A voucher specimen of the Pancalieri chemotype of Arthemisia absinthium is kept in Novara laboratories. 1,300 g of leaves and flowers, powdered, were extracted with acetone (3 × 7.5 L) in a vertical percolator at room temperature, affording 97 g (7.5%) of a dark green syrup. The acetonic extract was dissolved into the minimal amount of acetone at 45 C and then 97 g silica gel was added (ratio extract/silica 1:1), finally the suspension was evaporated. Then absinthin was isolated according to Beauhaire et al. (1980), whose structure elucidation and purity were confirmed from 1H NMR (Supplementary Figure S1).

The immunohistochemistry analysis was performed on µm-thick sections of FFPE with DAKO Autostainer (Dako) platform. After diagnostic procedures, remaining sections of tongue and locomotor skeletal muscle biopsies were baked for 30 min at 60°C, deparaffinised with xylene and rehydrated using EtOH washes of decreasing concentrations. For epitope retrieval, slides were treated in preheated citrate buffer and microwaved for 12 min at 650 W. The endogenous peroxidase activity was blocked by incubation in 3% H₂O₂ for 5 min. The incubation with primary antibody was performed for 1 h at RT, using polyclonal rabbit anti-human hTAS2R46 (dilution 1:1,500, OSR00173W, ThermoFisher). Subsequently, the reaction was revealed with Envision Dual Rabbit/Mouse detection system, using 3′3-diaminobenzidine tetrahydrochloride (DAB) as chromogen. The slides were counterstained with hematoxylin. When investigating the expression on the locomotor system, we used digastric, scalene and sternocleidomastoid muscle sections.

Total RNA was isolated by Trizol (ThermoFisher) from USCs, USC-SkMCs, primary myoblasts and myotubes, skeletal muscle biopsies from tongue and locomotor system. The amount and purity of total RNA were quantified at the spectrophotometer (Nanodrop, Thermo Fisher) by measuring the optical density at 260 and 280 nm. 1μg of total RNA was reverse-transcribed using a high-capacity SensiFAST™ cDNA Synthesis Kit (Bioline) according to the manufacturer’s instructions. For quantitative polymerase chain reaction (qPCR) gene specific primers (Supplementary Table S1) and TaqMan Expression Assay (hTAS2R46; Applied Biosystems) were used. TAS2Rs screening was performed with iTaq Universal SYBR Green Supermix (Biorad). For TaqMan assay, TaqMan Universal PCR MasterMix (2×) (without AmpErase UNG; Applied Biosystem) was used with a 7000 ABI Prism system (Applied Biosystems) was used. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and β-glucuronidase (GUSβ) were the endogenous controls; nicotinic acetylcholine receptor subunits (nACHRa4 and nACHRa9) were used as positive control, cytokeratin 10 (CK10) as negative control.

2 × 10⁵ cells were plated on 12 mm Ø glass dish, fixed in PAF 4% for 10′ and then incubated with the blocking buffer (3% BSA, 0.1% Triton X-100 in PBS) for 1 h at room temperature (RT). Cells were then incubated with polyclonal rabbit anti-human hTAS2R46 (OSR00173W, Thermo Fisher) primary antibody for 2 h at RT and then incubated for 45’ at RT in the dark with the secondary antibody goat anti-rabbit AlexaFluor 488 (Thermo Fisher). For nuclei and cytoskeleton staining DAPI and TRITC-phalloidin (Sigma-Aldrich) were added to the secondary antibody solution.

hTAS2R46 expression in USCs-SkMCs was silenced by lentiviral infection. Two lentiviral constructs targeting hTAS2R46 (TCRN0000014110 and TCRN0000014112) were obtained from TRCHs1.0 library (Dharmacon). Third-generation LVs were produced co-transfecting HEK293T packaging cells with plasmids pMDLg/pRRE, pMD2. VSVG, pRSV-Rev and transfer construct using the Lipofectamine 2000 (Invitrogen), as described previously (Talmon et al., 2019). USC-SkMCs were then transduced with the LV-shRNAs together and the silencing was assessed by qPCR.

For membrane potential analysis, SkMCs were loaded for 30 min at RT in the dark with a voltage-sensitive dye using the FluoVolt Membrane Potential Kit (Thermo Fisher), following manufacturer instructions, and the fluorescence were evaluated at the cyotofluorimeter (Attune NxT–Thermo Fisher). 20.000 were acquired to set basal conditions and then stimulated with absinthin (10 μM) and acetylcholine (100 μM) alone or combined. Data were expressed and analysed as delta of mean fluorescence intensity (MFI) before and after the stimulus.

