Endoplasmic Reticulum Stress Disrupts Mitochondrial Bioenergetics, Dynamics and Causes Corneal Endothelial Cell Apoptosis

Abstract

Purpose: Endoplasmic reticulum (ER) and mitochondrial stress are independently associated with corneal endothelial cell (CEnC) loss in many corneal diseases, including Fuchs’ endothelial corneal dystrophy (FECD). However, the role of ER stress in mitochondrial dysfunction contributing to CEnC apoptosis is unknown. The purpose of this study is to explore the crosstalk between ER and mitochondrial stress in CEnC.

Methods: Human corneal endothelial cell line (HCEnC-21T) and human corneal endothelial tissues were treated with ER stressor tunicamycin. ER stress-reducing chemical 4-phenyl butyric acid (4-PBA) was used in HCEnC-21T after tunicamycin. Fuchs’ corneal endothelial cell line (F35T) was used to determine differential activation of ER stress with respect to HCEnC-21T at the baseline. ER stress, mitochondrial-mediated intrinsic apoptotic, mitochondrial fission, and fusion proteins were determined using immunoblotting and immunohistochemistry. Mitochondrial bioenergetics were assessed by mitochondrial membrane potential (MMP) loss and ATP production at 48 hours after tunicamycin. Mitochondria dynamics (shape, area, perimeter) were also analyzed at 24 hours using transmission electron microscopy.

Results: Treatment of HCEnC-21T cell line with tunicamycin activated three ER stress pathways (PERK-eIF2α-CHOP, IRE1α-XBP1, and ATF6), reduced cell viability, upregulated mitochondrial-mediated intrinsic apoptotic molecules (cleaved caspase 9, caspase 3, PARP, Bax, cytochrome C), downregulated anti-apoptotic Bcl-2 protein, initiated mitochondrial dysfunction by loss of MMP and lowering of ATP production, and caused mitochondrial swelling and fragmentation with increased expression of mitochondrial fission proteins (Fis1 and p-Drp1). Fuchs’ CEnC (F35T) cell line also showed activation of the ER stress-related proteins (p-eIF2α, GRP78, CHOP, XBP1) compared to HCEnC-21T at the baseline. The 4-PBA ameliorated cell loss and reduced cleaved caspase 3 and 9, thereby rescuing tunicamycin-induced cell death but not mitochondrial bioenergetics in HCEnC-21T cell line.

Conclusions: Tunicamycin-induced ER stress disrupts mitochondrial bioenegetics, dynamics and contributes to the loss of CEnC viability. This novel study highlights the importance of ER-mitochondria crosstalk and its contribution to CEnCs apoptosis, seen in many corneal diseases, including FECD.

Corneal endothelium is a monolayer of hexagonal cells with limited proliferative capacity¹ lining the posterior surface of the cornea. It constantly pumps ions, thereby maintaining corneal transparency, and is a barrier between the corneal stroma and the aqueous humor. Fuchs’ endothelial corneal dystrophy (FECD) is a female-prevalent, age-related oxidative disorder of CEnCs with altered morphology.^2,3 It is characterized by extracellular matrix (ECM) deposition known as corneal guttae on the Descemet's membrane that causes progressive cellular edema, resulting in CEnC degeneration and apoptosis.^4,5 FECD affects 4% of the population in the United States over age 40, with a higher incidence in females.^6,7 The only treatment for FECD is corneal transplantation, posing a need for therapeutic interventions. In FECD, CEnC degeneration is primarily attributed to oxidant-antioxidant imbalance,³ mitochondrial,⁸ as endoplasmic reticulum (ER) stress⁹ apart from the genetic factors.¹⁰ 

ER and mitochondrial stress are two of the major stress factors contributing to CEnC degeneration in FECD. ER stress results because of alterations in cellular protein synthesis and folding mechanisms under pathological conditions. The homeostatic adaptive mechanisms for synthesizing new proteins in response to ER stress is called unfolded protein response (UPR), resulting in activation of three signaling protein pathways: Protein-kinase RNA-like ER kinase (PERK)/eukaryotic initiation factor 2 alpha (eIF2α)/activating transcription factor 4 (ATF4)/ C/EBP-homologous protein (CHOP) (PERK/eIF2α/ATF4/CHOP); inositol requiring protein 1α (IRE1α)/X-box binding proteins (XBP1) (IRE1α/XBP1); and activating transcription factor 6 (ATF6). With all pathways of ER stress, glucose-related protein 78 (GRP78), also known as BiP (Heat shock protein 70), is a major ER chaperone, binds three transmembrane ER stress sensors (IRE, PERK, and ATF6), and prevents their activation under unstressed conditions. However, under stress conditions, ER stress sensors are released, thereby disrupting their interactions with GRP78, leading to ER stress. Corneal guttae in FECD patients could result from excessive ECM deposition primarily because of sustained unfolded protein response (UPR)/ER stress.¹¹ Also, CEnCs of FECD patients demonstrate enlarged ER structure, including increased expression of GRP78, eIF2α, and CHOP.^9,12 Similarly, UPR has been involved in CEnC loss in alpha 2 collagen VIII transgenic knock-in mice, an animal model of early-onset FECD.¹³ 

Like ER stress, mitochondrial stress is also implicated in CEnC degeneration in FECD.⁸ Mitochondria, being a powerhouse of cells, regulate many key physiological processes.^14,15 Upon various endogenous or exogenous stimuli, such as DNA damage, oxidative or ER stress, ischemia, and more, the mitochondria-mediated intrinsic apoptotic pathway is activated with initial disruption of mitochondria membrane potential (MMP), followed by release of cytochrome C along with other proapoptotic proteins, formation of apoptosome, and activation of executioner caspase 3/6/7, resulting in apoptosis.¹⁶ Intrinsic apoptotic pathways further disrupt mitochondria quality control and alter mitochondrial dynamics with loss of mitochondrial fusion proteins, increased mitochondrial fragmentation proteins resulting in mitochondrial fission, and altered mitophagy. Maintaining mitochondrial function is central to aging processes and critical for degenerative loss in post-mitotic cells in other organs, like that in post-mitotically fixed CEnCs in FECD.¹⁷ In FECD, abnormal mitochondrial bioenergetics is associated with loss of MMP,¹⁷ abnormal ATP production, and increased mitochondrial reactive oxygen species after oxidative stress.¹⁸ Similarly, aberrant mitochondrial dynamics in FECD involve abnormal mitochondrial fusion or fission proteins leading to mitochondrial fragmentation, mitochondrial DNA damage, and excessive/constitutive activation of mitophagy.¹⁷ 

ER and mitochondrial stress have been independently associated with CEnC apoptosis in FECD. However, very few studies have suggested the crosstalk of ER and mitochondria stress in CEnC for FECD. Specifically, it is unknown how ER stress controls mitochondrial energetics and dynamics contributing to CEnC apoptosis, as seen in FECD. This is the first study demonstrating ER stress-induced disruption of mitochondrial pathways and their role in CEnC apoptosis. We hypothesize that ER stress disrupts mitochondrial bioenergetics, dynamics and activates mitochondrial-mediated intrinsic apoptotic pathways in CEnC. The study highlights the importance of ER-mitochondria crosstalk during stress response and opens windows for developing therapeutic interventions targeting ER and mitochondrial pathways simultaneously in FECD. 

