Peripheral sensory neurons and non-neuronal cells express functional Piezo1 channels

Abstract

Here, we present evidence showing Piezo1 protein expression in the primary sensory neurons (PSNs) and non-neuronal cells of rat peripheral nervous system. Using a knockdown/knockout validated antibody, we detected Piezo1 immunoreactivity (IR) in ∼60% of PSNs of rat dorsal root ganglia (DRG) with higher IR density in the small- and medium-sized neurons. Piezo1-IR was clearly identified in DRG perineuronal glia, including satellite glial cells (SGCs) and Schwann cells; in sciatic nerve Schwann cells surrounding the axons and cutaneous afferent endings; and in skin epidermal Merkel cells and melanocytes. Neuronal and non-neuronal Piezo1 channels were functional since various cells (dissociated PSNs and SGCs from DRGs, isolated Schwann cells, and primary human melanocytes) exhibited a robust response to Piezo1 agonist Yoda1 by an increase of intracellular Ca²⁺ concentration ([Ca²⁺]_i). These responses were abolished by non-specific Piezo1 antagonist GsMTx4. Immunoblots showed elevated Piezo1 protein in DRG proximal to peripheral nerve injury-induced painful neuropathy, while PSNs and SGCs from rats with neuropathic pain showed greater Yoda1-evoked elevation of [Ca²⁺]_i and an increased frequency of cells responding to Yoda1, compared to controls. Sciatic nerve application of GsMTx4 alleviated mechanical hypersensitivity induced by Yoda1. Overall, our data show that Piezo1 is widely expressed by the neuronal and non-neuronal cells in the peripheral sensory pathways and that painful nerve injury appeared associated with activation of Piezo1 in PSNs and peripheral glial cells.

Introduction

Piezo channels, including Piezo1 and Piezo2, are mechanosensitive channels (MSCs) that are expressed widely throughout the body of mammals.^1,2 These channels provide critical functions as constitutive PIEZO1 and 2 knockout mice die during embryogenesis and at birth, respectively.^3,4 In humans, mutations in the genes encoding PIEZO channels are not life-threatening but cause multiple hereditary human diseases. ⁵ Lack of painful responses to innocuous touch after skin inflammation is reported in human with the mutations of PIEZO2. ⁶ It is generally accepted that Piezo1 is a polymodal sensor of diverse mechanical forces primarily expressed in non-sensory tissues exposed to fluid pressure and flow, whereas Piezo2 is predominantly found in PSNs of sensory ganglia and cutaneous Merkel cells that respond to mechanical touch specifically. ⁷

Recent studies highlight the potential roles of Piezo1 in peripheral nociceptive mechanobiology. RNAscope identifies Piezo1 in DRG- and TG-PSNs of mice and moles, which is enriched in smaller-sized PSNs, and also detected in perineuronal glial cells of mice DRG.^8,9 Piezo1 is detected in rat and mouse DRG by immunoblots.^10,11 Piezo1 may contribute to mechanosensitive trigeminal and meningeal nociception related to trigeminal neuralgia pain and migraine. ¹² A new study reports that sensory neuron Piezo1 contribute to itch in mouse. ⁹ A search of the updated datasets in Human Protein Atlas (https://www.proteinatlas.org), the deposited data in GenBank GEO Profiles (https://www.ncbi.nlm.nih.gov/geoprofiles/), and published literature indicate presence of Piezo1 transcripts and protein in total DRG, trigeminal ganglia (TG), and isolated DRG/TG neurons and glial cells.^13–18 Application of Piezo1 agonist Yoda1 to dissociated PSNs and recording mechanically evoked currents both reveal functional Piezo1,^12,19 and intradermal Yoda1 injection causes allodynia and hyperalgesia in mice. ⁸ Piezo1 is also expressed in keratinocytes, where it contributes to cutaneous nociception by modulating sensory afferent firing.^20,21 Additionally, selective ablation of skin keratinocyte-Piezo1 in mice reduces responsiveness to non-noxious and noxious mechanical stimuli but spares responses to thermal stimuli. ²² Systemic administration of spider toxin GsMTx4, a nonspecific Piezo1 antagonist, relieves mechanical hypersensitivity in mice from both peripheral nerve injury and inflammation, ²³ while intra-articular injection of GsMTx4 suppresses elevated Piezo1 and relieves hypersensitivity in osteoarthritic animals. ²⁴ Although analgesic effects of GsMTx4 could be due to inhibition of multiple MSCs, ²⁵ the above observations together suggest a possibility that Piezo1 may contribute to peripheral nerve nociceptive mechanobiology.

Piezo1 protein expression and cell specificity in rat peripheral nerves have not been systematically studied. Here, we determined Piezo1 expression in rat peripheral sensory pathway. We examined Piezo1 protein expression in DRG, sciatic nerve, afferent peripheral and central terminals, spinal cord, and hindpaw skin. Functional Piezo1 was tested by Yoda1-stimulated increases of intracellular calcium ([Ca²⁺]_i). Piezo1 expression in DRG was characterized in rat models of peripheral nerve injury-induced painful neuropathy. Our data identify Piezo1 expression in the PSNs and non-neuronal glial cells. Peripheral PSNs and perineuronal glial cells express functional Piezo1 that is activated following neuropathic pain. These results suggest that Piezo1 may play roles in peripheral mechanosensation.

Materials and methods

Animals

Adult male and female Sprague Dawley (SD) rats (6–8 week old, Charles River Laboratories, Wilmington, MA) and male mice in C57BL/6 background with keratinocyte-selective Piezo1 null (k14 ^Cre /Piezo1 ^fl/fl ) and their wild type littermates at the age of ∼6 month-old ¹¹ were used. All animal experiments were performed with the approval of the Medical College of Wisconsin Institutional Animal Care and Use Committee in accordance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals. Animals were housed individually in a room maintained at constant temperature (22 ± 0.5°C) and relative humidity (60 ± 15%) with an alternating 12 h light-dark cycle, and were given access to water and food ad libitum throughout the experiment. All survival surgeries were completed in a sterile environment under a surgical microscope in animals anesthetized with isoflurane (2–5%), and all efforts were made to minimize suffering and the numbers of animal used. For tissue harvest euthanasia, animals were deeply anesthetized by isoflurane followed by decapitation with a well-maintained guillotine.

Rat pain models and sensory behavior testing

Peripheral nerve injury models

Two neuropathic pain rat models by spared nerve injury (SNI) and tibial nerve injury (TNI) were generated as we described previously. ¹¹ Sham-operated rats were subjected to all preceding procedures in the same manner as pain models but without nerve ligation and transection.

Mechanical allodynia and hyperalgesia

Behavioral tests were conducted between 9:00 AM and 12:00 AM, as we described previously. ²⁶ Experimenters were blinded to the treatment during all data acquisition. Sensory tests included eliciting reflexive behaviors induced at threshold intensity punctate mechanical stimulation (von Frey test), noxious mechanical stimulation (pin), cold stimulation (acetone), and heat stimulation (Hargreaves test), and were carried out as previously described. ²⁷

Tissue harvest for immunohistochemistry (IHC) and immunoblots

After transcardial perfusion with cold 100 mL PBS in rats and 40 mL in mouse, lumbar (L) 4 and 5 DRG lumbar spinal cord, and sciatic nerve segment proximal to the sciatic bifurcation were dissected, and fixed in Richard-Allan Scientific™ Buffered Zinc Formalin (ThermoFisher, Rockford, IL) overnight, followed by processing for paraffin embedment. The previously described histological protocol was adopted. ²⁶ For western blot experiments, L4/L5 DRG were removed, snap frozen in liquid nitrogen, and stored at −80°C for extraction of protein.

