Simvastatin Attenuates Glucocorticoid-Induced Human Trabecular Meshwork Cell Dysfunction via YAP/TAZ Inactivation

Abstract Purpose Impairment of the trabecular meshwork (TM) is the principal cause of increased outflow resistance in the glaucomatous eye. Yes-associated protein (YAP) and transcriptional coactivator with PDZ binding motif (TAZ) are emerging as potential mediators of TM cell/tissue dysfunction. Furthermore, YAP/TAZ activity was recently found to be controlled by the mevalonate pathway in non-ocular cells. Clinically used statins block the mevalonate cascade and were shown to improve TM cell pathobiology; yet, the link to YAP/TAZ signaling was not investigated. In this study, we hypothesized that simvastatin attenuates glucocorticoid-induced human TM (HTM) cell dysfunction via YAP/TAZ inactivation. Methods Primary HTM cells were seeded atop or encapsulated within bioengineered extracellular matrix (ECM) hydrogels. Dexamethasone was used to induce a pathologic phenotype in HTM cells in the absence or presence of simvastatin. Changes in YAP/TAZ activity, actin cytoskeletal organization, phospho-myosin light chain levels, hydrogel contraction/stiffness, and fibronectin deposition were assessed. Results Simvastatin potently blocked pathologic YAP/TAZ nuclear localization/activity, actin stress fiber formation, and myosin light chain phosphorylation in HTM cells. Importantly, simvastatin co-treatment significantly attenuated dexamethasone-induced ECM contraction/stiffening and fibronectin mRNA and protein levels. Sequential treatment was similarly effective but did not match clinically-used Rho kinase inhibition. Conclusions YAP/TAZ inactivation with simvastatin attenuates HTM cell pathobiology in a tissue-mimetic ECM microenvironment. Our data may help explain the association of statin use with a reduced risk of developing glaucoma via indirect YAP/TAZ inhibition as a proposed regulatory mechanism.


Introduction
The trabecular meshwork (TM) plays a central role in the conventional outflow pathway, which drains the aqueous humor from the anterior chamber to regulate outflow facility and intraocular pressure. [1][2][3] The bidirectional interactions between TM cells 4 and their extracellular matrix (ECM) are crucial for maintaining normal tissue function in the healthy eye. 5,6 In primary open-angle glaucoma, the most common form of glaucoma, 7 disruption of these interactions drive progressive fibrotic-like tissue remodeling. Key characteristics of this postulated process include increased TM contraction, actin stress fiber assembly, ECM deposition/crosslinking, and overall tissue stiffening. 8 These pathologic alterations lead to increased outflow resistance driving ocular hypertension, which provides further feedforward amplification and may ultimately push the TM to irreversibly fail. 9,10 Despite substantial scientific efforts over the past several decades devoted to understanding TM pathophysiology, the mechanisms underlying persistent tissue dysfunction in glaucoma remain elusive. Yes-associated protein (YAP) and transcriptional coactivator with PDZ-binding motif (TAZ) are the downstream mediators of the Hippo pathway that play important roles in tissue homeostasis and organ growth. 11,12 As mechanotransducers, activated YAP/TAZ translate biophysical stresses into biochemical signals by translocating to the nucleus from the cytoplasm and interacting with TEAD transcription factors. Through this mechanism, YAP/TAZ exert their function on regulating cellular gene expression, proliferation, and fate. 13 Imbalance or failure of this process is central to a variety of disorders. 14 YAP/TAZ hyperactivity is strongly associated with glaucomatous TM cell dysfunction. Our group 15 and others showed that various glaucoma-associated stressors (e.g. ECM stiffness, dexamethasone, transforming growth factor b2) increase YAP/TAZ activity independent of the canonical Hippo pathway. [16][17][18][19][20][21][22] Importantly, YAP1 was recently identified as one of forty-four previously unknown open-angle glaucoma-risk loci across European, Asian, and African ancestries, 23 suggesting a potential association with TM outflow dysfunction. Therefore, targeting YAP/TAZ signaling -directly or indirectly -may be of therapeutic value for treating outflow dysfunction in primary open-angle glaucoma.
Statins are widely used cholesterol-lowering oral medications for preventing and treating cardiovascular diseases. 24 They block the conversion of hydroxymethylglutaryl coenzyme A (HMG-CoA) to mevalonate by competitively inhibiting HMG-CoA reductase, the rate-limiting enzyme of the mevalonate pathway. 25 Emerging clinical reports suggest that ocular hypertensive/glaucoma patients have higher total cholesterol levels than patients without glaucoma. 26,27 Statin use has also been associated with a reduced risk of primary open-angle glaucoma development and progression in some studies, [28][29][30][31][32][33] possibly contributed to by statins' pleiotropic effects including anti-inflammatory and anti-oxidative benefits. Yet, mechanistic details underlying this potential protective effect in relation to tissues/cells of the outflow tract are incompletely understood.
The ability of statins to induce TM cell relaxation, as evidenced by changes in the actin cytoskeleton, and to increase aqueous humor outflow facility in porcine eye anterior segments was initially demonstrated using lovastatin. 34 By the same token, lovastatin was shown to cause marked changes in human TM cell morphology when cultured on conventional stiff 2D substrates including a loss of filamentous (F)actin stress fiber organization, 35 concomitant with a marked accumulation of cytosolic inactive Rho GTPase proteins. 36 Atorvastatin, another lipophilic statin with higher potency compared to lovastatin, 37 was found to reduce ECM protein expression in human TM cells, 38 as well as induce significant changes in cellular morphology and focal adhesions. 39 Importantly, it was shown that YAP/TAZ activity is controlled by the mevalonate pathway. In two independent small-scale library screens, statins including simvastatin and lovastatin were found to elicit strong YAP/TAZ inhibitory effects potently opposing nuclear localization and transcriptional responses in a Rho GTPase-dependent manner. 40,41 These observations suggest that statins may exert their beneficial clinical effects through modulating metabolic processes independent from cholesterol homeostasis. Therefore, targeting YAP/TAZ signaling with statins presents an intriguing avenue for mitigating TM cell pathobiology in glaucoma with translational potential.
Clinically, glucocorticoid exposure can result in ocular hypertension and may lead to the development of steroidinduced glaucoma, which has commonalities with primary open-angle glaucoma. 42,43 Treatment of TM cells with the synthetic glucocorticoid dexamethasone likewise induces cells to undergo a pathological phenotypic conversion. 44,45 Therefore, dexamethasone is widely used to reliably induce glaucoma-like TM cell pathobiology in vitro. In this study, we hypothesized that simvastatin attenuates dexamethasone-induced human TM cell dysfunction via YAP/TAZ inactivation -assessed by quantifying alterations in YAP/TAZ sub-cellular localization/activity, actomyosin cytoskeletal organization, and functional ECM contraction/remodeling -in a soft tissue-mimetic 3D ECM hydrogel. 46 Materials and methods

