TRPV4 regulates calcium homeostasis, cytoskeletal remodeling, conventional outflow and intraocular pressure in the mammalian eye

An intractable challenge in glaucoma treatment has been to identify druggable targets within the conventional aqueous humor outflow pathway, which is thought to be regulated/dysregulated by elusive mechanosensitive protein(s). Here, biochemical and functional analyses localized the putative mechanosensitive cation channel TRPV4 to the plasma membrane of primary and immortalized human TM (hTM) cells, and to human and mouse TM tissue. Selective TRPV4 agonists and substrate stretch evoked TRPV4-dependent cation/Ca2+ influx, thickening of F-actin stress fibers and reinforcement of focal adhesion contacts. TRPV4 inhibition enhanced the outflow facility and lowered perfusate pressure in biomimetic TM scaffolds populated with primary hTM cells. Systemic delivery, intraocular injection or topical application of putative TRPV4 antagonist prodrug analogs lowered IOP in glaucomatous mouse eyes and protected retinal neurons from IOP-induced death. Together, these findings indicate that TRPV4 channels function as a critical component of mechanosensitive, Ca2+-signaling machinery within the TM, and that TRPV4-dependent cytoskeletal remodeling regulates TM stiffness and outflow. Thus, TRPV4 is a potential IOP sensor within the conventional outflow pathway and a novel target for treating ocular hypertension.

Agonist-evoked calcium signals typically spread from the plasma membrane to the cell nucleus and remained elevated in the presence of the agonist. To determine the potential contribution of Ca 2+ -induced Ca 2+ release (CICR), we depleted endoplasmic reticulum Ca 2+ with cyclopiazonic acid (CPA; 5 μ M), an inhibitor of sarcoplasmic-endoplasmic reticulum ATPases (SERCAs). CPA alone elevated [Ca 2+ ] TM as Ca 2+ was released/ leaked from the stores but not pumped back in (n = 109 cells; P < 0.01). After store depletion by CPA and cytosolic Ca 2+ clearance by plasma membrane pumps/exchangers, the amplitude of GSK101 signals was reduced by 39.3 ± 7.99% (P < 0.05) (Fig. 3m). Thus, as observed in other cell types 22,26,27 , Ca 2+ release from internal stores amplifies TM plasmalemmal signals mediated by TRPV4.
Endogenous TRPV4 activation was proposed to involve phospholipase A2 (PLA2)-dependent generation of arachidonic acid (AA) and its eicosanoid metabolites 17 . Not all TRPV4-expressing cells utilize this 'canonical' mechanism for TRPV4 activation 19,28 ; yet, the PLA2 pathway is strongly expressed in the TM 29 and may play a role in glaucoma 30 . To test whether PLA2 signaling plays a role in TRPV4 function, we exposed cells to AA and blockers of cytochrome P450 (CYP450), the downstream enzyme that generates final eicosanoid activators of TRPV4 17,22 . 50 μ M AA (n = 159 cells) elevated [Ca 2+ ] hTM to a extent comparable to 25 nM GSK101 (n = 244 cells; Fig. 3n). AA-evoked calcium responses were blocked by HC-06 (1 μ M; P < 0.001; n = 94 cells) and by the pan-CYP450 inhibitor clotrimazole (CTZ, 10 μ M; n = 134 cells). Thus, TM TRPV4 channels are potently stimulated by the canonical PLA2 pathway.
Given that IOP-dependent increases in TM stiffness and contractility involve actin polymerization 9,10 , which is regulated by stretch and calcium ions 12,14 , we hypothesized that stretch-induced cytoskeletal remodeling might involve TRPV4. Consistent with this, substrate stretch potently stimulated the formation and thickening of phalloidin-Alexa488-tagged F-actin stress fibers (Fig. 4d-f). Stretch also increased phosphorylation of focal adhesion kinases (FAKs; Fig. 4g) and induced reorganization of vinculin, a key FA regulator shows TRPV4-ir within patches of the plasma membrane, whereas the cilium is immunonegative for TRPV4. Scale bar = 10 μ m. (b) Images from hTM (i) and pTM (ii, iii) cells show that TRPV4 may be localized to the ciliary base in a subset of TM cells (arrowheads). Scale bars = 50 μ m (i), 10 μ m (ii) and 5 μ m (iii). (c,d) Chloral hydrate (CH; 4 mM; 24 hrs) reduced the number and length of primary cilia within hTM cells and uncoupled ciliary bridges between adjacent cells. (e) Brightfield images of hTM cell morphology indicate that chloral hydrate is not TM-toxic. Scale bars = 20 μ m. (f,g) CH post-peak [Ca 2+ ] i plateau phase but has no effect on the amplitude of GSK101-evoked [Ca 2+ ] i signals.
