Edited by Yuh Nung Jan, Howard Hughes Medical Institute, University of California, San Francisco, CA, and approved July 29, 2016 (received for review June 23, 2016)
The Merkel disc is a main type of tactile end organ for sensing gentle touch and is essential for sophisticated sensory tasks, including social interaction, environmental exploration, and tactile discrimination. Despite recent studies showing that Merkel cells in Merkel discs are main sites of mechanotransduction in response to tactile stimuli, it remains unclear how Merkel cells transmit tactile signals to Aβ-afferent endings, leading to tactile sensations. Here we show that Merkel discs are serotonergic synapses in the epidermis, that tactile stimuli trigger serotonin release from Merkel cells to excite their associated whisker Aβ-afferent endings, and that this epidermal serotonergic transmission is critical to both electrophysiological and behavioral responses to tactile stimulation.
The evolution of sensory systems has let mammals develop complicated tactile end organs to enable sophisticated sensory tasks, including social interaction, environmental exploration, and tactile discrimination. The Merkel disc, a main type of tactile end organ consisting of Merkel cells (MCs) and Aβ-afferent endings, are highly abundant in fingertips, touch domes, and whisker hair follicles of mammals. The Merkel disc has high tactile acuity for an object’s physical features, such as texture, shape, and edges. Mechanisms underlying the tactile function of Merkel discs are obscured as to how MCs transmit tactile signals to Aβ-afferent endings leading to tactile sensations. Using mouse whisker hair follicles, we show herein that tactile stimuli are transduced by MCs into excitatory signals that trigger vesicular serotonin release from MCs. We identify that both ionotropic and metabotropic 5-hydroxytryptamine (5-HT) receptors are expressed on whisker Aβ-afferent endings and that their activation by serotonin released from MCs initiates Aβ-afferent impulses. Moreover, we demonstrate that these ionotropic and metabotropic 5-HT receptors have a synergistic effect that is critical to both electrophysiological and behavioral tactile responses. These findings elucidate that the Merkel disc is a unique serotonergic synapse located in the epidermis and plays a key role in tactile transmission. The epidermal serotonergic synapse may have important clinical implications in sensory dysfunctions, such as the loss of tactile sensitivity and tactile allodynia seen in patients who have diabetes, inflammatory diseases, and undergo chemotherapy. It may also have implications in the exaggerated tactile sensations induced by recreational drugs that act on serotoninergic synapses.
Sensory systems for detecting tactile stimuli have evolved from mechanosensitive free nerve endings in invertebrates to complicated tactile end organs in mammals, although mechanosensitive free nerve endings are conserved in mammals mainly for noxious mechanical stimuli. Tactile end organs, including Merkel discs, Pacinian corpuscles, Meissner’s corpuscles, and Ruffini endings, are highly specialized structures in the periphery of mammals (1, 2). These organs are crucial to the performance of sophisticated sensory tasks, such as environmental explorations, social interactions, and tactile discrimination (2). The Merkel disc, also known as Merkel cell-neurite complex, is a main type of tactile end organ highly abundant in human fingertips, whisker hair follicles, touch domes, and other tactile-sensitive spots throughout mammalian bodies (3, 4). Structurally, Merkel discs are composed of Merkel cells (MCs) and their associated Aβ-afferent nerve endings to form a structure of disc-shaped expansion (3, 5). Merkel discs have high tactile acuity and are very sensitive to skin indentation, pressure, hair movement, and other tactile stimuli. Tactile stimuli to Merkel discs in the touch domes of the skin and whisker hair follicles result in slowly adapting type 1 (SA1) responses, the characteristic Aβ-afferent impulses for tactile encoding (2, 5). Functionally, SA1 responses in fingertips and whisker hair follicles are essential for tactile discrimination of an object’s texture, shape, and other physical properties.
Despite the recent progress in demonstrating the essential role of MCs via Piezo2 channels in mechanotransduction and tactile behaviors in mammals (6⇓–8), it is currently not known how the mechanotransduction in MCs can subsequently result in Aβ-afferent SA1 impulses leading to tactile sensations. It has been hypothesized that Merkel discs may be sensory synapses and tactile signals are transmitted synaptically from MCs to Aβ-afferent nerve endings (3, 9, 10). However, it has long been believed that the first sensory synapses for somatosensory signals are located centrally in the dorsal horn of the spinal cord and brainstem, where sensory signals are transmitted synaptically using glutamate (Glu) as a principal neurotransmitter (11, 12). Thus, the hypothesis of Merkel discs being sensory synapses is challenged by the classic view about somatosensory transmission.
The idea that Merkel discs are sensory synapses has been further challenged by the finding that the generation of Aβ-afferent impulses by tactile stimulation is faster than the latency of chemical synapses (13). Interestingly, transcripts of synaptic release machinery, such as synapsin, synaptotagmin, and vesicular glutamate transporter 2, have been observed in MCs (14). Dense-core vesicles, which are thought to contain Glu, ATP, serotonin, substance P (SP), enkephalin (Enk), and other chemical messengers, have also been observed in MCs (3, 5). Although SA1 responses were found to be substantially suppressed by some antagonists for Glu and 5-hydroxytryptamine (5-HT) receptors at very high concentrations (15⇓–17), the inhibitory effects with the 5-HT receptor antagonists used in the previous studies were found to be a result of nonspecific effects. Moreover, previous studies have failed to show the presence of any excitatory receptors for the hypothetic transmitters at Aβ-afferent endings of Merkel discs (18). Alternatively, the chemical messengers, including Glu, ATP, and serotonin in MCs have been thought to be autocrine or paracrine to modify Merkel disc’s functions (3, 10, 19) rather than to be transmitters to directly elicit tactile impulses at Aβ-afferent endings in Merkel discs. Thus, molecular mechanisms underlying tactile signaling at Merkel discs remain largely obscured.
In the present study we set out to study whether Merkel discs are sensory synapses uniquely located in the epidermis to transmit tactile signals from MCs to Aβ-afferent endings, and if so, what is the transmitter released from MCs in response to tactile stimulation and how this transmitter initiates tactile impulses at Aβ-afferent endings to result in tactile sensations.
Robust SA1 whisker afferent impulses (SA1 responses) could be evoked following a small deflection of whisker hair in our mouse whisker hair follicle preparation (Fig. 1 A and B and Fig. S1 A and B). We performed patch-clamp recordings from MCs in the Merkel disc region to determine whether mouse MCs were mechanically sensitive and electrically excitable (Fig. 1C). We found that these cells were indeed mechanically sensitive and electrically excitable, as was evidenced by mechanically activated currents (MA) and membrane depolarization-evoked action potentials (AP), respectively (Fig. 1C and Fig. S1 C and E). To determine whether a chemical messenger might be used for transmitting MC activity into whisker afferent endings to drive SA1 responses, we tested possible candidates to see if focal application of each candidate could directly evoke whisker afferent impulses that mimicked SA1 responses (Fig. 1D). Eleven possible candidates, including six classic transmitters/chemical messengers and five neuropeptides, were tested by briefly (400 ms) puffing individual candidates via a sharp glass pipette to the Merkel disc region in the whisker hair follicle. Of the 11 candidates tested, only serotonin evoked robust whisker afferent impulses (Fig. 1D). In sharp contrast, the other 10 candidates, including ATP, Glu, norepinephrine (NE), acetylcholine (Ach), histamine (His), SP, vasoactive intestinal peptide (VIP), Enk, calcitonin gene-related peptide (CGRP), and cholecystokinin octapeptide-8 (CCK-8) did not induce any significant increase in whisker afferent impulses following their focal applications to the Merkel disc region of whisker hair follicles (Fig. 1D). As a control, the ability of these candidates to excite trigeminal ganglion (TG) neurons were examined, and in our hands we found that serotonin, ATP, Glu, Ach, and His, but not the remaining compounds, could directly excite some mouse TG neurons (Fig. S1 F and G).
Tactile SA1 responses can be mimicked by focal application of serotonin to Merkel discs where 5-HT3A receptors are identified at whisker afferent endings. (A) Experimental set-up and basic structures of a whisker hair follicle related to the present study. Whisker afferent impulses were recorded using a suction electrode. Tactile stimuli were delivered by deflecting the whisker hair. MC, Merkel cells; GM, glassy membrane; Rs, ring sinus; RRC, rete ridge collar. MCs are located at two places, underneath the Rs and around the RRC. (B) Sample trace (Upper) shows typical SA1 impulses elicited by a 38-µm displacement to deflect the whisker hair. Instantaneous frequency of SA1 impulse over time is shown in the Lower panel. (C, Left) Image shows MCs labeled by quinacrine vital staining in a fresh whisker hair follicle preparation. A patch-clamp recording electrode and a mechanical stimulation probe are indicated in the image. (Right) Sample traces show MA currents (Upper) and membrane responses (Lower) of a MC recorded in situ in a whisker hair follicle preparation. (D) Two sample traces show focally puffing serotonin (Upper) but not Glu (Lower) to Merkel disc region elicits whisker afferent impulses. Bar graph below the sample traces is summary data of whisker afferent impulses following the focal application of bath solution (control, n = 30), serotonin (1 mM, n = 12), ATP (1 mM, n = 12), Glu (1 mM, n = 6), NE (1 mM, n = 6), Ach (1 mM, n = 6), His (1 mM, n = 6), SP (358 µM, n = 6), VIP (30 µM, n = 6), Enk (0.9 mM, n = 6), CGRP (26 µM, n = 6), and CCK-8 (87 µM, n = 6). (E) Image shows that a 5-HT3AeGFP-positive whisker afferent fiber (arrow indicated) extends its processes (arrowheads indicated) from the Rs to the underneath MC region. (F) A 5-HT3AeGFP-positive whisker afferent fiber (arrow indicated) innervates the RRC Merkel cell region. The boxed region shows 5-HT3AeGFP terminal processes. (G) Image shows the presence of 5-HT-ir cells and 5-HT3AeGFP-whisker afferent endings in Merkel disc region. The surface section of a coronal cut follicle was used and the outer root sheath in this image is seen as a sheet of cells. Arrowhead indicates a 5-HT3AeGFP whisker afferent ending that is in close contact with a serotonin-ir cell in the Merkel disc region of a whisker hair follicle. Data represent the mean ± SEM; ***P < 0.001; compared with bath, one-way ANOVA with Bonferroni post hoc tests.