HTM cell isolation and culture
The use of human donor corneas was approved by the SUNY Upstate Medical University Institutional Review Board (protocol #1211036), and all experiments were performed according to the tenets of the Declaration of Helsinki for the use of human tissue. Primary human TM (HTM) cells were isolated from healthy donor corneal rims discarded after transplant surgery, as previously described, 15,[46][47][48] and cultured according to established protocols. 49,50 Five normal, previously characterized HTM cell strains (HTM05, HTM12, HTM14, HTM19, HTM36) were used in this study (Table 1). All HTM cell strains were validated with dexamethasone-induced (100 nM) myocilin expression in more than 50% of cells by immunocytochemistry and immunoblot analyses. Different combinations of HTM cell strains were used per experiment, depending on cell availability, and all studies were conducted between cell passages 3-7. HTM cells were cultured in low-glucose Dulbecco's Modified Eagle's Medium (DMEM; Gibco; Thermo Fisher Scientific, Waltham, MA, USA) containing 10% fetal bovine serum (FBS; Atlanta Biologicals, Flowery Branch, GA, USA) and 1% penicillin/streptomycin/glutamine (PSG; Gibco) and maintained at 37 C in a humidified atmosphere with 5% CO 2 . Fresh media was supplied every 2-3 days.

HTM hydrogel preparation
Hydrogel precursors MA-COL (3.6 mg/ml [all final concentrations]), SH-HA (0.5 mg/ml, 0.025% (w/v) photoinitiator), and SH-ELP (2.5 mg/ml) were thoroughly mixed in an amber color tube on ice. Thirty microliters of the hydrogel solution were pipetted onto Surfasil-coated (Fisher Scientific) 18 Â 18-mm square glass coverslips followed by placing 12-mm round glass coverslips onto the hydrogels to facilitate even spreading of the polymer solution. Hydrogels were crosslinked by exposure to UV light (OmniCure S1500 UV Spot Curing System; Excelitas Technologies, Mississauga, Ontario, Canada) at 320-500 nm, 2.2 W/cm 2 for 5 s, according to our established protocols. 15,[46][47][48] The coverslips were removed with fine-tipped tweezers and placed hydrogel-side facing up in polydimethylsiloxane-coated (PDMS; Sylgard 184; Dow Corning; Fisher Scientific) 24well culture plates before seeding HTM cells (1.5 Â 10 4 cells/cm 2 ) atop. To fabricate larger ECM hydrogels for immunoblot analyses, 500 ml of the hydrogel solution was pipetted into 6-well culture plates for uniform coverage across the entire well surface and crosslinked using a modified UV protocol (80 mW/cm 2 for 30 s) before seeding HTM cells (1.5 Â 10 5 cells/cm 2 ) atop. These adjusted settings were shown to yield ECM hydrogels with equivalent elastic modulus compared to standard hydrogels (Supplementary Figure 1). For cell-encapsulated hydrogels, HTM cells (1.0 Â 10 6 cells/ml) were thoroughly mixed with the hydrogel precursors on ice, followed by pipetting either 10 ll droplets of the mixture onto PDMS-coated 24-well culture plates, or 250 ml into custom 16 Â 1-mm PDMS molds and routine photocrosslinking.

HTM hydrogel treatments
HTM cells seeded atop ECM hydrogels were cultured in DMEM with 10% FBS and 1% PSG for 1-3 days until $80-90% confluent. Then, constructs were cultured in serum-free DMEM with 1% PSG and subjected to the following treatments for 3 days: 1) control (vehicle: 0.1% ethanol; 0.1% dimethyl sulfoxide (DMSO), both from Fisher Scientific), 2) dexamethasone (100 nM in ethanol; Fisher Scientific), 3) simvastatin (10 lM in DMSO; Sigma-Aldrich, St. Louis, MO, USA), 4) dexamethasone 1 simvastatin (100 nM dexamethasone; 10 mM simvastatin), and 5) dexamethasone 1 simvastatin 1 mevalonate-5-phosphate (100 nM dexamethasone; 10 mM simvastatin; 500 mM mevalonate-5-phosphate in water; Sigma-Aldrich). HTM cell-encapsulated ECM hydrogels were cultured in DMEM with 10% FBS and 1% PSG and subjected to the same treatments for 10 days. The co-treatment strategy was chosen to simulate a "prophylactic treatment approach." In another set of experiments, HTM hydrogels were first induced with dexamethasone for 5 days followed by different rescue treatments for 5 days with dexamethasone withheld. This sequential treatment strategy was chosen to simulate a "therapeutic treatment approach." HTM cell-encapsulated ECM hydrogels were cultured in DMEM with 10% FBS and 1% PSG and subjected to the following treatments for  Figure 2). The dexamethasone concentration was selected based on our previous study. 46 The simvastatin and mevalonate-5-phosphate concentrations were selected according to a previous report. 40 The netarsudil concentration was selected based on our recent study. 48 HTM hydrogel immunocytochemistry analysis HTM cells cultured atop ECM hydrogels in presence of the different treatments for 3 days were fixed with 4% paraformaldehyde (Thermo Fisher Scientific) at room temperature for 20 min, permeabilized with 0.5% Triton TM X-100 (Thermo Fisher Scientific), blocked with blocking buffer (BioGeneX, Fremont, CA, USA), and incubated with primary antibodies, followed by incubation with fluorescent secondary antibodies (Table 2); nuclei were counterstained with 4 0 ,6 0 -diamidino-2-phenylindole (DAPI; Abcam, Waltham, MA, USA). Similarly, cells were stained with DyLight TM 594 Phalloidin (Cell Signaling Technology, Danvers, MA, USA)/DAPI (Abcam) according to the manufacturer's instructions. Coverslips were mounted with ProLong TM Gold Antifade (Invitrogen; Thermo Fisher Scientific) on Superfrost TM microscope slides (Fisher Scientific), and fluorescent images were acquired with an Eclipse Ni microscope (Nikon Instruments, Melville, NY, USA).
All fluorescent image analyses were performed using FIJI software (National Institutes of Health (NIH), Bethesda, MD, USA). 52 The cytoplasmic YAP/TAZ intensity was measured by subtracting the overlapping nuclear (DAPI) intensity from the total YAP/TAZ intensity. The proportion of total YAP/TAZ intensity that overlapped with the nucleus (DAPI) was obtained to measure the nuclear YAP/TAZ intensity. YAP/TAZ nuclear/cytoplasmic (N/C) ratio was calculated as follows: N/C ratio ¼ (nuclear signal/area of nucleus)/(cytoplasmic signal/area of cytoplasm). YAP/TAZ N/C ratios and fluorescence signal intensities of TGM2, F-actin, aSMA, and p-MLC were measured in at least N ¼ 9 images from 3 experimental replicates per treatment group from 3 HTM cell strains with image background subtraction, followed by calculation of fold-change vs. control.

