tdTomato was expressed in multiple structures in the stomach and colon of TRPA1 Flp/1 R26 ai65f/1 mice (Fig. 9C-J; n ¼ 4 mice).Abundant tdT 1 fibers were found within the mucosa and the submucosal plexus.Some tdT 1 fibers were also observed in the myenteric plexus, but these did not appear to possess intraganglionic laminar endings (IGLEs).Many tdT 1 fibers expressed CGRP (Fig. 9C,F,G).No tdT 1 neuronal cell bodies were found.Robust tdTomato expression was observed in many large stellate-shaped cells within the stomach (Fig. 9C-E) and colonic mucosa (Fig. 9F-J), which were absent in R26 ai65f/ai65f mice (Fig. 9A,B).tdTomato expression did not overlap with either serotonin (marker for enterochromaffin cells; Bellono et al., 2017; Fig. 9D,E,  H,I) or cd11b (marker for most resident macrophages and dendritic cells; Cerovic et al., 2014; Fig. 9J).tdTomato expression was noted in nerve fibers innervating the bladder of TRPA1 Flp/1 R26 ai65f/1 mice (n ¼ 4 mice), as well as within some blood vessels (Fig. 9K).In the trachea of TRPA1 Flp/1 R26 ai65f/1 mice (n ¼ 4), we noted dense innervation by tdT 1 fibers, especially in the subepithelial layers of the trachealis muscle and annular ligaments (Fig. 10A).The tdT 1 fibers formed a dense plexus of free nerve endings that lacked defined arbors.Almost all tdT 1 fibers also expressed CGRP.In the lungs (n ¼ 3 mice), tdT 1 fibers densely innervated conducting airways of all diameters (Fig. 10B-D).Some of these tdT 1 fibers projected out into the alveolar region (Fig. 10C).Some pulmonary blood vessels were also innervated by tdT 1 fibers (Fig. 10D).Last, we observed tdTomato expression in the cells of Hensen in the cochlea of TRPA1 Flp/1 R26 ai65f/1 mice (n ¼ 3; Fig. 11A,B).tdTomato expression was absent from all other cell types including inner hair cells, outer hair cells, and spiral ganglion cells.No tdTomato expression was found in the cochlea of R26 ai65f/ai65f mice (n ¼ 2; Fig. 11C).

Chemogenetic activation of TRPA1-expressing nerves in the paw evokes nocifensive behaviors
We observed dense tdT 1 innervation of hairy skin and glabrous skin on the hindpaw from TRPA1 Flp/1 R26 ai65f/1 mice (Fig. 12A).tdTomato was expressed by free nerve endings (that lacked defined arbors) within the dermis and epidermis.We did not observe tdTomato expression in other cell types, including keratinocytes.To determine the effect of the activation of specific afferent populations within the skin on nocifensive/pain behaviors including flinching, we expressed the CNO-sensitive stimulatory DREADD receptor hM3Dq in TRPV1-expressing and TRPA1-expressing nerves.First, we found that intraplantar injection of CNO (1 mg) into the hindpaw of TRPV1-hM3Dq mice evoked a significant increase in flinching time compared with vehicle (p , 0.05, n ¼ 5 mice; Fig. 12B).To drive hM3Dq expression in TRPA1-expressing nerves, we injected the hindpaw of TRPA1 Flp/Flp mice with rAAV2-fDIO-hM3Dq (or rAAV2-fDIO-mCherry as a negative control).Four weeks later, we performed intraplantar injection of the hindpaw with vehicle, 1 mg of CNO, and 10 mg of CNO.In the TRPA1-hM3Dq mice (n ¼ 6), CNO at both concentrations evoked a significant increase in flinching time compared with vehicle (p , 0.05; Fig. 12C).Importantly, CNO had no effect on flinching time

Discussion
We developed the TRPA1 Flp mouse as a tool to detect and selectively manipulate TRPA1-expressing cells.Although selective TRPA1 agonists and antagonists have been extensively reported, there is disagreement in the literature regarding the cellular expression patterns of TRPA1 in mammals-especially in nonafferent cell typesin part because of the inconsistent/poor antibodies available.Here we used a genetic approach, replacing the endogenous TRPA1 stop codon with a 2A-Flp O cassette.As 2A is a self-cleaving peptide, we expect that the expression of Flp recombinase would match TRPA1 peptide expression on a one-to-one basis.Crossing the TRPA1 Flp mouse with Flp-sensitive reporter mice (e.g., R26 ai65f ) is then a powerful tool to visualize TRPA1 expression.