Primary cell culture and cell lines

DRG dissociated culture, neuron enrichment, and neuron-free glial cell isolation

²⁸ In brief, the L4/5 DRG rapidly harvested from the isoflurane-anesthetized rats (male) were incubated in 0.01% liberate blendzyme 2 (Roche Diagnostics, Madison, WI) for 30 min followed by incubation in 0.25% trypsin and 0.125% DNase for 30 min, both dissolved in Dulbecco’s modified Eagle’s medium/F12 (DMEM/F12) with glutaMAX (ThermoFisher). After exposure to 0.1% trypsin inhibitor and centrifugation, the pellet was gently triturated, and dissociated cells were plated onto poly-L-lysine (0.01 mg/mL) coverslips (Sigma-Aldrich, St Louis, MO) and maintained at 37°C in humidified 95% air and 5% CO₂ for 3 h in Neural basal media A (ThermoFisher) plus 0.5 μM glutamine, and were studied by Ca²⁺ imaging no later than 6 h after harvest. DRG neuron enrichment and neuron-free glial cell culture that is composed of satellite glial cells (SGC), Schwann cells (SCs), and resident microglia was established by a differential attachment protocol for DRG neuron enrichment and glial isolation, as described previously. ²⁸

Sciatic nerve (SN) Schwann cell isolation

The relevant steps required for nerve processing, enzymatic dissociation, and cell plating using the SN from adult male rat was previously described, with minor modifications.^28,29 In brief, bilateral sciatic nerves were harvested and attached adipose and muscular tissue stripped off using fine forceps. Subsequently, the desheathed inner nerve elements were collected for enzymatic digestion for 2 h as described in DRG dissociated culture. The end products of enzymatic digestion were filtered and subsequently collected by centrifugation. Isolated SCs were cultured in Schwann cell medium (Edina, MN) on poly-L-lysine (0.01 mg/mL) coated glass coverslips overnight before Ca²⁺ imaging or 4 days for immunoblot experiments.

Spinal cord dorsal horn (SDH) glia isolation

Spinal DH glial cell isolation was performed essentially the same as DRG dissociation. The rats (male, 1-month old) were killed by decapitation during anesthesia. The vertebral column was removed and cut into slices and collected in Petri dishes filled with cold DMEM/F12. After removing the vertebral arch, dorsal parts of the spinal cord were extracted, cleaned from surrounding spinal meninges, and collected in cold DMEM/F12. All SDH slices were transferred into an enzyme mixture for digestion, as DRG dissociation. After enzymatic digestion for 30 min, cells were dissociated mechanically, washed, and precipitated by centrifugation (2 min with 2000 rpm). The supernatant was removed, and cells were re-suspended in cultural medium consisting of Neurobasal A supplemented with 2% B27 and 10% FBS and cultured on poly-L-lysine (0.01 mg/mL) coated glass coverslips 24 h for Ca²⁺ imaging and 4 days for immunoblot experiments.

Cell lines

As a second source of Schwann cells, primary Schwann cells isolated from human spinal nerve were obtained from Neuromics (HMP303, Edina. Donor anonymous). Primary adult human epidermal melanocytes were purchased from ThermoFisher (C0245 C, female) and human U251 glioblastoma cells (astrocytes, male) from Sigma-Aldrich. Human SCs were cultured in Schwann cell medium (Edina), and human epidermal melanocytes cultured in Medium 254 with melanocyte growth supplement (ThermoFisher), according to manufacturer’s protocols. Neuronal NG108-15 cells (NG108), Neuro2A (N2A), F11 DRG neuronal-like cells, NSC34 spinal cord motor neuron-like cells, and C6 astrocytes were obtained from ATCC (Manassas, VA). N2A cells stably expressing CRISPR Cas9 nuclease (Cas9N2A) was obtained from Genecopoeia (Rockville, MD). Rat DRG-neuronal 50B11 cells (50B11) were used as reported previously. ³⁰ These cells were cultured by a standard protocol using Dulbecco’s modified Eagle’s medium (DMED) supplemented with 10% FBS and antibiotics (ThermoFisher), and were grown at 37°C and in 5% CO₂ in a humidified incubator.

Validation of Piezo1 antibody in CRISPR/Cas9-mediated Piezo1 knockout cells

Lentiviral expression plasmid pWPT-mCherry ³⁰ was used to express dual CRISPR guide RNAs (gRNAs) specific to mouse and rat Piezo1 (gRNA1: 5’-AGCATTGAAGCGTAACAGGG-3′, gRNA2: 5’-AGAGAGCATTGAAGCGTAAC-3′), designed by using online Cas9 crRNA Design Tool (Integrated DNA Technologies, Coralville, IA). In brief, DNA sequences encoding dual Piezo1 gRNA-scaffolds transcribed by U6 and H1 promoters, respectively, were synthesized (Genscript, Piscataway, NJ) and inserted into FseI site upstream of EF1α promoter to generate plasmid pWPT-mCherry-pPZ1gRNAs expressing dual gRNAs and mCherry (Supplement Figure 1). Lentivectors expressing mCherry (control) or dual Piezo1 gRNAs were packaged using pWPT-mCherry and pWPT-mCherry-PZ1gRNAs with packaging plasmid pCMVDR8.74 and envelop plasmid pVSV-g, and products titrated in the range of 1 × 10⁶ to 1 × 10⁷ transduction unit/ml, as previously reported. ³⁰ Cultured Cas9N2A cells grown to 50% confluence were infected by LV-mCherry-PZ1gRNAs or LV-mCherry (control) in the presence of 8 μg of polybrene (Sigma-Aldrich) per ml at an optimized multiplicity of infection 10. After infection at 37°C for 24 h, the medium was replaced. Piezo1 expression was analyzed 48h after transduction by immunoblots (see further).

Microfluorimetric Ca²⁺ imaging

Determination of [Ca²⁺]_i was performed using Fura-2 based microfluorimetry and imaging analysis as we previously described. ²⁸ Unless otherwise specified, the agents were obtained from Sigma-Aldrich. Dissociated DRG neurons, SGCs, sciatic nerve SCs, and DH glia from naïve male rats, as well as human melanocytes, NSC34 cells, 50B11 cells, and C6 astrocytes, were evaluated for the responses to Yoda1 at different concentrations (μM) in extracellular buffer to determine the dose-response relationship from at least three different cultures.

For measurement of [Ca²⁺]_i, cells cultured on coverslips were loaded with Fura-2-AM (5 μM, ThermoFisher) and maintained in 25°C Tyrode’s solution containing (in mM): NaCl 140, KCl 4, CaCl₂ 2, glucose 10, MgCl₂ 2, and 4-(2-hydroxyethyl)-1- piperazineethanesulfonic acid (HEPES) 10, with an osmolarity of 297–300 mOsm and pH 7.4. After 30 min, they were washed three times with regular Tyrode’s solution and left in a dark environment for de-esterification for 30 min and then mounted onto a 0.5 mL recording chamber that was constantly superfused by a gravity-driven bath flow at a rate of 3 mL/min. Agents were delivered by directed-microperfusion controlled by a computerized valve system through a 500 μm-diameter hollow quartz fiber 300 μm upstream from the chamber. This flow completely displaced the bath solution, and constant flow was maintained by delivery of bath solution when specific agents were not being administered. Solution changes were achieved within 200 ms. The fluorophore was excited alternately with 340 nm and 380 nm wavelength illumination (150 W xenon, Lambda DG-4; Sutter), and images were acquired at 510 nm using a cooled 12 bit digital camera (Coolsnap fx; Photometrics) and inverted microscope (Diaphot 200; Nikon Instruments) through a 20× or 40× Fluor oil-immersion objective. Cells were imaged to monitor Ca²⁺ responses to Yoda1 with different concentration and total DMSO concentration was kept equal to and below 1% for all tested Yoda1 concentrations. The [Ca²⁺]_i was evaluated as the ratio of emission in response to excitation at 340 and 380 nm, expressed as the 340/380 nm fluorescence emission ratio (R_340/380) that is directly correlated to the amount of intracellular calcium. ³¹ A ≥30% increase in R_340/380 from baseline after superfusion with Yoda1 was considered a positive response for all cells recorded. ²² Response of neurons to 50 mM KCl solution at the end of each protocol was used as a criterion for identifying viable neurons, and similarly for the response of glia to 10 μM ATP. Plasma membrane Ca²⁺-ATPase influence was eliminated by applying Tyrode’s solution with pH 8.8 during depolarization, while stable cytoplasmic [Ca²⁺]_c was maintained by simultaneously reducing bath Ca²⁺ concentration to 0.25 mM. ³²

Immunofluorescent staining

Previously established protocols were adopted. ¹¹ Sections were first immunolabeled with the selected primary antibodies in a humid atmosphere overnight at 4°C (Table 1). The fluorophore-conjugated (Alexa 488 or Alexa 594, 1:2000) secondary antibodies (Jackson ImmunoResearch, West Grove, PA) were used to reveal immune complexes. The immunostaining was examined, and images were captured using a Nikon TE2000-S fluorescence microscope (El Segundo, CA) with filters suitable for selectively detecting the green and red fluorescence using a QuantiFire digital camera (Optronics, Ontario, NY). NIH ImageJ software (http://rsbweb.nih.gov/ij/) was used for analysis.