We tested whether TRPV4 activation alone is sufficient for mimicking the effect of stretch on cytoskeletal remodeling by exposing the cells to GSK101 (5 nM; 1 hr). As illustrated in Fig. 4h,i, exposure to the TRPV4 agonist triggered stress fiber and FA remodeling that mirrored cytoskeletal changes induced by mechanical strain and was likewise antagonized by HC-06 (P < 0.05; N = 3 experiments; 3-5 slides/experiment). Simultaneous recording of [Ca 2+ ] i and mApple-actin dynamics confirmed that TRPV4-induced stress fiber formation takes place during the Ca 2+ elevation (Fig. 4j,k and Supplementary Video S1).
TRPV4 inhibition enhances outflow facility and lowers perfusate pressure in 3D pTM cell cultures. Convenience and accessibility encourage studies of TM cell behavior on 2D surfaces; however, in vivo these cells experience 3D microenvironments that alter many aspects of cell behavior, including mechanical constraints. To investigate the potential role of TRPV4 in the regulation of outflow, we used bioengineered   By showing that TRPV4 activity regulates perfusion pressure-dependent remodeling of the TM, outflow facility and perfusate pressure, these data suggest that inhibition of TRPV4 activity in the TM might promote drainage by the conventional outflow pathway.

TRPV4 channel activation contributes to pathological increases in IOP.
To understand the pathophysiological relevance of our in vitro observations in eye disease, we investigated the effects of pharmacological and genetic targeting of TRPV4 channels in a widely used mouse model of acute glaucoma 41 . Injecting the anterior eye with polystyrene microbeads (MBs) to obstruct fluid drainage (Fig. 6a,b) elicited a modest IOP increase in mouse eyes (to 20.04 ± 1.24 mm Hg), whereas intracameral injections of the vehicle (PBS) in the contralateral eye had no effect on IOP (9.8 ± 0.26 mm Hg) (Fig. 6b,c). Following IOP elevation, animals from MB-and PBS-treated cohorts were randomly assigned to control or treatment groups for daily intraperitoneal (IP) injections of PBS or HC-06 (10 mg/kg) 25 . Systemic injection of the TRPV4 antagonist lowered IOP to baseline levels (P < 0.0001; 2-way repeated measures ANOVA followed by Holm-Šídák tests in Fig. 6c; 2-way ANOVA followed by Tukey's test in Fig. 6c). Daily HC-06 treatment maintained low IOP levels in MB-treated eyes, which remained indistinguishable from vehicle-treated eyes. The IOP in age-matched, uninjected, WT eyes (12.84 ± 0.89 mm Hg; n = 25) was similar to the IOP in Trpv4 −/− eyes (12.40 ± 0.88 mm Hg; n = 50; p > 0.05), indicating that TRPV4 activity during physiological conditions is not a major regulator of IOP. MB microinjection failed to elevate IOP in 4/4 Trpv4 −/− animals above levels in PBS-injected eyes (Fig. 6d) showing that Trpv4 −/− eyes are resistant to pressure elevation induced by partial angle blockage. These data indicate that overactivation of the TRPV4 channel is necessary to sustain ocular hypertension.
