Increased synthesis and deposition of extracellular matrix proteins leads to endoplasmic reticulum stress in the trabecular meshwork

Increased synthesis and deposition of extracellular matrix (ECM) proteins in the trabecular meshwork (TM) is associated with TM dysfunction and intraocular pressure (IOP) elevation in glaucoma. However, it is not understood how ECM accumulation leads to TM dysfunction and IOP elevation. Using a mouse model of glucocorticoid (GC)-induced glaucoma, primary human TM cells and human post-mortem TM tissues, we show that increased ECM accumulation leads to endoplasmic reticulum (ER) stress in the TM. The potent GC, dexamethasone (Dex) increased the secretory protein load of ECM proteins in the ER of TM cells, inducing ER stress. Reduction of fibronectin, a major regulator of ECM structure, prevented ER stress in Dex-treated TM cells. Overexpression of fibronectin via treatment with cellular fibronectin also induced chronic ER stress in primary human TM cells. Primary human TM cells grown on ECM derived from Dex-treated TM cells induced ER stress markers. TM cells were more prone to ER stress from ECM accumulation compared to other ocular cell types. Moreover, increased co-localization of ECM proteins with ER stress markers was observed in human post-mortem glaucomatous TM tissues. These data indicate that ER stress is associated with increased ECM accumulation in mouse and human glaucomatous TM tissues.

Glaucoma is the second leading cause of irreversible vision loss affecting about 70 million people worldwide, and is more prevalent in African Americans 1 . Primary open angle glaucoma (POAG), the most common form of glaucoma is characterized by progressive loss of retinal ganglion cell (RGC) axons and irreversible loss of vision 2 . Although the exact cellular mechanisms that cause glaucoma are poorly understood, elevated intraocular pressure (IOP) is a major associated risk factor 3 . The maintenance of IOP in a narrow range is essential for survival of the neuroretina. The trabecular meshwork (TM), a specialized tissue located at the iridocorneal angle of the eye, maintains normal IOP by regulating aqueous humor outflow resistance. In POAG, there is increased resistance to aqueous humor outflow through the TM, thus elevating IOP [3][4][5] . Although glaucomatous TM damage is associated with various morphological and biochemical changes including extracellular matrix (ECM) protein accumulation [6][7][8][9][10][11][12] and loss of TM cells 13 , the exact pathological mechanisms that lead to this glaucomatous TM damage are not fully understood.
Glucocorticoids (GCs) are the mainstream treatment for the plethora of inflammatory disorders including ocular conditions such as dry eye, allergic eye disease, inflammation following eye surgery, uveitis, diabetic macular edema and many others 14 . Ocular hypertension is a serious side effect of glucocorticoid therapy. Topical or systemic GC administration can lead to ocular hypertension in about 30-50% patients depending on the route of administration, and sustained GC treatment can lead to secondary iatrogenic open-angle glaucoma if GCs are not withdrawn [15][16][17] . Although GC-induced glaucoma is a form of secondary iatrogenic open-angle glaucoma, its clinical presentations are similar in many ways to POAG 16 . Furthermore, GC responsiveness is significantly higher in POAG patients 18 . Similar to POAG, GC-induced glaucoma is also caused by increased resistance to aqueous humor outflow at the TM 16,19 . Therefore, a large number of in vitro and in vivo studies have examined GC-induced ocular hypertension to better understand glaucomatous TM damage 16,20,21 . GCs exhibit diverse effects on the TM including altered TM gene and protein expression, reduced TM phagocytic function, increased ECM synthesis

Results