Table 1.

Primary antibodies and IgG controls used in this study.

Antibody a	Host	Supplier/Cat# b	Dilution
IB4	—	LF/I21413	1.0 μg/mL (IHC)
Piezo1	Rabbit polyclonal	Alomone/APC087 Cat#ab53852, RRID: AB_881796	1:200 (IHC), 1:1000 (Wb)
Piezo1	Rabbit polyclonal	Proteintech/15939-1-AP	1:200 (ICC)
Piezo2	Rabbit polyclonal	Alomone/APC090	1:200 (IHC)
CGRP	Mouse monoclonal	SCB/sc57053	1:600 (IHC)
GFAP	Rabbit polyclonal monoclonal	Dako/Z0334	1:1000 (IHC)
GFAP (GA5)	Mouse monoclonal	CS/3655	1:400 (IHC)
Hmgcs1	Goat polyclonal	SCB/sc32422	1:400 (IHC)
NeuN	Mouse monoclonal	Millipore/MAB377	1:100 (IHC)
PKCγ	Mouse monoclonal	SCB/sc166385	1:200 (IHC)
Tubb3	Mouse monoclonal	SCB/sc80016	1:600 (IHC)
NKA1α	Mouse monoclonal	SCB/sc514614	1:1000 (IHC)
Cav3.2	Mouse monoclonal	SCB/sc136990	1:500 (IHC)
GS	Mouse monoclonal	SCB/sc74430	1:800 (IHC)
S100	Mouse monoclonal	TF/MA5-12969	1:1000 (IHC)
NF200	Mouse monoclonal	Sigma/N5389	1:1000 (IHC)
ACTA2	Mouse monoclonal	CS/56856	1:500 (IHC)
Syp	Mouse monoclonal	SCB/sc17750	1:200 (IHC)
Synpr	Mouse monoclonal	SCB/sc376761	1:200 (IHC)
CK14	Mouse monoclonal	SCB/sc53253	1:500 (IHC)
MBP	Goat polyclonal	SCB/sc13912	1:1000 (IHC)
GAP43	Mouse monoclonal	SCB/sc17790	1:200 (IHC)
Iba1	Rabbit polyclonal	Wako/019-19741	1:1000 (Wb)
GAPDH	Mouse monoclonal	Sigma/SAB1403850	1:5000 (Wb)
IgG control	Mouse	LF/31903	1:100∼400
IgG control	Rabbit	LF/MA5-16384	1:100∼1000

^aAntibody abbreviations: IB4, Isolectin IB4; Piezo1 and 2, Piezo type mechanosensitive ion channel component 1 and 2; CGRP, calcitonin gene-related peptide; GFAP, Glial fibrillary acidic protein; Hmgcs1, 3-hydroxy-3-methylglutaryl coenzyme A synthase-1; NeuN, Neuronal nuclear protein; PKCγ, Protein kinase C gamma; Tubb3, β3-Tubulin; NKA1α, sodium/potassium ATPase 1α; Cav3.2, T-type calcium channel 3.2α1H; GS, glutamine synthetase; S100, S100 calcium binding protein; NF200, neurofilament-200; ACTA2, α-Smooth muscle actin; Syp, synaptophysin; Synpr, synaptoporin; CK14, cytokeratin 14; MBP, myelin basic protein; P75NTR, neurotrophin receptor p75; Iba1, allograft inflammatory factor-1; GAPDH, glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase.

^bLF, Life Technologies, Carlsbad, CA; Alomone, Alomone Labs, Jerusalem, Israel; Proteintech, Rosemont, IL; SCB, Santa Cruz Biotechnology, Santa Cruz, CA; Millipore, Burlington, MA; CS, Cell signaling, Danvers, MA; Wako, Richmond, VA; Dako: Carpinteria, California; Sigma, Sigma-Aldrich, St Louis, MO.

Two different rabbit polyclonal Piezo1 antibodies were obtained from Alomone (Jerusalem, Israel) and from Proteintech (Rosemont, IL), respectively. The specificity of these two Piezo1 antibodies has been validated in knockout (KO) tissues, in cells with RNAi-mediated Piezo1 knockdown (KD), and they are widely used to detect Piezo1 expression by IHC and immunoblots.^33–38 The Alomone Piezo1 antibody was used for all IHC and immunoblots, while Proteintech Piezo1 antibody was used for ICC on human melanocytes since Alomone Piezo1 antibody does not detect human Piezo1 according to the datasheet from the vendor. To control for Piezo1 immunostaining, representative sections were processed in the same way as described above but using non-immune rabbit IgG instead of the Piezo1 primary antibody in the incubation. The specificities of the other antibodies used in this study have been previously confirmed, ¹¹ and the specificity of secondary antibodies was tested with omission of the primary antibodies, which has always resulted in absence of immunostaining. ¹¹

Quantification of immunostaining

Positive marker antibody immunostaining was defined as the cells having a fluorescence intensity greater than the average background fluorescence plus 2 standard deviations of the cells in a section of negative control (first antibody omitted) under identical acquisition parameters (n = 10 for different markers), identified by Hoechst counterstain at a different wavelength. ³⁹

The cross-sectional area and intensity (mean gray value) of Piezo1-labeled neurons for which nuclei were evident was measured using Adobe Photoshop CS6 (Adobe Systems Incorporated, San Jose, CA, USA). Neurons were divided into three size groups: small (<700 mM²), medium (700–1500 mM²), and (>1500 mM²) neurons as described previously.^28,30 Intensity correlation analysis (ICA) was performed to determine colocalization of Piezo1 with neuronal plasma membrane marker sodium/potassium ATPase 1 alpha (NKA1α) and SGC marker glial fibrillary acidic protein (GFAP), as previously described using an ImageJ 1.46r software plugin colocalization analysis module (http://imagej.nih.gov/ij).^28,40,41 Briefly, fluorescence intensity was quantified in the matched region of interest (the green and red colors varied in close synchrony) for each pair of images. Mean background was determined from areas outside the section regions and was subtracted from each file. On the basis of the algorithm, in an image where the intensities vary together, the product of the differences from the mean (PDM) will be positive. If the pixel intensities vary asynchronously (the channels are segregated), then most of the PDM will be negative. The intensity correlation quotient (ICQ) is based on the nonparametric sign-test analysis of the PDM values and is equal to the ratio of the number of positive PDM values to the total number of pixel values. The ICQ values are distributed between −0.5 and +0.5 by subtracting 0.5 from this ratio. In random staining, the ICQ approximates 0. In segregated staining, ICQ is less than 0; while for dependent staining, ICQ is greater than 0.

Immunoblot analysis of Piezo1 expression

Immunoblots were performed as described previously. ²⁶ Immunoreactive proteins were detected by Pierce enhanced chemiluminescence (ThermoFisher) on a ChemiDoc Imaging system (Bio-Rad) after incubation for 1 h with HRP-conjugated second antibodies (1:5000, Bio-Rad).

Piezo1 co-immunoprecipitation (Co-IP) and protein identification by LC-MS/MS

The DRG homogenates from naïve male rats were incubated on ice for 10 min followed by centrifugation at 12,000g for 20 min at 4°C to remove the insoluble fraction. The supernatant was transferred to a new tube, 50 μL of which was saved as input, and the rest of the supernatant was precleared by incubation with 1 mg in 1 × PBS pre-washed Dynabeads (Thermo Fisher Scientific) on a rotator at 4°C for 1h. To prepare the antibody-coupled beads, Piezo1 antibody (Alomone) was incubated together with washed Dynabeads (8 μg antibody per mg beads used) in 1x PBS buffer with 0.1 M citrate (pH 3.1) on a rotator at 4°C overnight. Controls were generated using washed beads not coupled to antibody but normal rabbit IgG. Following overnight incubation, the supernatant was collected and saved in a new tube. The Co-IP beads were washed three times in ice-cold 1xPBS buffer by gentle pipetting and directly boiled in 60 μL 1 × SDS sample buffer at 95°C for 10 min to elute protein complexes from the beads. Immunoprecipitated samples were separated by SDS-PAGE gel and stained with silver (ThermoFisher). The stained gel regions of interests were excised, and in-gel trypsin digested as previously described. ⁴² Extracted tryptic peptides were analyzed by nano reversed-phase liquid chromatography tandem mass spectrometry (nLC-MS/MS) using a nanoACQUITY (Waters Corporation, Milford, MA, USA) online coupled with an Orbitrap Velos Pro hybrid ion trap mass spectrometer (ThermoFisher). The precursor ions were selected automatically by the instrument. MS/MS spectra from the Piezo1 Co-IP experiment were searched and processed using the rat proteome database.