HTM hydrogel immunoblot analysis
HTM cells cultured atop ECM hydrogels in presence of the different treatments for 3 days were lifted from the hydrogel using 0.25% trypsin/EDTA (Gibco; Fisher Scientific). Care was taken not to contaminate the cellular fraction with ECM proteins from the hydrogel substrate. Pooled cells from 3 experimental replicates per group were lysed in lysis buffer (CelLytic TM M, Sigma-Aldrich) supplemented with Halt TM protease/phosphatase inhibitor cocktail (Thermo Fisher Scientific). Equal protein amounts (10 lg), determined by standard bicinchoninic acid assay (Pierce; ThermoFisher Scientific), in 4Â loading buffer (Invitrogen; Thermo Fisher Scientific) with 5% beta-mercaptoethanol (Fisher Scientific) were boiled for 5 min and subjected to SDS-PAGE using NuPAGE TM 4-12% Bis-Tris Gels (Invitrogen; Thermo Fisher Scientific) at 150 V for 120 min and transferred to 0.45 lm PVDF membranes (Sigma; Thermo Fisher Scientific). Membranes were blocked with 5% bovine serum albumin (Thermo Fisher Scientific) in trisbuffered saline with 0.2% Tween V R 20 (Thermo Fisher Scientific), and probed with primary antibodies followed by incubation with fluorescent secondary antibodies ( Table 2). Bound antibodies were visualized with an Odyssey CLx imager (LI-COR, Lincoln, NE, USA).

HTM hydrogel contraction analysis
Longitudinal brightfield images of HTM hydrogels subjected to the different treatments for 10 days were acquired at day 0 and day 10 with an Eclipse Ti microscope (Nikon). Construct areas from N ¼ 4-8 experimental replicates per treatment group from 3 HTM cell strains were quantified using FIJI software (NIH) and normalized to 0 d followed by normalization to controls.

HTM hydrogel cell viability analysis
The number of viable cells inside HTM hydrogels subjected to the different treatments for 10  . Sections were permeabilized with 0.5% Triton TM X-100, blocked with blocking buffer (BioGeneX) and incubated with a primary antibody against fibronectin (FN), followed by incubation with a fluorescent secondary antibody (Table 2). Slides were mounted with ProLong TM Gold Antifade (Thermo Fisher Scientific), and fluorescent images were acquired with an Eclipse Ni microscope (Nikon). Fluorescence signal intensity of fibronectin  Table  3). Transcript levels of fibronectin were normalized to GAPDH, and mRNA fold-changes were calculated relative to controls from N ¼ 4 experimental replicates per treatment group from 2 HTM cell strains using the comparative C T method. 54

Statistical analysis
Individual sample sizes are specified in each figure caption.
Comparisons between groups were assessed by unpaired t test, one-way or two-way analysis of variance (ANOVA) with Tukey's multiple comparisons post hoc tests, as appropriate. All data are shown with mean ± SD, some with individual data points. The significance level was set at p < 0.05 or lower. GraphPad Prism software v9.3 (GraphPad Software, La Jolla, CA, USA) was used for all analyses.