We further tested the efficacy of HC-06 by injecting it (in 1.85% DMSO and PBS, 2 μ l) directly into the anterior chamber of eyes with elevated IOP from MB injections (Fig. 6e,f). Single intraocular HC-06 injections induced a dramatic lowering of IOP that lasted for days (Fig. 6e), whereas injection of the vehicle did not lower IOP ( Fig. 6f; P < 0.0001; Holm-Šídák tests). To test whether IOP can be controlled through topical delivery, we designed and synthesized a putative prodrug analog (YX-02) of the antagonist HC-06 (Fig. 6g). Confirming efficacy as a TRPV4 antagonist, YX-02 inhibited 25 nM GSK101-induced [Ca 2+ ] hTM elevations with an IC50 of 0.74 ± 0.04 μ M (Fig. 6 g). A single eye drop of YX-02 decreased IOP to baseline levels for up to 24 hrs (Fig. 6 h; P < 0.01 to P < 0.0001; Holm-Šídák test). HC-06 or YX-02 administration evoked no obvious ocular inflammatory or toxicity response. 8 weeks after initiation of the IOP elevation protocol, wholemount counts of TuJ-1 + cells in the retinal ganglion cell layer (RGCL)/mm 2 were 3425.1 ± 251.3 in vehicle-treated eyes and 2370.1 ± 72.8 in MB-treated eyes, significantly lower (p < 0.05). MB-injected eyes of mice treated daily with systemic HC-06 were protected from the deleterious effects of IOP-elevation across all retinal quadrants (p < 0.01; Fig. 6i,j).

Discussion
Mechanical stimuli impact every cell in the vertebrate eye and sensitivity to force is essential for visual function. However, the mechanisms by which ocular cells sense and transduce mechanical forces are incompletely understood, affecting our insight into the physiology of IOP homeostasis and treatment of glaucoma. This question is particularly compelling with respect to the conventional outflow pathway because -possibly as a result of structural changes caused by mechanical stress -TM resistance to fluid flow increases dramatically during pathological IOP elevations [2][3][4][5] . In this study, we present evidence that TRPV4 is critically involved in the transduction of mechanical stress in TM cells and link overactivation of this force-sensitive channel to maintaining elevated IOP in a mouse model of glaucoma.
Emerging evidence supports a role for Ca 2+ -initiated signals in increased resistance of the conventional pathway to fluid outflow 9,13 ; however, neither the identity of the Ca 2+ channel involved nor the mechanism by which Ca 2+ signals promote these effects have been fully elucidated. We demonstrate the presence of Trpv4/TRPV4 transcripts, protein, currents and/or Ca 2+ signals in primary and immortalized cells from human TM preparations and in ex vivo mouse and human tissue. Our evidence suggests that TRPV4 mediates force-induced elevations in [Ca 2+ ] TM and reorganization of mechanically stressed TM. Consistent with this, force application to TM cells evoked sustained, stretch-dependent Ca 2+ elevations that were mimicked by GSK101 and suppressed by HC-06.
Mechanical stretch of TM cells triggered the formation, thickening and cross-linking of F-actin stress fibers, phosphorylation of FAK, and vinculin recruitment to focal adhesions. While this response might involve additional force sensors 42 , the sensitivity of stretch-induced cytoskeletal reorganization and/or [Ca 2+ ] i signals to BAPTA-AM, PLA2 blockers and HC-06 clearly shows that these processes involve TRPV4-dependent influx of Ca 2+ . Our data thus identify a novel Ca 2+ source which drives actomyosin remodeling associated with increased TM stiffness/contractility in response to mechanical stress 9,14,43 , whereas Ca 2+ -dependent proteins (such as K + and Cl − channels and/or myocilin) might exert additional modulating influence. Similar effects of TRPV4 activation on actin and contractility have been reported in other cell types that regularly experience hydrostatic pressure and cyclic strain [44][45][46] . We hypothesize that TRPV4-mediated Ca 2+ signals and/or TRPV4 coupling to actin 14,47 , integrin:collagen, β -catenin:E-cadherin 44 and/or TRPV4:RhoA/Rho kinase mechanisms 48,49 regulate the permeability of juxtacanalicular TM [46][47][48][49] . By analogy with similar roles for the channel in epithelial and endothelial tissues 24,[44][45][46]48 , we hypothesize that inhibition of TRPV4-induced [Ca 2+ ] i elevations suppresses the transfer of force from integrins to focal adhesions and actin, and disrupts stretch-induced polymerization of actin. Accordingly, we discovered that TRPV4 inhibition lowers the density of actin stress fibers (Figs 4 and 5), weakens β -catenin-containing cell-cell contacts and increases paracellular permeability 44 . These results suggest that the TM mechanosensor is integrated into dynamic modulation of stress fibers, which have been proposed to underlie increased TM rigidity, contractility and resistance to aqueous outflow 9,10,14 .
