Abstract
Somatosensory neurons mediate responses to diverse mechanical stimuli, from innocuous touch to noxious pain. While recent studies have identified distinct populations of A mechanonociceptors (AMs) that are required for mechanical pain, the molecular underpinnings of mechanonociception remain unknown. Here, we show that the bioactive lipid sphingosine 1-phosphate (S1P) and S1P Receptor 3 (S1PR3) are critical regulators of acute mechanonociception. Genetic or pharmacological ablation of S1PR3, or blockade of S1P production, significantly impaired the behavioral response to noxious mechanical stimuli, with no effect on responses to innocuous touch or thermal stimuli. These effects are mediated by fast-conducting Aδ mechanonociceptors, which displayed a significant decrease in mechanosensitivity in S1PR3 mutant mice. We show that S1PR3 signaling tunes mechanonociceptor excitability via modulation of KCNQ2/3 channels. Our findings define a new role for S1PR3 in regulating neuronal excitability and establish the importance of S1P/S1PR3 signaling in the setting of mechanical pain thresholds.
Introduction
Pain is a complex sensation. It serves to protect organisms from harmful stimuli, but can also become chronic and debilitating following tissue injury and disease. Distinct cells and molecules detect noxious thermal and mechanical stimuli. Thermal pain is detected by thermosensitive TRP channels in subsets of nociceptors1–4, and gentle touch is detected by Piezo2 channels in low-threshold mechanoreceptors (LTMRs)5–7. Aδ high-threshold mechanoreceptors (HTMRs) have been shown to play a key role in responses to painful mechanical stimuli8–10.
Recent studies have shown that there are at least two populations of HTMRs that mediate responses to noxious mechanical stimuli. The Npy2r+ subpopulation of HTMRs mediates fast paw withdrawal responses to pinprick stimulation and terminate as free nerve endings in the epidermis11. The Calca+ subpopulation of circumferential-HTMRs respond to noxious force and hair pulling, and terminate as circumferential endings wrapped around guard hair follicles12. Additionally, somatostatin-expressing interneurons of laminae I-III in the dorsal horn of the spinal cord receive input from nociceptors and are required for behavioral responses to painful mechanical stimuli13. Despite these advances in defining the cells and circuits of mechanical pain, little is known about the molecular signaling pathways in mechanonociceptors.
Here, we show that the bioactive signaling lipid sphingosine 1-phosphate (S1P) is required for mechanical pain. Mice lacking the S1P receptor S1PR3 display striking and selective deficits in behavioral responses to noxious mechanical stimuli. Likewise, peripheral blockade of S1PR3 signaling or S1P production impairs mechanical sensitivity. We found that S1P constitutively enhances the excitability of A mechanonociceptors (AMs) via closure of KCNQ potassium channels to tune mechanical pain sensitivity. The effects of S1P are completely dependent on S1PR3. Our findings uncover an essential role for S1P signaling in noxious mechanosensation.
Results
To identify candidate genes underlying mechanosensation, we previously performed transcriptome analysis of the sensory ganglia innervating the ultra-sensitive tactile organ (the star) of the star-nosed mole14. Immunostaining revealed the tactile organ is preferentially innervated by myelinated Aδ fibers14. While our original analysis focused on ion channels enriched in the neurons of the star organ, our dataset also revealed enrichment of several components of the S1P pathway, including S1PR3. Likewise, single-cell RNA seq of mouse DRG neurons revealed expression of S1pr3 in neurofilament H-expressing myelinated mechanoreceptors15, which includes Aβ as well as the Aδ neuronal populations. Thus, we set out to define the role of S1P signaling and S1PR3 in somatosensory mechanoreceptors.