USCs and USC-SkMCs were plated on a pre-coated glass coverslip with 0.1% gelatin and collagen type I, respectively. For evaluation of Ca²⁺ fluctuations, cells were loaded with 5 µM Fura-2 a.m. (ThermoFisher) in the presence of 0.02% Pluronic-127 and 10 µM sulfinpyrazone in Krebs–Ringer buffer (KRB; 135 mM NaCl, 5 mM KCl, 0.4 mM KH₂PO₄, 1 mM MgSO₄, 5.5 mM glucose, 20 mM HEPES, pH 7.4) containing 2 mM CaCl₂ (30 min, RT). Then, cells were washed and incubated with KRB for 30 min to allow the de-esterification of Fura-2 a.m. For measurements of mitochondrial Ca²⁺, cells were loaded with Rhod-2a.m. (ThermoFisher) in the presence of 0.2% Pluronic-127 in KRB containing 2 mM CaCl₂ (30 min, RT) and then incubated with KRB for 1 h to allow the de-esterification. The basal calcium was monitored for about 50 s and then cells were challenged with the following stimuli alone ore combined: acetylcholine (Ach, 100 μM; Sigma-Aldrich), absinthin (Abs 1, 10, 100 μM, 1 mM), cynaropicrin (10, 10 μM, 1 mM), strychnine (10, 10 μM, 1 mM), 3β-hydroxydihydrocostunolide (3-HDC, 1, 10, 100 μM, Sigma-Aldrich), H-89 dihydrochloride PKA inhibitor (10 μm, Sigma-Aldrich), ESI-09 (a pan-EPAC inhibitor, 10 μm, Sigma-Aldrich), and KB-R7943 carbamimidothioic-acid (KB, 10 μm, Sigma-Aldrich). During the experiments, coverslips were mounted into an acquisition chamber and placed on the stage of a Leica DMI6000 epifluorescent microscope equipped with S Fluor ×40/1.3 objective. Fura-2 a.m. was excited by alternating 340 nm and 380 nm using a Polychrome IV monochromator (Till Photonics, Germany), and the probe emission light was filtered through a 520/20 bandpass filter and collected by a cooled CCD camera (Hamamatsu, Japan). Rhod-2a.m. was excited at 552 nm and the fluorescence emission was recorded at 580 nm. The fluorescence signals were acquired and processed using MetaFluor software (Molecular Devices, United States). To quantify the differences in the amplitudes of Ca²⁺ transients, the Fura-2a.m. and Rhod-2a.m. fluorescences were expressed relative to the fluorescence intensity at the stimulation time (ΔF/F0).

USC-SkMCs and primary SkMCs (400.000 cells/mL) were embedded in collagen gel (mouse tail, final concentration of 8 mg/mL) with or without acetylcholine (100 μM), absinthin (10 μM), PKA inhibitor (H89, 10 µM) and EPAC inhibitor (ESI-09, 10 μM) and seeded in triplicate for each condition in 24-well plates. After solidification, the gels were incubated in DMEM with 5% FBS and stimuli. Cells were left to grow, and area of collagen disk was measured from 1 h to 72 h after cells reached the confluence.

Statistical analyses were performed using GraphPad Prism 5 (California, United States). Data are presented mean ± SEM of ‘n’ independent experiments performed in triplicate. Data were analysed by one-way ANOVA. To adjust for multiple testing and to assess differences between control group and each of the treatment we have applied the Dunn’s test. A value of p < 0.05 was considered statistically significant.

The expression of the 25 human bitter taste receptor subtypes was evaluated on skeletal muscle biopsies from the locomotor system and oral cavity. As shown by qPCR analysis (Figure 1A; Supplementary Table S1) several isoforms of the bitter taste receptor are expressed in skeletal muscle, among which TAS2R46 appears to be the most representative (Figure 1A). Immunohistochemistry analysis confirmed and highlighted that the TAS2R46 protein is well expressed all along the striated skeletal muscle fibres of both locomotor system (Figure 1B) and tongue (Figure 1C). The stain with the secondary antibody only (Figure 1E) demonstrates the specificity of the primary antibody and therefore the reliability of the receptor expression.

Both myoblasts from skeletal muscle biopsies and primary SkMC line were subcultured and differentiated into mature myotubes. hTAS2R46 expression level was evaluated in both cell types. As observed by immunofluorescence assay (Figure 2; Supplementary Figure S2) the receptor was detected in both myoblasts (Figures 2A, C) and myotubes (Figures 2B, D), with the difference that in myoblasts it was mainly localized in the perinuclear area while in the myotubes it was expressed predominantly on the cell surface. Moreover, levels of mRNA increased considerably during the differentiation process (Figure 2E; Supplementary Figure S2).

Functional characterization of primary myoblasts/myotubes is hampered by both technical and ethical issues. Therefore, we took advantage of USC-derived skeletal muscle cells (USC-SkMCs), a well characterized skeletal muscle cellular model established in our laboratory (Talmon et al., 2022b). As shown in Figure 3, expression of hTAS2R46 was demonstrated by immunofluorescence analysis also in this system (Figures 3A, B). We evaluated gene expression in both USC and USC-SkMCs and it is noteworthy the significant increase of hTAS2R46 expression after differentiation (Figure 3C; Supplementary Figure S2).