In the vagal ganglia of TRPA1 Flp/1 R26 ai65f/1 mice, we observed tdTomato expression in a major subset (;40%) of both nodose and jugular ganglionic neurons, consistent with previous reports of their sensitivity to TRPA1 agonists (Andrè et al., 2008;Nassenstein et al., 2008;Taylor-Clark et al., 2008b;Lin et al., 2015;Stanford et al., 2019).Most of these neurons also expressed the canonical nociceptive receptor TRPV1 and the neuropeptide CGRP, indicating that they were likely nociceptive in nature.Approximately 90% of dissociated tdT 1 vagal neurons expressed TRPA1 mRNA and were activated by the selective agonist AITC, confirming the effectiveness of the reporter.Interestingly, ;25% of dissociated tdT -vagal neurons also had TRPA1 expression/functionality, but this was much lower for tdT -neurons within the intact vagal ganglion (;7%).It is possible that dissociation selectively depletes tdTomato protein within the soma or otherwise renders tdTomato less detectable, although no evidence of this was found in previous studies of TRPV1 Cre R26 ai9 mice or P2X 2 Cre R26 ai9 neurons (Stanford et al., 2019;Kim et al., 2020b).Alternatively, acute dissociation and culture may have caused de novo TRPA1 expression in a subset of TRPV1 1 vagal neurons, possibly because of neurotrophic factors present in the FBS (Diogenes et al., 2007;Ciobanu et al., 2009).Importantly, AITC activated ;40% of dissociated vagal neurons-similar to the percentage of neurons expressing tdTomato observed in frozen sections of the vagal ganglia-arguing against dissociation-evoked de novo TRPA1 expression.Indeed, native tdTomato expression in vagal ganglia sections was only detectable in some neurons after signal augmentation using antibodies suggesting some native fluorescence was near the limits of detection.
Given that tdTomato expression denotes Flp-mediated recombination of the R26 ai65f locus at any time, the strong correlation of tdTomato with TRPA1 expression suggests that very few vagal neurons transiently express TRPA1 during development.
Some studies have identified a small subset of TRPA1 1 TRPV1 -vagal neurons in functional (Nassenstein et al., 2008;Taylor-Clark et al., 2008b;Stanford et al., 2019) and transcript studies (Mazzone et al., 2020) of dissociated neurons.Other studies of transcripts in dissociated neurons (Nassenstein et al., 2008;Brozmanova et al., 2011;Kupari et al., 2019;Meerschaert et al., 2020;Zhao et al., 2022), and single fiber recordings of vagal afferents innervating the airways or esophagus (Nassenstein et al., 2008;Brozmanova et al., 2011;Nesuashvili et al., 2013) argue that TRPA1 expression is confined to a subset of TRPV1 1 vagal neurons.Here, we found ;20% of tdT 1 neurons in vagal sections lacked immunoreactivity to TRPV1.However, we detected TRPV1 transcript in all tdT 1 vagal neurons and found that .95% of these neurons were capsaicin-sensitive.While there are caveats to interpreting TRP expression in dissociated cultures, it is probable that the high count of tdT 1 TRPV1 -neurons in vagal sections is partly because of inefficiencies of the TRPV1 antibody used in the present study.