SA1 impulses evoked by whisker hair deflections, mechanical sensitivity, and electrical excitability of MCs in mouse whisker hair follicles, and responses of mouse TG neurons to candidate chemical messengers. (A) Sample traces of SA1 impulses recorded from the whisker afferent bundle of an isolated mouse whisker hair follicle. SA1 impulses were induced by different displacements to deflect whisker hairs. From top to bottom, displacements are 3, 10, 25, and 38 μm. (B) Summary data for the experiments illustrated in A (n = 8). (C) Summary data of the relationship of MC MA current amplitudes and stimulation intensity (displacement distances) in normal bath solution or following serotonin application (100 μM, 20 min, n = 8). (D) Summary data of the V-I relationship of MCs in normal bath solution or following serotonin application (100 μM, 20 min, n = 8). (E) Summary data of AP firing frequency of MCs in responses to membrane depolarization in normal bath solution or following serotonin application (100 μM, 20 min, n = 8). (F) Sample images show Ca2+-imaging measurement of GCaMP3 fluorescent intensity changes in a TG neuron following the bath application of 1 mM ATP. (Upper) Baseline before ATP application. (Lower) Peak response following ATP application. The Ca2+-imaging experiment was performed on a whole-mount TG prepared from a GCaMP3 mouse. (G) Summary data of the changes in GCaMP3 fluorescent intensity (ΔF/F0) in TG neurons under the following conditions: application of bath (control, n = 7), KCl (50 mM, n = 7), serotonin (1 mM, n = 7), ATP (1 mM, n = 7), Glu (1 mM, n = 7), Ach (1 mM, n = 7), His (1 mM, n = 7), NE (1 mM, n = 7), SP (358 µM, n = 7), VIP (30 µM, n = 7), Enk (0.9 mM, n = 7), CGRP (26 µM, n = 7), and CCK-8 (87 µM, n = 7). Data represent the mean ± SEM. NS, no significant difference; **P < 0.01; ***P < 0.001; one-way ANOVA or paired Student’s t test.
Although serotonin evoked whisker afferent impulses, it had no effect on MC mechanical sensitivity and membrane excitability (Fig. S1 C–E), indicating that MCs are not the primary sites of serotonin’s action. Whisker afferent impulses evoked by focal serotonin application onto the Merkel disc region may be because of its direct action on whisker afferent endings. To begin to address this question, we used 5-HT3AeGFP transgenic mice, which showed 5-HT3AeGFP fluorescence in TG neurons (Fig. S2 A and B), to examine whether 5-HT3A was present at whisker afferent endings in the region of Merkel discs. The 5-HT3AeGFP-positive whisker afferent endings were indeed identified within whisker hair follicles using the 5-HT3AeGFP transgenic mice (Fig. 1 E–G). The 5-HT3AeGFP-positive afferent fibers terminated in the outer root sheath underneath the ring sinus (Rs) (Fig. 1E) and around the rete ridge collar (RRC) (Fig. 1F), the two Merkel disc regions in a whisker hair follicle (see schematic diagram Fig. 1A) (20). Moreover, 5-HT3AeGFP-positive afferent endings were found to have close contacts with cells that were serotonin-immunoreactive (ir) in the Merkel disc region (Fig. 1G).
Immunoreactivity of 5-HT3AeGFP, 5-HT3A, 5-HT2A, and 5-HT2B receptors in trigeminal afferent neurons. (A) Image shows 5-HT3AeGFP-positive neurons in a TG section. Some 5-HT3AeGFP-positive and -negative neurons in the field are indicated by arrows and arrowheads, respectively. (B) Cell size distribution of 5-HT3AeGFP-positive (green) and total (gray) TG neurons. (C) Images show immunoreactivity of 5-HT3A (Left), 5-HT2A (Center), and 5-HT2B (Right) in three TG sections. (D) Percent of positive (red) and negative (white) neurons as illustrated in C. (E) Cell size distribution of 5-HT3A-ir (Left, 337 positive cells in 1,246 total cells), 5-HT2A-ir (Center, 413 positive cells in 1,604 total cells), and 5-HT2B-ir (Right, 405 positive cells in 1,426 total cells) neurons in trigeminal sections. (F) Three images on top show FG retrogradely labeled neurons (indicated by arrowheads) in three different TG sections. FG was microinjected into whisker hair follicles 7 d before immunostaining experiments. The Lower three images show immunostaining for 5-HT3A, 5-HT2A and 5-HT2B on the same three TG sections. Arrowheads indicate the FG-labeled neurons that are immunoreactive for 5-HT3A (Left), 5-HT2A (Center), and 5-HT2B (Right). The overlay images are presented in Fig. 2I.
To understand the mechanisms underlying the long excitatory action after a brief focal application of serotonin, we performed patch-clamp recordings from whisker afferent neurons in a whole-mount trigeminal ganglion preparation. In this set of experiments, whisker afferent neurons were retrogradely labeled by microinjection of the fluorescent tracer DiI into individual hair follicles (Fig. 2A). All DiI-labeled whisker afferent neurons obtained from WT mice were found to respond to serotonin (100 μM) with inward currents (Fig. 2 B and C). Interestingly, most (75%) of these whisker afferent neurons responded to serotonin with an initial fast current (Ifast) followed by a slow current (Islow) (Fig. 2 B and C); the Islow current was long-lasting after the termination of serotonin application (Fig. 2B). The remaining 25% of whisker afferent neurons responded to serotonin with Islow currents only (Fig. 2C). We next determined whether serotonin-evoked inward currents were mediated by 5-HT3 receptors because they are the only ion channels in the 5-HT receptor family and all other 5-HT receptors are metabotropic receptors (21). In this set of experiments, 5-HT3A−/− mice were used and serotonin effects were tested on retrogradely labeled whisker afferent neurons. Because 5-HT3A subunits are essential for forming functional 5-HT3 receptors, genetic deletion of 5-HT3A subunits completely knocks out all functional 5-HT3 receptors (21, 22). Interestingly, we found that all retrogradely labeled whisker afferent neurons in 5-HT3A−/− mice still responded to serotonin with inward currents (Fig. 2 B and C). However, only Islow currents were observed (Fig. 2 B and C) and Ifast currents were completely lost (Fig. 2 B–D). This result indicates that Ifast currents seen in whisker afferent neurons of WT mice are mediated by 5-HT3 receptors, and Islow currents seen in whisker afferent neurons in both WT and 5HT3A−/− mice must be mediated by metabotropic 5-HT receptors.
Serotonin evokes inward currents in whisker afferent neurons through activation of both ionotropic and metabotropic 5-HT receptors. (A) Images show two DiI retrogradely labeled whisker afferent neurons (arrowheads indicated) in a whole-mount TG preparation. Both panels are the same field taken under DIC (Upper) and fluorescent (Lower) microscopy. (B) Two sample traces show inward currents evoked by 100 µM serotonin from a whisker afferent neuron of a WT (Upper) and a 5-HT3A−/− (Lower) mouse. The Upper trace shows a mixture of a fast current (Ifast) and a slow current (Islow) but the Lower trace displays only a Islow current. The horizontal bar above each trace indicates the duration (1 min) of serotonin application. (C) Percent of whisker afferent neurons showing mixed Ifast and Islow currents (solid bar) and Islow current only (open bar). WT, n = 16; 5-HT3A−/−, n = 10. (D) Amplitude of Ifast currents in whisker afferent neuron of WT mice (n = 12). No significant Ifast current was observed in any whisker afferent neuron of 5-HT3A−/− mice (n = 10). (E) First set of two bars, Islow currents evoked by the 5-HT2A agonist TCB-2 (10 µM) in whisker afferent neurons of WT mice (solid bar, n = 5) and 5-HT3A−/− mice (open bars, n = 7). Second set of two bars, Islow currents evoked by the 5-HT2B agonist BW ( 10 µM) in whisker afferent neurons of WT mice (solid bar, n = 5) and 5-HT3A−/− mice (open bars, n = 6). (F) Serotonin-evoked Islow currents in whisker afferent neurons of 5-HT3A−/− mice in the absence (open bar, n = 6), presence of 100 µM KT (n = 8) and 100 µM LY (n = 7). Islow currents were evoked by 100 µM serotonin in each test. (G) Single-cell RT-PCR results from two whisker afferent neurons; one shows 5-HT3A, 5-HT2A and 5-HT2B mRNAs (Left) and the other shows only 5-HT2A and 5-HT2B mRNAs (Right). (H) Summary data show the percent of cells that expressed all three 5-HT receptors (solid bar, n = 10) or only expressed 5-HT2A and 5-HT2B receptors (n = 9). (I) Immunoreactivity of 5-HT3A, 5-HT2A, and 5-HT2B receptors detected on retrogradely labeled whisker afferent neurons (arrows indicated). Data represent the mean ± SEM. NS, no significant difference; *P < 0.05; **P < 0.01; ***P < 0.001; unpaired Student t test or one-way ANOVA with Bonferroni post hoc tests.