Yoda1 and GsMTx4 sciatic nerve subepineural injection

Injection was performed in a blinded manner in which the operator was unaware of the content of the injectate. Yoda1 was dissolved in DMSO and GsMTx4 dissolved in saline, both to a 40 mM stock solution, stored at −20°C, diluted to desired concentrations with saline before use. It is reported that half of the total cross-section inside the sciatic nerve epineurium in human consists of non-neural connective tissue. ⁴³ Naïve male and female rats were randomized into three groups for each gender: saline, Yoda1, and Yoda1 plus GsMTx4 (0.05% DMSO) (both containing 0.05% DMSO); and sciatic nerve injection was performed as previously described.^44,45 Briefly, after appropriate anesthesia was obtained by inhalation of 2% isoflurane, the right sciatic nerves were exposed through a lateral incision of the middle thighs and division of the superficial fascia and muscle. Then 100 μL of the test dose of reagent was injected slowly, directly into sciatic nerve subepineural space (beneath the clear fascia surrounding the nerve but outside the perineurium) through a 33-gauge blunt nanofil needle (World Precision Instruments, Sarasota, FL, USA), proximal to the sciatic bifurcation. The needle remained in place at the injection site for 1 additional min, before it was slowly removed. The superficial muscle layer was sutured with 4-0 silk, and the wound was closed with metal clips. Ipsilateral hindpaw mechanical allodynia (vF) and hyperalgesia (Pin) were tested before and 15, 30, 45, 60, 120, and 180 min after the animal fully recovered from anesthesia, indicated by being fully alert with balance and gait, which typically required 3–5 minutes following termination of the anesthetic administration.

Statistics

Statistical analysis was performed with GraphPad PRISM 9 (GraphPad Software, San Diego, CA). The estimated numbers of animals needed for behavior were derived from our previous experience with similar experiments and the statistical analyses were done afterward without interim data analysis. ¹¹ The numbers of biological replicates (e.g. animals, cells, and immunoblot samples) were provided in the corresponding figures and legends. No data points were excluded. Significances of ICQs of Piezo1 immunocolocalization with NKA1α and GFAP were analyzed by means of the normal approximation of the nonparametric Wilcoxon rank test, as described previously.^40,41 The differences of the targeted gene expression by immunoblots and calcium imaging analyses were compared with one-way ANOVA, unpaired two-tailed t-test, or Mann-Whitney U test, where appropriate. Fisher’s exact test was used to compare the percentage. Mechanical allodynia (vF) and hyperalgesia (Pin) changes after sciatic nerve injection were compared to pre-injection baseline (BL) in naïve rats, with repeated measures two-way ANOVA and Tukey post-hoc for vF and Friedman’s tests and Dunn post hoc for Pin test. The vF and Pin area under the curves (AUC) in naïve rats after sciatic nerve injection were compared among groups by one-way ANOVA and Tukey post-hoc. Results are reported as mean and standard deviation of mean (SEM). Differences were considered to be significant for values at p < 0.05.

Results

Specificity of the Piezo1 antibody

We first determined the specificity of Piezo1 antibody by a CRISPR-Cas9 genome editing technology using N2A cells stably expression Cas9. Results showed that antibodies (from both vendors) recognized a clean ∼300 kDa band predicted as the canonical Piezo1 protein by immunoblot with comparable band density in control Cas9N2A cells and Cas9N2A cells expressing mCherry, while lentivector-mediated Piezo1-gRNA expression in Cas9N2A cells induced completely ablation of Piezo1 protein expression (Figure 1(a)). Piezo1 antibody recognized ∼300 kDa canonical Piezo1 band and additional ∼220 kDa and ∼70 kDa protein bands upon immunoblot (IB) of DRG lysates from naïve rats (Figure 1(b)). Two additional experiments were performed to further validate the specificity. First, Piezo1 Co-IP followed by IB was performed using rat DRG lysates, which showed that the ∼300 kDa band but not ∼220 kDa and ∼70 kDa protein bands was clearly detected in Piezo1 Co-IP samples (Figure 1(c)). Secondly, nLC-MS/MS spectra of ions matching 3 Piezo1 peptides ¹⁸⁸⁸GAAVVEAEHEEGEEGR¹⁰⁹³ (Xcorr2.44), ⁸⁹⁵GPVDPANWFGVR⁹⁰⁶ (Xcorr2.39), ²³⁷⁷QLQPDEEEDYLGVR²³⁹⁰ (Xcorr2.25) were detected in confidence, confirming Piezo1 identification on the excised band ranging 150∼320 kDa from the silver-stained 1D SDS-PAGE gel of Co-IP sample (Figure 1(d)). Representative MS/MS spectrum of ion of ¹⁸⁸⁸GAAVVEAEHEEGEEGR¹⁰⁹³ and fragments matched were shown (Figure 1(e) and (f)). IHC and neuronal size plotted vs. Piezo1 immunostaining density on DRG and TG sections from naïve rats (Figure 1(g) and (h)) revealed a similar PSN profile of Piezo1 expression, showing higher immunoreactive (IR) density in small- and medium-sized PSNs. No significant second antibody (donkey anti-rabbit Alexa-594 conjugated) staining was evident when both Piezo1 antibody was replaced by rabbit IgG (1:200) for IHC on DRG and spinal cord sections (Supplement Figure 2). Additionally, the keratinocyte Piezo1-IR profile was identified on IHC of hindpaw skin prepared from wild-type mice (WT), but the Piezo1 immunopositivity was barely detectable in the hindpaw skin prepared from the keratinocyte Piezo1 KO (K14 ^Cre+ /Piezo1 ^fl/f ) mice. The skin Piezo2 staining density in K14 ^Cre+ /Piezo1 ^fl/f mice was comparable to WT mice, indicating no effect of Piezo1 deletion in skin keratinocytes on Piezo2 expression (Figures 1(i)-(k)). These data provided further evidence for Piezo1 antibody used in this study to detect Piezo1 expression by IHC and immunoblot.

Figure 1.

Specificity of Piezo1 (PZ1) antibody. Immunoblot (IB) detects canonical PZ1 protein band at the mass size of ∼300 kDa with comparable densities in the lysates of Cas9N2A cells and mCherry-expressing Cas9N2A cells, but PZ1 is completely ablated in Cas9N2A cells transduced by LV-mediated PZ1-Crispr gRNAs (A). Canonical PZ1 is detected upon IB of DRG lysates, with additional bands at ∼200 and ∼70 kDa (B) which are detected in the reduced total lysate proteins loaded on the gel (C, left size) but only canonical PZ1 is clearly detected in PZ1 Co-IP samples (C, right size, asterisk denotes IgG heavy chain). Co-IP sample was size-separated, silver stained (D, asterisk denotes IgG heavy chain), and gel pieces ranging ∼320-150 kDa excised for mass spectrometry. Representative MS/MS spectrum of an ion from rat Piezo1 peptide ¹⁸⁸⁸GAAVVEAEHEEGEEGR¹⁰⁹³ (E, red) and m/z fragments matched (F, red). Representative IHC images show PZ1-IR (red) in the sections of DRG (G) and TG (H) and PZ1-expressing PSN cross-section area plotted versus intensity (right panels of G and H) from adult naïve rats. Dashed line indicates cutoff levels of background signals. IHC montage images of colabeling of PZ1 or PZ2 (red) with CK14 (green) on hindpaw skin sections from wide-type (WT) mice (I) and Piezo1-ko mice (J, K), as indicated. Epi and de denote epidermis and dermis. Scale bars: 50 μm for G, H; 25 μm for I-K. Gas9N2A, N2A cell stably expressing Gas9; IP-IB, immunoprecipitation-immunoblot; DRG, dorsal root ganglia; IHC, immunohistochemistry; CK14, cytokeratin.