The hypothetical mechanism depicted in Fig. 7 shows TRPV4 as a central mediator of stretch-and Ca 2+ -dependent gene/cytoskeletal remodeling in TM cells. We propose that TRPV4 stimulation drives the upregulation of crucial ECM components (collagen and fibronectin), linking diverse types of mechanical stress (IOP, swelling, stretch) to the internal cellular strain-sensing machinery (based on integrins and focal adhesions), and release of AA 50 . This model is supported by our discovery that TRPV4 activation upregulates SF formation and secretion of fibronectin while TRPV4 inhibition prevents strain-induced [Ca 2+ ] i elevations and TM remodeling, and reduces perfusate pressure in 3D scaffolds bioengineered with human TM. The dramatic IOP lowering induced by topical or systemic application of TRPV4 blockers in our animal glaucoma model was achieved by breaking the link between TRPV4 activation and activation of downstream calcium-dependent signaling pathways. Given that TRPV4 overactivation contributes to two other key hallmarks of glaucoma pathology -RGC injury and reactive gliosis, both of which are reduced when retinas are subjected to systemic TRPV4 inhibition 21-23 , we propose that therapy using TRPV4 antagonists might achieve the hitherto unachievable trifecta in glaucoma treatment by controlling IOP protecting RGCs from mechanical stress and suppressing inflammatory activation of retinal glia.
It is important to compare our findings with evidence that baseline IOP is slightly elevated in Trpv4 −/− mouse eyes and that activation of TRPV4, confined to the ciliary membrane of pTM cells, lowers basal IOP in mice and the Wpk rat 31 . Our data in contrast suggest that resting IOP is comparable in WT and Trpv4 −/− mouse eyes and is unaffected by intraocular injections of GSK101 24 , whereas pharmacological inhibition and genetic disruption of TRPV4 were remarkably effective in lowering IOP during partial outflow obstruction. The main implication of these results is that TRPV4 inhibition enhances the outflow facility when outflow resistance is increased. We hypothesize that this facilitation effect is due to decreased cytoskeletal/FA/ECM remodeling that is downstream from pressure-induced TRPV4 activation. Our data also suggest that the conventional outflow pathway retains enough hydraulic conductivity (possibly in collaboration with uveoscleral outflow) to maintain fluid transport in the presence of microbeads and HC-06. Although we cannot exclude the possibility that the dramatic IOP lowering observed in vivo included HC-06-dependent facilitation of uveoscleral drainage, hydraulic conductivity of the Schlemm's canal or attenuated aqueous production, the dramatic potentiation of the outflow facility observed in biomimetic pTM-populated scaffolds demonstrates that suppression of the putative TM mechanosensor is in itself sufficient to stimulate outflow facility.