tdTomato was expressed in only 3-12% of neurons within the DRGs of TRPA1 Flp/1 R26 ai65f/1 mice.Most of these neurons also expressed TRPV1 and CGRP.Following dissociation, ;85% of tdT 1 DRG neurons were activated by AITC compared with ;25% of tdT -vagal neurons.Given that few DRG neurons expressed tdTomato, this meant that most AITC-sensitive neurons were tdT -.This did not appear to be because of culture-induced de novo TRPA1 expression because similar percentages (35-45%) of DRG neurons in intact ganglia and dissociated cultures were activated by AITC.Thus, our TRPA1 reporter preferentially labels TRPA1-expressing DRG neurons but with low efficiency.Functional and transcript studies have shown that TRPA1 expression in the DRG correlates with TRPV1 expression.In addition to the expected tdT 1 TRPV1 1 population, our data also suggest evidence of a TRPA1 1 TRPV1 -subset in DRGs, and this is consistent with some reports (Malin et al., 2011;Usoskin et al., 2015;Patil et al., 2018) whereas others argue that TRPA1 is expressed exclusively in TRPV1 1 neurons (Story et al., 2003;Mishra et al., 2011;Pogorzala et al., 2013;Wilson et al., 2013;Weng et al., 2015).The tdT 1 TRPV1 -DRG neurons observed here were larger than tdT 1 TRPV1 1 DRG neurons but were much smaller than tdT -TRPV1 -neurons, which are likely to be non-nociceptive mechanoreceptors.
The limited reporter efficiency in the DRGs of TRPA1 Flp/1 R26 ai65f/1 mice was recapitulated in TRPA1 Flp/1 R26 FLTG/1 mice, implicating the Flp itself.Here we used Flp O , which is the thermostable Flp E recombinase enzyme whose nucleotide sequence has been optimized for mammalian codons (Raymond and Soriano, 2007).When expressed in HEK293 cells, Flp O recombines FRT sites as efficiently as Cre recombines LoxP sites (Fenno et al., 2014).Although only four Flp molecules are technically necessary for recombination of paired FRT sites, the reduced efficiency of Flp E compared with Flp O (which are identical proteins) indicates that physiologically relevant expression levels are rate limiting (Raymond and Soriano, 2007;Kranz et al., 2010).Thus, it is possible that the low but functionally relevant expression of TRPA1 in some TRPA1 1 DRG neurons fails to drive sufficient Flp expression to recombine the ROSA26 FRT sites for tdTomato expression.Indeed, the AITC-evoked responses of AITCsensitive neurons was smaller in tdT -than in tdT 1 DRG neurons, suggesting lower TRPA1 expression.Nevertheless, we noted that even when R26 ai65f tdTomato expression was evoked in sensory afferents (as observed in frozen sections), its detection more often required antibody-mediated signal augmentation in DRG tdT 1 neurons (28.9%) compared with vagal tdT 1 neurons (8.1%), suggesting lower tdTomato expression per neuron in DRGs independent of the efficiency of Flp recombination.
Comprehensive direct comparisons between Cre-and Flp-expressing mouse strains are limited (Daigle et al., 2018), but one study of corticotropin-releasing hormone (CRH) reporter strains showed reduced efficiency of Flp O recombination of R26 ai65f/1 compared with Cre recombination of R26 ai9/1 (Zhao et al., 2023): while the two strategies labeled identical CNS regions, more tdT 1 neurons were noted for the Cre reporter because it was more efficient evoking tdTomato expression in neurons expressing lower levels of CRH.Interestingly, the efficiency of reporting was increased using homozygous CRH Flp mice.It is possible that more TRPA1-expressing DRG neurons would express tdTomato using homozygous TRPA1 Flp mice.