S1PR3 mediates acute mechanical pain
We first examined a variety of somatosensory behaviors in mice lacking S1PR316 (Sipr3tm1Rlp/Mmnc; referred to herein as S1pr3−/−). We initially investigated baseline responses to mechanical stimuli. S1pr3−/− mice displayed a dramatic loss of mechanical sensitivity (Fig. 1A), as von Frey paw withdrawal thresholds were significantly elevated in S1pr3−/− mice relative to S1pr3+/+ and S1pr3+/− littermates (mean thresholds: 1.737 g vs. 0.736 and 0.610 g, respectively). Moreover, S1pr3−/− mice demonstrated decreased responses to a range of noxious tactile stimuli (2–6 g; Fig. 1B) and to noxious pinprick stimulation (Fig. 1C), but normal responsiveness to innocuous tactile stimuli (0.6–1.4g; Fig. 1B). S1pr3−/− mice exhibited normal tape removal attempts6 (Fig. 1D), righting reflexes (Fig. S1A), radiant heat withdrawal latencies (Fig. S1B), and itch-evoked scratching17 (Fig. S1C). These results demonstrate a selective role for S1PR3 in acute mechanical pain.
As a complement to our analysis of somatosensation in S1pr3−/− animals, we employed a pharmacological approach, using the S1PR3-selective antagonist TY 52156 (TY). Similar to the phenotype of knockout animals, intradermal injection of 500 μM TY into the mouse hindpaw (the site of testing) triggered a rapid and significant elevation in von Frey paw withdrawal thresholds (Fig. 1E) and decreased responsiveness to noxious (2–6 g), but not innocuous (0.6–1.4g), tactile stimuli (Fig. 1F). In contrast, blockade of S1PR1 with the selective antagonist W146 had no effect on mechanical thresholds (Fig. 1E). Overall, these data show that S1PR3 signaling sets mechanical pain sensitivity.
Endogenous S1P mediates acute mechanical pain
We next asked whether peripheral S1P was required for the S1PR3-dependent effects on mechanosensation. We decreased S1P levels via injection of the sphingosine kinase inhibitor SKI II to block local production of S1P18 or elevated S1P levels via intradermal injection of S1P and measured behaviors 30 minutes after injection. Decreasing local S1P levels significantly reduced mechanical sensitivity (Fig. 2A), comparable to the hyposensitivity phenotype observed in S1pr3−/− mice (Fig. 1A). Again, similar to what was observed in S1pr3−/− animals (Fig. S1B), peripheral blockade of S1PR3 or S1P production had no effect on thermal sensitivity (Fig. S1D). Surprisingly, injecting exogenous S1P (10 μM; maximum solubility in saline vehicle) had no effect on mechanical sensitivity (Fig. 2A-B). However, as previously reported19, S1P injection triggered thermal hypersensitivity (data not shown) demonstrating that the lack of effect on mechanical hypersensitivity is not due to problems with S1P delivery or degradation. Although quantification of native S1P levels in skin is inaccurate owing to avid lyase activity20, our data establish that baseline S1P levels are sufficient to maximally exert their effect on S1PR3-dependent mechanical pain, such that increased levels of S1P do not evoke mechanical hypersensitivity, but diminished S1P leads to mechanical hyposensitivity. These data show that constitutive S1PR3 signaling is required for normal mechanosensitivity.
S1PR3 is expressed in A mechanonociceptors
We next aimed to characterize the somatosensory neuron subtypes expressing S1pr3 as a window onto the role of S1PR3 in acute mechanical pain. We used in situ hybridization (ISH) with a specific S1pr3 probe to examine expression patterns of S1pr3. In our experiments, 43% of cells from wild-type dorsal root ganglia (DRG) expressed S1pr3 (Fig. 3A; top, Fig. 3E). We observed no reactivity in DRG isolated from S1pr3−/− mice (Fig. 3A; bottom), and no significant differences in the distribution of cell sizes between S1pr3−/− and S1pr3+/+ ganglia (p = 0.18; Kolmogorov-Smirnov type II test), suggesting no loss of major neuronal subtypes. While a previous study using antibody staining concluded that S1PR3 protein was expressed by all DRG neurons21, we found that anti-S1PR3 antibodies showed broad immunoreactivity in DRG from mice lacking S1PR3 (data not shown).