Figure 2.

DRG Piezo1 (PZ1): Double immunostaining (male). Representative IHC montage images of DRG sections show PZ1-IR (red) co-stained with PSN markers (green), including Tubb3 (A), CGRP (B), IB4 (C), Cav3.2 (D), NF200 (E), and NKA1α (F, F1). The panels in the right sides of A-E calculate the percentage of PZ1-neurons overlaid to Tubb3-neurons (A1), neurons positive for CGRP/PZ1 (B1) and IB4/PZ1 (C1) overlaid to PZ1-neurons, and PZ1-neurons overlaid to Cav3.2 (D1) and NF200 (E1) neurons. The numbers are the counted PZ1-IR neurons (red) and marker-labeled neurons (green). ICA analyzes colocalization between PZ1 and NKA1α and scatter plots for the region demarcated by the white dashed line in panel F1 show data clustered along both positive and negative axes for both Piezo1 and NKA1α (G). “A” is the intensity of Piezo1 while “a” is the average of these values, and “B” is the intensity of NKA1α while “b” is the average of these values. For this region, the ICQ value is 0.059 (P_{sign test}<0.001). (H-L) Show PZ1 (red) with a selection of glial cell markers (green), including GS (H), Hmgcs1 (I) with the squared region shown at high magnification (I1), S100 (J) with the squared region shown at high magnification (J1), and GFAP (K, K1). White arrowheads in panel I1, J1, point to the colabeled glial cells. ICA analysis for colocalization between PZ1 and GFAP for the region demarcated by the white dashed line in panel K1 shows scattered plot data clustered along both positive and negative axes for both PZ1 and GFAP (L). “A” is the intensity of PZ1 while “a” is the average of these values, and “B” is the intensity of GFAP while “b” is the average of these values. For this region, the intensity correlation quotient (ICQ) value is 0.122 (P_{sign test}<0.001). (M, N) Show rat DRG neuron enrichment culture at DIV 0.25 (M) and isolated DRG neuron-free glial cell culture at DIV4 (N). (O) Canonical PZ1 protein band (∼300 kDa) was clearly detected upon immunoblot in the lysates prepared from rat DRG tissue, enriched rat DRG neurons, purified DRG glia, and 50B11 and F11 DRG neuronal-like cells. Scale bars: 50 μm for all. DRG, dorsal root ganglia, IHC, immunohistochemistry; PSN, primary sensory neuron; Tubb3, β3-tubulin; GS, glutamine synthetase; Hmgcs1, 3-hydroxy-3-methylglutaryl coenzyme A synthase-1; ICA, Intensity correlation analysis.

Neuronal and glial Piezo1 expression in DRG

We next characterized the Piezo1 expression within DRG sections by double immunostaining of Piezo1 with various established and distinct neuronal markers including Tubb3 (pan neurons), nonpeptidergic IB4-biotin (nonpeptidergic small neurons), peptidergic CGRP (peptidergic neurons), NKA1α (neuronal PM), NF200 (large Aβ low-threshold mechanoreceptors, i.e. Aβ-LTMRs and proprioceptors), and Cav3.2 (Aδ- and C-LTMRs). ⁴⁶ Results revealed Piezo1-IR in ∼60% of the Tubb3-positive PSNs (Figure 2(a), A1) with higher IR density in smaller- and medium-sized PSNs in male rat DRG, a finding that is consistent with the results by RNAscope in mice. ⁸ Averaged 62% and 68% of Piezo1-positive neurons were IB4- and CGRP-positive, respectively (Figure 2(b), B1, C, C1), indicating that Piezo1 is expressed by unmyelinated (C-type) and thinly myelinated (Aδ-type) PSNs that convey multiple modality nociceptive signals generated at peripheral nerve terminals to neurons in lamina I-II of the spinal cord. ⁴⁷ Nociceptive neurons in adult rat DRG can be double positive for both IB4 and CGRP (30–40%), ⁴⁰ suggesting that there is a substantial proportion of Piezo1-IR neurons positive for both IB4/CGRP. Most (76%) of Piezo1-positive PSNs were also positive for Ca_V3.2, suggesting high expression of Piezo1 in the Aδ- and C-LTMR neurons (Figure 2(d), D1). ⁴⁶ About 66% of NF200-neurons were Piezo1 immunopositive in a less immunostaining density (Figure 2(e), E1), suggesting detection of Piezo1 expression in the Aβ-LTMRs and proprioceptive PSNs that transmit mechanoreceptive and proprioceptive signals via thickly myelinated afferents to spinal lamina III-V. Colabeling of Piezo1 with NKA1α revealed Piezo1 profiles in the PSN plasma membrane, especially those of larger diameter neurons (Figure 2(f)). A prior RNAseq study detects Piezo1 transcript in mice DRG large neurons expressing neurofilament and parvalbumin, but less enriched than Piezo2. ⁴⁸ Transcriptomic analysis also identifies Piezo1 in human DRG neurons, although less enriched than Piezo2. ¹⁷ To obtain further insight into plasma membrane localization of Piezo1, we examined the ICA of the images co-stained for Piezo1 and NKA1α. Overlaid images of Piezo1 with NKA1α showed colocalizations of the two patterns of immunopositivity (Figure 2(f)1), but the ICA plots of Piezo1 and NKA1α resulted, however, in a complex relationship in which the data tended to cluster along both positive and negative axes with ICQ 0.04–0.2 (p < 0.01∼ 0.001, n = 10), indicating partial localization of Piezo1 in neuronal PM (Figure 2(g)).

Piezo1 has been identified in satellite glial cells of mouse, rat, and human DRG by RNAseq4^49,50 or RNAscope. ⁸ Here we performed colabeling IHC using a selection of established markers for SCs and SGCs; the latter may represent a population of developmentally arrested SCs. ⁵¹ Glial markers used ⁵² include glutamine synthetase (GS), 3-hydroxy-3-methylglutaryl coenzyme A synthase 1 (Hmgcs1), glial fibrillary acidic protein (GFAP), and S100. IHC staining revealed Piezo1-IR in the GS-, Hmgcs1-, S100-, and GFAP-positive perineuronal glial population (Figures 2(h)-(k)), indicating Piezo1 expression in SGCs and/or SCs since both are composed of perineuronal glia and express those glial markers. ⁵² To further verify the presence of Piezo1 in the perineuronal glia, ICA was performed to analyze the immunocolocalization between Piezo1 and GFAP. Overlaid images of Piezo1 with GFAP showed colocalization of two immunopositivity (Figure 2 K1), and the ICA plots of Piezo1 and GFAP also clustered along both positive and negative axes with ICQ 0.06–0.26 (p < 0.01∼0.001, n = 10), indicating Piezo1 partial colocalization in perineuronal glia cells (L). ICC on rat dissociated DRG culture also identified Piezo1 expression in the somata of neurons and their neurites, as well as GFAP- and S100-positive glial cells (Supplement Figure 3). Immunoblots detected strong ∼300 kDa Piezo1 protein band in the lysates prepared from rat DRG, enriched DRG-PSNs, PSN-free rat DRG glial cells, and 50B11 and F11 rat DRG neuronal cells (Figure 2 M-O), adding further evidence of Piezo1 protein in DRG neurons and glia. Supporting our findings, existing transcriptional datasets corroborated our finding since Piezo1 transcripts and protein have been detected in the rat and human SCs ¹⁴ and SGCs,^8,49,50 as well as mice SCs. ¹⁸ This provides support for our finding that SGCs and SCs express Piezo1.

Detection of Piezo1 in sciatic nerve, and spinal cord, and skin

We next characterized Piezo1 expression in tissues innervated by PSNs. Results showed Piezo1-IR signals in the tissue sections of sciatic nerve, which were co-stained with neuronal markers of Tubb3, NF200, and NKA1α (Figure 3(a)-(c)), indicating that Piezo1 was actively transported along the peripheral afferent axons. Notably, sciatic nerve Piezo1 was also detected in SCs since Piezo1 signals were colocalized to SCs markers of MBP, S100, and GAP43 (Figure 3(d)-(f)). ICC colabeling of Piezo1 with S100 on human SCs and SCs isolated from rat sciatic nerve verified Piezo1 expression in SCs (Figure 3(g) and (h)). Immunoblots detected strong ∼300 kDa canonical Piezo1 protein bands in the lysates of primary cultured SCs isolated from both human and rat (Figure 3(i) and (J)).

