The role of Piezo1 in conventional aqueous humor outflow dynamics.

Controlling intraocular pressure (IOP) remains the mainstay of glaucoma therapy. The trabecular meshwork (TM), the key tissue responsible for aqueous humor (AH) outflow and IOP maintenance, is very sensitive to mechanical forces. However, it is not understood whether Piezo channels, very sensitive mechanosensors, functionally influence AH outflow. Here, we characterize the role of Piezo1 in conventional AH outflow. Immunostaining and western blot analysis showed that Piezo1 is widely expressed by TM. Patch-clamp recordings in TM cells confirmed the activation of Piezo1-derived mechanosensitive currents. Importantly, the antagonist GsMTx4 for mechanosensitive channels significantly decreased steady-state facility, yet activation of Piezo1 by the specific agonist Yoda1 did not lead to a facility change. Furthermore, GsMTx4, but not Yoda1, caused a significant increase in ocular compliance, a measure of the eye's transient response to IOP perturbation. Our findings demonstrate a potential role for Piezo1 in conventional outflow, likely under pathological and rapid transient conditions.

Introduction

Ocular health requires a stable, suitable level of intraocular pressure (IOP), which is determined by the rate of production of aqueous humor (AH) and its resistance to drainage from the eye (Chandler, 1977). Increased resistance to AH outflow leads to a sustained elevation of IOP, which is the major risk factor for glaucomatous optic neuropathy (Buckingham et al., 2008; Goel et al., 2010) and its attendant vision loss (Sihota et al., 2018).

The conventional AH outflow pathway, including the trabecular meshwork (TM) and Schlemm canal (SC), drains the majority of AH in humans (Weinreb et al., 2014). The TM contributes 50%–75% of the resistance to AH outflow in the conventional outflow pathway (Goel et al., 2010). It is mechanosensitive (Liton and Gonzalez, 2008; Ramos et al., 2009), and improved understanding of mechanotransduction pathways in this tissue would enable novel strategies for controlling IOP and improving glaucoma therapy. However, it is still poorly understood how the TM senses and responds to mechanical forces. Recently, mechanosensitive ion channels have been reported in the TM (Tran et al., 2014; Yarishkin et al., 2019). Specifically, TRPV4 (transient receptor potential cation channel subfamily V member 4) and TREK-1 (TWIK-related potassium channel-1) were shown to respond to mechanical force and to influence TM function by modulating calcium homeostasis, remodeling TM cell cytoskeleton, and changing TM extracellular matrix composition (Carreon et al., 2017; Ryskamp et al., 2016; Yarishkin et al., 2018).

Piezo ion channels have essential roles in transducing mechanical forces and, when compared with TRPV4 and TREK-1, are more sensitive to certain mechanical stimuli (Coste et al., 2010). For example, Piezo1, as a blood flow sensor, is required for vascular development and blood pressure regulation (Li et al., 2014; Ranade et al., 2014a; Retailleau et al., 2015). Piezo1 also has important roles in regulating the volume of red blood cells (Cahalan et al., 2015), homeostasis of epithelial cell numbers (Eisenhoffer et al., 2012), cell migration, mechanotransduction in cartilage (Lee et al., 2014), control of urinary osmolarity (Martins et al., 2016), neural stem cell fate (Pathak et al., 2014), and neuronal axon growth (Koser et al., 2016). Piezo2 acts as a key mechanotransducer in response to gentle touch on the skin (Ikeda et al., 2014; Ranade et al., 2014b; Woo et al., 2014), proprioception (Woo et al., 2015), and airway stretch and lung inflation (Nonomura et al., 2017). Genetic gain or loss of Piezo channel function is associated with several diseases such as hereditary xerocytosis (Lukacs et al., 2015), allodynia (Eijkelkamp et al., 2013), and distal arthrogryposis (Coste et al., 2013). However, it remains unknown whether Piezo channels can sense mechanical stimuli in the TM and whether they have a role in regulating AH outflow.

The goal of the present study was to investigate whether and how Piezo ion channels function in AH outflow. We first confirmed the functional expression of Piezo1 in the TM by examining protein expression and cellular electrophysiology. We then determined the role of Piezo1 in the steady-state outflow of AH from the eye.