Co-ISH revealed that a population of S1pr3+ neurons represents Aδ mechanonociceptors (AMs). These cells were medium-diameter (25.9 ± 4.5 μm) and expressed Scn1a (39.9% of all S1pr3+; Fig. 3B–C), a gene that encodes the Nav1.1 sodium channel which mediates mechanical pain in Aδ fibers10, as well as Npy2r (20.4% of all S1pr3+; Fig. 3B–C), a marker of a specialized subset of mechanonociceptive A fibers7. S1pr3 was expressed in 70.6% of Scn1a+ cells and 72% of Npy2r+ cells, comprising a majority of both these populations. Interestingly, a substantial subset of cells co-expressed S1pr3 and the mechanically sensitive channel Piezo2 (Fig. 3C), which is expressed in Aβ, Aδ, and C fibers6. The remaining S1pr3+ cells were small diameter (23.4 ± 8.2 μm), Trpv1+ C nociceptors (27.3% of total cells). To elucidate the role of S1PR3 in mechanical pain, we focused our analysis on Aδ mechanonociceptors, as C-fibers do not contribute to baseline mechanical thresholds22.
We also used an S1pr3mCherry/+ reporter mouse23 as an independent strategy to explore S1pr3 expression and localization. Analysis of this line detected expression of S1PR3-mCherry fusion protein in 48.3% of cultured DRG neurons (Fig. 3E), mirroring the results from ISH. Additionally, we observed S1PR3 expression in nerve fibers that innervate the skin using antibodies against mCherry in whole-mount immunohistochemistry (Fig. 3D). We observed overlap of S1PR3-expressing free nerve endings with NefH+ myelinated A fibers, but did not observe expression of S1PR3 in circumferential hair follicle receptors (Fig. 3D). These results suggest that S1PR3 is not expressed in the circ-HTMR subpopulation of mechanonociceptors identified by Ghitani et al. In addition to our mRNA and protein localization experiments, calcium imaging revealed that a subset of S1PR3+ neurons were activated by the Nav1.1-selective toxin Hm1a10 (Fig. 3E). We conclude that a subset of S1pr3+ neurons represents AMs that terminate as free nerve endings.
S1PR3 is required for nociceptive responses of high-threshold AM nociceptors
Given that the behavioral deficit of S1pr3−/− animals was restricted to noxious mechanosensation, we utilized ex vivo skin-nerve recordings to analyze the effects of genetic ablation of S1PR3 on AM afferents, which mediate fast mechanical pain sensation (Fig. 4A). We hypothesized that S1PR3 may play a role in AM afferent function. Strikingly, the median von Frey threshold to elicit firing in AM nociceptors was significantly higher in S1pr3−/− animals (3.92 mN) compared to littermate controls (1.56 mN; Fig. 4C). Additionally, S1pr3−/− AM nociceptors displayed reduced sensitivity in their force-response relation (slope for +/− versus −/−: 50 Hz/N versus 35 Hz/N), as well as attenuated firing over the noxious, but not innocuous, range of mechanical stimulation (Fig. 4D). Furthermore, S1pr3−/− AM nociceptors displayed a right-shifted cumulative response curve (50% effective force for +/− versus −/−: 33.7 versus 60.0 mN; Fig. 4E), consistent with the mechanonociceptive hyposensitivity observed in vivo. A recent report suggested that A-nociceptors are composed of two distinct neuronal populations that differ genetically, in mechanically evoked response patterns, and in conduction velocity11. Accordingly, we found that a proportion of AM nociceptors, characterized by adapting responses to static mechanical stimuli (Adapting AM), were selectively lost in S1pr3−/− animals (Fig. 4F–G). However, we found no differences in conduction velocity between genotypes (Fig. 4B). We conclude that S1PR3 is an essential regulator of both mechanical threshold and sensitivity in a distinct population of AM nociceptors.