Figure 3.

IHC of Piezo1 (PZ1) axonal component and Schwann cell expression (male). Representative IHC montage images show double immunostaining of Piezo1 (red) with a selection of neuronal markers (green), including Tubb3 (A), NF200 (B), and NKA1α (C); and PZ1 (red) with glia markers (green), including MBP (D), S100 (E), and GAP43 (F). Human Schwann cells (G) and isolated Schwann cells (H) from rat sciatic nerve show double staining of Piezo1 (red) with S100 (green). Scale bars: 50 μm for all. Immunoblots show detection of canonical PZ1 protein (∼300 kDa) in the lysates of primary cultured human and rat Schwann cells (I, J). GAP43, Growth associated protein 43; MBP, myelin basic protein.

The results presented above suggest that Piezo1 is actively transported along the peripheral processes of PSNs (Figure 3), thus subsequent experiments examined whether Piezo1 is also transported to the central terminals innervating the spinal cord dorsal horn (DH). Indeed, Piezo1-IR signals were present in the central terminal neuropils that synapse in the DH. These are likely transported along axons from DRG-PSNs since DH Piezo1-IR was highly overlaid to presynaptic markers of IB4, CGRP, synaptic vesicle protein synaptophysin (Syp), and synaptoporin (Synpr) (Figures 4(a)-(d), A1-D1), all of which are expressed in smaller-sized DRG neurons with active axonal transportation.^53,54 Additionally, we found Piezo1-IR in the intrinsic spinal cord neurons of both DH and ventral horn (VH). Colabeling of Piezo1 with NeuN and PKCγ verified Piezo1-IR in the spinal cord DH interneurons and VH motor neurons (Figures 4(e)-(g), E1-G1), in agreement with recent studies that showed abundant Piezo1 (and Piezo2) transcripts in the spinal somatostatin interneurons, DH nociceptive projection neurons, and VH motor neurons.^55–58 DH Piezo1-IR was colabeled with GFAP-positive astrocytes (Figure 4(h), (H)1), immunoblots detected strong ∼300 kDa canonical Piezo1 protein band in the lysates of primary cultured DH glial cells and NSC spinal cord neuronal cells, and colocalization of Piezo1 and GFAP was verified by ICC on spinal DH dissociated glia cultures (Supplemental Figure 4). Finally, Piezo1-IR was also detected by IHC in the brain neurons and GFAP-positive astrocytes; Piezo1 protein (∼300 kDa) detected in the lysates of NG108 cortex neuronal cells, C6-rat astrocytes, and human U251 astrocytes; and supportively, the functional Piezo1 was verified in the NG108 cells and C6 astrocytes by Yoda1 stimulation (Supplemental Figure 5).

Figure 4.

IHC of Piezo1 (PZ1) expression in spinal cord (male). PZ1-IR (red) is co-stained with a selection of dorsal horn presynaptic markers (green), including IB4 (A), CGRP (B), Syp (C), and Synpr (D), showing immunocolocalization (yellow) in the magnified merged images (A1-D1). D1 is the squared region of panel D. Piezo1-IR (red) and NeuN (green) on the DH (E) and VH (F) show immunocolocalization (yellow) of PZ1 with the squared region of panel E showing the magnified image (E1) and amplified image showing PZ1-labeled VH neurons (F1). Piezo1-IR (red) and PKCγ (green) on the DH (G) show immunocolocalization (yellow) in the magnified merged images (G1). (H) Positive PZ1 (red) in GFAP-positive (green) astrocytes on spinal cord DH sections, with the squared region shown at high magnification (H1). Immunoblots show detection of clean canonical PZ1 protein band (∼300 kDa) in the lysates of primary cultured rat spinal cord glial cells (I) and NSC SC-neurons (J). Scale bars in IHC images: 50 μm for all. Syp, synaptophysin; Synpr, synaptoporin, DH, spinal cord dorsal horn; VH, ventral horn.

Piezo1 is highly expressed in skin, ¹ and has been detected in keratinocytes ²² and sensory afferent lanceolate endings. ⁵⁹ We here characterized Piezo1 protein expression profile by IHC in rat hindpaw skin (Figure 5). We found Piezo1-IR in the Merkel cells of the hindpaw epidermis, which colabeled with CK14, a marker for the basal keratinocytes and Merkel cells.^60,61 IHC also revealed Piezo1-IR in Meissner’s corpuscles, lanceolate endings, onion-shaped Pacinian corpuscle, afferent terminal nerve bundles, and the endothelial cells of small blood vessels within dermis. We also detected the epidermal Syp-IR closely anear Piezo1 keratinocytes, suggestive of Piezo1 keratinocyte–sensory neuron synaptic-like contacts, ⁶² which will require further investigation. S100-IR was overlaid upon Piezo1-IR signals in the cutaneous nerve bundles, Meissner’s corpuscles, and SCs surrounding cutaneous afferent nerve bundles within the dermis, suggesting that Piezo1 was likely expressed by cutaneous SCs. Additionally, Piezo1-IR in the epidermal basal layer keratinocytes was colabeled with S100, which is a well-validated melanocyte (or pigment cells) marker but absent in Merkel cells. ⁶³ Colabeling of Piezo1 with S100 by ICC and immunoblot detected Piezo1 protein in the cultured human epidermal melanocytes, and Yoda1 stimulation verified functional Piezo1 expression in the human melanocytes (Figure 6). Together, our data indicate that Piezo1 protein is abundantly expressed in skin, including the Merkel cells and epidermal melanocytes, various cutaneous sensory corpuscles, sensory terminals, and microvascular networks. Glabrous Meissner’s corpuscles were colabeled with Piezo1, NF200, IB4/CGRP, and S100, supporting that the Piezo1-positive Meissner’s corpuscles are a multi-afferented mechanoreceptor consisting of Aβ and accessories of C/Aδ fibers axons, as well as nonmyelinating SCs.^64,54

Figure 5.

IHC delineation of Piezo1 (PZ1) expression in skin (male). Hindpaw glabrous skin sections display PZ1 (red) and CK14 (green), showing colabeling (yellow, empty arrowheads) in the merged image (A, B); the white arrowheads point to PZ1 labeled Meissner corpuscles. PZ1-IR in Meissner’s corpuscles (white arrowheads) is immunolocalized with NF200 (C), IB4 (D), and CGRP) (E). PZ1-IR in the hairy skin lanceolate endings is colabeled with NF200 (F, white arrowheads) and CGRP around hair follicles (G, white arrowheads point to lanceolate endings), as well as putative onion-like Pacinian corpuscle colabeled with NF200 (H). PZ1-IR is immunocolocalized with CGRP (I) and IB4 (J), NF200 (K), Cav3.2 (L), and S100 (M) in the nerve bundles within dermis. IHC reveals colabeling of PZ1 (red) with Syn (green) in the nerve bundles within the dermis (N, white arrowheads; red arrowheads point to PZ1-labeled endothelial cells of small vesicles). PZ1 (red) is colabeled with Syn (green) in the epidermis (O), with the squared region rotated clockwise 90-degree shown at high magnification (O1). PZ1 (red) is co-stained with ACTA2 (green) in the nerve bundles within the dermis (P, white arrowheads point to nerve bundles, red arrowheads to PZ1-labeled endothelial cells of small vesicles, and green arrowheads to ACTA2-labled vascular smooth muscles). IHC shows PZ1-IR colabeled with S100 in Meissner’s corpuscles (white arrowheads) (Q, R), with the squared region in panel Q shown at high magnification (Q1); and in epidermal basal layer cells (S and T, empty arrowheads; white arrowheads point to Meissner corpuscles). Scale bars: 50 μm for all. IHC, immunohistochemistry; Cav3.2, T-type calcium channel 3.2α1H; Syp, synaptophysin; ACTA2, α-Smooth muscle actin.

Figure 6.