Because activation of 5-HT2A receptors have been shown to increase neuronal excitability (23), we tested whether serotonin-evoked Islow currents in whisker afferent neurons were a result of the activation of 5-HT2A receptors. We found that application of the 5-HT2A selective agonist TCB-2 (10 μM) evoked Islow currents in whisker afferent neurons of both WT and 5-HT3A−/− mice (Fig. 2E). Interestingly, the Islow currents in 5-HT3A−/− mice were larger than those of WT mice (Fig. 2E), which is likely a result of compensatory up-regulation. Because 5-HT2B receptors are functionally similar to 5-HT2A receptors (24), we tested the 5-HT2B selective agonist BW-723C86 (BW, 10 μM) and found that it also evoked Islow currents in whisker afferent neurons of both WT and 5-HT3A−/− mice (Fig. 2E). We further tested effects of the 5-HT2A antagonist ketanserin (KT, 100 μM) and the 5-HT2B antagonist LY-266097 (LY, 100 μM) on serotonin-evoked Islow currents in whisker afferent neurons of 5-HT3A−/− mice; serotonin-evoked Islow currents were significantly inhibited by both KT and LY (Fig. 2F).
The involvement of 5-HT3, 5-HT2A, and 5-HT2B receptors in serotonin-evoked responses was further supported by single-cell RT-PCR experiments, which showed the expression of mRNAs of these three 5-HT receptors in retrogradely labeled whisker afferent neurons (Fig. 2 G and H). All individual whisker afferent neurons expressed mRNAs of both 5-HT2A and 5-HT2B receptors, and the majority of these neurons (60%) coexpressed 5-HT3A mRNAs (Fig. 2 G and H). This result is consistent with serotonin-evoked Ifast and Islow currents in whisker afferent neurons of WT mice (Fig. 2C). The expression of 5-HT3A, 5-HT2A, and 5-HT2B receptors in whisker afferent neurons was also consistent with immunostaining results, which showed presence of immunoreactivity for each of these three receptors in retrogradely labeled whisker afferent neurons (Fig. 2I and Fig. S2 C–F).
We determined whether ionotropic and metabotropic 5-HT receptors may play differential roles in evoking whisker afferent impulses because whole-cell inward currents mediated by 5-HT3 and 5-HT2A/5-HT2B receptors were distinct in whisker afferent neurons. In whisker hair follicles of 5-HT3A−/− mice, focal application of serotonin (1 mM, 400 ms) to Merkel disc regions evoked a delayed response; whisker afferent impulses were not evoked during serotonin application and were generated with delay after the termination of serotonin application, and the responses extended for a prolonged period (Fig. 3A). In contrast, focal application of the 5-HT3 agonist SR57227 (1 mM) did not evoke any impulses (Fig. 3A). We further tested 5-HT2A agonist TCB-2 (1 mM) and the 5-HT2B agonist BW (1 mM) to see if they could directly elicit whisker afferent impulses. We found that both TCB-2 and BW produced delayed increases of whisker afferent impulses but the effect of BW was very small (Fig. 3A). These results in 5-HT3A−/− mice indicate that functional metabotropic 5-HT receptors (5-HT2A/5-HT2B) are present at the whisker afferent endings in the Merkel disc region and their activation elicits delayed impulses.
Ionotropic and metabotropic 5-HT receptors play differential roles with synergism in driving whisker afferent impulses. (A, Left) Sample traces show whisker afferent impulses elicited by serotonin or 5-HT receptor subtype selective agonists focally applied to whisker hair follicle Merkel disc region of 5-HT3A−/− mice. From top to bottom, the tests are serotonin, the 5-HT3A agonist SR57227, the 5-HT2A agonist TCB-2, and the 5-HT2B agonist BW. (Right) Summary data of the impulse numbers induced by the following tests: serotonin (n = 7), SR57227 (n = 7), TCB-2 (n = 7), and BW (n = 12). (B) Similar to A except WT mice were used for the tests of serotonin (n = 7), SR57227 (n = 7), TCB-2 (n = 7), and BW (n = 12). (C) Similar to B except tests were performed with different agonist combinations as follows, SR57227+TCB-2+BW (n = 7), SR57227+BW (n = 7), SR57227+TCB-2 (n = 7), and TCB-2+BW (n = 7). For all panels, all compounds were puff-applied at the concentration of 1 mM for 400 ms. Data represent the mean ± SEM; *P < 0.05; **P < 0.01; ***P < 0.001; one-way ANOVA with Bonferroni post hoc tests.
In contrast to the results obtained from 5-HT3A−/− mice, focal application of serotonin to whisker hair follicles of WT mice induced immediate increases of whisker afferent impulses during serotonin application (Fig. 3B). There were also delayed and prolonged impulses after the termination of serotonin application (Fig. 3B). The 5-HT3 agonist SR57227 (1 mM) elicited immediate but brief increases of afferent impulses (Fig. 3B). In contrast to the 5-HT3 agonist, application of the 5-HT2A agonist TCB-2 (1 mM) or the 5-HT2B agonist BW (1 mM) induced a delayed increase of whisker afferent impulses; the effect of BW was very small but was significantly higher than control with bath application (Fig. 3B). Selective agonists for other 5-HT receptor subtypes (5-HT1, 5-HT2C, 5-HT4, 5-HT5, 5-HT6, and 5HT7) were tested and none of them elicited any whisker afferent impulses (Fig. S3).
Tests of 5-HT receptor subtype selective agonists for their ability to elicit whisker afferent impulses. The agonists tested were eltoprazine (n = 11, for 5-HT1), TCB-2 (n = 16, for 5-HT2A), BW (n = 11, for 5-HT2B), m-CPP (n = 11, for 5-HT2C), SR57227 (n = 20, for 5-HT3), cisapride (n = 11, for 5-HT4), 5-CT (n = 11, for 5-HT5), EMD (n = 11, for 5-HT6), and AS-19 (n = 11, for 5-HT7). Bath was applied as control (n = 11). Each agonist was tested at the concentration of 1 mM and was puff-applied for 400 ms to the Merkel disc region around the Rs. Data represent the mean ± SEM; *P < 0.05; ***P < 0.001; NS, no significant difference, one-way ANOVA.
The effects of serotonin were much greater than the simple summation of the effects induced by individually applying the 5-HT3 agonist SR57227, 5-HT2A agonist TCB-2, and 5-HT2B agonist BW (Fig. 3B), raising a possibility that ionotropic 5-HT receptors (5-HT3) may have synergistic effects with metabotropic 5-HT receptors (5-HT2A/5-HT2B) on generating whisker afferent impulses. Consistent with this idea, when a mixture made with SR57227, TCB-2, and BW was focally applied, the afferent impulses became greatly increased to the level comparable to those induced by serotonin (Fig. 3C). Removing any one of the agonist from the mixture significantly weakened its effects in eliciting whisker afferent impulses (Fig. 3C). These results suggest that although ionotropic and metabotropic 5-HT receptors have differential temporal response characteristics in generating whisker afferent impulses, they also produce synergistic effects in driving whisker afferent impulses when both ionotropic and metabotropic receptors are activated simultaneously.
To demonstrate that endogenous serotonin activates 5-HT3 receptors to drive SA1 impulses, we recorded SA1 impulses in 5-HT3A−/− mice in response to whisker hair movement and found that SA1 frequency was significantly less than those of WT mice from the beginning to middle (100–1,200 ms) of the SA1 response, but not during the later phase (1,300–2,600 ms) of the SA1 response (Fig. 4 A–C). Electrophysiological and mechanical transduction properties of MCs were not different between 5-HT3A−/− and WT mice (Fig. S4); membrane properties of whisker afferent neurons were also not different between 5-HT3A−/− and WT mice (Table S1). Pharmacological block of whisker afferent neuron 5-HT3 receptors in WT mice with the two 5-HT3 selective antagonists Y25130 (2 and 20 μM) (Fig. 4D and Fig. S5A) and MDL 72222 (10 µM) (Fig. S5A) also significantly inhibited SA1 impulses in a manner similar to genetic deletion of 5-HT3 receptors (Fig. 4C). To cross-validate the results obtained from 5-HT3A−/− mice and WT mice, we tested the effects of the two 5-HT3 antagonists at the above concentrations on SA1 responses in 5-HT3A−/− mice, and found that the inhibitory effects were completely lost (Fig. 4E and Fig. S5B). This result indicates that the inhibitory effects by these antagonists in WT mice are specifically mediated by 5-HT3 receptors. We determined whether block of 5-HT2A and 5-HT2B receptors could inhibit SA1 responses, and found that the 5-HT2A antagonist KT (1 and 10 μM) plus the 5-HT2B antagonist LY (10 and 100 µM) suppressed SA1 responses in the later phase SA1 responses (500–2,600 ms) but not in the very early phase (100–400 ms) SA1 responses in WT mice (Fig. 4F). The effects of these two antagonists at the concentrations used in the present study were unlikely because of a potential nonspecific effect of them on MCs, because MC electrophysiological and mechanical transduction properties were not significantly affected (Fig. S6). It is worth noting that previous studies used MDL-72222 and KT at much higher concentrations (up to 500 μM) and showed much stronger inhibitory effects on SA1 responses (15, 17), but we found that at these high concentrations the inhibitory effects on SA1 responses by MDL-72222 and KT were a result of their nonspecific actions (Fig. S7).