Methods

Cell Culture

Human corneal endothelial cell line (HCEnC-21T) (a kind gift from Ula V. Jurkunas, Harvard University) was maintained in Chen's medium containing Opti-MEM (cat no. 51985034; Thermo Fisher, Waltham, MA, USA), 8% fetal bovine serum (cat no. A3840002; Thermo Fisher), 0.1 gm calcium chloride (cat no. C7902; Sigma-Aldrich, St. Louis, MO, USA), 0.4 gm chondroitin sulfate (cat no. C9819, Sigma-Aldrich), gentamycin (0.5%) (cat no. 15750078; Thermo Fisher) antibiotic-antimycotic (1%) (cat no. 15240096; Millipore Sigma, Burlington, MA, USA), 25 µL of epidermal growth factor (cat no. GF144; Thermo Fisher) 50 µg of bovine pituitary extract (cat no. 50-753-3049; Fisher Scientific). Cells were grown on a T75 flask (cat no. 156499; Thermo Fisher) coated with FNC coating mix (cat no. 0407; Athena Enzyme Systems, Baltimore, MD, USA) passaged every two or three days and maintained in an incubator containing 5% CO₂ at 37°C. The culture of Fuchs’ corneal endothelial dystrophy cell line 35 (F35T, 1500 CUG repeats) (a generous gift from Dr. Albert S. Jun, Johns Hopkins University) was maintained in Opti-MEM (cat no. 51985034; Thermo Fisher) supplemented with 8% fetal bovine serum, 0.1 gm calcium chloride (cat no. C7902; Sigma-Aldrich), 0.4 gm chondroitin sulfate (cat no. C9819; Sigma-Aldrich), gentamycin (0.5%) (cat no. 15750078; Thermo Fisher), antibiotic-antimycotic (1%) (cat no. 15240096; Millipore Sigma), 25 µL of epidermal growth factor (cat no. GF144; Thermo Fisher), 100 µL of nerve growth factor (cat no. 13257-019; Thermo Fisher), and 10 mg of ascorbic acid (cat no. 255564; Sigma-Aldrich). Cells were grown on a T25 flask (cat no. 156340; Thermo Fisher) coated with FNC coating mix (cat no. 0407; Athena Enzyme Systems) passaged every two or three days and maintained in an incubator containing 5% CO₂ at 37°C. 

Cell Viability

For the MTT assay, HCEnC-21T (1 × 10⁴ cells/well) were seeded and grown in 96-well plates (cat no. 701003; Fisher Scientific,) until 70% to 80% confluency. Subsequently, after desired confluency, they were treated with various concentrations of dimethyl sulfoxide (DMSO) (0.02%, 0.1%, 0.2%) and tunicamycin (1, 5, 10 µg/mL) and incubated for 24 hours. To reduce ER stress, 4 phenyl butyric acid (4-PBA, 2.5 µM) was used. We performed MTT (3-(4,5-dimethylthiazol-2-yl)-2-5-diphenyltetrazolium bromide) assay (cat no. V13154; Thermo Fisher) for determining cell viability after different treatments. Briefly, the culture medium was removed, and MTT (5 mg/10 mL of Opti-MEM) was added in each well and incubated for four hours. After incubation, the medium was removed, 100 µL of DMSO was added to each well, and the plate was kept on a shaker for 10 minutes. The absorbance of the samples was recorded at 540 nm using a plate reader (Molecular SpectraMax M2; Molecular Devices, San Jose, CA, USA). We also determined cell death by using lactate dehydrogenase (LDH) released into the media. LDH release was calculated after tunicamycin treatment (1, 10 µg/mL) for 48 hours in HCEnC-21T cells using a Cytotoxicity Detection Kit (MAK380; Sigma-Aldrich) according to the manufacturer's protocol. The absorbance was measured at 450 nm using a plate reader (Varioskan; Thermo Fisher). 

Immunocytochemistry

HCEnC-21T (4 × 10⁴ cells/well) were seeded in 8 well-chamber slides (cat no. 154534PK; Thermo Fisher) to grow for reaching 70% to 80% confluency. For CHOP staining, cells were treated with different concentrations of DMSO (0.02%, 0.1%, 0.2%) and TUN (1, 5, 10 µg/mL) and incubated at 37°C cells for 24 hours. After treatment, cells were fixed with paraformaldehyde (PFA) (4%) (cat no. J19943.K2; Thermo Fisher) for 20 minutes, then blocked for one hour in blocking buffer (5% normal goat serum + 0.3% Triton X-100 [cat no. X100; Millipore Sigma]) and then incubated with mouse anti- CHOP antibody (cat no. L63F7, 1:1000; Cell Signaling, Danvers, MA, USA) overnight at 4°C. The next day, the cells were washed with ice-cold phosphate buffer saline solution (PBS) (cat no. 10010072; Thermo Fisher) three times and incubated with goat anti-mouse secondary antibody (Alexa Fluor 488, #ab150113, 1:500; Abcam, Cambridge, UK) for one hour. Cells were washed again with cold-ice PBS and stained nuclei with DAPI (cat no. P36962, Prolong Antifade Mountant with DAPI; Thermo Fisher). Slides were visualized on a confocal microscope (Leica STED 3X; Leica, Wetzlar, Germany). A similar immunocytochemistry method was used for staining mitochondria by cytochrome C (cat no. 12963, 1:1000; Cell Signaling) except for treatment (0.2%, 0.4% DMSO; 10, 20 µg/mL tunicamycin for six hours), fixation with 4% PFA for 10 minutes, permeabilization with 0.01% Triton X-100 for five minutes, and blocking (1% BSA + 10% normal goat serum + 0.1% Tween 20) for one hour. For quantifying CHOP⁺DAPI⁺ cells per visual field (×40 lens), 10 to 15 images were taken under each condition, and the total number of CHOP⁺ DAPI⁺ cells or fragmented mitochondria was counted and averaged. A similar procedure was followed for calculating the percentage of fragmented mitochondria. Quantification was masked to reduce the experimental bias. For the human corneal tissue immunocytochemistry, Descemet's membrane containing endothelial cells was stripped from human donor corneas and treated with DMSO (0.2%) and tunicamycin (10 µg/mL) and incubated for 24 hours. After treatment, the tissues were incubated with MitoTracker Deep Red FM (50 nM) (cat no. M22426; Thermo Fisher) for 30 minutes, then fixed in 4% PFA (cat no. J19943.K2; Thermo Fisher) for 30 minutes. After fixation, the tissues were permeabilized in 0.5% Triton-X 100 (cat no. X100; Millipore Sigma) in PBS (cat no. 10010072; Thermo Fisher) for 30 minutes and blocked in blocking buffer (10% goat serum, ab7481; Abcam) for one hour. The tissues were then incubated with mouse anti-CHOP antibody (cat no. L63F7,1:1000; Cell Signaling) overnight at 4°C. Tissues were washed with PBS the following day and incubated with goat anti-mouse secondary antibody (Alexa Fluor 488, cat no. ab150113, 1:500; Abcam) for one hour. After secondary antibody incubation, tissues were washed with PBS thrice and mounted using ProLong Diamond Antifade Mountant with DAPI (cat no. P36962; Thermo Fisher). The tissues were visualized on a confocal microscope (Leica STED 3X; Leica). 