Piezo1 in human melanocytes (female). Shown are ICC montage images of double labeling of PZ1 (A, red) with S100 (B, green), Piezo1/S100 colabeling (C, merged), showing morphology of melanocytes in phase image (D). Scale bars: 25 μm for all. Detection of canonical PZ1 protein (∼300 kDa) in the lysates of human melanocytes, asterisk denotes putative non-specific band (E). Representative [Ca²⁺]_i traces (F) and bar charts (G) summarize averaged [Ca²⁺]_i peak values evoked by different concentration of Yoda1 and Yoda1 plus GsMTx4 (1 μM, red text) as indicated in human melanocytes. ICC, immunocytochemistry; hMCs, human melanocytes.

Functional verification of Piezo1 expression in dissociated PSNs and nonneuronal cells of male rats

We next focused on testing the presence of functional Piezo1 in various cell populations by the [Ca²⁺]_i responses upon Yoda1 stimulation. Piezo1 channels are activated and robustly characterized by mechanical cell indention ¹ and high-speed pressure-clamp (HSPC) ⁶⁶ under the whole cell patch-clamp via heterologous overexpression system and by a heterocyclic compound Yoda1 chemically-stimulated microfluorimetric Ca²⁺ imaging. ⁶⁷ Because both Piezo1 and Piezo2 co-express in PSNs and non-neuronal cells,^8,26,68 the apparent similarity in Piezo1/2-mediated MA-currents may confound the results of either of Piezo homolog characterization. Although not fully mimicking mechanical opening dynamics of the channels, Yoda1, a synthetic small molecule agonist capable of selectively activating Piezo1 with micromolar affinity, is a useful research tool to confirm the presence of functional Piezo1 channels and to delineate the functional impact of Piezo1 on various cellular processes and reactions. ⁶⁷ Selectivity of Yoda1 is validated since genetic deletion of Piezo1 by Cas9/CRISPR system completely abolish Yoda1-induced Ca²⁺ responses,^69,70 Piezo1 knockdown by RNA interference suppresses such effects, ⁷¹ Yoda1 activates Piezo1 but not Piezo2 consistent with Yoda1 having a Piezo1 selective effect,^67,72 and activation by Yoda1 and by mechanical stimuli are closely coupled.^67,73 Here we tested the ability of Yoda1 to induce Ca²⁺ response in dissociated PSNs, DRG glia (composed of SGCs, SCs, and other non-neuronal populations resident within DRGs), sciatic nerve SCs, and DH glial cells of male naïve rats, as well as NSC-SC neurons, C6 astrocytes, F11 and 50B11 DRG neuronal cells, and primary human melanocytes. We observed that Yoda1 induced dose-dependent increases of [Ca²⁺]_i in all of these cell types (c.f. Figure 6–7 and supplemental Figure 5), with apparent variability in the magnitude of the [Ca²⁺]_i response. Relative sensitivity to Yoda1 stimulation is highest in human epidermal melanocytes, followed by rat SCs, isolated DH glial cells, 50B11 cells, and isolated SGCs and DRG neurons. Currently, no Piezo1-specific inhibitor is available, but activation of Piezo1 can be sensitively blocked by GsMTx4, which, although not uniquely selective antagonist to Piezo1, has been widely and coupled used to characterize Piezo1 function stimulated by Yoda1. Results (Figure 7) here showed that, in all cell types tested, Yoda1 responses were sensitively abolished by the GsMTx4 (0.5–1.0 μM). Among rat PSNs, the majority of the responders were small- and medium-sized (Figure 7(a)2), consistent with a previous report in mice ¹⁹ showing that Yoda1 induces inward currents more often in small- and medium-sized PSNs than in large neurons. Even within a neuronal size group and glial population, there were substantial variations in the magnitude of [Ca²⁺]_i response to Yoda1, suggesting functional diversity of Piezo1 in different neuronal and glial subpopulations. Taken together, our study showed that functional Piezo1 is widely expressed in neuronal and non-neuronal cells in peripheral sensory pathways, as well as glial cells in spinal cord of rats.

Figure 7.

Yoda1-stimualted functional Piezo1 in primary cultural cells and spinal cord/DRG neuronal cell lines. Representative [Ca²⁺]_i traces and bar charts summarize averaged [Ca²⁺]_i peak values evoked by different concentration of Yoda1 and Yoda1 plus GsMTx4 (1 μM, red text) as indicated in acute dissociated PSNs (A, A1) with dot plots (A2) showing correlation of Yoda1 responses to PSN sizes (small-medium size ≤40 diameter), DRG dissociated SGCs (B, B1, including other glia, e.g. SCs), primary cultural SCs isolated from naïve rat sciatic nerves (C, C1), dissociated spinal DH glia (D, D1), NSC spinal cord motor neurons (G, G1), 50B11 DRG neurons (H, H1), and F11 DRG neurons (I, I1). The numbers on the top of scattered plots with bars are the responders (numerators) out of total cells recorded (denominators) and % of responders (brackets). DRG, dorsal root ganglia; PSN, primary sensory neuron; DH, spinal cord dorsal horn.

Increased DRG Piezo1 protein levels following painful peripheral nerve injury

Piezo channels are activated after axon injury, ⁷⁴ inflammation promotes Piezo1 activity, and inflammatory signaling sensitizes Piezo1 that can also promote inflammation.^75,76 Since neuropathic pain is a comorbid pathology of nerve injury and inflammatory responses, we next attempted to determine whether neuropathic pain is associated with alteration of Piezo1 protein levels in rat DRG. We generated TNI and SNI neuropathic pain in male animals, confirmed by reduced the threshold for withdrawal from mild mechanical stimulation (vF) and hyperalgesia evident with noxious (Pin) mechanical stimulation when compared to baseline and sham-operated animals. L4/L5 DRG were harvested at the 28 days after TNI or SNI injury to evaluate the protein expression level of Piezo1 and microgliosis in whole DRG. The canonical Piezo1 (∼300 KDa) upon immunoblots was significantly increased in tissue lysates from the DRG ipsilateral to TNI and SNI, with microgliosis as shown that the Iba1 protein was significantly elevated, compared to controls (Figures 8(a)-(d)). Since Piezo1 is both neuronal and glial protein, the alterations in Piezo1 expression can be ascribed to the changes in either sensory neurons or glia or both. We additionally assessed whether TNI-induced neuropathic pain modifies [Ca²⁺]_i responses to Yoda1 in dissociated DRG neurons and glia (SGCs) of adult rats. Functional assessment (application of Yoda1) of the dissociated cultures from DRG ipsilateral to TNI showed that 5 and 10 μM of Yoda1-evoked [Ca²⁺]_i responses were significantly higher in both PSNs (Figure 8(e)-E2) and SGCs (Figure 8(f)-F2), compared to cells from the sham-operated DRG. Additionally, the responders (%) of both PSNs and SGCs from TNI rats to 10 μM Yoda1 stimulation were significantly increased, compared to the cells from the sham-operated DRG. These data suggest that both neuronal and non-neuronal Piezo1 be likely activated following peripheral nerve injury.

Figure 8.

Activation of Piezo1 in PSNs and SGCs following TNI in male rats. Scattered plots with bars summarize means ± SEM of mechanical allodynia (vF) and hyperalgesia (Pin) in TNI (A) and SNI (C) rats (n = 10 for each group), unpaired two-tailed Student’s t-test for vF and Mann–Whitney test for Pin. DRG homogenates were extracted from the DRG (pooled L4/L5) at 28 days after TNI (B) and SNI (D) or control (sham), and subjected to immunoblotting (IB) as shown in the representative IBs of Piezo1, Iba1, and Gapdh. The densitometry of canonical piezo1 (∼300 KDa, asterisks denote non-specific bands) and Iba1was analyzed and summarized in bar charts (right panels of B, D); unpaired two-tailed Student’s t-test. Comparison of [Ca²⁺]_i traces evoked by 1 and 10 μM of Yoda1 in PSNs (E) and SGCs (containing other glia population) (F), the scattered plots with bars showing the averaged [Ca²⁺]_i peak values evoked by 1 and 10 μM of Yoda1 in PSNs (E1) and SGCs (F1) from sham and TNI rats, unpaired two-tailed Student’s t-test. Shown in E2 and F2 summarize the % of responders versus non-responders evoked by 1, 5, and 10 μM of Yoda1 in PSNs (E2) and SGCs (F2) from sham and TNI rats, chi-square and Fisher’s exact tests. *p < 0.05, **p < 0.01, and ***p < 0.001. TNI, tibial nerve injury; SNI, spared nerve injury; DRG, dorsal root ganglia; Iba1, allograft inflammatory factor-1.