Results

Piezo1 is widely expressed in human and mouse iridocorneal angle tissues

The conventional AH drainage tissues largely control IOP and are subjected to significant mechanical deformations. We therefore interrogated these tissues to determine Piezo1 expression levels and distribution. Co-immunohistochemical (IHC) staining of anterior segments from three human donors revealed that Piezo1 was robustly expressed throughout the entire TM, with positive labeling observed in 72% of cells in the uveal meshwork, 59% of cells in the corneoscleral meshwork, and 64% of cells in the juxtacanalicular tissue (Figures 1A and 1B). We also examined Piezo1 expression in mouse TM tissue, which is composed of three to four trabecular beams in the anterior meshwork (Smith et al., 2001). Positive staining of Piezo1 could also be observed throughout the mouse TM (Figure 1C). Western blot analysis confirmed the expression of Piezo1 in mouse iridocorneal angle tissues (Figure 1C). Moreover, the expression of Piezo1 was confirmed in human (Figure 1D) and mouse (Figure 1E) cultured primary TM cells by IHC and western blot.

Figure 1

Expression of mechanosensitive ion channels in human and mouse trabecular meshwork

(A) Co-immunohistochemical (IHC) staining of Piezo1 (green) and ColIV (red, antibody labels basement membrane collagen in TM) or CD31 (red, antibody labels Schlemm canal endothelial cells). Typical results from n = 3 human eyes are shown.

(B) Quantification of percentage of cells expressing Piezo1 in subregions of the human TM, namely, the uveal meshwork (UM), the corneoscleral meshwork (CM), and the juxtacanalicular tissue (JCT).

(C) Upper panel: Piezo1 was visualized (green) in the TM of 2-month-old C57BL/6J mice. The region in the white frame is shown in magnified view at the right. Lower panel: The expression of Piezo1 in the iridocorneal angle was determined by western blot (n = 3).

(D) Piezo1 (green) in human primary TM cells was assessed by IHC staining (upper panel) and western blot (lower panel) (n = 3).

(E) Piezo1 was detected in mouse primary TM cells through IHC (green; upper panel) and western blot (lower panel) (n = 2). Nuclei were stained with DAPI (blue). Scale bars, 100 μm.

(F) Upper panel: Protocol for cyclic mechanical stretch. Lower panel: Increased expression of Piezo1 and TRP was detected in human primary TM cells subjected to cyclic mechanical stretch (n = 3). Horizontal lines indicate means; bars are standard deviation. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 by one-way ANOVA.

The endothelial cells of SC adjacent to the TM are also mechanosensitive and participate in conventional aqueous outflow (Stamer et al., 2015). Co-IHC staining indicated that Piezo1 expression co-localized with the SC endothelial cell biomarker CD31 in human tissues (Figure 1A). Furthermore, Piezo1 was detected in the ciliary muscle (Figure 1A) and non-pigmented epithelial cells of the ciliary body (Figure S1), which are crucial for the regulation of unconventional outflow and AH secretion, respectively. Furthermore, strong expression of Piezo1 was detected in cornea (Figure 1). The widespread expression of Piezo1 in outflow tissues suggests that Piezo1 may play a role in modulating AH dynamics in response to mechanical stresses.

Cyclic mechanical stretch increases Piezo1 expression in TM cells

The TM undergoes significant cyclic mechanical stretching due to the ocular pulse and other perturbations (Johnstone, 2004). We therefore asked whether Piezo1 present in TM cells was functionally involved in mechanotransduction in response to such stretching. We subjected cultured human primary TM cells (Ramos et al., 2009) to cyclic mechanical stretch with an amplitude of 20% at a frequency of 1 Hz, an experimental model that mimics acute sustained elevation of IOP (Hirt and Liton, 2017; Kerr et al., 2003), and assessed resulting changes in gene expression profiles. Piezo1 channel mRNA, but not Piezo2 message, was significantly upregulated after 3 h of stretch (Figure 1F), suggesting a functional role in the TM. Several TRP (transient receptor potential cation) channels, such as TRPA1, TRPC1, TRPC6, TRPV4, and TRPP2, were also upregulated after cyclic stretch, consistent with previous observations of their role in mechanotransduction (Ranade et al., 2015).