Simvastatin reduces pathologic YAP/TAZ nuclear localization and TGM2 levels in HTM cells
The transcriptional coactivators YAP and TAZ dynamically shuttle between the cytoplasm and nucleus to regulate gene expression; increased nuclear YAP/TAZ localization principally indicates enhanced transcriptional activity. 55 Our recent studies on human TM and Schlemm's canal cells support this notion in ocular cells. 15,56 Moreover, the mevalonate pathway was shown to promote YAP/TAZ nuclear accumulation and activity in non-ocular cells; 40  TAZ: $1.8) ( Figure. 1(A-D). These results were consistent with our previous study using TGFb2 to induce glaucomatous HTM cell dysfunction. 15 With simvastatin alone, we observed YAP/TAZ N/C ratios (YAP: $2.3; TAZ: $2.5) comparable to the control group. Importantly, co-treatment with dexamethasone þ simvastatin significantly decreased YAP/TAZ N/C ratios (YAP and TAZ: $2.2) compared to dexamethasone-induced HTM cells, restoring control levels.
When mevalonate-5-phosphate was supplemented to bypass the simvastatin-mediated HMG-CoA reductase inhibition, we observed significantly increased YAP/TAZ N/C ratios (YAP: $3.6; TAZ: $3.5) compared to dexamethasone þ simvastatin, indistinguishable from the dexamethasonetreated HTM cells. Different HTM cell strains significantly affected the results (p < 0.0001), in agreement with normal donor-to-donor variability, showing significant interaction with the different treatments (p < 0.05). Of note, the YAP and TAZ N/C ratio data acquired on ECM hydrogels closely mirrored results obtained using HTM cells on conventional glass coverslips that served as an additional control (Supplementary Figure 3). Next, we focused on the ECM crosslinking enzyme transglutaminase 2 (TGM2), whose expression is elevated in the TM of glaucomatous eyes. 57 After YAP/TAZ translocate to the nucleus, they interact with TEAD transcription factors to regulate the expression of glaucoma-related putative downstream effectors including TGM2. A previous report showed the inhibition of TGM2 expression in HTM cells with siRNA-mediated YAP knockdown; 58 this observation was recently confirmed in our own study. 15 Consistent with the YAP/TAZ nuclear localization data, dexamethasone treatment significantly increased TGM2 intensity in HTM cells ($3.5-fold) compared to controls ( Figure. 1(E,F)); simvastatin alone had no effect on TGM2. Co-treatment with dexamethasone þ simvastatin significantly decreased TGM2 intensity ($1.2-fold) compared to dexamethasone-induced HTM cells, restoring baseline levels. The addition of mevalonate-5-phosphate resulted in significantly increased TGM2 intensity ($2.9-fold) compared to dexamethasone þ simvastatin, showing comparable levels to the dexamethasone-induced group. While different HTM cell strains did not affect the results (p ¼ 0.3848), there was significant interaction with the different treatments (p ¼ 0.0070). Qualitative immunoblot analyses validated the TGM2 immunocytochemistry results showing overall very similar trends (Figure 1(G)).
Taken together, these data demonstrate that simvastatin prevents glucocorticoid-induced pathologic YAP/TAZ nuclear localization and concurrently reduces expression of the downstream effector TGM2 in HTM cells in a tissuemimetic ECM microenvironment.

Simvastatin reduces pathologic F-actin and aSMA stress fibers in HTM cells
The actin cytoskeleton is the principal force-generating machinery in the cell. 59 F-actin filaments are sensitive to mechanical stimuli 60 and have been demonstrated to be  (Figure 2(A,B)). Co- treatment with dexamethasone þ simvastatin significantly decreased F-actin intensity ($0.8-fold) compared to dexamethasone-treated HTM cells, reaching control levels. Supplementation of mevalonate-5-phosphate significantly increased F-actin intensity ($2.0-fold) compared to dexamethasone þ simvastatin, equivalent to dexamethasone-induced HTM cells. Different HTM cell strains significantly affected the results (p < 0.0001), showing significant interaction with the different treatments (p < 0.0001). Next, we looked at alpha-smooth muscle actin (aSMA)a hallmark of tissue fibrosis; 64 increased aSMA stress fibers have also been linked to fibrotic-like HTM cell pathobiology and outflow dysfunction in glaucoma. [65][66][67][68] Consistent with the F-actin data, exposure of HTM cells to dexamethasone significantly increased aSMA stress fibers/overall intensity ($7.5-fold) compared to controls, which only showed very weak staining ( Figure. 2(C,D)). Simvastatin alone had no effect on aSMA. Co-treatment with dexamethasone þ simvastatin significantly decreased aSMA intensity ($1.0-fold) compared to dexamethasone-induced HTM cells, restoring baseline levels. The addition of mevalonate-5-phosphate significantly increased aSMA stress fibers ($7.5-fold) compared to dexamethasone þ simvastatin, identical to dexamethasone-treated HTM cells. Different HTM cell strains significantly affected the results (p < 0.0001), showing significant interaction with the different treatments (p < 0.0001). Qualitative immunoblot analyses validated the aSMA immunocytochemistry results showing overall very similar trends ( Figure. 2(E)).
Taken together, these data demonstrate that simvastatin blocks glucocorticoid-induced pathologic actin stress fiber formation in HTM cells in a tissue-mimetic ECM microenvironment.

