Piezo2 expressing nociceptors mediate mechanical sensitization in experimental osteoarthritis

Abstract

Non-opioid targets are needed for addressing osteoarthritis pain, which is mechanical in nature and associated with daily activities such as walking and climbing stairs. Piezo2 has been implicated in the development of mechanical pain, but the mechanisms by which this occurs remain poorly understood, including the role of nociceptors. Here we show that nociceptor-specific Piezo2 conditional knock-out mice were protected from mechanical sensitization associated with inflammatory joint pain in female mice, joint pain associated with osteoarthritis in male mice, as well as both knee swelling and joint pain associated with repeated intra-articular injection of nerve growth factor in male mice. Single cell RNA sequencing of mouse lumbar dorsal root ganglia and in situ hybridization of mouse and human lumbar dorsal root ganglia revealed that a subset of nociceptors co-express Piezo2 and Ntrk1 (the gene that encodes the nerve growth factor receptor TrkA). These results suggest that nerve growth factor-mediated sensitization of joint nociceptors, which is critical for osteoarthritic pain, is also dependent on Piezo2, and targeting Piezo2 may represent a therapeutic option for osteoarthritis pain control.

Fig. 1. Piezo2 is depleted from nociceptors using the marker Na_V1.8.

a, b ScRNA-seq of naive L3-L5 dorsal root ganglion (DRG) cells demonstrates that Piezo2 is expressed by a subset of nociceptors (Na_V1.8 + neurons) in addition to large diameter neurons (NF; Na_V1.8− neurons); SCHW Schwann cells, SATG satellite glia, NF neurofilament, NP non-peptidergic nociceptors, PEP peptidergic nociceptors, TRPM8 transient receptor potential melastatin 8, TH tyrosine hydroxylase containing, IMM immune cells, VASC vascular cells. c–e Piezo2 was deleted in nociceptors by using the marker Na_V1.8. RNAscope was used to quantify the number of cells expressing Na_V1.8 and Piezo2. There is reduced co-expression of Na_V1.8 + / Piezo2 + in heterozygous Piezo2^CKOfl/+ (cyan circles) and homozygous Piezo2^CKOfl/fl (blue triangles) mice compared to control mice (black squares) (greater than 12 weeks of age). c One-way ANOVA with Dunnet’s multiple comparison test was used to compare numbers of cells between strains. Mean ± SEM. d Representative sections from control and e Piezo2^CKOfl/fl mice. Scale bar = 25 μm. Yellow arrow indicates example cells expressing Na_V1.8 and Piezo2. Red arrow indicates example cells only expressing Na_V1.8. Blue arrow indicates example cells expressing only Piezo2. n = 4 no Cre controls; n = 3 Piezo2^CKOfl/+; n = 5 Piezo2^CKOfl/fl. Adjustments to individual colour channels on merged images was performed using brightness and contrast tools. Adjustments were made to the entire image and have been applied across all images and controls. For c, source data are provided as a Source Data file.

Fig. 2. In vivo calcium imaging demonstrates a role for nociceptor expression of Piezo2 in mediating intracellular calcium responses to mechanical force applied to the knee joint.

a The L4 dorsal root ganglion (DRG) is imaged by two photon microscopy in anesthetized mice while mechanical stimuli are applied to the knee joint using an instrumented forceps. b, c Representative DRG images and d, e individual neuronal traces from control (Na_V1.8;GCaMP6s) and Piezo2^CKOfl/+ (Na_V1.8;GCaMP6s;Piezo2^fl/+) mice. Corresponds to Supplementary Movies 1 and 2. Heatmaps depict changes in fluorescence for responding neurons (each column is one responding neuron; each row is one frame). f The number of Na_V1.8+ neurons responding to each stimulus was quantified and compared between strains by unpaired two-tailed t-test (each dot = one mouse; for each mouse >225 neurons imaged in total) (control, n = 5 mice (black; 30 g = circles; 100 g = triangles); Piezo2^CKOfl/+, n = 4 mice (blue; 30 g = diamonds; 100 g = nabla)). Mean±SEM. g The peak area under the curve (AUC) of each neuron responding to each stimulus was quantified, averaged for each mouse, and compared among strains by unpaired two-tailed t-test. (same mice as in f: control, n = 5 mice; Piezo2^CKOfl/+, n = 4 mice). Mean ± SEM. h The relative frequency distribution of the areas (µm²) of responding neurons. For f–h, source data are provided as a Source Data file.

Fig. 3. Piezo2 depletion protects from CFA induced mechanical sensitization.

a Knee swelling was assessed in female wild-type (WT) (black squares) or homozygous Piezo2^CKOfl/fl (blue circles) mice given a single intra-articular injection of Complete Freund’s adjuvant (CFA) (i.a. 5 µg, 5 µL). Two-way repeated measures ANOVA. b Area under the curve analysis over the time course was used to assess knee swelling. Unpaired two-tailed t-test. c Knee hyperalgesia was assessed in the same mice as part a. Dashed line indicates maximum of the assay: 450 g. Two-way repeated measures ANOVA with Sidak’s post-test. d Area under the curve analysis over the time course was used to assess knee hyperalgesia. Unpaired two-tailed t-test. Mean ± SEM. For a–d WT (n = 5) and homozygous female Piezo2^CKOfl/fl (n = 7) mice. i.a. CFA intra-articular Complete Freund’s adjuvant. For a–d, source data are provided as a Source Data file.