Sciatic nerve injection of GsMTx4 inhibits Yoda1-induced mechanical hypersensitization in naïve rats

Our studies indicate that SCs of sciatic nerve express high level of Piezo1 and are especially sensitive to Yoda1 stimulation. As an initial test to determine the functional significance of sciatic nerve Piezo1, we evaluated the allodynia and hyperalgesia mechanical responses in naïve rats to sciatic nerve injections of either Yoda1 (20 μM) or Yoda1 (20 μM) combined with GsMTx4 (40 μM). Since Piezo1 function could be sex-dependent, ⁷⁷ both male and female rats were tested. Results showed that a single delivery of Yoda1 induced significant mechanical allodynia and hyperalgesia response during 15–60 min after injection, while the Yoda1-induced mechanical hypersensitivity was significantly blocked in both male and female rats when combined injection of Yoda1/GsMTx4 (Figures 9(a) and (b)). These Yoda-GsMTx4 coupled results suggest Piezo1-dependent effects.

Figure 9.

Reversal of Yoda1-induced mechanical hypersensitivity by GsMTx4. (A, B) Scatter and line plots show time courses for the group averages of sensitivity to vF (A1, B1) and Pin (A2, B2) following sciatic nerve injection of saline, Yoda1 (20 μM) or Yoda1 (20 μM) plus GsMTx4 (40 μM) in male and female naïve rats, as indicated, for comparison to BL (repeated measures two-way ANOVA and Tukey post-hoc for vF and nonparametric analyses by Friedman’s test with Dunn’s post hoc for Pin. Right scattered plots of A1-B2 show AUCs (15–60 min) comparison among groups, one-way ANOVA, and Tukey post-hoc. *p < 0.05, **p < 0.01, and ***p < 0.001.

Discussion

This report provides data showing the distribution and the potential roles of Piezo1 in peripheral sensory pathways. First, IHC using a KD/KO validated antibody identifies Piezo1 in DRG-PSNs (∼60%), preferentially in smaller-sized neurons, a result likewise to the recent reports by in situ hybridization or calcium imaging showing Piezo1 expression predominantly in the smaller DRG-PSNs but also in some larger PSNs of mice.^8,19 Secondly, Piezo1 expression in PSNs extends from the peripheral terminals in the skin to the central presynaptic terminals in the spinal DH. Additionally, Piezo1 is expressed by DH neurons, ventral horn motor neurons, astrocytes, and in peripheral non-neuronal cells, including perineuronal glia composed of SGCs and SCs, sciatic nerve and cutaneous SCs, and skin epidermal melanocytes, as well as Merkel cells. Finally, painful nerve injury increases DRG Piezo1 protein levels, which may involve neurons and glial cells since the responses of both PSNs and SGCs are increased following painful neuropathy of rats. These observations suggest that Piezo1 is a mechanotransducer possibly associated with pain pathology. ⁸

The spinal cord DH is a key site for integrating and transmitting somatosensory information. ⁷⁸ The GenBank GEO Profile and the published literature report detection of Piezo1 (and Piezo2) transcripts in the DH projection neurons, interneurons, and spinal cord motor neurons.^55–58 Piezo1 is detected in human induced pluripotent stem cell (iPSC)-derived spinal motor neurons, ⁷⁹ and the Human Protein Atlas shows consensus datasets of Piezo1 transcript in spinal cord. These data provide credent evidence of Piezo1 in spinal cord transcriptome, and we now confirm Piezo1 protein expression in the spinal cord. Cumulatively, Piezo1 may play unexplored role in spinal cord which is the essential sensory processing hub that organizes descending and ascending mechanosensory and pain signals. ⁷⁸

Peripheral non-neuronal cells play key roles in the induction and maintenance of persistent mechanical pain.^80,81 Here, we show that Piezo1 is expressed extensively in the peripheral nerve non-neuronal cells of various types. The skin acts as a complex sensory organ, ⁸² and epidermal keratinocytes along with Merkel cells, Langerhans cells, and melanocytes express sensor proteins that regulate the neurocutaneous system and participate in nociception and mechanotransduction. ⁸¹ Non-neuronal skin keratinocytes^1,22 and the sensory afferent lanceolate endings express Piezo1, ⁵⁹ which is required for skin mechanotransduction. Our data also reveal Piezo1 in Meissner’s corpuscles, lanceolate endings, Pacinian corpuscle in deeper dermis, and the afferent terminal nerve bundles. It is well-defined that not only Aβ-LTMR but also C- and Aδ-LTMR fibers innervate lanceolate endings and Meissner’s corpuscles that contribute to behavioral responses, perception of light touch, and nociception. ⁸³ Beside detection of Piezo1 in the skin epidermal Merkel cells, we provide evidence that the epidermal melanocytes also express functional Piezo1. Skin melanocytes are dendritic cells, derived from skin Schwann cell precursors that originate from neural crest cells (NCC) and regulate mechanosensation. ⁸⁴ Melanocytes form tight contacts with cutaneous nerves, respond to mechanical stretch, and appear around nerve fascicles after damage. ⁸⁵ Pressure-sensing Merkel cells are thought of NCC origin, ⁸⁶ but new observations show that Merkel cells originate from epidermal progenitors but not NCCs. ⁸⁷ Moreover, we found that peripheral nerve glia, including SGCs and SCs, express Piezo1, which is colocalized to S100-positive dermal afferent nerve bundles and terminal fibers, indicative of Piezo1 expression in the cutaneous nociceptive SCs.

Sensory neurons and SGCs/SCs, the two principal cell types of the peripheral nerves, interact intimately and are critical for normal peripheral nerve functions and pain pathogenesis during nerve injury and inflammation. Peripheral SCs can sense and transduce mechanical signals involving mechanosensation.^88,89 and the specialized cutaneous nociceptive SCs in skin modulate the sensitivity threshold for mechanosensation. ⁹⁰ New reports indicate that SGCs can promote regenerative growth into PSNs ⁵⁰ and may even have stem cell characteristics to differentiate to PSNs. ⁹¹ SCs in peripheral nerves are critical for both sensory and motor functions. SCs are physiologically exposed to mechanical stresses (i.e. tensile, compressive, and shear strains), ⁸⁸ SCs can sense and transduce mechanical signals involving mechanosensation, and determine the sensitivity threshold for mechanosensation.^90,92 Therefore, peripheral nerve SCs-Piezo1 may play hitherto unexplored roles in sensory processing and mechanotransduction; for instance, by regulating signaling within SCs by mechanical stimulation. Functional Piezo1 is expressed in mouse and human neural stem cells^93,94 and both Piezo1/2 express in human embryonic stem cell-derived NCC cells.^95,96 Thus, it is speculated that Piezos may play roles for mechanosensitive lineage choice and that Piezos expression is required throughout the differentiation process along NCC differentiated sensory neuronal or nonneuronal glial lineages in peripheral nervous system.

The decrease in Yoda1-induced mechanical hypersensitivity by GsMTx4 adds further indirect evidence for possible involvement of peripheral Piezo1, as well as other MSCs, activation in mechanical sensation. Acute anti-nociception to mechanical stimulation after intraperitoneal and intradermal application of GsMTx4 has been used in addressing possible involvement MSCs (including Piezo1) in animal models of chronic pain.^23,25 Since Piezo1 is also expressed in spinal cords, it is unclear whether peripheral or central block of Piezo1 is analgesia. Reversal of Yoda1-induced mechanical hypersensitivity by GsMTx4 in sciatic nerve suggests that a peripheral site of Piezo1 action is possibly involved in mechanosensation. Yoda1 and GsMTx4 directly into the large nerve trunks of the sciatic nerve could act on axonal Schwann cell channels and directly at the site of injection, or may spread to more central and peripheral sites of action by bulk flow and diffusion longitudinally within the sciatic and associated spinal nerves, or by axoplasmic transport. Development of more specific Piezo1 inhibitors and a genetic approach of conditional ablation of Piezo1 in Schwann cells in future study will help to elucidate the role of Piezo1 in peripheral mechanobiology and in pain pathology.

ORCID iD

Hongwei Yu https://orcid.org/0000-0003-3029-3644