Dex-induced IOP elevation is associated with reduced outflow facility and abnormal ECM accumulation in the TM. We first examined whether Dex treatment elevates IOP by increasing outflow resistance at the TM. As shown in Fig. 1A, topical ocular Dex treatment leads to sustained and significant IOP elevation starting from 3 weeks of treatment. The approximate increase of IOP in Dex-treated mice over Veh-treated mice was 1.0 mmHg at 1-week (P > 0.05), 2.0 mmHg at 2 weeks (P > 0.05), 3.5 mmHg at 3 weeks (P < 0.01), 4.2 mmHg at 4 weeks (P < 0.001) and 4.8 mmHg at 5 weeks of treatment (P < 0.001, n = 10, 2-way ANOVA). Using a constant flow infusion method, we measured outflow facilities in Veh and Dex-treated mice (Fig. 1B). A significant eye drops were applied to A/J mice 3 times a day for 5 weeks, and IOPs were recorded weekly. A significant IOP elevation was observed starting from 3-weeks of Dex treatment (n = 10 each group, 2-WAY ANOVA, **P < 0.01, ***P < 0.001). (B) Reduced outflow facility in Dex-treated mice. A significant reduction in outflow facility was observed in 7-weeks Dex-treated mice compared to Veh-treated mice (n = 8 each group, unpaired student's t-test). (C) Dex increased fibronectin and actin immunostaining in the TM of 7-weeks Dex-treated mice (n = 3) compared with Veh-treated mice (n = 3). White box indicates TM. CB = ciliary body, I = Iris, C = cornea. Scale bar = 50 μm. decrease (p = 0.0053) in outflow facility was observed in Dex-treated mice (0.0135 ± 0.001581 µl/min/mmHg) compared to Veh-treated mice (0.01963 ± 0.001349 µl/min/mmHg). There was ~30% decrease in outflow facility in Dex-treated mice compared to Veh-treated mice. Several studies have shown that Dex induces ECM deposition and actin changes in the TM 16,21,23,51 . To confirm this, we next examined whether Dex treatment leads to increased ECM deposition and actin changes in the TM of our mouse model of Dex-induced ocular hypertension (Fig. 1C). Immunostaining for fibronectin and actin (Fig. 1C) as well as collagen type I (Col I; Fig. 2) in Veh and Dex-treated mice demonstrated that Dex prominently increased fibronectin, actin and Col I in mouse TM tissues.

Dex increases secretory protein overload including ECM proteins in the ER of TM cells.
To determine whether increased ECM synthesis induced by Dex in TM cells exceeds the normal ER capacity, thereby inducing ER stress, we first examined whether Dex increases de novo protein synthesis using the SUnSET assay, a nonradioactive method of protein synthesis 52 . In this assay, puromycin, a structural analog of aminoacyl tRNAs gets incorporated into the nascent polypeptides and stops elongation. At lower concentrations, puromycin incorporation in newly synthesized proteins reflects the rate of de novo protein synthesis. GTM3 cells were treated with Veh or Dex for 3 days and incubated with puromycin for 30 min. Western blot analysis using a puromycin antibody revealed that Dex increased general protein synthesis in TM cells (Fig. 3A). Coomassie staining was used as a total protein loading control. As a negative control, TM cells incubated without puromycin demonstrated no puromycin reactive bands in the Western blot (Fig. 3A, lane 1). As a positive control, TM cells incubated with the protein synthesis inhibitor, cycloheximide reduced protein synthesis (lane 3) compared to control cells incubated with puromycin (lane 2). Compared to Veh-treated TM cells, Dex increased protein synthesis as evident from the intense band at the bottom of gel representing puromycin truncated proteins (lane 4 Vs lane 5). We next examined whether Dex increases the secretory protein load in the ER. Primary human TM cells (n = 2 cell strains) were treated with Veh or Dex for 3 days and incubated with puromycin for 30 min. Fixed cells were immunostained with a puromycin antibody and the ER marker calreticulin to examine de novo protein synthesis of secretory proteins in the ER. Immunostaining demonstrated that Dex dramatically increased the protein synthesis of secretory proteins from the ER of TM cells (Fig. 3B). Moreover, immunostaining for puromycin and GRP78 demonstrated that de novo protein synthesis is associated with ER stress (Fig. 3C). Immunostaining for puromycin and fibronectin demonstrated that Dex induces de novo synthesis of fibronectin (Fig. 3D).