S1PR3 modulates KCNQ2/3 channels to regulate AM excitability
To define the molecular pathway(s) by which S1PR3 modulates mechanonociceptor function, we next examined S1P signaling in cultured DRG neurons. Unlike the environment of sensory neurons in intact skin, the physiological Ringer’s solutions used in vitro lack S1 P. Thus for our in vitro experiments we tested the effects of exogenous S1P on neuronal excitability. We interrogated the molecular mechanism by which S1P signaling regulates mechanical pain using current-clamp recording of medium-diameter S1pr3mCherry/+ dissociated DRG neurons (membrane capacitance = 61.05 ± 1.92 pF), which ISH identified as putative mechanonociceptors (Fig. 3B–C). In these cells, 1 μM S1P application did not elicit firing in the absence of current injection (Fig. 5A, Fig. S2A). However, S1P dramatically lowered the threshold to fire action potentials (rheobase) in an S1PR3-dependent manner (Fig. 5A, Fig. S2B).
We then set out to determine how S1PR3 activity increases neuronal excitability using wholecell voltage clamp recording. While we found that S1P application had no effect on instantaneous sodium currents or steady state potassium currents (Fig. S2C–D), S1P significantly reduced slow, voltage-dependent tail current amplitudes (Fig. 5B; Fig. 5D (top)) in an S1PR3-dependent manner (Fig. 5B, center).
As tail currents in Aδ neurons are primarily mediated by KCNQ2/3 potassium channels24, we postulated that S1P might alter tail currents through modulation of these channels. To address whether KCNQ2/3 channels mediated S1P-dependent neuronal excitability, we examined several other aspects of the S1P-sensitive current. S1P triggered a robust increase in input resistance (Fig. 5E), consistent with the closure of potassium channels. Likewise, I-V analysis revealed that the current inhibited by S1P application was carried by potassium (Fig. 5F). Additionally, the lack of an effect of S1P on resting membrane potential (Fig. S2E) and steady state potassium current (Fig. S2D) were consistent with the electrophysiological properties of KCNQ2/3 channels in DRG neurons24–27. Finally, application of the KCNQ2/3-selective inhibitor linopirdine completely occluded the effects of S1P on tail current (Fig. 5C, Fig. 5d (bottom)).
These findings are consistent with S1P/S1PR3-dependent inhibition of KCNQ2/3 in somatosensory neurons.
We also found that the effect of S1P on KCNQ2/3 currents was mediated by low levels of S1P, exhibiting an IC50 of 48 nM with saturation at 100 nM (Fig. S2F). While S1P cannot be accurately measured in non-plasma tissues, this is similar to estimated levels of S1P in peripheral tissues28,29. Thus, our in vitro IC50 supports our behavioral observations that baseline S1P levels are sufficient to maximally exert their effect on mechanical pain. In summary, our electrophysiological and behavioral observations support a model in which baseline S1P/S1PR3 signaling governs mechanical pain thresholds through modulation of KCNQ2/3 channel activity in AM neurons (Fig. 6).
Discussion
This study identifies S1P/S1PR3 as a key pathway that tunes mechanical pain sensitivity. Recent studies have identified distinct populations of AM nociceptors that are required for mechanical pain11,12. Likewise, it was discovered that a subset of somatostatin-expressing spinal interneurons are required for mechanical pain transduction13. While these papers delineate the cells and circuitry of mechanical pain, the molecular underpinnings of mechanonociception in these neurons were unknown. We now demonstrate that S1PR3 is required for normal mechanosensitivity in a majority of mechanonociceptors, including the Npy2r+ population, recently identified by Arcourt et al., that innervates the epidermis and encodes noxious touch11.
While the transduction channels that detect noxious force remain enigmatic, we show that S1PR3 signaling modulates KCNQ2/3 channels to regulate excitability of mechanonociceptors (Fig. 6). GPCR-mediated inhibition of KCNQ2/3 potassium channels is a well-known mechanism by which neuronal excitability is regulated30. Other studies have shown that KCNQ channels mediate excitability of Aδ fibers24 and that opening KCNQ2/3 channels directly with retigabine alleviates pain in vivo25,31,32. Our results not only complement previous work implicating KCNQ2/3 channels in pain, but also define the upstream mechanisms that promote the regulation of KCNQ2/3 channels to tune mechanical pain thresholds. Our data thus highlight S1PR3 as a novel and attractive target for the treatment of mechanical pain and describe a new signaling pathway by which S1P regulates AM nociceptor excitability.