Importantly, our data demonstrate that both mechanical stress and TRPV4 activation strengthen the cells' intrinsic tensile apparatus composed of actin stress fibers, focal adhesions and extracellular matrix. Because pressure-induced increases in TM resistance to aqueous outflow require cell stiffening and actin upregulation 14,43,[51][52][53][54] , it is possible that targeting TRPV4 counters the effect of mechanical stress by 'protecting' the downstream transduction mechanism. Cyclic stretch increased secretion of fibronectin in pTM-populated nanoscaffolds, in pulmonary cells (which continually experience) tensile stretch 55 and in glaucomatous TM 7 . We conjecture that in vivo IOP elevations impose tensile strain onto ECM, inducing influx of Ca 2+ through stretch-sensitive TRPV4 channels. The resulting increase in [Ca 2+ ] i triggers changes in gene expression, reorganization of the TM cytoskeleton and secretion of ECM that ultimately impede fluid passage through the delicate meshwork formed by TM processes (Fig. 7). Because TRPV4-incompetent cells lack the mechanism that couples force to the cytoskeletal/matrix they cannot sustain IOP elevations in response to microbead injection. Consistent with our model, elevated pressure disproportionally increases cytosolic Ca 2+ levels 11,13 and induces TM stiffening in POAG TM cells 9,10,43,54 . Stretch tended to be a more effective facilitator of stress fiber formation than GSK101 even when both stimuli were applied at a ~half-maximal dose (6% stretch vs. 25 nM) for [Ca 2+ ] i elevations. Because HC-06 and BAPTA-AM were equally effective in antagonizing the effects of stretch/TRPV4 agonists on actin remodeling, we propose that the former activates auxiliary (yet TRPV4-dependent) downstream signaling mechanisms (possibly involving the Rho signaling pathway), which are less effectively stimulated by chemesthesis alone. This is consistent with the observation that TRPV4-activating stimuli promote channel opening through different amino acid residues of the protein 18,56 .
While our report suggests a critical role for TM TRPV4 in sustaining IOP elevations, other ocular sites may also be impacted by TRPV4 antagonists or channel deletion. For example, HC-06 could affect the function of nonpigmented epithelial cells of the ciliary body 24,57 , corneal endothelial cells 58 , retinal ganglion cells 23 as well as inflammatory signals mediated through glial cells 22,28 ; however, the absence of antagonist/agonist/ knockout effects on baseline IOP argues against a major role for TRPV4 in steady-state fluid secretion or drainage. An important question to be addressed in future studies is whether and how these TRPV4-expressing ocular tissues might respond to physiological fluctuations in IOP induced by saccades, changes in Figure 7. Model of trabecular meshwork signaling in response to mechanical stress. Mechanical stress (e.g., pressure, swelling, and tissue distension) stretches the plasma membrane and activates TRPV4 and a Ca 2+ -and stretch-sensitive phospholipase A2 (PLA2). The product, arachidonic acid (AA), is a substrate for cytochrome P450, which drives synthesis of eicosanoid metabolites (EETs), the final activators of TRPV4. Stretch might activate PLA2 simultaneously with TRPV4; alternatively, stretch-induced TRPV4 activation could stimulate Ca 2+ -dependent PLA2s which amplify the initial TRPV4 signal (horizontal blue arrows). TRPV4 activation may be additionally augmented by TRPV1 channels and/or cell swelling, mediated through aquaporin 1 (AQP1) channels [67][68][69] . Stretch regulates the outflow resistance through Ca 2+ -dependent actin remodeling, focal adhesion stabilization, actomyosin contractility, fibronectin production and TM gene expression. Acting in concert, these signaling components account for many known aspects of the TM response to mechanical and inflammatory challenges, including the response to the response to chronic IOP elevations. posture, etc. TRPV4 channels could also play a role in the modulation of aqueous humor transfer through transcellular or paracellular routes in the inner wall of Schlemm's canal, possibly by regulating tight/adherens junctions of cells which functionally resemble TRPV4-containing vascular endothelia 8,46,59 . However, our data show that primary cilia are not required for TRPV4 signaling in immortalized or primary human TM cells, and are thus consistent with evidence against a role for primary cilia in cytosolic Ca 2+ homeostasis 35 and TRPV4 signaling 33,60 . Antibody staining, optical imaging and use of deciliating agents showed that the large majority of TRPV4 channels are localized to the TM cell plasma membrane and may be functionally coupled to CICR and, possibly, store-operated calcium signaling 61 .
Our observations that PLA2 antagonists inhibit stretch-induced Ca 2+ signals and that TRPV4 blockers suppress AA-induced [Ca 2+ ] TM increases suggest that the channel is activated through the canonical pathway. It remains to be determined whether PLA2 functions as the primary force sensor or a TRPV4-mediated rise in [Ca 2+ ] TM might stimulate translocation, membrane binding and phosphorylation of cPLA2, which cleaves phospholipids at the sn-2 position to generate AA, the precursor of eicosanoid activators of TRPV4 (Fig. 7). The involvement of proinflammatory eicosanoids could, over time, exacerbate the pathological response to mechanical stress. It is thus of interest that PLA2 is conspicuously upregulated in POAG TM 29 and that mutations in the TM-specific B1 isoform of cytochrome P450 were associated with early-onset glaucoma 30 . Further consistent with our model, PLA2 blockers mirror the effect of TRPV4 antagonists on IOP by inhibiting actin stress fiber formation and linkage to the ECM 62 .