We used the TRPA1 Flp/1 R26 ai65f/1 mouse to map TRPA1expressing afferents.We observed dense innervation of the gut, bladder, skin, and airways by tdT 1 fibers, but no intrinsic tdT 1 neurons were found in these organs.Most of the tdT 1 fibers innervating the esophagus mucosa were derived from the vagal ganglia.Abundant tdT 1 fibers were also found within the mucosa and the submucosal plexus of the stomach and colon.Many tdT 1 fibers in the gut expressed CGRP, but no evidence was found of tdT 1 fibers with IGLEs.In the lungs, we observed tdT 1 fibers innervating conducting airways and blood vessels, similar to vagal TRPV1 1 fibers previously described (Su et al., 2021;Kim et al., 2022).The observation that some of the tdT 1 fibers here also penetrated the alveolar space argues that these were projected from nodose neurons (Kim et al., 2022).In the skin, we found abundant tdT 1 -free nerve endings within the dermis and epidermis that resembled TRPV1 1 and CGRP 1 fibers (Cavanaugh et al., 2011;McCoy et al., 2012;Le Pichon and Chesler, 2014).In the brainstem medulla, abundant tdT 1 fibers were found terminating in the nTS, Pa5, and Sp5, consistent with the innervation patterns of vagal and trigeminal TRPV1 1 fibers (Kim et al., 2020a;Su et al., 2021).In the spinal cord, tdT 1 fibers mostly innervated dorsal horn lamina I and II, mimicking previous studies of nociceptive peptidergic and nonpeptidergic Ad -and C-fibers projected from DRG neurons (Cavanaugh et al., 2011;McCoy et al., 2012;Le Pichon and Chesler, 2014) that synapse with spinal neurons involved in pain processing (Todd, 2010).AITC has previously been shown to evoke glutamatemediated excitatory postsynaptic potentials in nTS and dorsal horn neurons, implicating the presence of TRPA1expressing presynaptic terminals (Kosugi et al., 2007;Sun et al., 2009).
Despite the relatively low efficiency of the Flp reporter in DRGs, chemogenetic activation of TRPA1-expressing afferents innervating the hindpaw evoked flinching, similar in quality to chemogenetic activation of TRPV1-expressing afferents (albeit at lower intensity).Such responses are consistent with the expression of TRPA1 in nociceptive DRG afferents, and multiple studies showing TRPA1 agonists evoke pain (Bautista et al., 2006;Kwan et al., 2006;McNamara et al., 2007;Patil et al., 2020).
The use of our TRPA1 Flp reporter to map TRPA1 expression was somewhat hampered by the Flp-independent expression of tdTomato in R26 ai65f mice blood vessels, particularly in the gut and bladder, partially in the skin and trachea and virtually absent in the lung and CNS.FRT-stop-FRT elements (which precedes the tdTomato sequence in the ROSA26 knockin) are not completely "leak proof" (Dabrowski et al., 2015), and thus may account for the tdTomato expression in these blood vessels.We are not aware of any other previous reports that investigate peripheral tissues in R26 ai65f mice, but Flp-independent tdTomato expression does not appear to occur in the CNS (Yao et al., 2020).
We found no expression of tdTomato, other than in afferent-like fibers and some blood vessels, in the skin, trachea, or lungs of TRPA1 Flp/1 R26 ai65f/1 mice, and thus we can provide no evidence for TRPA1 expression in keratinocytes, airway epithelium, or airway smooth muscle cells.We found many large tdT 1 stellate-shaped cells within the stomach and colonic mucosa; however, these expressed neither serotonin nor cd11b, indicating that they were neither enterochromaffin cells nor resident macrophages/dendritic cells, respectively.The role of TRPA1 in the function of these unidentified cells is presently not known.We found tdTomato expression in a small subset of cortical neurons within the second/third layers of the somatomotor cortex and the piriform area, but no tdTomato expression was found in astrocytes, Schwann cells, or vascular endothelium.Last, in the cochlea we observed robust tdTomato expression in Hensen's cells, which are members of a group of supporting cells for the inner and outer hair cells in the organ of Corti, but not in the hair cells themselves.Initial reports using antibodies reported TRPA1 expression in inner hair cells (Corey et al., 2004;Nagata et al., 2005), but the in situ hybridization signal (Corey et al., 2004) appeared adjacent to the outer hair cells-consistent with the location of Hensen's cells.Recently, AITC-evoked Ca 21 transients have been shown in Hensen's cells (Vélez-Ortega et al., 2023).