Interestingly, the neurons that innervate the ultra-sensitive tactile organ of the star-nosed mole are highly enriched in transcripts for S1PR3 and KCNQ channels, as well as for a variety of other potassium channels14. While it is difficult to directly examine the physiological basis for heightened mechanosensitivity in the star-nosed mole, S1PR3-dependent modulation of KCNQ may represent an important mechanism underlying the high tactile sensitivity of the star organ.
Outside of the nervous system, S1P signaling via S1PR1 allows for the continuous circulation of lymphocytes between blood, lymph, and peripheral tissues33. Our findings that S1P plays a key role in somatosensation is in line with recent studies showing that sensory neurons co-opt classical immune pathways to drive chronic itch or pain34,35. What distinguishes this study from the others is that S1P signaling is critical for acute mechanical pain, even in the absence of inflammation. In the immune system, disruptions in S1P levels or S1PR1 signaling results in significant immune dysfunction and disease36–38. Paralleling the role of S1PR1 in the immune system, we propose that S1PR3 might constitutively regulate the excitability of a variety of neurons and that aberrant S1P signaling may trigger nervous system dysfunction and disease. For example, S1P has been proposed to constitutively modulate synaptic transmission and excitability of hippocampal neurons in slice recordings via S1PR339–42, and S1PR3-deficient animals display learning and memory deficits39. However, these previous studies, unlike ours, did not identify a molecular mechanism by which S1P signaling alters neuronal activity. Abnormal S1P signaling is also linked to a host of neurological disorders, including thermal hypersensitivity19, multiple sclerosis43, Huntington44, Parkinson’s45, and Alzheimer’s disease46. In addition to expression in the somatosensory system, S1PR3 is expressed in brain regions relevant to these disorders (Allen Brain Atlas) and may thus represent an important mechanism by which S1P signaling regulates excitability.
The relationship between lipids and ion channel activity in the context of mechanotransduction is well-established and thought to occur through direct channel-lipid interactions47,48. Here, we highlight the importance an indirect pathway in which a lipid activates a GPCR to modulate mechanoreceptor activity. Our findings complement very recent work demonstrating the diverse and complex roles that GPCRs play in somatosensory signaling49. Our study demonstrates a crucial role for S1P signaling in the peripheral nervous system and highlights the potential of S1PR3 as a target for new therapies for mechanical pain.
Materials and Methods
Behavioral studies & mice
S1pr3mcherry/+ and S1pr3−/− mice were obtained from Jackson Laboratory and backcrossed to C57bl6/J. Wherever possible, wild-type/heterozygous (S1pr3) littermate controls were used in behavioral experiments. Mice (20–25 g) were housed in 12 h light-dark cycle at 21°C. Mice were singly housed one week prior to all behavioral experiments and were between 8–10 weeks at the time of the experiment. All mice were acclimated in behavioral chambers on 2 subsequent days for 1 hour prior to recording for itch behavior, von Frey, and radiant heat.
Itch and acute pain behavioral measurements were performed as previously described50,51. Mice were shaved one week prior to itch behavior. Compounds injected: 500 μM TY 52156 (Tocris), 50 μM SKI II (Tocris), 10 μM S1P (Tocris, Avanti Polar Lipids), 50 mM chloroquine (Sigma), and 27 mM histamine (Tocris) in PBS with either DMSO-or Methanol-PBS vehicle controls. Pruritogens were injected using the cheek model (20 μL) of itch, as previously described52. Behavioral scoring was performed while blind to experimental condition and mouse genotype. All scratching behavior videos were recorded for 1 hour and scored for the first 30 minutes. Bout number and length were recorded.
For radiant heat and von Frey hypersensitivity behavior, drugs were injected intradermally into the plantar surface of the hindpaw (20 μL). Radiant heat assays were performed using the IITC Life Science Hargreaves test system. Mechanical threshold was measured using calibrated von Frey monofilaments (Touch Test) on a metal grate platform. Von Frey was performed as previously described53,54 using the up-down method while blinded to compound injected and genotype. Valid responses for both von Frey and radiant heat included fast paw withdrawal, licking/biting/shaking of the affected paw, or flinching. For radiant heat and von Frey, mice were allowed to acclimate on platform for 1 hour before injection.