In summary, our data bring together previously unrelated aspects of force sensing by TM cells and place them within a novel mechanistic framework that may provide insight into the molecular linkage between mechanotransduction, Ca 2+ homeostasis and reorganization of the conventional outflow pathway. Given that IOP reduction represents a highly effective treatment of both hypertensive and normal-tension versions of the disease 1 , the novel prodrug analog described here may provide safe and effective glaucoma therapy by targeting the pressure sensor, a major currently unmet need in the clinical treatment of glaucoma 2 . The potential for protecting RGCs from mechanical stress by TRPV4 inhibition 21,63 offers additional impetus to explore possible solutions for combined IOP-lowering and neuroprotective treatments to prevent vision loss in glaucoma. Animals. C57BL/6J mice were from JAX (Bar Harbor, ME) and Trpv4 −/− mice were obtained from Dr.

Study approval. Experiments followed recommendations of the NIH Guide for Care and Use of Laboratory
Wolfgang Liedtke (Duke University). Mice were maintained in a pathogen-free facility with a 12-hour light/dark cycle and ad libitum access to food and water. Temperature was set at ~22-23 °C. Up to 5 adult mice of the same gender were housed in a single cage. Animals were regularly monitored during and following surgical procedures and were promptly euthanized in rare instances of declining health. Male and female mice were included in the study. No sex differences in the data were noted, so their data were pooled. Mice were 3 to 6 months in age. Sample sizes were based on pilot experiments. An independent co-author randomized group assignment and coded conditions to blind experimenters to conditions involving animals.
TM cell culture and isolation. Primary hTM cells, isolated from the juxtacanalicular and corneoscleral regions of the human eye (ScienCell Research Laboratories; Carlsbad, CA) were grown in Trabecular Meshwork Cell Medium (ScienCell, Catalog#6591) at 37 °C and 5% CO 2 . Confluent cells showed the flattened phenotype that is typical of cultured hTM. To test whether hTM cells exhibit molecular and physiological (steroid sensitivity) characteristics typical of primary TM cultures, we measured expression of α B-crystallin, TIMP3, aquaporin 1 and smooth muscle actin genes and dexamethasone (DEX)-induced expression of myocilin ( Supplementary Fig. S1). Key physiological features (e.g., responses to TRPV4 agonists) were replicated in primary TM (pTM) cells isolated from two donors (35 and 40 years old) and cultured following established protocols. Human corneal rims used to culture primary TM cells were obtained by Dr. Balamurali Ambati (Moran Eye Institute, University of Utah). Human tissues were used in concordance with the tenets of the WMA Declaration of Helsinki and the Department of Health and Human Services Belmont Report. Salts and reagents were purchased from Sigma or ThermoFisher unless specified otherwise. mRNA and protein analysis. Total RNA was extracted from hTM cells using E.Z.N.A total RNA kit (Omega). RNA was reverse transcribed using the XLT cDNA super mix kit (Quanta). Amplified Trpv4 mRNA was referenced to α-tubulin signals for each sample. For immunoblots, hTM cells were dissociated from culture flasks and centrifuged at 2000 rpm for 3 minutes. The cell pellet was washed with PBS and homogenized in 20 mM Tris buffer (pH 7.4) containing: 0.1 M NaCl, 0.2 mM EDTA, 0.2 mM PMSF, 50 mM NaF and the proteinase inhibitor cocktail (Santa Cruz). Protein was separated by 10% SDS-PAGE gels and transferred to polyvinylidene difluoride membranes (Bio-Rad). Membranes were blocked with 5% skim milk in PBS containing 0.1% tween 20 and incubated at 4 °C overnight with primary antibodies against myocilin (rabbit polyclonal, 1:500; Sigma 276-290), FAK (rabbit polyclonal 1:1000; Cell Signaling Technology 3285S) or pFAK-(Tyr397) (rabbit polyclonal 1:1000; Cell Signaling Technology 3283S). Secondary antibodies were conjugated to horseradish peroxidase and visualized using on X-ray film (Thermo Scientific) by using ECL solution (Pierce) and a developer machine (AFP Imaging Corp).