Immunoblotting

HCEnC-21T cells were treated either with DMSO or Tunicamycin for 6 or 24 hours for analyzing ER and mitochondrial stress proteins and for 48 hours for analyzing mitochondrial oxidative phosphorylation (OxPhos) complexes. F35T cells were cultured and grown as mentioned in the methods and collected without any treatment for analysis of ER stress markers. Cells were lysed using RIPA lysis and extraction buffer (cat no. 89900; Thermo Fisher) containing protease and phosphatase inhibitor cocktail (cat no. 78440, 1:1000; Thermo Fisher) after treatment (DMSO: 0.02, 0.1, 0.2%; tunicamycin: 1, 5, 10 µg/mL, 4-PBA: 2.5 µM) at six or 24 hours and spun in a centrifuge at 10,000g for 10 minutes at 4°C. The supernatant was collected and assessed for quantifying protein concentration (A53227; Thermo Fisher). Protein 30 µg mixed with appropriate LDS sample buffer (cat no. NP0007; Thermo Fisher) and reducing agent (cat no. NP0009; Thermo Fisher) was subjected to 4% to 12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (cat no. NP032; Thermo Fisher), followed by transfer to polyvinylidene difluoride membrane (cat no. 88518; Thermo Fisher) at 30 V for 90 minutes. The blots were blocked in blocking buffer (25 mM Tris, 0.15 M NaCl, 0.05% Tween 20, 5% skim milk) for one hour at room temperature and then treated with primary antibody (GRP78 cat no. 3183, PERK cat no. 3192, p-elF2α cat no. 3398, p-PERK (cat no. 3179), eIF2α (cat no. 5324), CHOP cat no. 2895, ATF6 cat no. 65880, IRE1α cat no. 3294, XBP-1s cat no. 27901, Bcl2 cat no. 4223, Bax cat no. 41162, cytochrome C cat no. 11940, cleaved caspase 9 cat no. 20750, cleaved caspase 3 cat no. 9664, PARP cat no. 9532, DRP1 cat no. 8570, pDRP1 cat no. 4867, OPA1 cat no. 80471, Mfn2 cat no. 11925, Actin cat no. 3700; all from Cell Signaling), OxPhos antibody cocktail cat no. 45-8099, Thermo Fisher) overnight at 4°C. The next day, the blots were washed three to five times in TBS-Tween 20 (25 mM Tris, 0.15 M NaCl, 0.05% Tween 20) (cat no. 28360; Thermo Fisher) and HRP-conjugated secondary antibody (anti-mouse IgG cat no. 7076S, Anti-rabbit IgG cat no. 7074S; Cell Signaling) was added for one hour with similar subsequent washing. The immunoreactive bands were visualized using a Femto (cat no. 34094; Thermo Fisher) or pico chemiluminescent substrate (cat no. 34579; Thermo Fisher). The band intensity was measured and quantified using ImageJ software. 

Measurement of MMP

MMP was measured using the TMRE-Mitochondrial Membrane Potential Assay Kit (ab113852; Abcam). HCEnC-21T (1 × 10⁴ cells/well) were seeded in 96-black well flat bottom plates (cat no. 3916; Corning, Inc., Corning, NY, USA) until 70% to 80% confluency, then treated with DMSO (0.2%) and tunicamycin (10 µg/mL) and incubated for 48 hours. Carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP, 20 µM) was used as positive control and was added to untreated cells 10 minutes before TMRE staining. After treatment, the culture medium was removed, and 200 nM TMRE in Opti-MEM (cat no. 11058021; Thermo Fisher) was added to each well and incubated for 15 minutes. After incubation, the TMRE stain was removed, and 100 µL of DMSO was added per well. Fluorescence (ex/em: 549/575 nm) was measured using a plate reader (Varioskan LUX; Thermo Fisher). 

ATP Production

ATP levels were detected using the ATP Assay Kit (ab83355; Abcam) according to the manufacturer's protocol. HCEnC-21T (1 × 10⁴ cells/well) were seeded in 96-well plates and treated with DMSO (0.2%) and tunicamycin (10 µg/mL) when confluency reached 70% to 80%. HCEnC was incubated with treatment for 48 hours. The colorimetric signal was measured at 570 nM using a plate reader (Varioskan LUX; Thermo Fisher). 

Single-Cell RNA Seq

HCEnC-21T cell line was cultured as mentioned above, grown to 80% to 90% confluency, and then treated with DMSO (0.02%) and tunicamycin (1 µg/mL) for 24 hours. Single-cell RNA seq experiments were performed per the manufacturer's protocol using the 10X Genomics Chromium system at Genomic Core Facility, Mount Sinai, NY. RNA-seq data analysis of tunicamycin and DMSO-treated cells was performed using Partek Flow bioinformatics software (Partek, Inc., St. Louis, MO, USA). Cells were filtered using single-cell quality assurance/quality control and normalized using Partek's automatic normalization tool. Genes with zero expression in 99% of all cells were filtered out, as recommended by Partek's scientific team. To demonstrate differential expression of pathways related to ER and mitochondria, gene set enrichment analysis (GSEA) of differential pathway expression was performed between the tunicamycin and DMSO-treated HCEnC-21T cells. Pathways were filtered based on false discovery rate adjusted P value or q value (q < 0.05) and fold change (FC) < −2 or > 2. Relevant pathways that demonstrated significantly enriched gene sets were plotted on a bar graph based on normalized enrichment score (q < 0.05). Oxidative phosphorylation leading edge genes for tunicamycin and DMSO-treated HCEnC-21T cells were plotted by heatmap. To confirm the validity of our RNA seq data, we performed permutation tests as described¹⁹ using R programming language. There are two groups: (a) control with 6254 cells and (b) tunicamycin treatment with 6445 cells. Each cell in control and treatment is an independent data point. We calculated the absolute difference between the control and treatment mean expression values for each gene, thus defining the observed (test) statistic. We then performed 100,000 random permutations of the data for each gene and compared the resultant permutation statistics for each permutation to the calculation for our observed/test statistic. The P value was calculated for each gene by assessing how many of the permutation statistics were greater than our test statistic. We accounted for multiple tests (one for each gene) using the Bonferroni multiple correction method for which we divided the original P value by the number of tests being performed. 

Transmission Electron Microscopy

HCEnC-21T (4 × 10⁴ cells/well) were seeded in 8 well Permanox Lab-Tek Chamber Slides (cat no. 70413; Fisher Scientific), grown to 80% confluency, and treated with 10 µg/mL of tunicamycin. Cells were fixed with 2% paraformaldehyde/2.5% glutaraldehyde in 0.1M sodium cacodylate solution (cat no. 15960-01, Electron Microscopy Sciences (EMS); Fisher Scientific) at 4°C, rinsed in 0.1 M sodium cacodylate buffer, then post-fixed with 2% osmium tetroxide/1.5% potassium ferricyanide in 0.1 M sodium cacodylate, and en bloc stained with 2% uranyl acetate in dH₂O. Cells were dehydrated in an ethanol series (25% EtOH/dH₂O up to 100% EtOH), infiltrated through an ascending ethanol/resin series (cat no. 14120, EMS Embed 812 Kit; Fisher Scientific) and placed in pure resin overnight. Chambers were separated from the slides, and a modified BEEM embedding capsule (EMS, cat no. 70000-B; Fisher Scientific) was placed over defined areas containing cells. Capsules were filled with a drop of pure resin and placed in a vacuum oven to polymerize at 60°C for several hours. Epon resin was added to fill the capsules and polymerized for 72 hours in the vacuum oven.  Immediately after polymerization, capsules were snapped from the substrate with pliers to dislodge the cells from the slide. Semithin sections (0.5 µm) were obtained using a Leica UC7 ultramicrotome (Leica), counterstained with 1% toluidine Blue, placed under a coverslip, and viewed under a light microscope to identify successful dislodging of cells. Ultra-thin sections were collected on copper 300 mesh grids using a Coat-Quick adhesive pen. Sections were counter-stained with 1% uranyl acetate and lead citrate and imaged on an HT7500 transmission electron microscope (Hitachi High-Technologies, Tokyo, Japan) using an AMT NanoSprint12 12-megapixel CMOS transmission electron microscopy (TEM) Camera (Advanced Microscopy Techniques, Danvers, MA, USA). Final images (20–30 images per condition) were collected, and the brightness, contrast, and size of these images were adjusted using Adobe Photoshop CS4 software version CS4 11.0.1 (Adobe, Inc., San Jose, CA, USA). Swollen mitochondria and ER structures were counted under control and treated conditions. Mitochondria area and parameter were calculated using the ImageJ software. 

Statistical Analysis

We used Student's t-test (unpaired) and one-way ANOVA (Tukey's multiple comparison test) for experiments having a group of two and more than two groups, respectively, using GraphPad Prism 8 (V8, La Jolla, CA, USA). The level of significance was chosen at *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 with mean ± SEM. All the experiments were repeated with three to nine technical replicates. 