Mechanical stimulation activates Piezo1 currents in human and mouse TM cells

To further examine the function of Piezo1 in the TM, we mechanically stimulated human and mouse primary TM cells and performed whole-cell patch-clamp recordings. We first confirmed the TM phenotype in both human and mouse TM cells, verifying robust expression of TM biomarkers and dexamethasone-induced myocilin secretion, assayed by RT-PCR, IHC, and western blotting (Figures S2 and S3), as in our previous publication (Yu et al., 2019). We observed that mechanical stimulation caused currents with fast activation and inactivation kinetics (Figure 2A). Current density in human TM cells increased in response to 6-μm cellular indentation, and a similar (although more rapid) response was observed in mouse TM cells (Figure 2B), following which Yoda1 (10 μM) or GsMTx4 (2.5 μM) were delivered to cells. Treatment of TM cells with Yoda1 (10 μM), a specific agonist for Piezo1 that affects its sensitivity and inactivation kinetics (Syeda et al., 2015), resulted in a significant increase in the mechanosensitive peak current amplitude by about 1.6-fold (−481.7 ± 94.6 versus −297.7 ± 56.1 pA) for human TM cells, and by 2.2-fold (−641.0 ± 200.6 versus −292.7 ± 44.3 pA) for mouse TM cells in response to 6-μm mechanical indentation (Figures 2C and 2D). Furthermore, Yoda1 significantly slowed current inactivation kinetics by about 5-fold (228.1 ± 21.0 versus 46.0 ± 4.5 ms) in human TM cells and by 4-fold (64.5 ± 13.2 versus 16.2 ± 1.9 ms) in mouse TM cells (Figures 2E and 2F). We note that Yoda1 has a half maximal effective concentration [EC50] of 17 and 27 μM for mouse and human Piezo1, respectively (Syeda et al., 2015), similar to the 10 μM used in this study.

Figure 2

Mechanical activation of Piezo1 current in human and mouse primary TM cells

(A) Cells were subjected to a series of mechanical stimuli consisting of sequential 1 μm indentations. Representative mechanosensitive inward currents recorded in hTM cells are shown (n=13-15 cells, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, by two-tailed t test). (B) Representative mechanosensitive inward currents recorded in mTM cells are shown (n=6-21 cells, ∗∗∗∗p < 0.0001, by two-tailed t-test). (C) Left panels show representative mechanosensitive inward currents elicited by 6 μm mechanical stimulation in control hTM cells (black), hTM cells treated with 10 μM Yoda1 (green), or hTM cells treated with 2.5 μM GsMTx4 (red). Right panels show quantification of normalized mechanosensitive currents in the presence of Yoda1 or GsMTx4 (n=7-13 cells, ∗∗∗p< 0.001, ∗∗p< 0.01, by two-tailed t-test). (D) Similar to (C), showing representative mechanosensitve inward currents in mTM cells treating with Yoda1 or GsMTx4 (n=7-13 cells, ∗∗∗p < 0.001, ∗p < 0.05, by two-tailed t-test). (E) Left panels show inactivation current traces following 6 μm mechanical stimulation in the absence (black) or presence (green) of Yoda1 in hTM cells. Right panels show time constants of mechanically-activated current inactivation in control or Yoda1 treated cells (n=6-12 cells, ∗∗∗p< 0.001, by two-tailed t test). (F) Similar to (E), showing time constants of current inactivation in control or Yoda1 treated mTM cells (n=6-12 cells, ∗∗p< 0.01, by two-tailed t test). hTM, human trabecular meshwork cells; mTM, mouse trabecular meshwork cells. The dotted line indicates the zero current level.