Simvastatin reduces pathologic p-MLC levels in HTM cells
Myosin light chain is a master regulator of cell contractility when phosphorylated by Rho kinases via the generation of pulling forces from actomyosin filament contraction. 69-71 Increased phospho-myosin light chain (p-MLC) is strongly associated with a pathologic hypercontractile TM cell phenotype akin to activated myofibroblasts. 4,47,68 In contrast, a decrease in p-MLC has been shown to increase aqueous outflow facility in perfusion studies. 72 Statin-mediated inhibition of the mevalonate pathway directly affects downstream Rho GTPase signaling; 40 therefore, we hypothesized that simvastatin decreases glucocorticoid-induced pathologic p-MLC levels in HTM cells. Dexamethasone treatment significantly increased p-MLC intensity in HTM cells ($4.2-fold) compared to controls; simvastatin alone had no effect on p-MLC levels (Figure 3(A,B)). Co-treatment with dexamethasone þ simvastatin significantly decreased p-MLC intensity ($0.8-fold) compared to dexamethasone-induced HTM cells, restoring control levels. The addition of mevalonate-5-phosphate significantly increased p-MLC intensity ($4.3-fold) compared to dexamethasone þ simvastatin, indistinguishable from dexamethasone-treated HTM cells. Different HTM cell strains significantly affected the results (p < 0.0001), showing significant interaction with the different treatments (p < 0.0001).
From this experiment, we conclude that simvastatin prevents glucocorticoid-induced pathologic myosin light chain phosphorylation in HTM cells in a tissue-mimetic ECM microenvironment to prevent cells from acquiring a hypercontractile myofibroblast-like phenotype.

Simvastatin reduces pathologic contraction, stiffening, and FN deposition in HTM cells
The TM undergoes increased fibrotic-like contraction and aberrant ECM deposition that together contribute to pathologic tissue stiffening in primary open-angle glaucoma. 8 To better approximate the 3D tissue architecture in the juxtacanalicular TM region that is critical for outflow regulation, 9 HTM cells were encapsulated in ECM hydrogels for tissuelevel functional studies. We hypothesized that simvastatin decreases glucocorticoid-induced pathologic contraction, stiffening, and fibronectin deposition in HTM cell-laden hydrogels. Consistent with the p-MLC data and supported by our previous study, 46 exposure of HTM cells to dexamethasone significantly increased hydrogel contraction ($78.1%) compared to controls (Figure 4(A,B)). Simvastatin alone had no effect on HTM hydrogel contraction. Co-treatment with dexamethasone þ simvastatin significantly decreased hydrogel contraction ($103.9%) -or in other words relaxed the constructs -compared to dexamethasoneinduced HTM hydrogels, restoring baseline levels. Supplementation of mevalonate-5-phosphate significantly increased hydrogel contraction ($86.2%) compared to dexamethasone þ simvastatin, identical to dexamethasone-treated HTM hydrogels. Different HTM cell strains significantly affected the results (p < 0.005); yet, there was no significant interaction with the different treatments (p ¼ 0.2411). To rule out that hydrogel contractility was influenced by the cell number, HTM cell viability inside the 3D ECM network was assessed. We observed no differences between the different groups (Supplementary Figure 4).
Having shown that the supplementation of mevalonate-5phosphate fully offsets the inhibitory effects of simvastatin using a variety of techniquesthereby unequivocally confirming that the results seen are mediated by the mevalonate pathway, we dropped the group for subsequent experiments. To assess the functional consequences of increased hydrogel contraction on tissue-level construct stiffness, we next performed oscillatory rheology analyses. Consistent with our previous report, 46 dexamethasone treatment induced significant hydrogel stiffening ($1.9-fold) compared to controls; simvastatin alone had no effect on hydrogel stiffness ( Figure  4(C)). Co-treatment with dexamethasone þ simvastatin significantly decreased hydrogel stiffness ($1.1-fold) -or softened the constructs -compared to dexamethasone-induced HTM hydrogels, restoring control levels. Different HTM cell strains did not affect the results (p ¼ 0.9700), and there was no significant interaction with the different treatments (p ¼ 0.7135).
Lastly, we looked at the deposition of fibronectin (FN) in HTM hydrogels. FN is a major ECM player and signaling component in the native tissue, 73 and has been detected at elevated levels in the TM of glaucomatous eyes. 74 Significance was determined by two-way ANOVA using multiple comparisons test; shared significance indicator letters ¼ non-significant difference (p > 0.05), distinct letters ¼ significant difference (p < 0.05).
Furthermore, it has long been known that HTM cells express high FN levels upon glaucomatous stimulation. 75,76 Exposure to dexamethasone significantly increased FN intensity ($2.1-fold) compared to controls; simvastatin alone did not affect FN levels ( Figure. 4(D,E)). Co-treatment with dexamethasone þ simvastatin significantly decreased FN intensity ($1.5-fold) compared to dexamethasone-induced HTM hydrogels, with levels remaining significantly higher than controls. Different HTM cell strains significantly affected the results (p < 0.01); yet, there was no significant interaction with the different treatments (p ¼ 0.3233). FN transcript analyses by qRT-PCR confirmed the immunohistochemical data ( Figure. 4(F)).
Taken together, these data demonstrate that simvastatin blocks glucocorticoid-induced pathologic ECM contraction and stiffening, and concurrently reduces extracellular FN deposition in HTM cells in a tissue-mimetic ECM microenvironment.