To determine whether the Dex-induced excessive ECM protein load also induces ER stress in TM cells, we examined co-localization of ECM proteins with the ER stress marker KDEL. Both fibronectin and collagens constitute major ECM components of TM; therefore, we examined levels of fibronectin and collagen as general representatives of the trabecular meshwork ECM as described previously 53 . Increased co-localization of fibronectin ( Fig. 4A) or Col I (Fig. 4B) with KDEL (ER stress marker) in Dex-treated cells demonstrated increased ECM synthesis and induction of ER stress in TM cells. We further examined whether Dex-induced fibronectin and Col I interact with ER stress proteins by performing co-immunoprecipitation studies (Fig. 4C). Total lysates were immunoprecipitated using KDEL antibody and blotted for ECM proteins. Lysates incubated with IgG antibody was used as a negative control and demonstrated no bands. A weak interaction of fibronectin and Col I with ER stress proteins (GRP78 and GRP94) was observed in Veh-treated cells, but this interaction increased prominently after Dex treatment. Note that although no band is observed for GRP94 in input samples at the low exposure shown in Fig. 3C, GRP94 band was observed in input samples at higher exposure. These data suggest that Dex-induced ECM synthesis is associated with ER stress in TM cells.

Reduction of fibronectin prevents Dex-induced ER stress in TM cells.
Previous studies have shown that fibronectin acts as a major regulator of ECM synthesis and deposition [54][55][56] . Therefore, we explored whether reduction of fibronectin prevents Dex-induced ER stress in GTM3 cells. GTM3 cells were transfected with or without CRISPR-Cas9 targeting fibronectin for 24 hours and then treated with Veh or Dex for 2 days. Knockdown of fibronectin in Dex-treated cells further reduced ER stress as evident from reduced levels of ER stress markers, GRP78 and CHOP (Fig. 5A). Immunostaining for fibronectin and KDEL revealed that fibronectin knockdown reduced fibronectin protein levels and also decreased ER stress in GTM3 cells treated with Dex (Fig. 5B).

Cellular fibronectin induces ER stress in human TM cells. Since increased extracellular ECM is known
to be associated with human glaucomatous TM tissues, we next examined whether the extracellular deposition of ECM leads to ER stress in primary human TM cells. Cellular fibronectin is an insoluble fibronectin isoform that forms fibril networks and regulates ECM-cell interactions 57,58 . A recent study demonstrated that the cellular fibronectin isoform EDA was increased in post-mortem human glaucomatous TM tissues 7 . As shown in Fig. 6A, treatment of primary human TM cells with cellular fibronectin induced ER stress as evident from increased GRP78, ATF4, and CHOP further suggesting that extracellular deposition of ECM can induce ER stress in TM cells. Immunostaining showed increased extracellular fibronectin and ER stress induction in primary human TM cells treated with cellular fibronectin (Fig. 6B). Transduction of GTM3 cells with a lentiviral fibronectin expression vector also increased fibronectin and KDEL, indicating overexpression of fibronectin alone can induce ER stress in human TM cells (Fig. 6C).
Decellularized Dex-induced extracellular ECM deposition is sufficient to induce ER stress in primary human TM cells. We next examined whether Dex-induced extracellular ECM leads to ER stress induction in human primary TM cells by culturing TM cells on ECM derived from Dex-treated TM cells. Human primary TM cells (n = 2 cell strains) grown in chamber slides were treated with Veh or Dex for 5 days. The ECM was decellularized from slides to remove TM cells. Immunostaining for fibronectin along with DAPI in decellularized slides demonstrated that Dex increased extracellular fibronectin staining while the absence of DAPI nuclear staining further supported successful decellularization (Fig. 7A). We also observed increased staining for Col I in Dex-treated ECM compared to Veh-treated ECM (data not shown). Fresh human primary TM cells were added to Veh or Dex-treated decellularized ECM and cells were cultured for an additional 5 days without any treatment (Fig. 7B). Interestingly, immunostaining revealed increased fibronectin and KDEL staining in TM cells cultured on the decellularized ECM from Dex treated cells, indicating that the Dex-induced extracellular ECM alone leads to ER stress in TM cells.