We next used the Piezo1 antagonist GsMTx4 (2.5 μM), observing a significant reduction of Piezo1 current amplitude (57.5% ± 7.1% decrease in human primary TM cells and 59.6% ± 7.1% decrease in mouse primary TM cells; Figures 2C and 2D) after mechanical indentation. We note that the delivered concentrations of GsMTx4 (2.5 μM) were in the effective range, because GSMTx4 inhibits Piezo1 at micromolar concentrations (Bae et al., 2011). We further down-regulated Piezo1 expression by infecting with an Ad5 virus carrying Piezo1 short hairpin RNA (shRNA) (shRNA-Piezo1) and green fluorescent protein (GFP; Figure S5), and chose the successfully infected HTM cells, i.e., GFP-positive cells, to record mechanically activated (MA) current by whole-cell patch-clamp recordings (Figure S6). As shown in Figure S6, Piezo1 knockdown led to a significant (66.0% ± 7.3%) decrease in MA current amplitude due to 6-μm mechanical indentation. As Piezo1 was not completely knocked down by Ad5-shRNA-Piezo1 infection (Figure S5C), we can state that complete Piezo1 knockdown would cause an MA current reduction of more than 66%. These results demonstrate that the Piezo1 channel is functional in both human and mouse primary TM cells.

GsMTx4 leads to a significant decrease in steady-state aqueous outflow facility

To examine the role of mechanosensitive channels (MSCs) in regulating conventional outflow, we delivered GsMTx4, an inhibitor of cationic MSCs, into the anterior chamber of mouse eyes and quantified its effect on aqueous outflow facility. Outflow facility is the numerical inverse of tissue hydrodynamic flow resistance and is the most important measure of TM function. GsMTx4 (3.3 or 10 μM) was perfused unilaterally, whereas the contralateral eye received control (vehicle) solution. Importantly, our methodology measures quasi-steady outflow facility, i.e., the facility after the eye has equilibrated in response to a pressure perturbation. Although inhibition of MSCs by GsMTx4 (3.3 μM) did not change facility significantly (control: versus treated: , , ), GsMTx4 at a higher concentration (10 μM) significantly decreased the steady-state facility to from (, ; Figure 3). As shown in a “cello plot,” the facility in 10 μM GsMTx4-treated eyes changed on average by −56% [-67, −41] relative to control (Figure 3). We note that drugs delivered into the anterior chamber mix with AH already present within the eye, so that the drug concentration delivered to the meshwork is less than the perfused concentration. Our calculations indicate that, over the timescale of a perfusion experiment, the drug concentration delivered to the TM is approximately half of that perfused into the eye (data not shown). Thus we chose to deliver GsMTx4 at concentrations of 3.3 and 10 μM to achieve micromolar concentrations at the target tissue. We conclude that suppression of MSC function by GsMTx4 has a significant effect on AH outflow.

Figure 3

GsMTx4 significantly reduced steady-state outflow facility

(A) Representative flow-pressure curves from two pairs of enucleated mouse eyes receiving GsMTx4 (left: 3.3 μM; right: 10 μM; red) and the correspondingly contralateral vehicle-treated control eyes (blue). Data points are fitted with an existing relationship (see text), with the shaded region showing associated 95% confidence limits (Sherwood et al, 2019).

(B) Left: “Cello plot” of steady-state outflow facility in GsMTx4 (3.3 μM)-injected eyes and their contralateral controls. Each data point shows the reference facility, in one eye, with error bars showing standard deviation. Shaded regions indicate the best estimates of the sample distribution, with the geometric mean and two-sigma shown by the thick and thin horizontal lines, respectively. Dark central bands show the 95% confidence interval on the mean. No significant difference was detected due to 3.3 μM GsMTx4 (control: vs. treated:,). Middle: “Unity plot”, showing reference outflow facility in each 3.3 μM GsMTx4-treated eye cross-plotted against facility in the contralateral control eye. Each data point represents one pair of eyes. Filled ellipses indicate 95% confidence intervals from the regression fitting, and outer ellipses indicate additional uncertainty due to hardware noise. The unity line is shown in blue and a linear regression through the data points is shown in red, with its 95% confidence interval in grey. Right: Plot of the relative difference between treated and contralateral control eyes, showing that 3.3 GsMTx4μM led to only a mild (29% [-14, 94]) change in facility that was not statistically different from zero.