Simvastatin rescues pathologic contraction in HTM cells
All experiments to delineate the effects of simvastatin on HTM cell pathobiology up to this point were conducted by simulating a "prophylactic treatment approach," i.e. dexamethasone and simvastatin were co-delivered from the start. Next, to mimic a "therapeutic treatment approach," HTM cell-encapsulated hydrogels were treated for 5 days with dexamethasone to establish a pre-contracted baseline, followed by treatment with simvastatin for 5 days with dexamethasone withheld. Using this design, we assessed the rescue potential of simvastatin in direct comparison with the FDA-approved Rho kinase inhibitor netarsudil, shown to increase aqueous outflow via reducing TM contractile tone. 77,78 We hypothesized that simvastatin rescues glucocorticoid-induced pathologic contraction of HTM cell-laden hydrogels to a similar degree as clinically-used netarsudil. Dexamethasone treatment significantly increased hydrogel contraction ($82.1%) compared to controls (Figure. 5). Sequential treatment with simvastatin fully rescued dexamethasone-induced HTM hydrogel contraction ($96.4%), restoring baseline levels. For netarsudil, we observed an even more potent rescuing effect ($112.6%), consistent with our previous data, 48 that exceeded both the simvastatin and control groups. While different HTM cell strains did not affect the results (p ¼ 0.2469), there was significant interaction with the different treatments (p < 0.0001). HTM cell viability analysis demonstrated that hydrogel contractility was not affected by the cell number; we observed no differences between the different groups (Supplementary Figure 5).
From this data we conclude that simvastatin rescues glucocorticoid-induced pathologic ECM contraction in a tissue-mimetic ECM microenvironment, albeit to a lower extent compared to a clinically-used Rho kinase inhibitor.