Results

Induction of ER Stress by Tunicamycin in Human Corneal Endothelial Cell Line in Vitro and Human Corneal Endothelial Tissues Ex Vivo

Over the last decade, research has suggested ER stress's involvement in the pathophysiology of FECD.^9,11,20 We first investigated the differential activation of ER stress-related proteins in Fuchs’ corneal endothelial cell line (F35T) compared to normal HCEnC-21T cell line. Immunoblotting data showed increased expression of GRP78, CHOP, p-eIF2α, and XBP1 in F35T compared to HCEnC-21T cell line under the baseline without tunicamycin treatment (Fig. 1A). These data confirm the involvement of ER stress in FECD. We then investigated whether tunicamycin can activate all three known pathways of UPR/ER stress (PERK-eIF2α-CHOP, IRE1α-XBP1, and ATF6) in HCEnC-21T cell line. Because the baseline ER stress markers were high in the F35T cell line, we did not use tunicamycin in this cell line. As mentioned in the introduction, GRP78, being a gatekeeper, remains bound with all the ER stress sensors (IRE1α, ATF6, and PERK) under normal physiological conditions, which are activated and released after stress. Figure 1B demonstrated the increased expression of GRP78 protein at 24 hours in 1, 5, and 10 µg/mL tunicamycin-treated HCEnC-21T cell line compared to 0.02%, 0.1%, and 0.2% DMSO-treated control groups suggesting the activation of an important gatekeeper protein (GRP78) of UPR. For activation of the PERK-eIF2α-CHOP pathway of UPR, PERK activates eIF2α, leading to attenuation of global protein synthesis, which in turn activates the CHOP-mediated apoptotic pathway. We demonstrated increased expression of p-PERK specifically in 5 µg/mL tunicamycin-treated HCEnC-21T cell line compared to 0.1% DMSO-treated control groups with no change in PERK expression, increased expression of p-eIF2α in 1, 5, and 10 µg/mL tunicamycin-treated HCEnC-21T cell line compared to 0.02%, 0.1%, and 0.2% DMSO-treated control groups, as well as increased expression of CHOP in 1, 5, and 10 µg/mL tunicamycin-treated HCEnC-21T cell line compared to 0.02%, 0.1%, and 0.2% DMSO-treated control groups (Fig. 1B) at 24 hours. Some ER stress markers such as p-PERK were not upregulated at lower doses of tunicamycin. However, all tunicamycin doses demonstrated activation of pro-apoptotic CHOP molecules. Concerning IRE1α-XBP1 pathway activation, XBP1 gets spliced and activated by IRE1α on ER stress to initiate and regulate UPR transcriptional programs. Figure 1C demonstrates the appearance of a spliced form of XBP1 at 24 hours in 1, 5, and 10 µg/mL tunicamycin-treated HCEnC-21T cell line compared to 0.02%, 0.1%, and 0.2% DMSO-treated control groups. Another UPR/ER stress sensor ATF6 is cleaved on ER stress, as demonstrated by Figure 1D at 24 hours in 1, 5, and 10 µg/mL tunicamycin-treated-treated HCEnC-21T cell line compared to 0.02%, 0.1%, and 0.2% DMSO-treated control groups. We also confirmed CHOP expression in normal HCEnC-21T cell line at 24 hours after tunicamycin treatment (1, 5, and 10 µg/mL). Figure 1E demonstrated the increased number of CHOP⁺DAPI⁺ cells (marked by white arrows) at 24 hours in 1, 5, and 10 µg/mL tunicamycin-treated HCEnC-21T cell line compared to 0.02%, 0.1%, and 0.2% DMSO-treated control groups. Specifically, there was a 3.9-, 4-, and 5.6-fold increase in the number of CHOP⁺DAPI⁺ cells at 24 hours in 1, 5, and 10 µg/mL tunicamycin-treated HCEnC-21T cell line compared to 0.02%, 0.1%, and 0.2% DMSO-treated control groups. To further validate the cell line data, we confirmed the induction of CHOP at 24 hours after tunicamycin treatment (10 µg/mL) in human corneal endothelial tissue ex-vivo using immunohistochemistry. (Fig. 1G). These data suggest the activation of the critical proteins in the three major UPR/ER stress pathways (PERK-eIF2α-CHOP, IRE1α-XBP1, and ATF6) by tunicamycin in the human corneal endothelial cell line, as well as human corneal endothelial tissues and differential activation of ER stress proteins in Fuchs’ cell line (F35T) compared to normal HCEnC-21T cell line at baseline. 

Figure 1.

 

Tunicamycin activates ER stress in HCEnC-21T cell line. Representative immunoblot showing the activation of the activation of (A) ER stress-related proteins (GRP78, p-eIF2α, eIF2α, CHOP, and XBP1) in Fuchs’ CEnC (F35T) cell line compared to normal HCEnC-21T cell line, (B) PERK-eIF2α-CHOP pathway with increased protein expression of p-PERK, GRP78, p-eIF2α, and CHOP, (C) IRE1α-XBP1 pathway demonstrating cleavage and activation of XBP1, and (D) ATF6 pathway also showing cleavage of ATF6 in HCEnC-21T cell line at 24 hours after treatment with tunicamycin (1, 5, and 10 µg/mL) compared to DMSO control groups (0.02%, 0.1%, and 0.2%). (E) Immunostaining showing an increased number of CHOP⁺DAPI⁺ cells, also demonstrated by white arrows. (F) Quantification as a bar graph at 24 hours in tunicamycin-treated HCEnC-21T cell line (1, 5, and 10 µg/mL) compared to DMSO-treated groups (0.02%, 0.1%, and 0.2%). (G) Immunostaining showing an increase of CHOP release (green) and mitochondrial fragmentation as observed by Mitotracker staining (red) post tunicamycin treatment (10 µg/mL) compared to DMSO control group (0.2%) in human corneal tissue. (n = 5, ****P < 0.0001, one-way ANOVA with Tukey's multiple comparison test; scale bar: 50 µm).

Induction of Caspase-Mediated Apoptosis and Reduction of Cell Viability by Tunicamycin in Human Corneal Endothelial Cell Line

ER stress often results in apoptosis and has been implicated in the pathophysiology of many ocular diseases,²¹ including FECD.⁹ Thus we demonstrated the activation of caspase-mediated apoptosis and diminished cell survival in the tunicamycin-treated HCEnC-21T cell line. Specifically, caspase-3 is a member of the caspase family of 13 aspartate-specific cysteine proteases, playing a critical role in the execution of an apoptotic program, and primarily activates/cleaves PARP, a mediator of DNA repair.²² Here we showed increased cleaved PARP and caspase-3 protein expression in the tunicamycin-treated HCEnC-21T cell line compared to DMSO-treated control groups (Fig. 2A). Specifically, there was a 1.9-, 1.7-, and 1.5-fold increase in the expression of cleaved PARP for 1, 5, and 10 µg/mL of tunicamycin-treated HCEnC-21T cell line, respectively, compared to 0.02%, 0.1%, and 0.2% of DMSO-treated control groups (Fig. 2B). Similarly, cleaved caspase 3 protein expression increased 2.5-, 2.6-, and 2.7-fold in 1, 5, and 10 µg/mL of tunicamycin-treated HCEnC-21T cell line, respectively, compared to 0.02%, 0.1%, and 0.2% of DMSO-treated control groups (Fig. 2C). To confirm decreased cell viability or increased apoptosis after tunicamycin-induced ER stress, we also quantified cell viability using an MTT assay. The cell viability decreased significantly at 24 hours for 1, 5, and 10 µg/mL tunicamycin-treated HCEnC-21T cell line, compared to 0.02%, 0.1%, and 0.2% of the DMSO-treated control groups, respectively (Fig. 2D). We also confirmed cell death by measuring LDH release, which significantly reduced cell viability after tunicamycin (1 and 10 µg/mL) treatment (Fig. 2E). These findings suggest that tunicamycin-induced ER stress activates apoptosis in CEnCs by inducing caspase 3 and PARP, ultimately reducing cellular viability. 