(C) Similar to panel (B), showing the effects of 10 μM GsMTx4. Treated eyes exhibited a significantly lower facility than contralateral control eyes (vs., which corresponded to approximately a -56% [-67, -41] reduction. ∗∗∗p < 0.001, by two-tailed t-test on log-transformed data.

Piezo1 activation did not influence steady-state aqueous outflow facility

Because GsMTx4 is active against several MSCs, the earlier results do not specifically demonstrate a role for Piezo1 in influencing AH dynamics, and we thus wished to more specifically perturb the function of Piezo1 in conventional outflow tissues. To this end, a specific agonist for Piezo1, Yoda1, was perfused into mouse eyes at nominal concentrations of 20 or 40 μM, and the steady-state facility was measured. As before, the contralateral eye was used as a paired control. As Yoda1 is hydrophobic, we were concerned that it might bind to tubing in the iPerfusion system, leading to non-effective delivery of Yoda1 to the eye. Thus, solution delivered through 20 cm of polyethylene (PE) tubing, similar to that used in the iPerfusion system, was collected and analyzed by mass spectrometry. Yoda1 was detected in both samples, indicating that Yoda1 can be delivered to the eye in perfusion experiments (Figure S4). Furthermore, we were concerned that Piezo1 in TM cells may not be abundant. However, Syeda et al. (2015) showed that Piezo1 activity can be monitored using calcium-sensitive fluorophores and that Yoda1 administration alone induces robust Ca2+ responses in cells with abundant Piezo1. We thus measured Yoda1-induced intracellular Ca2+ level in TM cells, finding that Yoda1 administration alone induced robust Ca2+ responses in TM cells (Figure S7), suggesting the presence of abundant Piezo1 in these cells. However, as shown in Figure 4, Yoda1 (20 or 40 μM) did not affect steady-state outflow facility ( for 20 μM Yoda1 and for 40 μM Yoda1). Moreover, Yoda1 had no significant effect on the flow-pressure nonlinearity parameter (data not shown). Even after accounting for intracameral drug dilution effects (see above), these observations demonstrate that Piezo1 activation does not play a role in directly influencing steady-state AH outflow.

Figure 4

Activation of Piezo1 by Yoda1 did not influence steady-state AH outflow facility

(A) Two sets of representative flow-pressure curves from pairs of enucleated mouse eyes receiving Yoda1 (left: 20 μM; right: 40 μM; red) and their correspondingly contralateral vehicle-treated control eyes (blue). Data points are fitted with an existing relationship (see text), with the shaded region showing associated 95% confidence limits.

(B) Yoda1 (20 μM) treatment did not lead to a significant change in steady-state facility when compared with vehicle control (versus, , ).

(C) Similar to (B), for 40 μM Yoda1 ( versus, , ). p values were calculated based on a two-tailed t test. Refer to Figure 3 for detailed interpretation of plots.

GsMTx4 affects the eye’s response to a pressure perturbation

As Piezo1-induced MA currents are rapid and transient (Figure 2), we next investigated the role of Piezo1 in the eye's rapid response to pressure changes, quantified through ocular compliance, a quantity that is based on transient pressure-flow data after an IOP perturbation (Figure 5A). Interestingly, GsMTx4 (10 μM) increased ocular compliance ( versus , , ; Figure 5B). However, 20 μM Yoda1 showed no significant effect on ocular compliance ( versus , , ; Figure 5C). These observations indicate that MSCs affect the eye's transient pressure-flow response after a pressure perturbation, but that we could not detect a specific role for Piezo1 in this response.

Figure 5

GsMTx4 significantly increases ocular compliance

(A) Two sets of representative compliance-pressure curves generated from eyes receiving either 10 μM GsMTx4 (left, red) or 20 μM Yoda1 (right, red), and the corresponding contralateral vehicle control eyes (blue). Data points are fitted with an existing relationship (see text) and shaded regions show 95% confidence bounds.

(B) 10 μM GsMTx4 led to a significant increase in ocular compliance vs. vehicle control (vs.,).

(C) 20 μM Yoda1 did not change ocular compliance vs vehicle control (vs.,). ∗∗p < 0.01. p values were calculated based on a two-tailed t test. Refer to Figure 3 for interpretation of plots.