Discussion
Simvastatin, a member of the cholesterol-lowering statin drug class, inhibits HMG-CoA reductase that catalyzes the production of mevalonate. The anabolic mevalonate cascade provides key isoprenoid metabolites for diverse cellular processes including cholesterol synthesis and post-translational membrane targeting of Rho GTPases. 79,80 Furthermore, it was recently demonstrated that the mevalonate pathway has a profound impact on the function of the transcriptional regulators YAP and TAZ in different cancer cells. 40,63,81 These studies mechanistically linked the mevalonate pathway to (i) geranylgeranyl pyrophosphate (GGPP)mediated Rho GTPase activation and F-actin fiber assembly rather than the squalene/cholesterol arm of the mevalonate cascade, and (ii) reduction of YAP/TAZ inhibitory phosphorylation and sustained YAP/TAZ transcriptional activity via nuclear accumulation independent of canonical Hippo-LATS1/2 kinase activity. With the increasing evidenceincluding from our own laboratory -of aberrant YAP/TAZ activity in human TM cells isolated from glaucoma eyes or induced with glaucoma-associated stressors, [15][16][17][18][19][20][21][22] in conjunction with the recent discovery of YAP as a potential "risk gene" for open-angle glaucoma 23 that frequently involves impaired TM outflow function, we here tested the hypothesis that simvastatin attenuates glucocorticoid-induced glaucomatous TM cell pathobiology via targeting YAP/TAZ activity in a tissue-mimetic ECM microenvironment. 46 Simvastatin was identified among more than 600 FDAapproved compounds to have potent inhibitory effects on YAP/TAZ activity in human breast cancer cells, and to efficiently rescue Drosophila eye overgrowth induced by the YAP orthologue Yki. 40 In another high-throughput library screen of more than 13,000 small-molecule compounds, simvastatin was found to exhibit strong YAP inhibitory effects via the same GGPP/Rho/actin signaling axis in human lung fibroblasts, and to reduce fibrotic markers in the bleomycin mouse model of pulmonary fibrosis. 82 Together, these data Significance was determined by two-way ANOVA using multiple comparisons test; shared significance indicator letters ¼ non-significant difference (p > 0.05), distinct letters ¼ significant difference (p < 0.05). support the notion that simvastatin has intriguing potential to modulate pathologic YAP/TAZ activity in the context of TM cell dysfunction that has been associated with fibroticlike ECM remodeling, tissue contraction and stiffening in glaucoma. 7,8 We found that simvastatin potently decreased dexamethasone-induced YAP/TAZ nuclear localization in HTM cells, the main mechanism to regulate their function, 11 concurrent with a reduction of the downstream effector TGM2 known to be expressed at increased levels in the TM of glaucoma eyes. 57 These data were in agreement with our recent study in which we showed that genetic or pharmacologic YAP/TAZ inactivation potently blocked TGFb2induced HTM cell pathobiology in the same ECM hydrogel environment. 15 Of note, we observed nearly identical trends for YAP and TAZ. This suggests that simvastatin may regulate the two paralogues in a similar manner in HTM cells consistent with their acknowledged functional redundancy. 83 We also found that the simvastatin-mediated YAP/TAZ inactivation in HTM cells was fully negated when mevalonate-5-phosphate was supplemented, comparable to what has been reported in non-ocular cells. 40 While we did not inhibit distinct enzymes here to identify the specific metabolic intermediate involved in YAP/TAZ regulation in HTM cells, evidence from previous studies strongly support that protein geranylgeranylation is responsible for the positive effect of the mevalonate pathway on YAP/TAZ activity. 