Figure 2.

 

Tunicamycin activates the caspase-mediated apoptotic pathway and reduces cell viability in HCEnC-21T cell line. (A) Representative immunoblot showing cleaved (cle) PARP and caspase-3 in HCEnC-21T cell line at 24 hours after treatment with tunicamycin (1, 5, and 10 µg/mL) compared to DMSO control (0.02%, 0.1%, and 0.2%). Bar graph demonstrating increased expression of (B) cleaved PARP and (C) cleaved caspase 3 normalized by actin in tunicamycin-treated HCEnC-21T cell line compared to DMSO control. (For cleaved PARP, n = 4, **P < 0.01, *P < 0.05. For cleaved caspase 3, n = 4, **P < 0.01, ***P < 0.001, one-way ANOVA with Tukey's multiple comparison test.) (D) Bar graph showing reduced cell viability in HCEnC-21T cell line at 24 hours after treatment with tunicamycin (1, 5, and 10 µg/mL) compared to DMSO control (0.02%, 0.1%, and 0.2%) (n = 6, tunicamycin 1 µg/mL, **P < 0.01, tunicamycin 5 µg/mL, ***P < 0.001, tunicamycin 10 µg/mL, **P < 0.01, one-way ANOVA with Tukey's multiple comparison test). (E) Reduction in cell viability measured by LDH release in HCEnC-21T cell line at 48 hours after treatment with tunicamycin (1 and 10 µg/mL) compared to DMSO control (0.02% and 0.2%). (n = 6, tunicamycin 1 µg/mL, *P < 0.05, tunicamycin 10 µg/mL, **P < 0.01, one-way ANOVA with Tukey's multiple comparison test.)

Activation of Mitochondria-Mediated Intrinsic Apoptosis Pathway by Tunicamycin in Human Corneal Endothelial Cell Line

Upon various intracellular stressors such as DNA damage, oxidative or ER stress, the mitochondria-mediated intrinsic apoptotic pathway starts with activation of cleaved caspase 9, followed by increased expression of proapoptotic cleaved caspase 3, Bax, and cytochrome C release with a subsequent decrease in antiapoptotic protein, Bcl2. To investigate activation of mitochondria-mediated intrinsic apoptotic pathway post tunicamycin-induced ER stress, we showed increased expression of proapoptotic cleaved caspase-9, as well as decreased expression of antiapoptotic Bcl-2 protein at 24 hours in tunicamycin-treated HCEnC-21T cell line (1, 5, and 10 µg/mL) compared to DMSO-treated control groups (0.02%, 0.1%, and 0.2%) (Fig. 3A). Specifically, there was a 1.61-, 1.53-, and 1.46-fold increase in cleaved caspase 9 protein expression for 1, 5, and 10 µg/mL tunicamycin-treated HCEnC-21T cell line, respectively, compared to 0.02%, 0.1%, and 0.2% of DMSO-treated control groups (Fig. 3B). Concerning antiapoptotic Bcl-2 protein expression, there was a significant 1.21-, 1.39-, and 1.35-fold decrease in Bcl-2 protein expression at 24 hours for 1, 5, and 10 µg/mL of tunicamycin-treated HCEnC-21T cell line, compared to 0.02%, 0.1%, and 0.2% of DMSO-treated control groups, respectively (Fig. 3C). Proapoptotic cytochrome C and Bax protein expression did not change at 24 hours after 1, 5, and 10 µg/mL dose of tunicamycin (data not shown). Thus we determined the cytochrome C and Bax expression at six hours after a 20 µg/mL dose of tunicamycin in the HCEnC-21T cell line. cytochrome C and Bax increased 1.33- and 1.5-fold at six hours in 20 µg/mL of tunicamycin-treated HCEnC-21T cell line compared to 0.4% DMSO-treated control group (Fig. 3D). Using immunocytochemistry, we further demonstrated the release of cytochrome C in the cytosol at six hours in 20 µg/mL of tunicamycin-treated HCEnC-21T cell line compared to the DMSO-treated control group (0.4%) (Fig. 3E). These findings suggest that tunicamycin-induced ER stress activates mitochondrial-mediated intrinsic apoptotic pathways in CEnCs. 

Figure 3.

 

Tunicamycin induces mitochondria-mediated intrinsic apoptotic pathways in HCEnC-21T cell line. (A) Representative immunoblot demonstrating increased cleaved caspase-9 expression and decreased expression of Bcl-2 in HCEnC-21T cell line at 24 hours after treatment with tunicamycin (1, 5, and 10 µg/mL) compared to DMSO control (0.02%, 0.1%, 0.2%). Bar graph exhibiting increased expression of (B) cleaved caspase 9 and decreased expression of (C) Bcl-2 normalized by actin in tunicamycin-treated HCEnC-21T cell line compared to control. (For cleaved caspase 9, n = 4, *P < 0.05 for all doses of tunicamycin compared to DMSO control, one-way ANOVA with Tukey's multiple comparison tests. For Bcl-2, n = 4, tunicamycin 1 µg/mL, *P < 0.05, n = 4 tunicamycin 5 µg/mL,***P < 0.001, n = 4 tunicamycin 10 µg/mL,**P < 0.01.) (D) Representatives immunoblot and a bar graph showing increased expression of Bax, Cyt C at six hours for tunicamycin-treated HCEnC-21T cell line (20 µg/mL) compared to DMSO control group (0.4%). (For cytochrome C, n = 3, *P < 0.05, for Bax, n = 3, **P < 0.01, unpaired Student's t-test.) (E) Immunostaining showing the release of cytochrome C (green) from the mitochondria (mitotracker red) to the cytoplasm shown by white arrow and fragmented mitochondria by white arrowhead in tunicamycin treated HCEnC-21T cell line (20 µg/mL) compared to DMSO control group (0.4%), Scale bar: 50 µm.

Altered Mitochondrial Bioenergetics with Mitochondrial Depolarization, Impaired ATP Production, and Oxidative Phosphorylation Genes by Tunicamycin in Human Corneal Endothelial Cell Line