The pinprick assay was conducted on a von Frey testing platform (IITC). The mouse hindpaw was poked with a 31 g syringe needle without breaking the skin to induce fast acute mechanical pain. Each paw was stimulated 10 times with the needle, and the % withdrawal (fast withdrawal, licking/biting/shaking of paw, squeaking, and/or flinching) was calculated from the total number of trials.
The tape assay was conducted according to previously described methods6. Number of attempts to remove a 3 cm piece of lab tape was recorded for 10 minutes after manual tape application to the rostral back. Scorer and experimenter were blinded to genotype.
For righting reflex measurements, age-matched S1pr3−/− and +/+ P6-7 neonates were used. Briefly, pups were overturned one at a time on the home cage lid while experimenter was blinded to genotype. The time to righting was measured to the nearest 1/10th of a second with a stopwatch.
All behavior experiments were carried out using age-matched or littermate cohorts of male mice. Mice were tested in 4-part behavior chambers (IITC Life Sciences) with opaque dividers (TAP Plastics) with the exception of righting reflex measurements. Itch behavior was filmed from below using high-definition cameras. All experiments were performed under the policies and recommendations of the International Association for the Study of Pain and approved by the University of California, Berkeley Animal Care and Use Committee.
In situ hybridization (ISH)
Fresh DRG were dissected from 8–12 week old mice, flash frozen in OCT embedding medium, and sectioned at 14 μm onto slides. ISH was performed using Affymetrix Quantigene ViewISH Tissue 2-plex kit according to manufacturer’s instructions with Type 1 and Type 6 probes. The following probes against mouse mRNAs were created by Affymetrix and used for ISH: S1pr3, Scn1a, and Piezo2. Slides were mounted in Fluoromount with No. 1.5 coverglass. Imaging of ISH experiments and all other live- and fixed-cell imaging was performed on an Olympus IX71 microscope with a Lambda LS-xl light source (Sutter Instruments). Images were analyzed using ImageJ software. Briefly, DAPI-positive cells were circled and their fluorescence intensity (AFU) for all channels was plotted against cell size using Microsoft Excel software. Co-labeling analysis was performed using ImageJ. Intensity thresholds were set based on the negative control (no probe) slide. Cells were defined as co-expressing if their maximum intensities exceeded the threshold for both the Type 1 and Type 6 probe channels.
Immunohistochemistry (IHC) of DRG
8-12 week old mice were deeply anesthetized and transcardially perfused with ice-cold PBS followed by ice-cold 4% PFA. DRG were dissected and post-fixed in 4% PFA for one hour. DRG were cryo-protected overnight at 4°C in 30% sucrose-PBS, embedded in OCT, and then sectioned at 12 μm onto slides. Briefly, slides were washed 3x in PBST (0.3% Triton X-100), blocked in 2.5% horse serum + 2.5% BSA PBST, and incubated overnight at 4°C in 1:1000 primary antibody in PBST + 0.5% horse serum + 0.5% BSA. Slides were washed 3X in PBS then incubated 1-2 hours at RT in 1:1000 secondary antibody. Slides were washed 3X in PBS and mounted in Fluoromount with No. 1.5 coverglass. Primary antibodies used: Rabbit anti-S1PR3, Mouse anti-NF200, Chicken anti-Peripherin (Abcam). Secondary antibodies used: Goat anti-Mouse Alexa 350, Goat anti-Chicken Alexa 547, Goat anti-Rabbit Alexa 488 (Abcam). Cells labeled with anti-peripherin were circled to define regions of interest.