days and stained for KDEL and fibronectin (A) or Col I (B). A prominent co-localization of fibronectin or Col I with KDEL was observed in Dex-treated TM cells indicating induction of ER stress and excessive intracellular ECM overload in primary TM cells. (C) GTM3
cells were treated with Veh or Dex for 3 days and total lysates were subjected to immunoprecipitation using KDEL antibody and immunoblotted for ECM proteins, GRP78 and GRP94. IgG antibody was used as a negative control. Total cellular lysates were used as input.    However, ER stress induction was only evident in TM cells ( Fig. 8A-C). We further examined chronic ER stress markers using Western blot analysis (Fig. 8D). Dex treatment increased fibronectin in both GTM3 and ARPE-19 cells. Interestingly, Dex treated GTM3 cells demonstrated induction of chronic ER stress markers including ATF4 and CHOP compared to Veh-treated cells. However, these markers were unaltered in ARPE-19 cells. Similarly, treatment of GTM3 or ARPE-19 cells with cellular fibronectin increased intracellular fibronectin but ER stress was only evident in GTM3 cells (Fig. 8E). These findings demonstrate that TM cells are more prone to ER stress induction from Dex-induced ECM synthesis and deposition.

Increased levels and co-localization of fibronectin and Col I with ER stress marker in human
glaucomatous TM tissues. Since we observed increased co-localization of fibronectin with ER stress marker in primary human TM cells treated with Dex compared to Veh-treated TM cells, we next examined whether similar co-localization of fibronectin and Col I with ER stress is present in human glaucomatous TM tissues. Immunostaining for fibronectin and KDEL revealed that fibronectin and KDEL are noticeably increased in human post-mortem TM tissues from POAG donors compared to normal donor eyes (Fig. 9A,B and SI1). In addition, increased co-localization of fibronectin with KDEL was observed in these glaucomatous human TM tissues (Fig. 9A,B). Out of 6 pairs of human eyes immunostained, we observed a strong co-localization of fibronectin with KDEL in 5 glaucomatous eyes. Similarly, we observed increased Col I and its co-localization with ER stress markers in all glaucomatous TM tissues compared to age-matched normal donor tissues (Fig. 10A,B  and SI2). Importantly, we observed that both ECM and ER stress markers were consistently increased in all glaucomatous TM tissues (SI1 and SIII). These studies suggest that increased ECM accumulation is associated with ER stress in the glaucomatous TM tissues.

Discussion
Excessive ECM accumulation is considered a major feature of the glaucomatous TM tissues. However, it is not yet understood how this ECM accumulation leads to TM dysfunction and IOP elevation. Here, we demonstrate that GC-induced ocular hypertension and reduced outflow facility is associated with increased ECM accumulation in the TM. We further demonstrate that Dex increases secretory protein load of ECM proteins, inducing ER stress in primary human TM cells. Dex treatment increases interactions between the ER stress marker GRP78 with fibronectin and Col I in TM cells. Reduction of fibronectin prevents ER stress in Dex-treated TM cells. The expression of cellular fibronectin and ER stress markers is higher in the TM of open angle glaucoma eyes 7,49 ; treatment of cultured primary TM cells with exogenous cellular fibronectin also induced ER stress in cultured primary TM cells. Interestingly, ECM derived from Dex-treated cells was sufficient to induce ER stress in normal TM cells. We further demonstrate that TM cells are sensitive to ER stress from increased ECM protein accumulation compared to other ocular cell types. Importantly, increased levels of fibronectin and Col I and their co-localization with an ER stress protein was observed in human glaucomatous TM tissues further suggesting that increased synthesis and deposition of ECM is associated with ER stress in the TM.