ER stress can disrupt mitochondria bioenergetics by initiating the loss of MMP and decreased ATP production, leading to the reduction of mitochondrial oxidative phosphorylation genes and mitochondrial complexes in many pathological conditions. We first investigated whether there are any baseline differences in MMP between normal and Fuchs’ cell lines. There were no differences in MMP for HCEnC-21T and F35T cell lines (Supplementary Fig. S1A). To investigate the effect of ER stressor tunicamycin on mitochondrial bioenergetics in corneal endothelial cells, we measured MMP and ATP in HCEnC-21T cell line at 48 hours after tunicamycin (10 µg/mL) treatment and compared to DMSO control (0.2%) (Figs. 4A, 4B). The bar graphs demonstrate a significant decrease in MMP (Fig. 4A) and ATP levels (Fig. 4B) in the tunicamycin-treated HCEnC-21T cell line compared to the DMSO-treated control group. These data suggest that the impaired mitochondrial bioenergetics are probably initiated by altered ATP production and MMP loss after ER stress. We also used an ER stress-reducing agent (4-PBA) to test whether tunicamycin-mediated altered mitochondrial bioenergetics could be rescued. The supplementary data showed that 4-PBA did not prevent MMP loss following tunicamycin (10 µg/mL) at 24 hours in the HCEnC-21T cell line (Supplementary Fig. S1B), as well as Fuchs’ CEnC (F35T) cell line without tunicamycin (Supplementary Fig. S1A). Similarly, 4-PBA did not increase ATP production in HCEnC-21T cell line at 24 hours after tunicamycin (10 µg/mL) (Supplementary Fig. S1C). Altered mitochondrial bioenergetics after stress often results in impaired mitochondrial OxPhos genes and loss of mitochondrial OxPhos complexes. To investigate this hypothesis, we first perform transcriptome analysis using single-cell RNA-seq in HCEnC-21T cell line treated with either DMSO or tunicamycin. GSEA demonstrated upregulated genes set enriched in the N-glycan biosynthesis, protein export, protein processing in ER, and ferroptosis pathways in tunicamycin-treated HCEnC-21 T cell line compared to DMSO control (Fig. 4C). This suggests that tunicamycin induces disruption in protein processing and contributes to apoptosis. Moreover, ubiquitin-mediated proteolysis, glycolysis/gluconeogenesis, TGF-beta, and oxidative phosphorylation were downregulated in the tunicamycin group compared to control (Fig. 4C). Specifically, we showed a downregulation of oxidative phosphorylation genes in the tunicamycin-treated group compared to the DMSO control group. This was demonstrated by a heatmap of leading-edge genes generated using a normalized enrichment score (q < 0.05) by Partek software (Fig. 4D). To confirm bioinformatics analysis, we performed permutation tests that suggested that there were significant differences in the expression for oxidative phosphorylation genes between tunicamycin and DMSO treated groups, with P < 0.001 similar to adjusted P value or q value (q < 0.05) generated using Partek software. Most of our oxidative phosphorylation genes are associated with electron transport chain/mitochondrial OxPhos complexes. To confirm bioinformatics analysis about downregulation of OxPhos genes and its relation of mitochondrial OxPhos complexes after ER stress, we performed immunoblotting for OxPhos complexes and demonstrated decreased expression of mitochondrial complexes I, II, III, and IV in tunicamycin (10 µg/mL)-treated HCEnC-21T cell line compared to the DMSO (0.2%)-treated control group at 48 hours and also reduced expression of complexes II and III with 1 µg/mL tunicamycin treatment compared to DMSO (0.02%) (Fig. 4E). All these results confirm that the dysfunctional mitochondrial bioenergetics demonstrated by MMP loss and decreased ATP production and loss of mitochondrial OxPhos genes and complexes under tunicamycin-induced ER stress. 

Figure 4.

 

Tunicamycin alters mitochondrial bioenergetics by depolarizing MMP, lowering ATP levels, and decreasing oxidative phosphorylation genes and complexes in HCEnC-21T cell line. (A) Bar graph showing loss of MMP measured by TMRE Assay and (B) lowering of ATP levels; at 48 hours after treatment with tunicamycin (10 µg/mL) in HCEnC-21T cell line compared to DMSO control suggesting mitochondrial depolarization. (n = 3, tunicamycin 10 µg/mL, *P < 0.05, ****P < 0.0001, unpaired t-test.) (C) Transcriptome analysis demonstrated the upregulation of common pathways related to ER stress (glycan biosynthesis, protein processing, and export) and apoptosis (ferroptosis), as well as the downregulation of proteasome and ubiquitin-mediated proteolysis, glycolysis, and mitochondrial oxidative phosphorylation using GSEA analysis. (D) Heat map of 50 oxidative phosphorylation genes showing downregulation in tunicamycin treated HCEnC cell line compared to DMSO treated group (adjusted P value or q < 0.05 and fold change (FC) <−2 or >2). (E) Representative immunoblot showing decreased expression of mitochondrial complexes I, II, III, IV in tunicamycin treated HCEnC-21T cell line (10 µg/mL) compared to DMSO control group (0.2%) at 48 hours. Tunicamycin 1 µg/mL also reduces complexes II and III expression at 48 hours.

Disruption of Mitochondrial Dynamics with Mitochondrial Swelling and Fragmentation by Tunicamycin in Human Corneal Endothelial Cell Line as Well

Tunicamycin-induced ER stress disrupts mitochondrial dynamics in metabolic diseases such as diabetes²³ and skeletal muscle-related diseases.²⁴ To investigate tunicamycin-induced changes in mitochondrial dynamics in HCEnC-21T cell line, we first performed TEM. We found swollen mitochondria (marked by white arrows) at 24 hours in tunicamycin-treated HCEnC-21T cell line (10 µg/mL) compared to the DMSO-treated control group (0.2%) (Fig. 5A). There was a significant 3.6-, 1.5-, and 1.3-fold increase in the number of swollen mitochondria, mitochondrial area, and parameter, respectively, for the tunicamycin-treated HCEnC-21T cell line compared to the DMSO-treated control group (Figs. 5B, 5C), suggesting the altered mitochondrial morphology after ER stress. ER stress usually results in altered ER morphology, as well shown by white arrowhead (Fig. 5A) and a 5-fold increase in the number of swollen ER after tunicamycin-treated HCEnC-21T cell line compared to the DMSO-treated control (Fig. 5B). Mitochondrial swelling ultimately results in its fragmentation after ER stress. Thus we investigated mitochondrial fragmentation after tunicamycin-induced ER stress in HCEnC-21T cell line. Figure 5D demonstrated normal, well-defined tubular mitochondria in DMSO-treated and fragmented (dotted donut-shaped) mitochondria in the tunicamycin-treated HCEnC-21T cell line. Specifically, there were 11.7- and 9.25-fold increases in % mitochondrial fragmentation in the 10 and 20 µg/mL tunicamycin-treated HCEnC-21T cell line, respectively, compared to 0.2% and 0.4% DMSO-treated groups (Figs. 5E, 5F). We also showed mitochondrial fragmentation (mitotracker staining) in human corneal endothelial tissues ex vivo after treatment of tunicamycin (10 µg/mL) at 24 hours) (Fig. 1G), which further validates our mitochondrial fragmentation data in the cell line. To further understand the mechanism of mitochondrial fragmentation, we investigated the two important proteins (Fis1 and Drp1) primarily involved in mitochondrial fission and dynamics after stress. We demonstrated increased expression of Fis1 (1.67-fold) and pDrp1 (1.53-fold) for the tunicamycin-treated HCEnC-21T cell line (1 µg/mL) compared to the DMSO control (0.02%) at six hours (Fig. 5G), suggesting activation of Fis1-Drp1 mediated mitochondrial fragmentation. Apart from increased mitochondrial fragmentation proteins after ER stress, there are often decreased mitochondrial fusion proteins in the ER or other stress-induced mitochondrial fragmentation events. Two common mitochondrial fusion proteins, OPA-1, and MFN-2, often mediate mitochondrial fusion and regulate mitochondrial dynamics after stress. Interestingly, we did not observe any differences in the expression of MFN-2 and OPA-1 at six hours for tunicamycin-treated HCEnC-21T cell line (1 µg/mL) compared to DMSO-treated control groups 0.02%. Decrease or degradation of mitochondrial fusion proteins might be delayed, which we may detect with a higher dose of tunicamycin or later time points in a lower dose of tunicamycin in HCEnC-21T cell line. 

Figure 5.

 

Tunicamycin disrupts mitochondrial dynamics by inducing mitochondrial swelling and fragmentation in HCEnC-21T cell line. (A) TEM showing mitochondrial swelling (white arrows) and enlarged ER (white arrowhead) at 24 hours in tunicamycin-treated HCEnC-21T cell line (10 µg/mL) compared to DMSO control group (0.2%). (B) Bar graph representing an increased number of swollen ER and mitochondria per visual field in tunicamycin-treated HCEnC-21T cell line compared to DMSO-treated control group (n = 3, ****P < 0.0001, unpaired Student's t-test, scale bar: 1 µm). (C) Bar graph also representing increased mitochondria area and parameter in tunicamycin-treated HCEnC-21T cell line compared to DMSO-treated control group (n = 4, ***P < 0.001, unpaired Student's t-test). (D) Representative immunostaining of mitochondria by cytochrome C (cyt c) demonstrating dotted donut-shaped mitochondria fragmentation (white arrow) at six hours after tunicamycin treatment (10 and 20 µg/mL) compared to DMSO control group (0.2% and 0.4%) in HCEnC-21T cell line. Bar graph representing an increased percentage of mitochondrial fragmentation at six hours in (E) 10 µg/mL and (F) 20 µg/mL tunicamycin-treated HCEnC-21T cell line compared to their 0.2% and 0.4% DMSO control groups, respectively (n = 3, ****P < 0.0001, unpaired Student's t-test). (G) Representative Western blot along with bar graph demonstrating increased expression of Fis1 and phosphor-Drp1 (pDRP1) at six hours in tunicamycin-treated HCEnC-21T cell line (1 µg/mL) compared to DMSO control group (0.02%) (n = 3, *P < 0.05, unpaired Student's t-test). (H) Western blot showing no differences in mitochondrial fusion proteins (OPA-1 and MFN-2) expression at six hours in tunicamycin-treated HCEnC-21T cell line (1 µg/mL) compared to DMSO control group (0.02%).

