Synergy between Piezo1 and Piezo2 channels confers high-strain mechanosensitivity to articular cartilage

Abstract

Diarthrodial joints are essential for load bearing and locomotion. Physiologically, articular cartilage sustains millions of cycles of mechanical loading. Chondrocytes, the cells in cartilage, regulate their metabolic activities in response to mechanical loading. Pathological mechanical stress can lead to maladaptive cellular responses and subsequent cartilage degeneration. We sought to deconstruct chondrocyte mechanotransduction by identifying mechanosensitive ion channels functioning at injurious levels of strain. We detected robust expression of the recently identified mechanosensitive channels, PIEZO1 and PIEZO2. Combined directed expression of Piezo1 and -2 sustained potentiated mechanically induced Ca(2+) signals and electrical currents compared with single-Piezo expression. In primary articular chondrocytes, mechanically evoked Ca(2+) transients produced by atomic force microscopy were inhibited by GsMTx4, a PIEZO-blocking peptide, and by Piezo1- or Piezo2-specific siRNA. We complemented the cellular approach with an explant-cartilage injury model. GsMTx4 reduced chondrocyte death after mechanical injury, suggesting a possible therapy for reducing cartilage injury and posttraumatic osteoarthritis by attenuating Piezo-mediated cartilage mechanotransduction of injurious strains.

Keywords: Piezo; cartilage; cartilage injury; chondrocyte; mechanotransduction.

Fig. 1.

Piezo1 (P1) and Piezo2 (P2) are robustly expressed in chondrocytes. (A and B) mRNA expression was measured by real-time quantitative PCR (RT-qPCR) using the ΔΔC_t method for analysis of relative gene expression levels in bladder, lung, skin, trigeminal ganglion (TG), and joint cartilage (hip, knee) of 4-wk-old male mice. Organs (except cartilage) were sampled from single mice (n = 6–10) and mRNA abundance was averaged. Cartilage was pooled (n = 3 pools from three, three, and four mice) to generate sufficient amounts of starting material. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a housekeeping gene for normalization, and expression in lung was assigned a numerical value of “1.” (C) RT-qPCR Ct values of Piezo1, Piezo2, and GAPDH, based on RNA isolated from (C) porcine chondrocytes (n = 6 pigs) and (D) human chondrocytes (n = 4 subjects). (E) Piezo1- and Piezo2-specific immunolabeling of chondrocytes in porcine cartilage tissue. (Scale bar, 10 µm.)

Fig. 2.

Directed coexpression of Piezo1 and Piezo2 potentiates mechanically induced Ca²⁺ signals and transmembrane currents. (A) N2A transfected cells were mechanically stimulated by compressive mechanical loading (∼400 nN, using a flat AFM probe) while recording intracellular Ca²⁺. N2A cells with directed expression of (B) GFP, (C) Piezo1, (D) Piezo2, and (E) Piezo1 and Piezo2. Stimulated N2A cells overexpressing Piezo1 or Piezo2 (B and C) show rapidly decaying modest spikes (∼30 nM Ca²⁺ influx). Note in contrast the robust Ca²⁺ signal (∼500 nM) in N2A cells expressing both Piezo1 and Pieo2 (E). (F) Maximal [Ca²⁺]i, prestimulation subtracted (ΔCa²⁺). (G) Stepwise negative pressure induced transmembrane electrical currents in cell-attached mode: negative pipette pressure 0 to −100 mmHg, ∆−10 mmHg for 500 ms, holding potential of −65 mV. N2A cells were directed to express (H) nontransfected, (I) Piezo1, (J) Piezo2, and (K) Piezo1 and Piezo2. (L) Maximum amplitude of membrane stretch-evoked current is shown. (M) Maximum average plateau currents between ∼350 and 450 ms are shown (blue shades in K). For L and M, note the potentiation of the signal for coexpression of PIEZO1 and PIEZO2. Bars represent the mean ± SEM; the number of cells tested (n) is shown in the bars in F and L. Significantly different from all other bars: ^##P < 0.005, ^###P < 0.0005, ANOVA, LSD post hoc.

Fig. 3.