Discussion

The TM and inner wall of SC, centrally involved in determining IOP and hence of great interest in the study of glaucoma, experience large mechanical deformation and are known to be mechanosensitive. Several mechanosensing systems/pathways responders have already been identified in these tissues, including integrin deformation (Filla et al., 2017), nitric oxide signaling (Cavet et al., 2014), caveolin signaling (Elliott et al., 2016), and mechanosensitive ion channels (Tran et al., 2014). Mechanosensitive ion channels are of interest due to their ability to respond very rapidly to mechanical stimuli (Suchyna, 2017), important in view of the dynamic mechanical environment within the eye due to blinks, saccades, and the ocular pulse (Turner et al., 2019).

Recent investigations have demonstrated a functional role for other stretch-activated ion channels in the outflow pathway, including TRPV4's role in regulating calcium homeostasis and TM cytoskeletal remodeling (Carreon et al., 2017; Ryskamp et al., 2016; Tran et al., 2014) and the role of TREK1 in influencing TM tensile homeostasis and extracellular matrix components (Yarishkin et al., 2019). Here we extend these earlier findings to study the fast-acting Piezo1 channel. We observed that Piezo1 was expressed in human and mouse iridocorneal angle tissues, including the TM, SC, ciliary muscle, and ciliary body (Figures 1 and S1), all of which are involved in AH dynamics (Borras, 2003; Civan and Macknight, 2004; Crawford and Kaufman, 1987; Johnson et al., 1992; Stamer et al., 2015). We focused on the TM due to its critical role in mediating AH outflow and maintaining IOP homeostasis (Goel et al., 2010; Tran et al., 2014; WuDunn, 2009). Not only was Piezo1 expressed in TM but also our electrophysiological findings showed that it is a functionally active and rapid mechanosensor in the TM, with MA currents induced within microseconds that were strongly influenced by Piezo1 shRNA (Figures S5 and S6), GsMTx4, and Yoda1 (Figure 2). Furthermore, when challenged by a physiologically appropriate cyclic mechanical stretch, both Piezo1 and TRP family members were upregulated in TM cells within hours (Figure 1F). These results demonstrate that ion channels, including Piezo1, are functional within TM cells and thus have the potential to be mechanotransducers of AH dynamics.

Surprisingly, we did not, however, find that perturbation of Piezo1 affected steady-state aqueous outflow facility. Specifically, we used the iPerfusion system, which can monitor microvolumetric changes of AH dynamics in the mouse eye (Sherwood et al., 2016), to investigate the effects on conventional outflow of antagonists and agonists against Piezo1 (Figure 3). As described in a newly published study (Yarishkin et al., 2020), we also observed a facility effect of GsMTx4, but unfortunately this inhibitor of cationic ion channels (Bae et al., 2011) can act on many channels, such as TRPs, ASICs, G-protein receptors, and Nox2 (Inoue et al., 2009; Kurima et al., 2015; Suchyna, 2017; Sukharev et al., 1993). For example, TRP channels in muscle were investigated by using GsMTx4 to inhibit whole-cell currents (Friedrich et al., 2012) and GsMTx-4 peptide was used to block the activation of TRPC6 channels (Spassova et al., 2006). Regarding our electrophysiological findings in TM cells treated with Piezo1 shRNA (Figures S5 and S6), Piezo1 was primarily responsible for inducing MA currents in TM cells, because when Piezo1 was partially knocked down using shRNA, MS currents were reduced by 66%. We infer that GsMTx4, which induced a 57% MA current reduction, acts by interfering with Piezo1 activity. We thus cannot unambiguously determine the role of Piezo1 in AH outflow from these results. Therefore, we next used Yoda1, a specific agonist for Piezo1, yet did not observe any change in steady-state outflow facility due to Yoda1 (Figure 4).