To assess changes in cytoskeletal remodeling in hTM by GSK101, cells were plated on type I collagen-coated glass coverslips for 24 hrs and were then treated with DMSO (control group), 5 nM GSK101, 1 μ M HC-06 or 1 μ M HC-06 and 5 nM GSK101 (HC-06 was added 30 min before GSK101) for 1 hr at 37 °C. To examine cytoskeletal remodeling, cultures were fixed in 4% PFA, washed twice in PBS, permeabilized with 0.1% Triton X-100, washed twice in PBS and exposed to blocking solution. F-actin was labeled with phalloidin-Alexa 488 nm or phalloidin-Alexa 594 nm, either alone or in the presence of another antibody. Cells were washed twice in PBS and mounted with DAPI-Fluoromount-G (Southern Biotech). The distribution of F-actin and vinculin was determined in confocal (Olympus CV1200) stacks using NeoFluor 40x water immersion objectives. Formation of stress fibers was quantified by averaging the overall fluorescence within the field of view of 30 ROIs per experiment. Final data contains averaged data from at least 3 separate experiments. Experimenters were blinded to the condition during imaging and analysis.
Electrophysiology. Macroscopic whole-cell currents were recorded using a Multiclamp 700B amplifier and Cell stretch assay. hTM cells were plated on flexible silicon membranes coated with type I/IV collagen or Pronectin (protein polymer incorporating multiple copies of the RGD attachment epitope from human fibronectin) and grown to 80% confluence. Cells were placed into a FlexJunior chamber controlled by the Flexcell-5000 Tension system (Flexcell, Hillsborough, NC) and stimulated with cyclic biaxial stretch (6%, 2 Hz) for 1 hour at RT. 1 μ M HC-06 or the vehicle were added 1 hr prior stretching 65,66 . To chelate cytosolic Ca 2+ , cells were loaded with 100 μ M BAPTA-AM for 30 minutes prior stretching. Then samples were used to analyze cytoskeletal remodeling as described above. For determination stretch-induced Ca 2+ influx, cells were loaded with Fura-2-AM for 30-60 minutes and then stimulated with cyclic biaxial stretch (2-14%, 2 Hz). A 5 minute stretch period was chosen to optimize capture of the stretch-evoked fluorescent response by adjusting for the change in focal plane (which disrupted calcium imaging for several seconds as indicated by breaks in response trace in Fig. 4a). Cells were imaged with a Nikon E600FN upright microscope. Excitation light was provided by a xenon lamp within a Lambda DG4 (Sutter Instruments) controlled by Nikon Elements.
Bioengineered TM: 3D culture of primary human TM cells on scaffolds. Primary TM cells were isolated from donor tissue rings discarded after penetrating keratoplasty. Before use, all pTM cell strains were characterized for expression of α B-crystallin and α -smooth muscle actin. pTM cells were initially plated in 75 cm 2 cell culture flasks with 10% fetal bovine serum (FBS) (Atlas Biologicals, Fort Collins, CO) in Improved MEM (IMEM) (Corning Cellgro, Manassas, VA) with 0.1 mg/mL gentamicin. Fresh medium was supplied every 48 hrs. Cells were maintained at 37 °C in a humidified atmosphere with 5% carbon dioxide until 80-90% confluence, at which point cells were trypsinized using 0.25% trypsin/0.5 mM EDTA (Gibco, Grand Island, NY) and subcultured. All studies were conducted using cells before the 5th passage. To create 3D pTM constructs, 40,000 pTM cells were seeded on individual microfabricated SU-8 scaffolds placed in a 24-well plate and cultured in 10% FBS-IMEM for 14 days. Medium was changed every 2-3 days. On day 14, 3D pTM constructs were serum-starved for 24 hrs and then treated with GSK101 (25 nM), HC-06 (1 μ M) or the vehicle for 24 hrs.