Fig. 4. Piezo2 plays a role in mechanical sensitization in two mouse models of osteoarthritis.

a Weight-bearing asymmetry was assessed in wild-type (WT) (n = 5; black squares) and Piezo2^CKOfl/fl (n = 5; blue circles) male mice after DMM surgery. Dashed line indicates 0 g which would represent mice putting equal weight on both hind legs; negative numbers indicate mice are putting more weight on the uninjured left rear leg. Two-way repeated measures ANOVA with Sidak post-test. Mean ± SEM. DMM destabilization of medial meniscus. b Knee hyperalgesia was assessed in mice from part a. Dashed line indicates maximum of the assay: 450 g. Ipsilateral operated right leg data is plotted. Contralateral unoperated left legs: 4 weeks (mean ± SEM): wild-type (425 ± 2) and Piezo2^CKOfl/fl (437 ± 7). 8 weeks: wild-type (439 ± 2) and Piezo2^CKOfl/fl (445 ± 5). 12 weeks: wild-type (437 ± 2) and Piezo2^CKOfl/fl (432 ± 7). 16 weeks: wild-type (441 ± 2) and Piezo2^CKOfl/fl (446 ± 4). Two-way repeated measures ANOVA with Sidak post-test. An independent experiment is shown in Supplementary Fig. 6. c Hind paw mechanical allodynia was assessed in naive control mice (littermate no cre, n = 5, black squares) or Piezo2^CKOfl/fl (n = 8, blue circles) at 1.5 years of age. Histology shown in Supplementary Fig. 7. Unpaired two-tailed t-test on log-transformed data. Mean ± SEM. For a–c, source data are provided as a Source Data file.

Fig. 5. Silencing Piezo2+ neurons transiently reverses knee hyperalgesia.

a Piezo2-Cre^+/−;Pdi^fl/+ mice express the inhibitory DREADD receptor on dorsal root ganglia (DRG) neurons. Representative images showing immunofluorescence for the HA-tagged DREADD receptor. Scale bar = 25 µm. Piezo2-Pdi mice (n = 6, 22–31 weeks old). b Knee hyperalgesia in male Piezo2-Cre^+/−;Pdi^fl/+ inhibitory DREADD mice 9 weeks after DMM surgery was assessed before, 1, 2, and 4 h after intra-articular injection of saline (n = 10 mice; black squares) or CNO (n = 9 mice; blue circles). Dashed line indicates maximum of the assay: 450 g. Two-way repeated measures ANOVA with Sidak post-test. Mean ± SEM. DMM destabilization of medial meniscus, CNO Clozapine-N-oxide. HA hemagglutinin. ‘Designer Receptors Exclusively Activated by Designer Drugs’ = DREADD. For b, source data are provided as a Source Data file.

Fig. 6. A subset of sensory neurons co-expresses Na_V1.8, Piezo2, and Ntrk1 in both mouse and human dorsal root ganglia (DRG).

RNAscope used to identify cells expressing Na_V1.8, Ntrk1 and Piezo2 with DAPI in a, b wild-type mouse DRG (n = 3 mice; two male and one female) or c, d Na_V1.8, NTRK1 and PIEZO2 with DAPI in human DRG (n = 3 donors; one male and two females); a, b Representative sections of mouse DRG. White arrows indicate example cells expressing Na_V1.8, Ntrk1 and Piezo2. c, d Representative sections of human DRG. White arrows indicate example cells expressing Na_V1.8, NTRK1 and PIEZO2. Scale bar = 50 μm (a, c) and 25 μm (b, d). e, f Number of cells expressing Na_V1.8, Ntrk1 and Piezo2 in mouse DRG (white circles) or Na_V1.8, NTRK1 and PIEZO2 in human DRG (black squares) determined by RNAscope. Mean ± SEM (e); percentages of whole (f). g Relative frequency distribution of cell subset diameters in mouse and h human DRG. Distributions assessed by two-tailed Kolmogorov–Smirnov test. For a–d, adjustments to individual colour channels on merged images was performed using brightness and contrast tools. Adjustments were applied to the entire image and have been applied across all images and controls. For a, b, n = 3 independent repeats, for c, d, n = 3 independent repeats. For e–h, source data are provided as a Source Data file.

Fig. 7. Piezo2 depletion protects from NGF induced knee swelling and mechanical sensitization.

a Knee swelling was assessed in wild-type (WT) or heterozygous Piezo2^CKOfl/+ male mice given repeated intra-articular injections of recombinant murine NGF (i.a. 500 ng, 5 µL, 2x/week; WT = white squares; Piezo2^CKOfl/+ = white circles) or vehicle (5 µL, 2x/week; WT = black squares; Piezo2^CKOfl/+ = blue circles) for 8 weeks; n = 5 mice/group. Repeated measures two-way ANOVA with Sidak’s post-test: WT NGF vs. vehicle, p < 0.0479 from day 7 onward. Piezo2^CKOfl/+, no differences between NGF vs. vehicle. b For the same mice as in a, area under the curve analysis over the time course was used to assess knee swelling. Two-way ANOVA with Sidak’s post-test. (Piezo2^CKOfl/+ vehicle vs. Piezo2^CKOfl/+ NGF: p > 0.99). c Knee hyperalgesia was assessed in the same mice as part a. Dashed line indicates maximum of the assay: 450 g. Repeated measures two-way ANOVA with Sidak’s post-test. d For the same mice as in a, area under the curve analysis over the time course was used to assess knee hyperalgesia. Two-way ANOVA with Sidak’s post-test. (Piezo2^CKOfl/+ vehicle vs. Piezo2^CKOfl/+ NGF: p = 0.1243). e RNA was extracted from the L3-L5 DRGs of a separate cohort of mice after 3 injections of NGF or vehicle and qPCR was performed (n = 3 WT mice/group; n = 5 Piezo2^CKOfl/+ mice/group). Unpaired two-tailed t-test. Mean ± SEM. DRG dorsal root ganglia, NGF nerve growth factor, VEH vehicle. For a–e, source data are provided as a Source Data file.