To evaluate whether hTAS2R46 was functional in skeletal muscle, we treated cells with absinthin, its most selective and potent agonist (Talmon et al., 2019), and two non-specific bitter tastants, strychnine and cynaropicrin. Absinthin was unable to induce a measurable increase of cytosolic calcium by itself at the lowest concentration tested (10 μM), although at higher concentrations a dose-dependent cytosolic calcium release could be observed (Supplementary Figure S3). The same trend and rank order of potency were observed treating cells with strychnine, while cynaropicrin gave a stronger effect reaching the maximum calcium peak at 100 μM (Supplementary Figure S3).

We then evaluated the effect of the absinthin on the membrane potential and on calcium variations induced by one of the major pro-contractile neurotransmitters, acetylcholine (100 µM). As shown in Figure 4, acetylcholine induced a rapid membrane potential transient that was reverted by co-stimulation with absinthin. Moreover, Fura-2a.m. analysis demonstrated that acetylcholine induced a rapid increase in cytosolic calcium (Figures 5A, B) which was counteracted by absinthin, already at the lowest concentration tested (1–100 μM, Figure 5A). The effect of absinthin on acetylcholine-induced Ca²⁺-rises was strictly dependent on hTAS2R46 activation since it was reversed by 3HDC, a specific antagonist of this receptor (Brockhoff et al., 2011) (Figure 5B), and was completely abolished in USC-SkMCs silenced for TAS2R46 expression (Figure 5C; Supplementary Figure S4).

We previously demonstrated that in airway smooth muscle cells absinthin counteracts histamine-induced calcium rise, in a similar manner to what observed with acetylcholine in skeletal muscle, stimulating mitochondrial Ca²⁺-uptake in a cAMP/EPAC dependent way (Talmon et al., 2019). To test whether this mechanism was present also in skeletal muscle, we recorded the cytosolic calcium oscillation in USC-SkMCs stimulated with acetylcholine alone or combined with absinthin, in the presence of a PKA inhibitor (H-89, 10 µM) or a specific EPAC inhibitor (ESI-09, 10 µM). As shown in Figure 6, both the co-stimulation with H-89 and ESI-09 abolished the ability of absinthin to reduce acetylcholine-induced calcium increases (Figure 6), demonstrating therefore a direct involvement of the cAMP/EPAC cascade in the downstream signalling of hTAS2R46. Since Epac1 is localized on the mitochondrial inner membrane and matrix and can control the activity of the mitochondrial calcium uniporter (MCU) (Faza et al., 2017; Wang et al., 2016), we analysed the calcium movements in USC-SkMCs co-stimulated with acetylcholine, absinthin and the MCU inhibitor KB-R7943 (carbamimidothioic acid, KB, 10 µM). Again, KB-R7043 treatment resulted in a complete inhibition of the effect of absinthin (Figure 6, yellow line and bar).

The above data would suggest that absinthin signals via cAMP/EPAC to the mitochondrial MCU to increase the uptake of Ca²⁺. If this were the case, it could be expected that the reduction in cytosolic calcium would be paralleled by a simultaneous increase in calcium within the mitochondria. We then evaluated mitochondrial calcium levels by Rhod2-AM analysis. As expected, acetylcholine determined a mild increase in mitochondrial calcium that was significantly augmented by the co-stimulation with absinthin (Figure 7, red and blue lines and bars). The effect was completely reverted by 3HDC, confirming that the effect is strictly dependent on hTAS2R46 (Figure 7; light blue line and bar), and by the MCU inhibitor (Figure 7; yellow line and bar). Moreover, the effect of absinthin was counteracted by the incubation with H-89 and ESI-09, confirming the involvement of cAMP in the downstream cascade and perfectly aligning the effects obtained on cytosolic Ca²⁺ to those on mitochondrial Ca²⁺.

The above results, alongside providing further evidence for a novel signalling pathway that decreases Ca²⁺-signalling via increasing mitochondrial uptake, also suggest that hTAS2R46 activation might influence acetylcholine-induced contraction. We therefore tested this hypothesis in the collagen contraction assay. Skeletal muscle cells are spontaneously endowed of contraction and therefore even in absence of stimuli (Figure 8, no stimuli) it possible to measure a slight reduction of the collagen disc diameter. Interestingly, absinthin, alone, did not stimulate the contraction of USC-SkMCs, as demonstrated by a contraction kinetics superimposable to that of unstimulated cells (Figure 8A). As expected, acetylcholine, amplified and fastened the contraction process as early as 3 h after the addition of the neurotransmitter (Figure 8A), significantly shrinking the collagen disk, in comparison to control cells (Figure 8C). The co-stimulation with absinthin and acetylcholine slowed down the kinetic of the acetylcholine-induced contraction, and after 6 h the collagen area was still significantly larger than that of the acetylcholine-stimulated cells (Figures 8A, B). To further understand the machinery involved in the transduction signal, we performed the collagen contraction assay on primary SkMCs in the presence of both EPAC and PKA inhibitors: both compounds confirmed the ability of absinthin to contrast acetylcholine-induced contraction and strengthens the hypothesis that TAS2R46 mediates muscle dilation via cAMP (Figures 8D, E). Indeed, the EPAC inhibitor ESI-09 partially reverted absinthin effect between 3 and 4 h after contraction induction (Figure 8E).