A recent study demonstrated that systemic Dex treatment in mice leads to reduced outflow facility and increased fingerprint-like materials in the TM 42 . Similar ECM changes were also observed in TM tissues of humans treated with steroids 20,22,23,59 . Consistent with these findings, we show that GC-induced ocular hypertension is associated with increased aqueous humor outflow resistance and increased ECM accumulation in the mouse TM. Therefore, it is likely that Dex-induced abnormal ECM deposition in the TM is responsible for this increased outflow resistance.
Treatment of human TM cells with cellular fibronectin induced chronic ER stress proteins including ATF4 and CHOP, which are known to be associated with cell dysfunction/death. We have previously demonstrated increased ATF4 and CHOP levels in the glaucomatous TM tissues from POAG patients 49 and in anterior segment tissues of mouse models of myocilin or GC-induced glaucoma 41,60 . Chronic ER stress can cause detrimental effects on TM function. In addition, the overwhelmed ER may compromise the quality of secreted ECM proteins, which are more likely to form abnormal depositions in the ECM as observed in human TM tissues from GC-induced glaucoma and POAG donors. Consistent with this, we have recently demonstrated that ER stress induced by mutant myocilin causes intracellular retention of selected ECM proteins 53 .
Although we focused our studies to selected ECM proteins, it is likely that accumulation of other ECM proteins is also associated with ER stress in the TM. Dex has been shown to induce expression of several ECM proteins 14,16,23 , which are normally processed in the ER. Consistent with this, we observed that Dex increases de novo Figure 9. Increased co-localization of fibronectin with ER stress marker, KDEL in post-mortem human glaucomatous TM tissues. Post-mortem human TM tissues from age-matched normal (n = 6) and glaucoma (n = 6) were stained with fibronectin (green) and KDEL (red) and co-localization was examined in anterior segment tissues. Representative images are shown at lower (A) scale bar is 100 microns) and higher (B) scale bar is 25 microns) magnification. Both fibronectin and KDEL was increased in human glaucomatous TM tissues. In addition, increased co-localization of fibronectin with KDEL was observed in 5 out of 6 human glaucomatous TM tissues examined.
protein synthesis of secretory proteins, including selected ECM proteins. We observed increased load of fibronectin and Col I in the ER of TM cells, which may further compromise ER homeostasis, inducing ER stress in TM cells. Genetic knock down of fibronectin alone was able to partially prevent Dex-induced ER stress. Since we hypothesize that Dex-induced ER stress is result of a cumulative load of several ECM proteins that are increased by Dex, a single knock down of ECM proteins other than fibronectin may not abrogate Dex-induced ER stress. We chose fibronectin knock down for multiple reasons. First, fibronectin is one of the predominant ECM protein in the TM and it is prominently induced by Dex. Second, fibronectin is known to regulate the assembly of other ECM proteins in TM cells. A recent study has shown that disruption of fibronectin matrix prevents Dex-induced deposition of collagen type IV, fibrillin and laminin into ECM of TM cells 54 . This would result in overall decrease in Dex-induced overload of other ECM proteins, thereby preventing ER stress in TM cells. Third, fibronectin was increased and co-localized to ER stress proteins in the human glaucomatous TM tissues. To evaluate the effects of other ECM on ER stress, we have utilized the ECM derived from Dex-treated cells. Primary TM cells grown on ECM derived from Dex-treated cells exhibited ER stress indicating that Dex-induced extracellular ECM is sufficient to induce ER stress. Nonetheless, it should be noted that our study is limited to examining the effects of fibronectin alone in Dex-induced ER stress.
It is likely that Dex-induced excessive intracellular ECM load is a result of increased synthesis load, decreased proteosomal degradation, and inefficient secretion. In addition, our studies demonstrate that increased extracellular fibronectin deposition alone is sufficient to induce ER stress in TM cells as evident from the following findings. First, treatment of human primary TM cells with cellular fibronectin, which forms extracellular fibronectin deposits induces ER stress. Second, induction of ER stress was evident when normal TM cells were grown on decellularized ECM from Dex-treated cells. These findings are important for POAG pathology since several studies have shown extracellular deposition of ECM material 14,23 as well as cellular fibronectin 7 in glaucomatous TM tissues. Increased co-localization of fibronectin with ER stress in human TM tissues from POAG donor eyes further supports that ER stress is associated with increased ECM in the TM. It is interesting to note that ER stress induction was only observed in the TM region of POAG donor eyes despite the increase in both fibronectin and Col I in other regions of anterior segment tissues. This further supports our hypothesis that TM is more sensitive to ER stress from ECM accumulation.