Synergism of ionotropic and metabotropic 5-HT receptors is essential for electrophysiological whisker tactile responses. (A) Sample trace shows SA1 impulses recorded from whisker hair follicles of a WT mouse. Arrowhead-indicated region is expanded on right. (B) Similar to A except SA1 impulses were recorded from a 5-HT3A−/− mouse. (C) Summary of SA1 impulses of WT (n = 30) and 5-HT3A−/− (n = 30) mice. (D) SA1 impulses of WT mice in the absence (control, n = 10) and presence (n = 10) of 2 µM Y25130. (E) Similar to D except using 5-HT3A−/− mice and 20 µM Y25130 (n = 6). (F) SA1 impulses of WT mice in control and the presence of KT+LY (n = 6). (G) Similar to F except the blocker mixture also included Y25130 (n = 7). (H) Similar to F except whisker afferent impulses were recorded from 5-HT3A−/− mice (n = 6). Data represent the mean ± SEM; *P < 0.05; **P < 0.01; ***P < 0.001; two-way ANOVA with Bonferroni post hoc tests.
Electrophysiological and mechanical transduction properties of MCs in WT and 5HT3A−/− mice. (A) Sample traces show membrane responses and APs in response to depolarizing current steps in a MC of a WT mouse (Left) and a MC of a 5-HT3A−/− mouse (Right). (B, Left) Summary data show V-I relationship of MCs of WT (n = 11) and 5-HT3A−/− mice (n = 9). (Right) Summary data of action potential frequency of WT MCs (open bar, n = 11) and 5-HT3A−/− MCs (closed bar, n = 9). Patch-clamp recordings were performed under the current-clamp mode and depolarizing steps were applied from -60 to +220 pA in increments of 20 pA. (C, Left) Two sets of traces show MA currents recorded from a WT MC (Upper) and a 5-HT3A−/− MC (Lower) in response to displacement steps from 0.5 to 3 µm in increments of 0.5 µm. (Right) Summary data of MA current amplitudes elicited at different displacement distances in WT Merkel cells (n = 7) and 5-HT3A−/− Merkel cells (n = 7). Data represent the mean ± SEM. NS, no significant difference, two-way ANOVA or unpaired Student’s t test.
Membrane properties of whisker afferent neurons in WT and 5-HT3A−/− mice
Pharmacological block of ionotropic 5-HT3 receptors impairs the earlier phase of whisker SA1 responses. (A) Summary data of SA1 impulses of WT mice at initial (100 ms), middle (1,000 ms), and end (2,600 ms) time points during the course of 2,600-ms whisker displacement (also see Fig. 4D). SA1 impulses were tested in the absence (control, open bars, n = 10), and presence of either 2 µM Y25130 (black bar, n = 10), 20 µM Y25130 (gray bars, n = 10), or 10 µM MDL-72222 (horizontal hatched bars, n = 6). (B) Similar to A except experiments were performed using whisker hair follicles from 5-HT3A−/− mice and tests are control (n = 6, open bars), 2 µM (n = 6, black bars), and 20 µM (n = 6, gray bars) Y25130, and 10 µM MDL (n = 6, horizontal hatched bars). Data represent the mean ± SEM; *P < 0.05; **P < 0.01; ***P < 0.001; NS, no significant difference, comparing with control, one-way ANOVA.
MC electrophysiological properties and mechanical transduction are not affected by the 5-HT2A antagonist KT (10 µM), the 5-HT2B antagonist LY (10 µM), and the 5-HT3 antagonist Y25130 (100 µM). (A) Sample traces show membrane responses and AP in response to depolarizing current steps in a MC. The tests were performed in the absence (control, Left) and presence (Right) of 10 µM KT + 10 µM LY + 100 µM Y25130. Depolarizing steps were from −60 to +220 pA in increments of 20 pA for both traces. (B, Left) Summary data (n = 10) show V-I relationship under current-clamp recordings in the absence (control) and presence of 10 µM KT + 10 µM LY + 100 µM Y25130. (Right) Summary data (n = 10) of AP frequency in the absence (control, open bar) and presence (closed bar) of 10 µM KT + 10 µM LY + 100 µM Y25130. (C, Left) Two sets of MA currents elicited from MCs in the absence (control) and presence of 10 µM KT + 10 µM LY + 100 µM Y25130. Displacement steps were applied from 0.5 to 3 µm in increments of 0.5 µm. (Right) Summary data (n = 6) of MC MA current amplitudes elicited at different displacements in the absence (control) and presence of 10 µM KT + 10 µM LY + 100 µM Y25130. All MCs were from WT mice. Data represent the mean ± SEM. NS, no significant difference, two-way ANOVA or paired Student’s t test.
High concentrations of KT and MDL-72222 used in previous studies produce nonspecific effects. Previous studies showed that the 5-HT3 receptor antagonist MDL-72222 at the concentrations of 100 and 500 μM almost completely blocked SA1 responses and that the 5-HT2A antagonist KT at the concentration of 500 μM also greatly suppressed SA1 responses (15⇓–17). However, our results shown in the figure indicate that the previous studies are a result of the nonspecific effects of these two drugs at very high concentrations. (A and B) The 5-HT3 receptor antagonist MDL-72222 at the concentration of 100 μM blocked SA1 responses in both WT (A, n = 8) and 5-HT3−/− (B, n = 6) mice. In A and B, the Left panel shows sample traces of SA1 responses in the absence (Upper, control) and presence (Lower) of 100 μM MDL-72222. The Right panel shows summary data of SA1 impulse frequency over time in control (black circles) and in the presence of MDL-72222 (red circles). SA1 impulses were completely blocked by MDL-72222 in both WT (A) and 5-HT3−/− (B), indicating that the inhibitory effects on SA1 responses by MDL-72222 at this or higher concentrations are not mediated by 5-HT3 receptors. (C) The 5-HT2A antagonist KT at the concentration of 500 μM inhibits Piezo2-mediated mechanotransduction in MCs, the upstream key event for driving SA1 responses. (Left) Two sample traces show MA currents in MCs in the absence (control, black trace) and presence (red trace) of 500 µM KT. (Right) Summary data show MC MA currents were significantly suppressed in the presence of 500 µM KT (red circles, n = 6). Data represent the mean ± SEM; *P < 0.05; **<0.01; ***P < 0.001, two-way ANOVA (A and B) or paired Student’s t test (C).
We investigated whether ionotropic and metabotropic 5-HT receptors had synergism to drive the majority of SA1 impulses. In whisker hair follicles of WT mice, the combination of low concentrations of the 5-HT3 antagonist Y25130, 5-HT2A antagonist KT, and 5-HT2B antagonist LY produced pronounced inhibition of SA1 responses (Fig. 4G). At combined higher concentration, these ionotropic and metabotropic serotonin receptor antagonists almost completely inhibited SA1 responses (Fig. 4G). MC electrophysiological and mechanical transduction properties were not significantly affected by these antagonists (Fig. S6). In 5-HT3A−/− mice, combined application of 5-HT2A and 5-HT2B receptor blockers (KT+LY) at a low concentrations produced pronounced inhibitory effects, and at high concentrations almost completely blocked SA1 impulses (Fig. 4H). These results indicate that endogenous serotonin drive the majority of electrophysiological tactile responses through the synergism of metabotropic and ionotropic 5-HT receptors.
Endogenous serotonin may be released in whisker hair follicles responding to whisker hair deflections. To test this idea, we determined whether SA1 responses elicited by whisker hair deflections could be diminished by botulinum toxin type A (BoNTA), a neurotoxin that specifically cleaves synaptosomal-associated protein (25 kDa) (SNAP-25) to prevent vesicle fusion and neurotransmitter release (25). As shown in Fig. 5A, SA1 impulses were significantly suppressed by BoNTA. Moreover, using ELISA, we detected serotonin release from whisker hair follicles in a tactile stimulation-dependent manner (Fig. 5B). Serotonin may be released from MCs because mRNAs of tryptophan hydroxylase 1 (an enzyme for serotonin synthesis) were expressed in MCs (Fig. 5C) and serotonin-ir cells identified in the MC region (Fig. 5C), where these cells were closely contacted by 5-HT3AeGFP whisker afferent endings (Fig. 1G).
Tactile stimulation releases serotonin from MCs in whisker hair follicles. (A) Two sets of sample traces on Left show whisker afferent SA1 responses recorded from whisker hair follicles of WT mice in normal bath (control, Upper) and in the presence of 0.5 nM BoNTA for 2 h (Lower). (Right) Summary data of the changes of whisker afferent SA1 impulses recorded over time in normal bath solution (control, n = 6, black circles) and in the presence of 0.5 nM BoNTA (n = 8, red circles). (B) Serotonin detected in bath solution by ELISA following mechanical stimulation that deflects whisker hairs. Bars from left to right are serotonin concentrations detected following increased tactile stimulation numbers. In each experiment, 10 follicles were used and a tactile stimulus is a displacement at 38 µm. Eight sets of tests (n = 8) were performed for each condition. (C, Upper) RT-PCR shows the expression of tryptophan hydroxylase 1 (Tph1) mRNAs in MCs. Similar results were observed in two other sets of RT-PCR experiments. (Lower) Image shows the present of 5-HT-ir cells (arrowheads indicated) in Merkel disc region of the out root sheath. The middle section of a coronal cut follicle was used and the out root sheath (MC region) is a thin layer at the border below the Rs. (D) Image shows a Merkel disc region of a fresh whisker hair follicle used for real-time detection of serotonin release from MCs. MCs were vital-stained by quinacrine and a MC is indicated by an arrowhead. The MC is in close contact with the tip of a carbon fiber electrode (star indicated) and the same carbon fiber electrode was used for both mechanical stimulation and amperometric detection of serotonin release. (E) Sample trace of oxidation current events detected by a carbon fiber electrode following mechanical stimulation of a Merkel cell by a 10-μm displacement. Two traces on the bottom are at an expanded scale for the two parts indicated by dashed lines. (F) Percent of MCs showing multiple release events (n = 75) or only a single release event (n = 47). (G) Peak amplitude (serotonin concentrations) of the first event of multiple release event cells (n = 40) and single release event cells (n = 32). (H) Event-amplitude distribution of multiple release events. The first event is excluded from the plots. Greater than 65% of the release events were in the size range of 0.4–0.6 μM serotonin. (I, Left) Sample traces show the detection of mechanical stimulation-evoked serotonin release in normal bath (control, Top), in a low Ca2+ bath solution (Middle), and recovery after returning to normal bath (Bottom). (Right) Summary data (n = 10) of serotonin release in control, low Ca2+ bath solution and recovery. The value of the biggest event in each test was used for the summary data. Data represent the mean ± SEM; *P < 0.05; ***P < 0.001; one-way ANOVA with Bonferroni post hoc tests or paired Student’s t test.