40,63,81,82 It was shown that only farnesyl diphosphate synthase or geranylgeranyl transferase inhibition were able to reproduce the effect of simvastatin on YAP/TAZ nuclear localization and transcriptional activity, whereas inhibition of squalene synthase -the enzyme that catalyzes the first step of sterol biosynthesis -or farnesyl transferase had no effect. 40 YAP/TAZ activity requires actomyosin cytoskeletal integrity/tension and involves contractile as well as adhesive structures. 84,85 We found that simvastatin effectively blocked glucocorticoid-induced F-actin stress fiber formation in HTM cells. This was in agreement with our recent study using verteporfin to inhibit YAP/TAZ 15 and data demonstrating that the mevalonate pathway regulates YAP/TAZ through F-actin. 63 Previously, HTM cells treated with lovastatin or geranylgeranyl transferase inhibitor were shown to exhibit decreased actin stress fiber organization and increased accumulation of unprenylated (i.e. inactive) RhoA and RhoB 35 in conventional 2D cultures, lending further support for protein geranylgeranylation as being the nexus of statin-mediated effects on HTM cell cytoskeletal organization. We also observed that simvastatin treatment potently prevented aSMA fiber formation, an accepted indicator of a fibrotic-like HTM cell phenotype. 68 These data were consistent with evidence that simvastatin exhibits anti-fibrotic potential in human lung fibroblasts by targeting the GGPP/Rho/actin signaling axis. 82 The Rho/ROCK pathway is a master regulator of the actin cytoskeleton that has been strongly associated with HTM cell contractility via phosphorylation of myosin light chain. 68 We showed that simvastatin potently decreased dexamethasone-induced p-MLC levels in a tissue-mimetic ECM environment to prevent HTM cells from acquiring a hypercontractile myofibroblast-like phenotype. Consistent with the p-MLC data, we further demonstrated that simvastatin efficiently blocked pathologic contraction and stiffening of HTM cell-encapsulated ECM hydrogels that more accurately simulate the native tissue architecture. These functional alterations were found to coincide with a reduction in FN mRNA expression and protein deposition that is known to play a role in the development of ocular hypertension/glaucoma. 73,74 These cumulative findings are in line with our previous observations that YAP/TAZ inhibition with verteporfin counteracts tissue-level functional impairments, 15 further strengthening the argument that aberrant YAP/TAZ activity in cells of the outflow tract may contribute to glaucoma pathogenesis. Importantly, our data suggest that supplementation of simvastatin can both prevent (i.e. cotreatment) and rescue (i.e. sequential treatment) dexamethasone-induced TM cell pathobiology contingent on cell-ECM interactions. We acknowledge that simvastatin fell short of matching the potent contraction-reversing effects of the FDA-approved ROCK inhibitor netarsudil, the active ingredient in Rhopressa TM , that increases outflow through the stiffened TM via reducing tissue contraction as a function of ECM-focal adhesion and actin stress fiber disassembly. 77,78,[86][87][88] This could be explained by the direct vs. indirect modes of action on cellular contractility. Potential cotreatment regimens could be explored in future studies.
In conclusion, we demonstrated that YAP/TAZ inactivation with simvastatin attenuates dexamethasone-induced HTM cell dysfunction in a tissue-mimetic ECM microenvironment. Our data may help explain the association of statin use with a reduced risk of developing glaucoma by proposing indirect YAP/TAZ inhibition as a regulatory mechanism and highlight the therapeutic potential of localized simvastatin therapy to treat outflow tissue dysfunction in glaucoma.