Perfusion of bioengineered TM. Samples were securely placed in a perfusion chamber 37,38 and media was perfused through the culture/scaffold for 6 hrs per flow rate (2, 4, 8 and 16 μ l/min). Samples were perfused in an apical-to-basal direction with perfusion medium consisting of Dulbecco's modified Eagle's medium (DMEM) (Cellgro) with 0.1% gentamicin (MP) containing GSK101 (25 nM), HC-06 (1 μ M) or the vehicle. The temperature was maintained at 34 °C. Pressure was continuously monitored and recorded. After perfusion, samples were fixed and stained for confocal imaging. The "outflow facility" of our bioengineered 3D pTM model, Δ flow/Δ pressure, was calculated from the inverse of the slope of the pressure versus flow graph per unit surface area. These experiments were performed with the experimenter blinded to the compound identity. Microbead occlusion glaucoma model. Mice were anesthetized with an intraperitoneal (IP) injection of ketamine/xylazine (90 mg/10 mg / kg of body weight) and eye drops were used to numb the eyes (0.5% proparacaine hydrochloride) and dilate the pupils (1% tropicamide ophthalmic solution USP; Bausch & Lomb). IOP was elevated unilaterally as in 64 by injecting 2 μ l of polystyrene microbeads (7.8 μ m FluoSpheres; Bangs Laboratories) with a blunt tip, Hamilton syringe (Hamilton Company) into the anterior chamber after making a guide hole using a 30.5 gauge needle, gently depressing the cornea to displace aqueous humor and drying the eye. Microbeads were injected over 60 s and the needle remained for an additional 60 s before injecting a small bubble to seal the cornea and prevent microbead outflow. The contralateral eye was injected with PBS. 0.5% Erythromycin ophthalmic ointment USP (Bausch & Lomb) was applied after the procedure. Injected eyes were visually examined for MBs at multiple time points during experiments. Following injections, intracameral MBs were stably localized to the angle. This procedure and IOP measurements were performed in a hood within the Moran Eye Center vivarium.

Mouse IOP measurements.
A TonoLab rebound tonometer was used to measure IOP of Avertin-treated ( Fig. 6b-c) and awake mice ( Fig. 6d-f,h) 64 between 10 AM and noon. For awake mice, 0.5% proparacaine hydrochloride was applied prior to measurements. IOP was determined from the mean of 10 to 20 tonometer readings.
In vivo TRPV4 inhibition and generation of prodrugs. HC-067047 (Sigma Aldrich, St. Louis, MO) was dissolved in 1.85% DMSO and PBS, then it or the vehicle was administered via a 10 mg/kg intraperitoneal injection (IP) once daily for 8 weeks. For topical application, the prodrug was identical to HC-06 except an isopropyl ester group replaced the trifluoromethyl group. To make the prodrug (Fig. 6f), 2-methyl-1-(3-morpholinopropyl)-5-phenyl-1H-pyrrole-3-carboxylic acid (200 mg; 0.610 mmol) in 5 mL dichloromethane (DCM) was added thionyl chloride (220 μ L; 3.05 mmol) and dimethylformamide (20 μ l). After stirring for 3 hrs at RT, the reaction mixture was evaporated and dried under vacuum pressure. To the residue was added DCM (5 ml), isopropyl 3-aminobenzoate (218 mg; 1.20 mmol) and N,N-diisopropylethylamine (0.42 ml), and the reaction mixture was stirred overnight. Water was added and the mixture was extracted with DCM. The organic phase was washed with water, brine, dried by anhydrous sodium sulfate, concentrated and purified by chromatography to provide the prodrug YX-02 ( Unless specified, an unpaired t-test was used to compare two means and an ANOVA along with Tukey's multiple comparisons test was used to compare three or more means 22,23,69 . P > 0.05 = NS, P < 0.05 = * , P < 0.01 = * * , P < 0.001 = * * * and P < 0.0001 = * * * * .