In view of the abundance of Piezo1 in TM, this finding was surprising, and we considered several possible explanations. One possibility is that Yoda1 at nominal concentrations of 20 and 40 μM functions ineffectively in the eye due to absorption of this hydrophobic compound to tubing upstream of the eye, a particularly critical concern in view of the small flow rates and volumes used when perfusing mouse eyes, and the dilution issue in the anterior chamber. However, mass spectrometry showed that Yoda1 did successfully pass through tubing similar to that used in the ocular perfusion experiments (Figure S4). Furthermore, our electrophysiological results (Figure 2) showed that Yoda1, at nominal concentrations of 10 μM, delivered through PVC-based tubing effectively activated Piezo1 in the TM cells. As the PE-based tubing used in our ocular perfusion system has relatively low adsorption to hydrophobic drugs (Jin et al., 2017; Syeda et al., 2015), we conclude that the absence of effects seen with Yoda1 were likely not due to delivery issues into the eye. Furthermore, we observed Yoda1-increased MA currents (Figure 2) and Yoda1-induced intracellular calcium levels (Figure S7), confirming that Yoda1 (carrier-free, Sigma-Aldrich, SML 1558) is able to cause a downstream reaction.

A second possibility relates to the timescale of our measurements. We hypothesized that Piezo1 is specifically responsible for rapid adjustments in aqueous outflow, consistent with its rapid response in cultured TM cells (Figure 2). Recalling that facility data described above were obtained under quasi-steady conditions, we therefore analyzed transient pressure-flow data in mouse eyes, as quantified through ocular compliance. Usually ocular compliance is thought to depend exclusively on the biomechanical properties of the corneoscleral shell, but can also be affected by transient changes in aqueous outflow during the measurement period (Sherwood et al., 2019). We found that GsMTx4 significantly increased ocular compliance (Figure 5), but Yoda1 did not. Previous investigations have suggested that Piezo1 can significantly regulate extracellular matrix and reinforce tissue stiffening (Chen et al., 2018), which could affect ocular compliance through an effect on the sclera. However, this seems implausible on the timescale of our experiments, and thus changes in outflow dynamics mediated by MSCs are a more likely explanation for the effect of GsMTx4 on ocular compliance. Despite these findings, our observation of dynamic changes in outflow appears not to be mediated by Piezo1, because Yoda1 did not affect ocular compliance. However, this may be due to a limitation of our measurement capabilities: the transient analysis of flow-pressure data that we report requires data collection over a timescale of several minutes, which is much longer than the timescale for Piezo1-derived MA currents (<1 s, Figure 2). A more complete elucidation of Piezo1's role in the TM will require development of functional assays with response times of order a second or less, comparable with the timescale of blinks, saccades, and the ocular pulse.

A final possible explanation for our findings is that activation of Piezo1 in the TM of normal mice, as used in our study, is essentially maximal during facility measurement. In this situation, Yoda1 would not be able to further increase Piezo1 activity and thus would not influence facility. It is possible that the effects of Yoda1 would only become evident under pathological conditions, such as occur in ocular hypertension, and conducting such studies would be of great interest.

In summary, our results confirm the role of mechanosensitive ion channels in regulation of conventional steady-state AH outflow. Piezo1, as one such channel, is widely expressed and is functional in TM, yet its specific activation appears to have little effect on quasi-steady outflow facility. These results suggest that Piezo1 may be important in mechanoregulation of AH dynamics in pathological situations and in response to very rapid ocular changes, such as commonly occur due to ocular saccades and other events.

Limitations of the study

Our study demonstrates the potential role for Piezo1 in conventional outflow, likely under pathological and rapid transient conditions. However, the transient analysis of flow-pressure data that we report requires data collection over a timescale of several minutes, which is much longer than the timescale for Piezo1-derived MA currents (<1 s). A more complete elucidation of Piezo1's role in the TM will require development of functional assays with response times of order a second or less, comparable with the timescale of blinks, saccades, and the ocular pulse.

Resource availability

Lead contact

Further requests for resources and materials should be direct to the Lead Contact, Wei Zhu (wzhu@qdu.edu.cn).

Materials availability

Requests for materials should be direct to the Lead Contact.

Data and code availability

The raw data are available on Mendeley Data (https://doi.org/10.17632/5v2mbp2sdx.1).

Methods

All methods can be found in the accompanying Transparent methods supplemental file.