To our knowledge, this is first study that demonstrates induction of ER stress from increased ECM accumulation. It is not clear how TM cells communicate with the ECM to induce ER stress. Cells can communicate with ECM via fibronectin-integrin interactions to regulate cellular responses [61][62][63] . Fibronectin is known to act as major regulator of ECM assembly 56,64 . Therefore, fibronectin may act as a major regulator of Dex-induced ECM accumulation and ER stress induction in TM cells. Consistent with this, a recent study has demonstrated that treatment of primary human TM cells with cellular fibronectin increased expression of other ECM proteins including collagen I and laminin 65 . In the present study, we observed that knockdown of fibronectin reduced Dex-induced ER stress in TM cells. Moreover, treatment of TM cells with cellular fibronectin was sufficient to induce ER stress. It is therefore likely that fibronectin regulates several other ECM proteins in the TM. It is interesting to note that TM cells are more sensitive to ER stress from Dex-induced ECM accumulation compared to other ocular cells. Considering the unique function of TM cells, it is possible that TM cells are incapable of handling increased protein accumulation, inducing chronic ER stress proteins such as CHOP, which may further exacerbate chronic ER stress.
In conclusion, we demonstrate that increased ECM synthesis and deposition leads to ER stress in the TM, which may be associated with TM dysfunction and IOP elevation. We further present evidence that TM cells are more sensitive to ER stress from Dex-induced ECM accumulation. Importantly, ER stress is associated with increased ECM accumulation in human TM tissues from POAG donors. It is possible that targeting ER stress pathway may provide strategies to reduce ECM accumulation in the glaucomatous TM.

Materials and Methods
Antibodies. Antibodies were purchased from the following sources: fibronectin (catalog # Ab2413, Abcam, . Both KDEL and GRP78 were used as ER stress markers. Mouse KDEL antibody was primarily used for immunostaining since it works well on fixed cells and formalin fixed human tissues and can be combined with rabbit fibronectin or collagen I antibody. We observed that KDEL primarily recognizes GRP78 and GRP94 in Western blot analysis of human TM cells. Therefore, we used KDEL as an ER stress marker. Rabbit GRP78 antibody was used along with mouse puromycin antibody for immunostaining. For Western blot analysis of ER stress in cells, we primarily utilized the GRP78 antibody. Dex treatment of mice. Topical 0.1% Dex phosphate (Bausch & Lomb Inc.) or Veh eye drops containing sterile phosphate-buffered saline (PBS) were applied to 3-month-old A/J mice thrice daily for 7 weeks (Fig. 1) as described previously 41 . At the end of the treatment, whole eyes were fixed & sectioned for immunostaining. Intraocular Pressure (IOP) Measurement. Intraocular pressure (IOP) was determined in behaviorally trained conscious mice using a TonoLab rebound tonometer (Colonial Medical Supply, Franconia, NH, USA) as previously described 66 . Day-time IOPs were measured between 10 AM to 2 PM. Approximately an average of 6 individual IOP measurements were taken to calculate the final IOP value for each eye. All IOP measurements were recorded in a masked manner.

Aqueous Outflow Facility (C).