We then determined whether mechanical transduction in MCs directly coupled to serotonin release. Mechanical stimulation evoked serotonin release from MCs, as was evidenced by the oxidation currents recorded by carbon fiber electrodes (Fig. 5 D and E; also see Fig. S8 A and B). In contrast, no significant oxidation current could be detected when the mechanical stimulation was applied to non-MCs (Fig. S8 C and D). In most MCs, the first release events exhibited the highest amplitudes, and were followed by many smaller release events (Fig. 5 E–H). Some MCs only showed single release events with large amplitudes (Fig. 5 F and G). For cells showing multiple release events, most events had small amplitudes within a narrow range (Fig. 5H). Serotonin release was dependent on Ca2+ entry and no release event was detected in a low extracellular Ca2+ bath condition (Fig. 5I). These results indicate that serotonin release is coupled with mechanical transduction in MCs.
Detection of serotonin release from MCs using the amperometry with carbon fiber electrodes. (A) Calibration for the measurement of transient serotonin following a brief puff-application. A set of transient oxidation currents detected by a carbon fiber electrode following brief puff applications of serotonin onto the electrode tip in a bulk bath solution. For detecting serotonin transiently present at the electrode tip, an oxidation voltage step of 550 mV was applied for 10 s. Serotonin was puff-applied with a glass electrode for a short duration of 100 ms (arrow indicated) during the continuous oxidation voltage of 550 mV. (B) Summary data of the averaged calibration values of oxidation currents with serotonin concentrations up to 5 μM (n = 6). (C) Top and Middle traces show examples of oxidation currents following mechanical stimulation of a MC of a WT mouse and a MC of a 5-HT3A−/− mouse, respectively. Bottom trace shows no detectable oxidation currents following the same mechanical stimulation to a non-MC of a WT mouse. (D) Summary data of the peak concentrations of serotonin release from MCs of WT mice (n = 12) and 5-HT3A−/− mice (n = 10), and the lack of serotonin release from non-MCs of WT mice (n = 9). In this set of experiments, MCs were prelabeled by vital staining with quinacrine. Detections were performed from MCs underneath the Rs and from non-MCs in the region between the RRC and Rs. Data represent the mean ± SEM. NS, not significantly different, unpaired Student’s t test.
Finally, to determine whether the epidermal serotonergic transmission plays a role in behavioral responses to tactile stimuli, we measured behavioral responses elicited by gently touching a whisker hair (Fig. 6A). WT mice responded to the tactile stimuli with a high percent of avoidance reactions in control animals (Fig. 6B). Microinjection of the metabotropic 5-HT receptor antagonists KT + LY or the ionotropic 5-HT receptor antagonist Y25130 separately did not produce a measurable change in tactile behavioral responses (Fig. 6B). In contrast, microinjection of all three antagonists together resulted in the significant suppression of tactile behavioral responses (Fig. 6B). In 5-HT3A−/− mice, the metabotropic 5-HT receptor antagonists KT + LY could effectively impair tactile responses (Fig. 6C). These behavioral results are consistent with the electrophysiological recordings showing that substantial suppression of SA1 responses could be observed only when both the ionotropic and metabotropic 5-HT receptors were blocked (Fig. 4 G and H).
Ionotropic and metabotropic 5-HT receptors in whisker hair follicles are involved in behavioral whisker tactile responses. (A) Diagram illustrates tactile behavioral tests. Dashed arrow indicates the direction of tactile stimulation to move the whisker hair using a tactile stimulation filament. (B) Behavioral responses of WT mice to whisker hair tactile stimulation following hair follicle microinjection of saline (control), KT+LY, Y25130, or KT+LY+Y25130. (C) Behavioral responses of 5-HT3A−/− mice to whisker hair tactile stimulation following microinjection of saline or KT+LY, n = 6 for each experiment in B and C. The testing drugs were administered at 100 μM with a total volume of 1 µL. Data represent the mean ± SEM; *P < 0.05, Student’s t test.
In the present study we show that serotonin meets the criteria of being a transmitter conveying tactile signals from MCs to whisker afferent endings and establish Merkel discs as true sensory synapses uniquely located in the epidermis. This tactile synapse is distinct from the first sensory synapses that transmit other somatosensory signals, such as temperatures and pain in that the latter are located within the CNS and Glu is the principal transmitter (11, 12). The epidermal serotonergic synapses provide novel molecular insights into the tactile functions of the Merkel disc, an important tactile end organ that is crucial to complicated tactile tasks performed by whisker hair follicles of nonprimate mammals and fingertips of humans and other primates (2, 26).
We show that MCs in mouse whisker hair follicles are mechanical transducers responding to mechanical stimulation with MA currents. The MA currents in the present study have the electrophysiological properties identical to the Piezo2-mediated currents in MCs of rat whisker hair follicles (8) and of mouse touch domes in the skin (6, 7). We further show that mouse MCs situated in our whole-mount whisker hair follicle preparations are electrically excitable and fire slow APs in response to membrane depolarization. This interesting property is not shared by other types of cells in the skin and also is not observed in mouse MCs grown in culture (6, 7). Importantly, we have found that the mechanotransduction in MCs triggers serotonin release. This finding provides key evidence indicating that Merkel discs are sensory synapses using a chemical messenger to transmit tactile signals. This idea of serotonergic synapses for tactile transmission is further supported by our findings of the presence of tryptophan hydroxylase 1 for serotonin synthesis and serotonin immunoreactivity in MCs. The kinetics and other properties of mechanical stimulation-triggered serotonin release events shown in our study strongly suggest that MCs use a synaptic mechanism of vesicular exocytosis to release serotonin. This idea is consistent with the previous observation of the presence of dense core vesicles in MCs (3). The idea of synaptic serotonin release from MCs is further strengthened by our findings that the release is Ca2+-dependent and that disruption of synaptic release machinery with BoNTA significantly attenuates SA1 responses.
In our experiments detecting serotonin release from MCs, a large amplitude event of serotonin release usually occurs immediately after mechanical stimulation, and it is often followed by multiple small release events that continually occur for some period. The first large release event may represent synchronized exocytosis of multiple vesicles caused by large Ca2+ entry via both Piezo2 channels and voltage-gated Ca2+ channels. Multiple vesicular release has been often observed in the synapses formed between hair cells and their afferent endings (27). For those small release events, the event sizes are in a very narrow range of 0.4 μM to 0.6 μM, which may represent quantal releases from individual vesicles. The detected serotonin molecules by our amperometry approach are most likely those that leaked out from synaptic clefts into the bulk bath solution surrounding MCs. Therefore, the measured serotonin concentrations may be substantially lower than actual serotonin concentrations in synaptic clefts of Merkel discs. The synchronized release events at initial stimulation and the multiple small release events over time would generate strong initial actions and prolonged weaker actions over time on whisker afferent endings, respectively. This may be one of the mechanisms underlying the characteristics of SA1 impulses of Merkel discs in response to tactile stimulation.
Serotonin released from MCs acts as a transmitter rather than as an autocrine or a paracrine because it directly elicits impulses in whisker Aβ-afferent endings. We demonstrate that the excitatory action of serotonin is mediated by both ionotropic 5-HT receptors (5-HT3) and metabotropic 5-HT receptors (5-HT2A/5-HT2B) expressed at whisker afferent endings innervating Merkel discs in whisker hair follicles. Consistently, we show that focal application of serotonin and the agonists to 5-HT3 and 5-HT2A/5-HT2B receptors to whisker afferent endings directly elicits whisker afferent impulses that mimic tactile-evoked SA1 responses in a manner of fast onset and prolonged excitation. The involvement of 5-HT3 and 5-HT2A/5-HT2B receptors in the tactile synaptic transmission at Merkel discs is further evidenced by our finding that genetic deletion and pharmacological block of these receptors significantly impair SA1 responses. In our pharmacological tests, we have validated the specificity of 5-HT receptor antagonists at low concentrations used in the present study. However, we have found that high concentrations of KT (500 µM) and MDL-72222 (100 and 500 µM), which were used in previous studies (15, 17), produced strong nonspecific effects on SA1 responses. This finding indicates that the previous studies with these two antagonists (15, 17) are not valid evidence of serotonergic transmission at Merkel discs. We have demonstrated that the ionotropic and metabotropic 5-HT receptors at whisker afferent endings exhibit the synergism in initiating electrophysiological tactile responses (SA1 impulses). The synergistic effects drive the majority of whisker afferent SA1 impulses, as is evidenced by the large impairment of SA1 impulses after blocking these 5-HT receptors pharmacologically and genetically. In fact, SA1 impulses in static phase appear to be entirely mediated by the serotonergic transmission because these impulses are completely abolished after blocking these 5-HT receptors pharmacologically and genetically. Future studies using mice whose 5-HT3 and 5-HT2A/5-HT2B receptors are genetically deleted together would help to confirm this point or to reveal other possibilities, such as the involvement of other 5-HT receptor subtypes or other receptor families in mediating SA1 responses.