Rescue of Cell Viability and Decreased Mitochondrial-Mediated Apoptotic Mediator Proteins by 4-PBA After Tunicamycin Treatment in Corneal Endothelial Cell Line

The 4-PBA is a well-established ER stress-reducing compound that increases ocular cell survival and decreases apoptotic proteins in many in vitro²⁵ and in vivo ocular disease models.^26,27 Thus we first investigated the effect of 4-PBA on cell viability. The cell viability decreased 2.4-, 2.1-, and 2.2-fold for 1, 5, and 10 µg/mL of tunicamycin-treated HCEnC-21T cell line, respectively, compared to 0.02%, 0.1%, and 0.2% DMSO-treated control groups (Fig. 6A), as demonstrated earlier in Figure 2D. The cell viability increased 1.6-, 1.7-, and 1.8-fold after 4-PBA (2.5 µM) and 1, 5, and 10 µg/mL treated-tunicamycin compared to 1, 5, and 10 µg/mL tunicamycin treatment alone in HCEnC-21T cell line at 24 hours (Fig. 6A), indicative of activation of cell survival/protective mechanisms by 4-PBA. Moreover, there were no significant differences in the cell viability between DMSO and 4-PBA treated HCEnC-21T cell line (Fig. 6A), suggesting minimum toxicity of 4-PBA and DMSO at selected doses. We then investigated the effect of 4-PBA on an important mitochondrial-mediated apoptotic molecule, cleaved caspase 9, and general apoptotic, and mitochondrial-mediated molecule cleaved caspase 3 and PARP after tunicamycin. Cleaved caspase 9 and 3 protein expressions increased by 4.8- and 2.7-fold in 1 µg/mL tunicamycin-treated HCEnC-21T cell line compared to 0.02% DMSO-treated control group (Figs. 6B, 6D, 6E), suggesting activation of mitochondria and general cellular apoptotic pathway as shown earlier (Fig. 2A, Fig. 3A). The 4-PBA reduced cleaved caspase 9 by 1.5-fold and cleaved caspase 3 by 1.7-fold after tunicamycin treatment compared to tunicamycin alone (1 µg/mL) at 24 hours in HCEnC-21T cell line (Figs. 6B, 6D, 6E), indicative of inhibition of mitochondrial and general apoptotic pathways by 4-PBA. Concerning cleaved PARP, tunicamycin (1 µg/mL)-treated HCEnC-21T cell line did increase cleaved PARP expression (2.6-fold) compared to the DMSO (0.02%)-treated control group (Figs. 6B, 6C), as previously shown (Figs. 2A, 6B). However, there were no significant differences in cleaved PARP expression between tunicamycin+ 4-PBA treated HCEnC-21T cell line compared to tunicamycin alone at 24 hours (Figs. 6B, 6C), suggesting 4-PBA does not involve PARP in inhibiting apoptosis. Moreover, there were no changes in the expression of cleaved PARP, caspase 9, and 3 between the 4-PBA and DMSO group, suggesting minimum toxicity for 4-PBA and DMSO (Figs. 6B–E). 

Figure 6.

 

The 4-PBA rescues cell viability and reduces mitochondrial-mediated intrinsic apoptotic protein, cleaved caspase 9 and general apoptotic mediator, cleaved caspase 3 in tunicamycin treated HCEnC-21T cell line. (A) Bar graph showing increased cell viability in tunicamycin (1, 5, and 10 µg/mL) and 4-PBA (2.5 µM) treated HCEnC-21T cell line compared to tunicamycin treatment (1, 5, and 10 µg/mL) alone and reduced cell viability in HCEnC-21T cell line at 24 hours after treatment with tunicamycin (1, 5, and 10 µg/mL) compared to DMSO control (0.02%, 0.1%, and 0.2%) (*P < 0.05, one-way ANOVA with Tukey's multiple comparison test). (B) Representative Western blot and (C, D, E) bar graph showing decreased cleaved caspase 9 and 3 but no change in cleaved PARP protein expression in tunicamycin (1 µg/mL) and 4-PBA (2.5 µM) treated HCEnC-21T cell line compared to tunicamycin treatment (1 µg/mL) alone and increased cleaved caspase 3 and 9 and PARP at 24 hours in tunicamycin-treated (1 µg/mL) HCEnC-21T cell line compared to DMSO-treated group (0.02%) (for cleaved PARP, n = 9, ***P < 0.001, for cleaved cas 9, n = 9, ****P < 0.0001, for cleaved cas 3, n = 9, **P < 0.01, one-way ANOVA with Tukey's multiple comparison test).

Discussion

In this study, we have investigated ER stress induction in HCEnC-21T and F35T²⁸ under baseline unstressed conditions and the impact of tunicamycin-induced ER stress on mitochondrial bioenergetics, dynamics, and cellular apoptosis in the HCEnC-21T cell line. Specifically, we have demonstrated the upregulation of all three ER stress pathways (PERK/eIF2α/ATF4/CHOP, IRE1α/XBP1, ATF6) after tunicamycin treatment of HCEnC-21T cell line. Moreover, tunicamycin-induced ER stress initiated MMP loss, decreased ATP production and oxidative phosphorylation gene expression, induced mitochondrial swelling, and fragmentation with upregulation of mitochondria fission proteins along with induction of the mitochondrial-mediated intrinsic apoptotic pathway in HCEnC-21T cell line. ER stress-reducing agent, 4-PBA, rescued cell viability and decreased expression of important intrinsic apoptotic proteins. 

Our findings about the induction of ER stress by tunicamycin have been supported by many studies in vitro^11,29 and in vivo³⁰ for developing ocular disease phenotypes. Specifically, there is induction of CHOP and degeneration of photoreceptors in the retina,²⁹ activation of PERK, GRP78, CHOP, and XBP-1 in the lacrimal gland leading to inflammation and dry eye syndrome,³⁰ and activation of ER stress molecules and elevation of endothelial permeability and leakage of extracellular superoxide dismutase leading to retinal endothelial dysfunction³¹ after intravitreal injection of tunicamycin in vivo. Our study demonstrated that tunicamycin-induced ER stress reduces cell viability and activates proapoptotic cleaved caspase 3 in CEnCs in vitro, supported by the Muller cell line, MIO-M1 cell line,³² and retinal pigmental cell line (ARPE-19),³³ neurons,³⁴ cardiomyocytes,³⁴ intestinal epithelial cells,³⁵ and auditory cells.³⁶ Similarly, we demonstrated cleaved PARP after tunicamycin supported by studies in hepatoma,³⁷ mouse embryonic fibroblasts,³⁸ and choriocarcinoma cells.³⁹ We have not investigated the crosstalk of ER stress pathway (i.e., PERK/ATF4/CHOP with IRE1α/XBP1 or ATF6) after tunicamycin, which is a potential limitation of this study and will be explored in the future. 