Overall, we find limited expression of tdTomato in nonafferent cell types in TRPA1 Flp/1 R26 ai65f/1 mice.While the inhibition or knockout of TRPA1 reduces nociceptive responses to electrophilic irritants, oxidative stress, and inflammatory pain models (Bautista et al., 2006;Kwan et al., 2006;Dai et al., 2007;McNamara et al., 2007;Eid et al., 2008), a role for TRPA1 has also been reported for other physiological and pathophysiological processes such as endothelium-dependent cerebral artery dilation (Earley et al., 2009), anxiety (de Moura et al., 2014), cognition and memory (Lee et al., 2016a,b), which are unlikely to be mediated by sensory nerves.It is possible that these are because of off-target effects of TRPA1 knockout/inhibitors.Alternatively, it is possible that some tdT -cells in our reporter express functionally relevant levels of TRPA1 without sufficient Flp expression to drive tdTomato expression.Finally, de novo expression of TRPA1 in nonafferent cell types under pathologic conditions (not tested here) could explain the role of TRPA1 in some processes.
In summary, we describe a TRPA1 flp mouse that provides a new tool for the selective identification and manipulation of TRPA1-expressing cells.Our data confirm that TRPA1 is expressed by a subset of sensory neurons, which largely but not exclusively overlap with the canonical nociceptive TRPV1 channel.We also find limited TRPA1 expression in other cell types, but most notably in subepithelial cells in the mucosa of the stomach and colon.Recently, a TRPA1 cre mouse line has been reported (Matsuo et al., 2021).When crossed with a Cre-sensitive enhanced yellow fluorescent protein (YFP) expressing mouse, YFP 1 neurons were identified in vagal and trigeminal ganglia, with some YFP 1 fibers innervating the nTS and Sp5 regions in the medulla, consistent with our findings here.

Figure 1 .
Figure 1.Development of the knock-in TRPA1 Flp mouse.A, Targeting strategy for the replacement of the TRPA1 TAG stop codon with a 2A-Flp cassette.Homology arms (blue and red lines) were generated for the TRPA1 gene [exons and 39 untranslated region (UTR) in orange].Homology arms of targeting vector include a neomycin resistance gene (Neo R ) flanked by SDA sites for positive selection.A DTA fragment gene was placed in a nonhomologous region of targeting vector as a negative selection for nonhomologous recombination.B, PCR of TRPA1 gene in offspring from a pairing of heterozygous TRPA1 Flp mice.As expected, offspring have a Mendelian distribution of mutant (i.e., TRPA1 Flp , at 250 bp) and wild-type (at 188 bp) alleles.

Figure 2 .
Figure 2. Subsets of sensory neurons in TRPA1 Flp R26 ai65f mice express tdTomato.A, B, Representative slices of native tdTomato expression (in red) compared with antibodies against tdTomato (DsRed; green) and TRPV1 (blue) in vagal ganglia (A) and thoracic DRGs (B).Scale bars, 50 mm.C, D, Euler diagram showing the overlap of neurons expressing native tdTomato and immunoreactivity for tdTomato (DsRed) and TRPV1 in vagal ganglia (C) and thoracic DRGs (D).E, Percentage of vagal and DRG neurons (n ¼ 1442 and n ¼ 2635, respectively) that express tdTomato and/or TRPV1.F, G, Violin plots of the cell body area (in square micrometers) of neurons expressing tdTomato and/or TRPV1 in the vagal ganglia (F) and thoracic DRGs (G).*Significant difference in cell body area (p , 0.05).