We have noticed that several initial impulses in the dynamic phase (period during displacement step ramp) of SA1 responses are not significantly suppressed after blocking these 5-HT receptors. These initial impulses may be a result of the direct mechanotransduction at whisker afferent endings because Piezo2 channels may be expressed on these Aβ-afferent endings at Merkel discs (6, 7). However, serotonergic transmission should also significantly contribute to the generation of SA1 impulses in dynamic phase because SA1 impulses in dynamic phase are significantly suppressed when 5-HT3 and 5-HT2A/5-HT2B receptors are inhibited. Thus, the serotonergic transmission with the synergism of ionotropic and metabotropic 5-HT receptors appears to contribute to most, if not all, SA1 impulses in mouse whisker hair follicles under our experimental conditions. However, chemical messengers including ATP, Glu, SP, and enkephalin may be also present in the vesicles of MCs (3, 5), raising a possibility that these chemical messengers may play some roles in tactile transmission at Merkel discs. The serotonergic transmission at Merkel discs shown by our in vitro experiments is consistent with tactile behavioral responses in the present study; our tactile behavioral tests demonstrate that blocking 5-HT3 and 5-HT2A/5-HT2B receptors together compromises tactile behavioral responses. Because tactile stimuli to whisker afferents may be partially transduced by Aβ-afferent endings in Merkel discs and also by other afferent endings in whisker hair follicles (e.g., lanceolate endings), it is not unexpected that some degree of tactile behavioral responses will remain even when the serotonergic transmission at Merkel discs is completely blocked.
Serotoninergic synapses for tactile transmission at Merkel discs provide a novel mechanism of somatosensory processing. The serotonergic transmission in the epidermis, probably like that in the CNS, can be regulated by factors affecting serotonin uptake and release. This raises an interesting issue as to whether serotonin uptake inhibitors, such as cocaine, methamphetamine, and other recreational drugs in this category, may act at the epidermal serotonergic synapses to alter tactile sensations. It would also be also interesting to know whether the epidermal serotonergic transmission may be altered under pathological conditions in patients with diabetes, tissue inflammation, and undergoing chemotherapy, because tactile dysfunctions including mechanical allodynia and reduced tactile sensitivity are commonly observed in these patients.
Additional experimental details are provided in SI Materials and Methods.
Wild-type C57BL/6 mice, 5-HT3AeGFP, 5-HT3A−/− mice, and GCaMP3 mice were used in the present study. Animal care and use conformed to NIH guidelines for care and use of experimental animals. Experimental protocols were approved by the Institutional Animal Care and Use Committee at the University of Alabama at Birmingham. Details are provided in SI Materials and Methods.
Whisker hair follicle preparations and whisker afferent fiber recordings were performed using our previously described method (8). To record whisker afferent impulses elicited by whisker deflections (SA1 responses) or induced by focal applications of candidate chemical messengers, compound action potentials conducted on whisker afferent fibers were recorded using a suction electrode. Details are provided in SI Materials and Methods.
Whisker afferent neurons were retrogradely labeled by microinjection of DiI or Fluoro-Gold (FG) solution into hair follicles. Seven days after the microinjection, TG were obtained for whole-cell patch-clamp recordings from the labeled neurons. To determine whole-cell currents evoked by serotonin or other 5-HT receptor agonists, neurons were held at −73 mV, testing compounds were bath-applied for 1 min, and currents were recorded for 10 min. To determine membrane excitability, step current pulses from −100 to +900 pA (20 pA per step, 250-ms duration) were injected into cells through patch-clamp electrodes. Details are provided in SI Materials and Methods.
Whisker hair follicles were dissected out from whisker pad and capsule of each hair follicle removed. Whisker hair follicles (without capsules) were then affixed in a recording chamber. After removing Rs and the glassy membranes, MCs were stained with quinacrine and patch-clamp recordings were made from quinacrine-stained cells. Membrane and AP properties of MCs were determined under the whole-cell current clamp mode. Voltage-evoked and MA currents were recorded with MCs voltage-clamped at –75 mV. Details are provided in SI Materials and Methods.
Hair deflection was used as a tactile stimulus to elicit whisker afferent SA1 impulses. Unless otherwise indicated, hair deflection was induced by a 38-µm forward step to push the hair follicle for the duration of 2.62 s. MC mechanical sensitivity was tested using a method described in our previous study (8). Details are provided in SI Materials and Methods.
The testing compounds include serotonin, ATP, Glu, NE, Ach, His, SP, VIP, Enk, CGRP, CCK-8, eltoprazine, TCB-2, BW, m-CPP, SR57227, cisapride, 5-carboxamidotryptamine maleate (5-CT), EMD 386088 hydrochloride, AS-19, KT, LY266097, Y25130, MDL-7222, and BoNTA. Details are provided in SI Materials and Methods.
DiI-labeled whisker afferent neurons or quinacrine-stained MCs were individually aspirated into micropipettes, and the tips of the aspiration micropipettes were broken into microcentrifuge tubes to release individual cells into lysis buffer. After denaturation (1.5 min at 65 °C), RT-PCR was performed using a multiplex strategy with external primers and internal primers. Details are provided in SI Materials and Methods.
For immunostaining of whisker afferent neurons, the following antibodies were used: rat anti-5HT2A (28), rat anti-5HT2B (29), and rat anti-5HT3A (30). For staining on afferent endings in whisker hair follicles, whisker hair follicles of 5-HT3AeGFP mice were used. The following antibodies were used: chicken anti-GFP and rat antiserotonin. Details are provided in SI Materials and Methods.
Serotonin release from whisker hair follicles in responses to whisker hair displacements was detected using ELISA kit based on the company’s instruction. Serotonin release from individual MCs were detected by amperometry (31) using a carbon fiber electrode. To detect serotonin release from a MC in response to mechanical stimulation, a +550-mV oxidation potential was applied to the carbon fiber electrode and mechanical stimulation was applied to the MC by directly touching it using the same carbon fiber electrode. Details are provided in SI Materials and Methods.
The whisker tactile test was performed in a blinded manner in that one examiner conducted drug treatments and animal grouping (e.g., WT group vs. 5-HT3A−/− group), and another examiner who did not know the grouping performed behavioral whisker tests on these animals. To test whisker tactile responses, a single whisker hair (D1 whisker) was displaced up to 2 mm in a caudal-rostral direction by the tactile stimulation filament, and the whisker tactile test was performed 20 times with an interval of 1 min between trials. A positive behavioral whisker tactile response was considered when the testing animal exhibited an avoidance reaction to the tactile stimulation. Details are provided in SI Materials and Methods.
Data are presented as mean ± SEM. Statistical significance was evaluated using Student’s t test, one-way or two-way ANOVA with Bonferroni post hoc tests for multiple groups, *P < 0.05, **P < 0.01, and ***P < 0.001. Details are provided in SI Materials and Methods.
Wild-type C57BL/6 mice were obtained from Harlan Laboratories. The 5-HT3AeGFP and 5-HT3A−/− mice were transferred from the Rocky Mountain Taste and Smell Center at the University of Colorado Medical School, which were originally obtained from the GENSAT project [STOCK Tg(Htr3a-EGFP) DH30Gsat/Mmnc; RRID:IMSR_MMRRC:000273]. The 5-HT3A−/− mice were originally obtained from Jackson Laboratory (B6.129 × 1-Htr3atm1Jul/J; RRID:IMSR_JAX:005251). GCaMP3 mice were obtained from the laboratory of Xinzhong Dong, The Johns Hopkins University, Baltimore, MD. Mice at the age of 4–8 wk were used for whisker afferent fiber recordings and patch-clamp recordings from whisker afferent neurons. Mice aged 10–20 d were used for patch-clamp recordings and amperometric recordings from MCs. Animal care and use conformed to NIH guidelines for care and use of experimental animals. Experimental protocols were approved by the Institutional Animal Care and Use Committee at the University of Alabama at Birmingham.
Whisker hair follicle preparations and whisker afferent fiber recordings were performed using our previously described method (8). In brief, whisker hair follicles with attached afferent fiber bundles were dissected out and anchored in a recording chamber. The whisker hair follicles were submerged and perfused in oxygenated Krebs solution that contained: 117 mM NaCl, 3.5 mM KCl, 2.5 mM CaCl2, 1.2 mM MgCl2, 1.2 mM NaH2PO4, 25 mM NaHCO3, and 11 mM glucose bubbled with 95% O2 and 5% CO2, had pH of 7.3, and osmolarity of 325 mOsm, and was maintained at 24 °C. Unless otherwise indicated, the end of the follicle capsule was cut open to facilitate drug diffusion into the MC region of the whisker hair follicle. To record whisker afferent impulses elicited by whisker deflections (SA1 responses) or induced by focal applications of candidate chemical messengers, compound APs conducted on whisker afferent fibers were recorded using a suction electrode, signals were amplified using a Multiclamp 700A amplifier, and sampled 10 KHz with low-pass filter set at 1 KHz.