High-strain AFM indentation induced Ca²⁺ influx and its suppression by Piezo knockdown suggest Piezo-mediated mechanotransduction in articular chondrocytes. (A) Schematic diagram of AFM indentation and a representative trace of mechanically activated Ca²⁺ influx of control chondrocytes (F = 400 nN, 1 µm/s ramp speed, dotted blue line indicates compression). (B) Cell strain increased significantly with applied force over a range of 10–500 nN (logarithmic regression of strain vs. ln(force), P < 0.0001). (C, Upper) Sequential images of the side view of a chondrocyte being compressed with an AFM cantilever, showing a smooth lateral expansion as the cell is compressed vertically. (Scale bar, 5 µm.) (Lower) Force curve from the same cell, showing force increase dependent on compression (x axis showing compression in micrometers). Note that we could not exert more than 300 nN force when using the AFM-PRISM microscope setup. (D) AFM-mediated Ca²⁺ influx of chondrocytes transfected with control siRNA (Left) and with Piezo1-targeting siRNA (Right) (50 nM siRNA each). (E) The average maximal Ca²⁺ influx and the mRNA level of Piezo1 determined by RT-qPCR. (F) AFM-mediated Ca²⁺ influx curves of chondrocytes transfected with control siRNA (Left) and with Piezo2-targeting siRNA (Right) (15 nM siRNA each); the average maximal Ca²⁺ influx and the mRNA expression level of Piezo2 are shown in G. Note robust attenuation of mechanically activated Ca²⁺ influx of chondrocytes subjected to Piezo1 or Piezo2 knockdown. GAPDH was used for normalization, ΔΔC_t method. Bars represent the mean ± SEM; the number of cells tested (n) is shown in bars. **P < 0.005, ***P < 0.0005, unpaired t test.

Fig. 4.

Characteristics of mechanically evoked Ca²⁺ transients in primary chondrocytes. (A–J) Representative traces of mechanically activated Ca²⁺ influx of chondrocytes (400 nN force, 1 µm/s ramp speed), specifically treated with (A) vehicle-control (cont), (B) thapsigargin (thaps) 1 μM, (C) EGTA 10 mM, (D) cytochalasin-D (cyto-D) 2 μM, (E) ruthenium red (RR) 1 μM, (F) verapamil ∼0.1–0.5 μM, (G) dynasore (dyn) 5 μM, (H) GsMTx4 (Gs) 20 μM, (I) GsMTx4 40 μM, and (J) GsMTx4 2 μM and dynasore 5 μM. (K) Average inactivation time (t_50%) of control (shown in A) and GsMTx4 20 μM (shown in H). **P < 0.005, unpaired t test. (L) Average maximum [Ca²⁺]i (ΔCa²⁺, prestimulation subtracted). Bars represent the mean ± SEM; the number of cells tested (n) is shown in the bars. Significantly different from control, not different from each other (ANOVA, ^###P < 0.0005, Dunnett’s post hoc).

Fig. 5.

Surface-labeled Piezo1 channels are less abundant in the cytoplasm in response to inhibition of dynamin GTPase in Piezo1/2 cotransfected N2A cells. (A) Schematic representation of bungarotoxin binding site (BgTx-bs) engineered into the first extracellular loop of mPiezo1 labeled with bungarotoxin (BgTx), conjugated to Alexa Fluor 555. These channels are fully functional (Fig. S8). (B) Representative confocal micrographs of N2A cells transfected with Piezo1-BgTx-bs (see A), exposed in vivo to BgTx-Alexa-555 (red) for 15 min, and labeled postfixation with phalloidin-CF350 (blue). Top Row shows a vehicle control-treated cell; Bottom Row, a cell treated with dynamin GTPase inhibitor dynasore (20 µM); cells were treated for 3 h. (Scale bar, 10 µm.) (C) Exemplary cytoplasmic ROI (confined by the yellow dotted line, note nuclear sparing). Bar diagram shows relative comparison of the mean fluorescence intensity of cytoplasmic BgTx labeling, background subtracted. Averaged n of quantified cells is given in the bars, which indicate mean ± SEM (error bars); *P = 0.014, unpaired t test.

Fig. 6.

GsMTx4-mediated PIEZO inhibition is chondroprotective in a cartilage explant injury model. (A) Our explant model in a schematic. (B and C) Live/dead (green/red) staining shows a zone of death around biopsy wound edge in representative control and 40 µM GsMTx4-treated samples. (Scale bar, 200 µm.) (D) Mean zone of death is significantly decreased in cartilage treated with GsMTx4. Bars represent the mean ± SEM; numbers in bars represent experimental repeats. Significantly different from control **P < 0.005, unpaired t test.