Aqueous outflow facility (C) was measured in 7-week Dex-treated mice by constant flow infusion as previously described 67,68 . Briefly, animals were anesthetized (ketamine 100 mg/kg; xylazine 10 mg/kg, I.P.), and the tip of a 30G steel needle was carefully placed into the anterior chamber. The needle was connected by tubing to a flow-through pressure transducer (BLPR2, World Precision Instruments (WPI), Sarasota, FL USA). The opposing end of the transducer was connected to a 1mL syringe filled with sterile filtered PBS loaded into an infusion pump (Microdialysis SP101i, WPI). A virtual chart recorder (LabScribe2, WPI) continuously displayed pressure signals. Following equilibration (20-30 min), eyes were infused at a flow rate of 100 nL/min. When pressure had stabilized (10-15 min), flow rates were increased sequentially (200 to 500 nL/ min, in 100 nL/min increments). Three stabilized pressures at each flow rate were recorded, and the average value calculated as mean stabilized pressure. C was calculated as the reciprocal of the slope of a plot of Mean Stabilized Pressure as the ordinate and Flow Rate the as abscissa. Decellularization. Decellularization was performed as described previously 72,73 . Briefly, primary human TM cells (n = 2 strains) were grown on 4-well chamber slides. After reaching approximately 70% confluency, cells were treated with either Veh or Dex (100 nM) for 5 days. Following the treatment, cells were detached by treating with 2 mL of 0.2% Triton X-100 in water at 37 °C for 10 min and detached cells were washed off with 1X PBS, (3 times). 2 mL of 0.3% ammonium hydroxide solution was added slowly to the wells and incubated for 5 min at 3 °C. Nuclear staining with DAPI was done to ensure a complete decellularization, and increased deposition of ECM was confirmed by immunostaining for fibronectin. Following the complete decellularization process, untreated primary human TM cells were re-plated on the same 4-well chamber slides and cultured to confluence. Immunostaining was used to analyze the ER stress (KDEL) in the TM cells induced by the decellularized ECM.

Effect of Dexamethasone on different ocular cells. Primary human TM cells as well as primary
human corneal fibroblasts and primary human retinal pigment epithelium cells (obtained from ScienCell Research laboratories, Carlsbad, USA) were grown on 4-well chamber slides. After reaching an approximately 70% confluence, cells were treated with either Veh or Dex (100 nM) for 3 days. Dex-induced ECM (FN) and ER stress (KDEL) were analyzed by immunostaining. Transformed GTM3 and ARPE-19 cells were grown to approximately 70% confluence in 6-well plates and treated for 3 days either with Veh or Dex or cellular fibronectin (10 ug/ ml). Following the treatment, medium was removed and cells were washed with 1X PBS and lysed using 1X lysis buffer, supplemented with 1X protease inhibitor cocktail. Cell lysates were subjected to Western blotting to analyze ECM (FN) and ER stress (KDEL and CHOP) markers. GAPDH was used as the loading control.
Immunostaining. Mouse eyes from Veh or Dex-treated groups were enucleated and fixed in freshly prepared 4% PFA for 3 hours, cryo-protected by keeping fixed eyes in 30% sucrose overnight before OCT embedding. 10 µ thin sections were made using a cryostat (Leica Inc, Buffalo Grove, IL, USA). The sections were allowed to dry at room temperature prior to use. The sections were then permeabilized using 0.1% Triton-X-100 in PBS for 15 min and blocked with 10% normal goat serum for an hour. Slides were incubated overnight with primary antibody (1:100 dilution) in 10% (v/v) normal goat serum, and then washed 3 times with PBS followed by 2-hour incubation with appropriate Alexa Fluor secondary antibodies (1:200; Life technologies, Grand Island, NY, USA). Sections were mounted with DAPI-mounting solution. Images were captured using Keyence microscope (Itasca, IL, USA). Cultured TM cells were fixed in 4% PFA for 15 minutes and permeabilized with 0.1% Triton-X-100 in 1xPBS for 10 minutes followed by 30 minutes blocking with 10% goat serum. The cells were then incubated with appropriate primary & secondary antibodies similarly as described above.