Figure 3 .
Figure 3. Overlap of TRPV1 and CGRP with tdTomato in sensory neurons from TRPA1 Flp R26 ai65f mice.A, Representative slices of tdTomato (red), TRPV1 (blue), and CGRP (green) expression in vagal ganglia (top) and thoracic DRGs (bottom).Scale bars: top, 100 mm; bottom, 50 mm.The vagal ganglion is visually divided by a dotted gray line into the nodose and jugular portions, each with a higher magnification insert.B, Percentage of nodose, jugular, and DRG neurons (n ¼ 988, n ¼ 326, and n ¼ 1093, respectively) that

Figure 7 .
Figure 7. tdTomato expression within the CNS of TRPA1 Flp R26 ai65f mice.A, Sagittal section of brain and brainstem showing tdTomato expression (black) in occasional neurons within the somatomotor cortex (CTX) and in fibers within the Sp5.B, Coronal section of brain (2.5 mm caudal of bregma) showing tdT 1 neurons in CTX.C, Coronal section of medulla (0.4 mm caudal of obex) showing tdT 1 fibers in the nTS and Sp5.D, Coronal section of thoracic spinal cord showing tdT 1 fibers innervating lamina I and II (LI-LII) of the dorsal horn.E, High magnification of superficial lamina of the dorsal horn showing tdTomato (red), the nonpeptidergic afferent marker IB4 (blue), and the peptidergic afferent marker CGRP (green).Note that IB4 also labels some CNS blood vessels in mice(Laitinen, 1987).tdT 1 fibers were found within the CGRP 1 lamina (yellow arrow) and within the deeper IB4 1 lamina (pink arrow).Scale bars: A, B, 2 mm; C, 1 mm; D, 500 mm; E, 100 mm.

Figure 9 .
Figure 9. tdTomato expression in the stomach, colon, and bladder of TRPA1 Flp R26 ai65f mice.A, B, tdTomato expression (red) and DAPI staining (blue) in the stomach (A) and colon (B) of control R26 ai65f/ai65f mice.C-K, tdTomato expression (red) in tissues from TRPA1 Flp R26 ai65f mice.C, Stomach with tdTomato and CGRP (green) expression and DAPI (blue).White arrows identify tdT 1 fibers, which almost exclusively also express CGRP.D, Stomach with tdTomato and 5-HT (green) expression and DAPI (blue).E, Higher magnification of white box identified in D. F, Colon with tdTomato expression and DAPI (left), with overlay of CGRP expression (green; right).White arrows identify tdT 1 fibers, gray arrows identify tdT 1 stellate-shaped resident cells.G, Higher magnification of white box identified in F. H, Colon with tdTomato and 5-HT (green) expression and DAPI (in blue).I, Higher magnification of white box identified in H. J, Colon with tdTomato and CD11b (green) expression and DAPI (blue).K, Bladder with tdTomato expression.White arrows identify tdT 1 fibers.Scale bars: A, B, C, F, H, J, 100 mm; D, E, G, I, K, 50 mm.

Figure 10 .
Figure 10.tdTomato expression in fibers innervating the trachea and lungs of TRPA1 Flp R26 ai65f mice.A, tdTomato (red) and CGRP (in green) expression in the trachea (whole mount).B, Lung slice showing tdT 1 fibers (red) innervating a large conducting airway labeled by E-cadherin (green).C, Lung slice showing tdT 1 fibers (red) projecting from a small conducting airway (labeled by E-cadherin; green) into the alveolar regions.D, Lung slice showing tdT 1 fibers (red) innervating a pulmonary blood vessel that expresses a smooth muscle actin (aSMA; blue) but not E-cadherin (green).Scale bars: A-C, 100 mm; D, 50 mm.

Table 1 :
Antibodies used in this study

Table 2 :
Primers used for single-cell RT-PCR in mouse neurons