Concerning the mitochondrial-mediated intrinsic apoptotic pathway, we demonstrated increased expression of proapoptotic cleaved caspase 9, Bax, cytochrome C, and decreased antiapoptotic protein, Bcl-2, in tunicamycin treated HCEnC-21T cell line. Tunicamycin treatment increases cleaved caspase 9 expressions in human leukemia cell lines (U937, HL-60),⁴⁰ auditory cell line (HCI-OC1),⁴¹ melanoma cell line (Me1007), cytochrome C release in primary mouse hepatocytes,⁴² Bax and cytochrome C in rat hippocampal neurons,⁴³ human umbilical vein endothelial cell lines (HUVECs, A549).⁴⁴ Similarly, the antiapoptotic protein, Bcl-2 decreased in the melanoma cell line (Mel-RM, MM200),⁴⁵ small lung cancer cell lines (NCI-H446, H69),⁴⁶ human tumor cell line (SW620),⁴⁷ and primary rat cardiomyocytes⁴⁸ after tunicamycin. We have observed decreased MMP and ATP levels and oxidative phosphorylation genes in HCEnC-21T cell line indicative of altered mitochondrial bioenergetics supported by similar studies in human adipocytes cell line (Chub-S7),²³ skeletal muscle cell line,²⁴ fibroblasts cells⁴⁹ and neuroblastoma cell line.⁵⁰ We have the role of specific ER stress pathways in regulating mitochondrial dynamics or bioenergetics under acute or chronic pharmacological or physiological ER stress chemical, which is a limitation of our study and will be investigated in the future. For ER stress-induced mitochondrial dynamics alterations, our study suggested swelling of mitochondria with the increased area, perimeter, and mitochondrial fragmentation, supported by similar studies in human trabecular meshwork stems cells and trabecular network cells,⁵¹ adipocyte cell line (Chub-S7),²³ mouse embryonic fibroblasts,^23,49,51,52 skeletal muscle cell line,²⁴ and neuroblastoma cell line (SH-SY5Y).⁵⁰ However, we also demonstrated increased expression of Fis-1 and p-DRP1, two important mitochondrial fragmentation proteins, but no change in the mitochondrial fusion proteins (OPA1, MFN2) after ER stress supported by a similar study in an adipocyte cell line (Chub-S7) at 48 and 72 hours after tunicamycin in the human skeletal muscle cell line²⁴ at 24 hours after tunicamycin. 

The 4-PBA has been used to reduce ER stress and, ultimately, cell loss in many disease models, such as retina ischemia-reperfusion injury,⁵³ glucocorticoid-induced ocular hypertension,⁵⁴ spinal cord injury,⁵⁵ and a myocilin mouse model of primary open-angle glaucoma⁵⁶ diabetic corneal endothelial injury.⁵⁷ Our study also suggests that 4-PBA prevents cell death with reduced cleaved caspase 3 and 9 in vitro. Other studies using different stressors support these results in Fuchs’. Okumura et al.²⁰ demonstrated that 4-PBA reduces extracellular matrix proteins and ER stress and rescues cell loss post-TGF-β treatment with reduced cleaved caspase 3 in Fuchs’ cell line.²⁰ Specifically, the same group showed that specific knockdown of CHOP reduces mitochondrial membrane potential and cleaved caspase 9 expression (intrinsic mitochondrial apoptotic pathway) after MG132-mediated ER stress in Fuchs’ cell line,¹¹ which also supports our cleaved caspase 9 results after 4-PBA. However, we could not rescue cleaved PARP after applying an ER stress-reducing agent, 4-PBA. This was surprising because 4-PBA decreases cleaved PARP expression after tunicamycin-induced ER stress in the hepatic cell line (HepG2 cell line)³⁷ and colonic epithelial cell line (LS174T).⁵⁸ It might be possible that 4-PBA might not rescue all the substrates of caspase-mediated apoptosis, such as PARP cleavage, after tunicamycin-mediated ER stress. Moreover, we did not observe the rescue of altered mitochondrial bioenergetics (ATP and MMP loss) after 4-PBA, which is a limitation of our study and will be investigated in the future. This can be due to the dose of 4-PBA/tunicamycin or time points used for assessing the effects, which needs to be explored further to gain insights into the crosstalk for chemical-induced ER stress and mitochondrial dysfunction. We will also investigate different ER stressors such as thapsigargin and ER stress-reducing agents such as trehalose and tauroursodeoxylcholic acid or physiological stressors such as Ultraviolet A irradiation⁵⁹ in the future. Few studies suggested reduced mitochondrial dysfunction after 4-PBA application under tunicamycin-induced ER stress. For example, He et al.⁶⁰ demonstrated tunicamycin-induced systolic and diastolic heart dysfunction with loss of mitochondria biogenesis and fusion genes, and use of 4-PBA prevented the mitochondrial dysfunction. However, many studies suggested that 4-PBA does prevent ER-induced mitochondrial dysfunction under pathological conditions such as aging heart condition⁶¹ and Parkinson's disease.⁶² Chen et al.⁶³ demonstrated that hyperglycemia induced ER and mitochondrial bioenergetic dysfunction in corneal endothelial and 4-PBA ameliorated hyperglycemia-induced mitochondrial bioenergetic deficits.⁶³ One of other potential limitations of this study is the unknown downstream effectors of ER stress-mediated mitochondrial apoptotic or altered bioenergetics or dynamics pathways. Also there are no in vivo experiments in this study, which limits the applicability of the findings in preclinical or clinical settings. 

The impact of ER-mitochondria crosstalk has been reported in many diseases, such as neurodegenerative,^64,65 cardiac diseases,⁶⁶ and metabolic diseases such as diabetes and kidney diseases.^67,68 ER and mitochondria stress independently contribute to ocular pathologies such as age-related macular degeneration,⁶⁹ retinitis pigmentosa,⁷⁰ glaucoma,⁷¹ diabetes retinopathy,⁷² and more. However, very few studies have examined the role and implications of ER-mitochondrial crosstalk in ocular diseases.⁷³ Matsunaga et al.⁷⁴ demonstrated the neuroprotective role of humanin, a small mitochondria-encoded peptide in retinal pigmented cells after tunicamycin-induced ER stress, by upregulating mitochondrial glutathione, suggesting intraorganellar crosstalk in mediating stress. However, the detailed molecular mechanisms of humanin-mediated ER-mitochondrial crosstalk remain elusive. Similarly, Yumnamcha et al.⁷⁵ showed that the pre-treatment of mitochondrial-targeted antioxidant, MitoQ, or inhibitor of the mitochondrial apoptotic molecule, Bid attenuated MMP loss and cell death in a retinal pigmental cell line (ARPE-19) after tunicamycin-induced ER stress. Our study provides exciting insights into crosstalk of ER and mitochondrial stress proteins, which can be screened and targeted simultaneously using high-throughput drug screening and open new avenues for drug discovery. 

Acknowledgments

The authors sincerely thank Allison Sowa, Bill Janssen, and Shilpa Dilipkumar for electron microscopy processing and imaging at The Microscopy CoRE and Advanced Bioimaging Center, Kristin Grant Beaumont, Sanjana Shroff, Anna Borges at Genomics core facility, Icahn School of Medicine at Mount Sinai. We also thank Arham Akhyer and Jannet Pawar (high school students) for their help in performing experiments and Ula V. Jurkunas (Schepens Eye Research Institute, Harvard University) for providing HCEnC-21T cell line and Albert Jun (Wilmer Eye Institute, John Hopkins University) for providing F35T cell line. 

Supported by NIH/NEI (4R00EY031339), Mount Sinai Seed Money, New York Eye and Ear Foundation, Sarah K de Coizart Charitable Trust/Foundation awarded to V.K. and Challenge Grant from Research to Prevent Blindness awarded to ophthalmology department. 

Disclosure: S. Qureshi, None; S. Lee, None; W. Steidl, None; L. Ritzer, None; M. Parise, None; A. Chaubal, None; V. Kumar, None