Mice were anesthetized by isoflurane and anesthesia maintained by continuous administration of isoflurane via a nose cone. Whisker hairs were lifted by forceps to expose the top part of the hair follicle. A sharp glass microelectrode with an ∼5-µm tip was filled with 0.25% DiI or 4% (wt/vol) FG solution and then attached to a 25-µm microsyringe. The microsyringe was mounted on the injection holder of a stereotaxic apparatus. The glass microelectrode was inserted vertically along the whisker hair into the top part of the hair follicle. DiI or FG solution at 1 µL was then injected into the hair follicle. Five to 10 hair follicles were injected in each mouse.
Seven days after the microinjection, animals were killed and TG were dissected out. Under a dissection microscope, connective tissues on the surface of TG were removed carefully using a pair of forceps. The TG were then affixed in a recording chamber with a tissue anchor and perfused in normal Krebs solution saturated with 95% O2 and 5% CO2 and at 24 °C. The recording chamber was mounted on the stage of an Olympus IX50 microscope equipped with IR-DIC and fluorescent imaging systems. The ganglions were exposed to 0.05% dispase II plus 0.05% collagenase in Krebs solution for 5–10 min, and the enzymes were then washed off with Krebs solution. DiI-labeled cells were identified under the fluorescent microscope.
Whole-cell patch-clamp recordings were performed on randomly selected DiI-labeled neurons. Electrodes for patch-clamp recordings were fabricated from thin-wall capillaries. The resistance of recording electrodes was ∼6 MΩ when filled with internal solution containing: 105 mM K-Gluconate, 35 mM KCl, 2.4 mM MgCl2, 0.5 mM CaCl2, 5 mM EGTA, 10.0 mM Hepes, 5.0 mM Na2ATP, 0.33 mM GTP-Tris salt; pH 7.35 (adjusted with NaOH) and 320 mOsm (adjusted with sucrose). Signals were amplified using an Axopatch 200B amplifier (Axon Instruments) with a low-pass filter set at 2 kHz and digitized at 10 kHz. Junction potential between bath and electrode solution of −13 mV was corrected for in the data analysis.
After gaining whole-cell access and in voltage-clamp mode, membrane capacitance and input resistance were measured at the beginning of the recording using a seal test. To determine whole-cell currents evoked by serotonin or other 5-HT receptor agonists, neurons were held at −73 mV (voltage command at −60 mV), testing compounds were bath-applied for 1 min, and currents were recorded for 10 min. To determine membrane excitability, step current pulses from −100 to +900 pA (20 pA per step, 250-ms duration) were injected into cells through patch-clamp electrodes. Neuronal excitability was characterized by AP threshold, rheobase, and number of APs evoked in response to depolarizing current injection. Other passive and active electrophysiological properties were assessed including cell capacitance, input resistance, AP amplitude, and width.
Mice were anesthetized with isoflurane and killed by decapitation. Whisker hair follicles were dissected out from whisker pad and capsule of each hair follicle removed under a dissection microscope. Whisker hair follicles (without capsules) were then affixed in a recording chamber with a tissue anchor and perfused with oxygenated Krebs solutions at 24 °C. The recording chamber was mounted on the stage of an Olympus IX50 microscope that was equipped with IR-DIC and fluorescent imaging systems. The whisker hair follicles were exposed to 0.05% dispase II and 0.01% collagenase in Krebs solution for 8∼11 min, then the enzymes were washed off with Krebs solution. Rs and the glassy membranes were removed using a glass electrode controlled by a micromanipulator. The whisker hair follicles were then incubated with 0.3 µM quinacrine in Krebs solution for 15 min to stain MCs. The whisker hair follicles were continuously perfused with oxygenated Krebs solution at a flow rate of 1.5 mL/min. Quinacrine-labeled MCs were identified using the fluorescent imaging system.
Patch-clamp recordings were made at room temperature from quinacrine-stained cells. Recording electrodes were filled with an internal solution containing: 135 mM K-gluconate, 5 mM KCl, 0.5 mM CaCl2, 2 mM MgCl2, 5 mM EGTA, 5 mM Hepes, 5 mM Na2ATP, and 0.5 mM GTP-Tris salt; the pH of the solution was adjusted to 7.3 with KOH. Signals were amplified and filtered at 2 kHz using the Multiclamp 700A amplifier and sampled at 4 kHz using pCLAMP 10 software (Molecular Devices). Membrane and AP properties of MCs were determined under the whole-cell current clamp mode. Step current pulses were injected into cells through patch-clamp electrodes from −60 pA to 220 pA in increments of 20 pA with a pulse duration of 200 ms. Voltage-evoked and MA currents were recorded with MCs voltage-clamped at −75 mV. For voltage-evoked currents, voltage steps were applied from −135 mV to 65 mV in increments of 20 mV and duration of 100 ms.
Hair deflection was used as a tactile stimulus to elicit whisker afferent SA1 impulses. In the present study, we anchored the whisker hair follicles in a recording chamber by affixing whisker hair shaft onto the bottom of the recording chamber and perfused them with Krebs solution. A fire-polished blunted glass probe was used for delivering mechanical stimuli. The probe was attached onto the capsule surface at the whisker hair follicle center and controlled by a piezo device. When the mechanical probe displaced the whisker hair follicle, it generated a whisker hair shaft deflection. This modified tactile stimulation method improved consistency of whisker SA1 responses. Unless otherwise indicated, hair deflection was induced by a 38-µm forward step to push the hair follicle for the duration of 2.62 s; the step had a 56-ms ramp at the speed of 0.68 µm/ms (dynamic phase) before reaching the 38-µm step (static phase).
MC mechanical sensitivity was tested using a method described in our previous study (8). In brief, a fire-polished blunted glass probe was used for the mechanical stimulation. This probe was connected to a computer-programmable piezo device (E-625 LVPZT; Physik Instrumente). The tip of the glass probe was ∼3 μm in diameter. It was positioned at an angle of 30° to the surface (the outer root sheath layer) of the hair follicle section. The distance from the probe tip to the surface of the hair follicle tissue was set in such a way that the tip would contact the surface if the probe had one step (0.5 μm) forward movement. The stepwise forward movement of the probe was delivered by the piezo device. MCs were displaced indirectly by the probe (indirect displacement stimulation), which was achieved by displacing nonrecorded cells so that mechanical force was transmitted across two adjacent cells (∼15 µm) to the recorded MCs.
The capsule of whisker hair follicle was cut to open a small hole at the Rs segment of the whisker hair follicle. Under the visual guidance with a 40× objective, a puff electrode with a testing compound was inserted through the opening into the Merkel disc region underneath the follicle Rs. Candidate chemical messengers and agonists were focally puff-applied to the MC region with duration of 400 ms and puff pressure of 5 mm Hg, and their actions to elicit whisker afferent impulses were determined by whisker afferent fiber recordings. The testing compounds include serotonin, ATP, Glu, NE, Ach, His, SP, VIP, Enk, CGRP, CCK-8, eltoprazine, TCB-2, BW, m-CPP, SR57227, cisapride, 5-CT, EMD, and AS-19. In a different set of experiments, some of these compounds were tested for their ability to directly excite neurons in trigeminal ganglion neurons of GCaMP3 mice using the Ca2+ imaging technique.
To test effects of 5-HT receptor agonists (SR57227, TCB-2, BW) and antagonists (KT, LY) on retrogradely labeled whisker afferent neurons, testing compounds were bath-applied to the recorded cells. Antagonists were always preapplied for 2 min and then coapplied with serotonin. To test the effects of receptor agonists and antagonists on SA1 responses, the following compounds were tested: KT (5-HT2A antagonist), LY (5-HT2B antagonist), Y25130, and MDL-7222 (5-HT3 antagonists). Each antagonist was preapplied for 30 min and also applied during displacement stimulation. To test the effect of BoNTA on SA1 responses, 0.5 nM BoNTA was bath-applied to whisker hair follicle preparations and SA1 responses were measured every 20 min for up to 3 h. To test the effects of serotonin and 5-HT receptor antagonists on electrophysiological and mechanical transduction properties of MCs, MC membrane properties and mechanically activated currents were tested before and following the bath application of serotonin (100 μM) or 5-HT receptor antagonists.
To isolate individual whisker afferent neurons for RT-PCR of serotonin receptors, whisker afferent neurons were retrogradely labeled by microinjection of DiI into whisker hair follicles as described above. Ipsilateral TG were dissected out 7 d after DiI injection, placed in the patch-clamp recording chamber, and incubated in a solution containing 0.5% dispase and 0.5% collagenase for 10 min. DiI-labeled cells were then individually aspirated into micropipettes, and the tips of the aspiration micropipettes were broken into microcentrifuge tubes to release individual TG neurons into lysis buffer. To isolate individual MCs for RT-PCR of tryptophan hydroxylase, the capsule of each whisker hair follicle was first removed and MCs were labeled by quinacrine. After incubation in a bath solution containing 0.5% dispase and 0.5% collagenase for 10 min, quinacrine-labeled MCs were individually aspirated into micropipettes and released into microcentrifuge tube that contained lysis buffer. After denaturation (1.5 min at 65 °C), reverse transcription was performed using oligo (dT)12–18 primer and 50 U SuperScript II (Invitrogen) under the following conditions: 10 min at room temperature, 20 min at 37 °C, and 10 min at 65 °C. cDNA was stored at −80 °C until multiplex PCR was performed. Primers were designed using the National Center for Biotechnology Information Primer-BLAST tool and each sequence was queried using two databases (Mouse G+T) optimized to test for highly similar sequences to determine specificity. The specificity of all primers (Integrated DNA Technologies) was confirmed using gel electrophoresis. Cells processed without reverse-transcriptase and cell-free bath aspirates were included as negative controls.