Whole mount skin IHC
Staining was performed according to Marshall and Clary et al55. Briefly, 8–12 week old mice were euthanized and the back skin was shaved, depilated, and tape-stripped. The removed skin was fixed overnight in 4% PFA, then washed in PBS (3X for 10 minutes each). Dermal fat was scraped away with a scalpel and skin was washed in PBST (0.3% Triton X-100; 3X for two hours each) then incubated in 1:500 primary antibody (Rabbit anti DsRed Polyclonal antibody; Clontech #632496) in blocking buffer (PBST with 5% goat serum and 20% DMSO) for 5.5 days at 4°C. Skin was washed as before and incubated in 1:500 secondary anti-body (Goat antiRabbit Alexa 594; Invitrogen #R37117) in blocking buffer for 3 days at 4°C. Skin was washed in PBST, serially dried in methanol: PBS solutions, incubated overnight in 100% methanol, and finally cleared with a 1:2 solution of benzyl alcohol: benzyl benzoate (BABB; Sigma) before mounting.
Cell culture
Cell culture was carried out as previously described17. Briefly, neurons from dorsal root ganglia (2–8 week old adults) or trigeminal ganglia (P0) were dissected and incubated for 10 min in 1.4 mg ml-1 Collagenase P (Roche) in Hanks calcium-free balanced salt solution, followed by incubation in 0.25% standard trypsin (vol/vol) STV versene-EDTA solution for 2 min with gentle agitation. Cells were then triturated, plated onto Poly D-Lysine coated glass coverslips and used within 20 h. Media: MEM Eagle’s with Earle’s BSS medium, supplemented with 10% horse serum (vol/vol), MEM vitamins, penicillin/streptomycin and L-glutamine.
Calcium imaging
Ca2+ imaging experiments were carried out as previously described17. Cells were loaded for 60 min at room temperature with 10 μM Fura-2AM supplemented with 0.01% Pluronic F-127 (wt/vol, Life Technologies) in a physiological Ringer’s solution containing (in mM) 140 NaCl, 5 KCl, 10 HEPES, 2 CaCl2, 2 MgCl2 and 10 D-(+)-glucose, pH 7.4. All chemicals were purchased from Sigma. Acquired images were displayed as the ratio of 340 nm/ 380 nm. Cells were identified as neurons by eliciting depolarization with high potassium Ringer’s solution (75 mM) at the end of each experiment. Responding neurons were defined as those having a > 15% increase from baseline ratio. Image analysis and statistics were performed using automated routines in Igor Pro (WaveMetrics). Fura-2 ratios were normalized to the baseline ratio F340/F380 = (Ratio)/(Ratio t = 0).
Ex vivo skin-nerve electrophysiology
Touch-evoked responses in the skin were recorded after dissecting the hind limb skin and saphenous nerve from 7–10 week old mice, according to published methods56,57. The skin was placed epidermis-side-up in a custom chamber and perfused with carbogen-buffered synthetic interstitial fluid (SIF) kept at 32 °C with a temperature controller (model TC-344B, Warner Instruments). The nerve was kept in mineral oil in a recording chamber, teased apart, and placed onto a gold recording electrode connected with a reference electrode to a differential amplifier (model 1800, A-M Systems). The extracellular signal was digitized using a PowerLab 8/35 board (AD Instruments) and recorded using LabChart software (AD Instruments).
For these studies, we focused on A-mechanonociceptors (AMs). To identify responses from these afferents in mutant and control genotypes, we used a mechanical search paradigm with a fine glass probe. Afferents were classified as AMs according to the following criteria: (1) Aδ conduction velocity (approximately, 1 to (≤ 12 m/s−1), (2) medium-sized receptive fields, (3) sustained response to mechanical indentation9,57,58.
Touch-sensitive afferents that did not meet these criteria were not analyzed further. Responses were classified as Adapting AMs if the ratio of mean firing rate in the dynamic phase of stimulation (first 0.2 s) to the static phase of stimulation (last 4.8 s) was greater than 2, and Non-Adapting AMs if the ratio was less than or equal to 2. Non-responders (Fig. 4g) responded to suprathreshold mechanical stimulation with von Frey monofilaments (tip diameter <0.5 mm), but not to maximal controlled mechanical stimulation (256 mN, tip diameter 2 mm). All recordings and analyses were performed blind to genotype.