Western blot analysis. The protein samples (30 ug total protein) were run on denaturing 4-12% gradient polyacrylamide ready-made gels (NuPAGE Bis-Tris gels, Life technologies, Grand Island, NY, USA) and transferred onto PVDF membranes. Blots were blocked with 10% non-fat dried milk for 1 hour then incubated overnight with specific primary antibodies at 4 °C on a rotating shaker. The membranes were washed thrice with PBST and incubated with corresponding HRP-conjugated secondary antibody for 90 minutes. The proteins were then visualized using ECL detection reagents (SuperSignal West Femto Maximum Sensitivity Substrate; Life technologies, Grand Island, NY, USA).
Surface sensing of translation (SUnSET) Assay. GTM3 cells were grown to approximately 70% confluence in 6-well tissue culture plates. Cells were treated with either vehicle control (ethanol, 0.1%) or Dex (100 nM) in serum free medium for 3 days. Cells without any treatment acted as a control. Following 3 days, one of the controls was treated with cycloheximide (10 ug/ml) for 16 hrs. The conditioned medium was removed from all the wells and incubated with serum free medium containing puromycin (10 ug/ml) for 30 min as described previously 52 . Following 30 min incubation, puromycin containing medium was removed and cells were washed twice with 1x PBS. Cells were lysed with ice cold 1x lysis buffer, containing protease and phosphatase inhibitors cocktail. The obtained cell lysates were subjected to Western blot analysis. Rabbit Anti puromycin antibody (1:10,000, Sigma) was used as the primary antibody and goat anti rabbit-HRP conjugated was used as the secondary antibody for analyzing puromycin incorporated into the newly synthesized proteins. Actin on the same blot was used as loading control and Coomassie brilliant blue staining was performed to examine total proteins. Co-immunoprecipitation. GTM3 cells were treated with Veh or Dex for 3 days and lysates were immunoprecipitated with anti-KDEL antibody using Dynabeads ™ Co-Immunoprecipitation Kit (Life technologies, Grand Island, NY, USA). KDEL or IgG antibodies (2.5 ug) were diluted in 200 ul of lysis buffer and incubated with 50 ul of Dynabeads protein G for 10 minutes at room temperature. The beads-Ab complexes were washed with lysis buffer 3 times according to manufacturer protocol using a magnet separator. Total protein lysates (100 ug) were added to the above complexes for 10 minutes at room temperature. The beads-Ab-Ag complexes were then washed 3 times with lysis buffer containing 0.5% Triton-x-100 and precipitated proteins were analyzed by SDS/ PAGE and Western blotting following our standard protocol. Note that cell lysates from Veh-treated cells (10 ug) was loaded in SDS/PAGE gel as an input.
Immunohistochemistry of human TM tissues. Age-matched (between 70-80 yrs) normal (n = 6) and POAG (n = 6) donor eyes were stained for fibronectin, Col I and the ER stress maker KDEL as described previously 49 . The eyes were obtained and managed in compliance with the Declaration of Helsinki. Briefly, eyes were obtained from the Lions Eye Institute for Transplant and Research (Tampa, FL) within 6 hours of death and fixed in 10% formalin. This is an accredited eye bank where donor or family permission is received and the eye tissues can be used for transplant or research. The eye tissues were fixed at the Lions Eye Institute before sending to our lab. The slides containing the anterior segment sections were deparaffinized in xylene and dehydrated twice with 100%, 95%, 70%, 50% ethanol for 5 minutes each. It should be noted that these sections were not subjected to antigen retrieval because we observed strong fibronectin staining in the cornea that obscured staining in TM region when sections were processed in boiling citrate buffer. Sections were incubated with 1% Triton containing 10% goat serum for 2 hours. Sections were incubated overnight at 4 °C with mouse KDEL and rabbit fibronectin or collagen I antibodies (1:300 in blocking buffer). Slides were washed with PBS and incubated with Alexa donkey-anti-rabbit 568 and Alexa donkey anti-mouse 488 (1:500 in PBS, Thermo Fisher Scientific Inc., Pittsburgh PA) for 2 hrs. Sections were washed with PBS and mounted with mounting medium containing DAPI (Vector Labs, Inc. CA). Sections incubated with no primary antibody served as a negative control.