RT-PCR was performed using a multiplex strategy in 25-µL reactions containing 1× GoTaq reaction buffer (Promega), 0.4 µM external primer mix, 0.2 mM dNTPs, and 1 U GoTaq DNA polymerase (Promega) using an iCycler thermal cycler (Bio-Rad). Primer pairs for targeted genes are listed below.
External primers for 5-HT2A: F: TGCTCGCCGATGACAACTT, R: ATGGACTGCATCGTCCTTCG, expected size: 262 bp; Internal primers for 5-HT2A, F: CACTTCAGAAAGAAGCCACCTTG, R: AGGATAATTTGGCCCGAGTGC, expected size 60 bp.
External primers for 5-HT2B: F: TCGGGGGTGAATCCTCTGAT, R: AACTTCGGAGCCTCATTGGG, expected size: 235 bp; Internal primers for 5-HT2B, F: ATTTCGGGAAGCATTTGGCAG, R: TTTACTGACTTTGTGGCTCGGT expected size 60 bp.
External primers for 5-HT3A: F: CCCCCTCATTGGTGTCTATTT, R: GGTTTCCCATGGCTGGAAGA, expected size: 263 bp; Internal primers for 5-HT3A, F: GACCATCTTCATTGTGCGGC, R: TGCCTCAGCCAGTCAGGTA, expected size 69 bp.
External primers for Tph1: F: GAACTCCTAGGCCACGTTCC, R: AGGGCTTGACTTTGGCATGT, expected size: 244 bp; Internal primers for Tph1, F: GGGCTGTGCAAACAAGATGG, R: GGCATGTCCAGAAAGTGCAT, expected size 93 bp.
External primers for Tph2: F: CAATCGAGTTCGGCCTTTGC, R: CTGCGTGTAGGGGTTGAAGT, expected size: 275 bp; Internal primers for Tph2, F: TGCAAGCAAGAGGGTCAACT, R: CTGCAAGCAGGTCGTCTTTG, expected size 126 bp.
External primer pairs for 5-HT2A, 5-HT2B, and 5-HT3A, or for Tph1 (tryptophan hydroxylase 1) and Tph2 (tryptophan hydroxylase 2), were combined in the first reaction with 2 µL of first-strand cDNA and processed using the following conditions: 10 min at 95 °C, 26 cycles of 30 s at 94 °C, 30 s at 55 °C, 30 s at 72 °C, then 10 min at 72 °C. Products were diluted 1:10 and used as template in a second reaction in which each primer pair was run individually using the same thermal cycler conditions described above for 35 cycles. Transcript expression for 5-HT2A, 5-HT2B, and 5-HT3A receptors and for Tph1 and Tph2 transcript were evaluated by gel electrophoresis on a 2% (wt/vol) agarose gel and imaged using a Gel Logic imager and 1D 3.6 imaging software (Kodac).
For immunostaining of whisker afferent neurons, mice were killed by decapitation under isoflurane anesthesia and TGs were dissected out. TGs were fixed in 4% (wt/vol) paraformaldehyde (PFA) in PBS at 4 °C overnight and then cryoproteced in 30% (wt/vol) sucrose in PBS at 4 °C for 2 d. The tissues were embedded in Tissue Freezing Medium (TFM, Triangle Biomedical Sciences) and rapidly frozen with powdered dry ice. TGs were sectioned at 20 µm by a cryostat. The sections were then treated with 50%, 70%, and 50% (vol/vol) ethanol in distilled water 10 min each. After washing three times in PBS for 10 min, they were blocked with PBS containing 10% (vol/vol) normal goat serum (NGS, Jackson ImmunoResearch) for 30 min at room temperature, then incubated with primary antibody with PBS containing 5% (vol/vol) NGS at 4 °C overnight. The sections were washed three times in PBS, incubated with secondary antibody 1:500 diluted in PBS containing 5% (vol/vol) NGS at room temperature for 1 h, washed three times in PBS, then mounted with mounting medium. The following antibodies were used: rat anti-5HT2A (ASR-033, Alomone Labs; 1:500) (28), rat anti-5HT2B (ASR-035, Alomone Labs; 1:500) (29), and rat anti-5HT3A (ASR-031, Alomone Labs; 1:500) (30). Alexa 594-conjugated goat anti-rabbit secondary antibody (A-11076, Invitrogen; 1:1,000) was used to visualize the signals.
For staining on afferent endings in whisker hair follicles, 5-HT3AeGFP mice were deeply anesthetized and perfused with 4% (wt/vol) PFA. Whisker hair follicles were dissected out from whisker pads and postfixed in the fixative at 4 °C overnight. The tissues were cryoprotected in 20% (wt/vol) sucrose in PBS, embedded, and cut as described above. The sections were washed three times in PBS for 10 min, blocked with PBS containing 2% (vol/vol) NGS, 1% BSA, and 0.3% Triton X-100 for 1 h at room temperature, then incubated with primary antibody diluted in blocking solution at 4 °C overnight. The next day, sections were washed three times in PBS and incubated with secondary antibody diluted in blocking solution. The following antibodies were used: chicken anti-GFP (GFP-1010, Aves Lab; 1:2,000) and rat antiserotonin (MAB352, Millipore; 1:100). Alexa 488-conjugated goat anti-chicken secondary antibody (ab150169, Abcam; 1:500) and Alexa 594-conjugated goat anti-rabbit secondary antibody were used.
Ten whisker hair follicles were dissected out from mice and anchored in a recording chamber that contained 120 µL Krebs solution. The end of each capsule was cut open to facilitate serotonin diffusion out. Whisker hairs were deflected by displacing hair follicle (38 μm, 1 Hz) using a piezo-controlled probe. After the mechanical stimulation, the Krebs solution was collected and centrifuged at 1,000 × g for 15 min. Serotonin in the supernatant (100 µl) was detected using ELISA kit based on the company’s instructions (Enzo Life Sciences) with optical density being measured at 405 nm by a Titertrek ELISA plate reader (Flow Laboratories).
Amperometric detection of serotonin release (31) was performed by using a 5-µm carbon fiber electrode (ALA Scientific Instruments). Signals were amplified by a NPI VA-10× amplifier (ALA Scientific Instruments), sampled at 10 KHz with a low-pass filter set at 2 KHz with Digidata 1440A interface and pClamp 9.2 software (Molecular Devices). Carbon fiber electrodes were cut freshly before each experiment and calibrated with known concentrations of serotonin in each use. To detect serotonin release from a MC, the tip of the carbon fiber electrode was positioned near the cell and at an angle of ∼30° to the surface. Unless otherwise stated, a +550-mV oxidation potential was applied to the carbon fiber electrode for 5 s. During the oxidation step, mechanical stimulation was applied to the MC by directly touching it using the same carbon fiber electrode. The movement of the carbon fiber electrode was controlled by a manipulator that was computer-programmed to deliver a 10-µm displacement at 1 mm/s. Basal currents during the application of oxidation potential (DC shift) was obtained by placing carbon fiber electrode in a position away from MCs. Oxidation currents in each experiment were adjusted by subtracting the DC shift. The same procedures were performed in non-MCs. To determine Ca2+-dependent serotonin release, the experiments were performed in normal Krebs solution and in a low Ca2+ (20 µM) Krebs solution.
The whisker tactile test was performed in a blinded manner in that one examiner conducted drug treatments and animal grouping (e.g., WT group vs. 5-HT3A−/− group), and another examiner who did not know the grouping performed behavioral whisker tests on these animals. To begin the tests, mice aged 3 mo were placed in a cage and habituated for 10 min. During habituation and subsequent experiments, the testing room only had a red light on so that animals could not see the examiner and the tactile stimulation filament. After the habituation, a single whisker hair (D1 whisker) was displaced up to 2 mm in a caudal-rostral direction by the tactile stimulation filament, and the whisker tactile test was performed 20 times with an interval of 1 min between trials. A positive behavioral whisker tactile response was considered when the testing animal exhibited an avoidance reaction to the tactile stimulation. Effects of KT (100 μM), LY (100 μM), and Y25130 (100 μM) on the behavioral tactile responses were tested by microinjecting these drug mixtures (1 µL) into the D1 hair follicles. The drugs were microinjected into D1 whisker hair follicle with the animals under anesthesia by isoflurane. After the microinjection, isoflurane was immediately discontinued to allow the animals to recover from anesthesia for 45 min. The whisker tactile tests were then performed to examine the effects of microinjected drugs.
Data were analyzed using Clampfit software. Data are presented as mean ± SEM. Statistical significance was evaluated using Student’s t test; one-way or two-way ANOVA with Bonferroni post hoc tests for multiple groups, *P < 0.05, **P < 0.01, and ***P < 0.001.
We thank Drs. A. MacDermott, S. Kinnamon, and A. Randich for comments on an earlier version of this manuscript; Drs. S. Kinnamon and E. Larson for transferring 5-HT3A−/− and the 5-HT3AeGFP mice to us from the Rocky Mountain Taste and Smell Center, a National Institute on Deafness and other Communication Disorders-funded P30 core facility (Grant P30DC004657 to Diego Restrepo) at the University of Colorado; and Dr. Xinzhong Dong (Johns Hopkins University) for providing us with GCaMP3 mice. This work was supported by NIH/National Institute of Dental and Craniofacial Research Grants DE018661 and DE023090 (to J.G.G.).
Author contributions: J.G.G. designed research; W.C., H.K., R.I., J.L., and J.J.D. performed research; W.C., H.K., and J.G.G. analyzed data; and J.G.G. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1610176113/-/DCSupplemental.