Mechanical responses were elicited with von Frey monofilaments and a force controlled custom-built mechanical stimulator. Mechanical thresholds were defined as the lowest von Frey monofilament to reliable elicit at least on action potential. Force controlled mechanical stimuli were delivered using a computer controlled, closed-loop, mechanical stimulator (Model 300C-I, Aurora Scientific, 2 mm tip diameter). Low-pass filtered, 5-second long, length control steps (square wave) simultaneously delivered with permissive force control steps (square wave) were generated using LabChart software (AD Instruments). An arbitrarily selected force step-and-hold protocol (8, 32, 4, 64, 128, 16, 256 mN) was delivered to all fibers. The period between successive displacements was 60 seconds.
Conduction velocity was measured by electrically stimulating identified receptive fields. Spike sorting by principal component analysis (PCA) and density based clustering, and data analysis was performed off-line with custom-made software in MATLAB. Statistics were performed in Prism.
In vitro electrophysiology
Electrophysiological experiments were carried out as previously described17. Briefly, recordings were collected at 5 kHz and filtered at 2 kHz (Axopatch 200B, pClamp software). Electrode resistance ranged between 1.5–5 MΩ. Internal solution contained 140 mM KCl, 2 mM MgCl2, 1 mM EGTA, 5 mM HEPES, 1 mM Na2ATP, 100 μM GTP, and 100 μM cAMP (pH 7.4). Bath solution was physiological Ringer’s solution. The pipette potential was canceled before seal formation. Cell capacitance was canceled before whole cell voltage-clamp recordings. Experiments were carried out only on cells with a series resistance of less than 30 MΩ. Analysis of electrophysiology data was performed in pClamp and IgorPro.
Statistical analyses
All statistical analyses, except for skin nerve data (see above), were performed using IgorPro software or Microsoft Excel. Values are reported as the mean ± SEOM where multiple independent experiments are pooled and reported (for whole cell electrophysiology), and mean ± SD where one experiment was performed with multiple wells (for calcium imaging) or mice (for behavior). For comparison between two groups, Student’s unpaired 2-tailed t-test was used. A paired t-test was employed only for measurements within the same biological replicate and after a given treatment. For single-point comparison between >2 groups, a one-way ANOVA followed by Tukey Kramer post hoc test was used. For the time course comparison between 2 groups, 2-way ANOVA was used and single comparison p-values were derived using Tukey’s HSD. For comparing distributions, a type II Kolmogorov-Smirnov test was used. Number of mice or samples required to attain significance was not calculated beforehand, and where multiple statistical tests were performed, a Bonferroni correction was applied. In figure legends, significance was labeled as: n.s., not significant, p ≥ 0.05; *p < 0.05; **p < 0.01; ***p < 0.001.
Author Contributions
R.Z.H, D.M.B., and R.B. conceived experiments and wrote the manuscript. R.Z.H. performed behavioral, immunostaining, whole cell electrophysiology, calcium imaging, and ISH, and analyzed data and made figures. T.M. also performed ISH experiments. E. A. L. & B.H. conceived and analyzed ex vivo recordings. B.U.H. and S. C. performed ex vivo recordings. All authors contributed to the final version of the manuscript.
Competing interests
The authors declare no competing interests at this time.
Acknowledgements
We thank Z. Rifi (UC Berkeley) for assistance with scoring itch behavior, R. P. Dalton (UC Berkeley) for assistance with confocal microscopy, and P. Lishko and M. Miller (UC Berkeley) for advice regarding lipid stability and usage. We would also like to thank R. Clary (Columbia) for whole mount skin staining protocols and D. Julius (UC San Francisco) for the gift of Hm1a spider toxin. We are grateful to all members of the D.M.B Laboratory (UC Berkeley) for constructive feedback and criticism. The National Institutes of Health grants NS077224 (to D.M.B and R.B.), AR059385 (to D.M.B.), AR051219 (to E.A.L), NS105449 and GM007367 (to B.U.H), and NS063307 (to the Neurobiology Course at the Marine Biological Laboratory); and a Howard Hughes Medical Institute Faculty Scholars grant (to D.M.B) supported this work.