Mitochondrial Permeability Transition Pore in Inflammatory Apoptosis of Human Conjunctival Epithelial Cells and T Cells: Effect of Cyclosporin A

Abstract

Purpose.: To investigate the role of mitochondrial permeability transition pore (MPTP) and effect of cyclosporin A (CsA) on inflammatory apoptosis of human conjunctival epithelial cells (IOBA-NHC) and T cells.

Methods.: IOBA-NHC and Jurkat cells were stimulated with IFNγ, TNFα, αFas, or PMA/αCD3, in the presence or absence of CsA. MPTP was determined using the calcein-cobalt technique. Mitochondrial membrane potential (ΔΨm) was measured with JC-1. Apoptosis was quantified by Annexin V/PI staining. Apoptosis mediators were evaluated by flow cytometry or Western blot.

Results.: In IOBA-NHC, TNFα, and IFNγ induced MPTP opening, ΔΨm loss, and increased cell apoptosis. This was accompanied by upregulation of Fas/FasL; Bax; and caspase-3, -8, and -9 activation. Addition of CsA prevented IOBA-NHC from cell death by blocking MPTP opening, ΔΨm loss, Fas/FasL, and caspase activation. In PMA/αCD3-activated Jurkat T cells, MPTP opening and ΔΨm loss were increased along with cell apoptosis and upregulated Fas/FasL/caspase expressions. CsA further promoted T-cell apoptosis, ΔΨm loss, and upregulation of Fas/FasL/caspase.

Conclusions.: Inflammation induces aberrant MPTP opening, resulting in an increased apoptosis in conjunctival epithelial cells. CsA protected IOBA-NHC from cell death by blocking both intrinsic and extrinsic apoptosis pathways. CsA promoted T-cell apoptosis via upregulating Fas/FasL and caspase activities with a minimal effect on MPTP. The findings suggest that the differential effect of CsA on T cells versus ocular surface resident epithelial cells may contribute to its therapeutic efficacy in treating ocular inflammation such as dry eye disease.

Introduction

Apoptosis occurs in autoimmune and inflammatory diseases not only as a consequence but as an integral part of disease mechanism. ^1,2 Dry eye (DE) is an immune-mediated inflammatory disorder of the ocular surface, ³ with a histologic hallmark of progressive lymphocytic infiltration prominently by CD4⁺ T cells, and increased apoptotic cell death in the ocular resident epithelium of the lacrimal functional unit. ^4,5 Growing evidence suggests that aberrant apoptosis contributes to DE pathogenesis and is a common phenotype in ocular inflammation. ^6–8 Expressions of Fas/Fas ligand (FasL), Bcl-2 family proteins, and caspase activations were found altered at gene and/or protein levels, ^8–12 suggesting that cell death may proceed via either or both the intrinsic (mitochondria-mediated) and extrinsic (initiated by “death” receptors) apoptosis pathways. ^13,14 Mitochondria play an essential role in controlling cell death. The main form of apoptosis in most vertebrates goes through a mitochondria-mediated pathway. ¹⁵ The mitochondrial permeability transition pore (MPTP) serves as a checkpoint to determine the fate of cells. ¹⁶ MPTP is an inner mitochondrial membrane high-conductance channel, whose open–closed transition is required for maintaining an electrochemical gradient for respiration and adenosine 5′-triphosphate (ATP) production, which is modulated by the mitochondrial membrane potential (ΔΨm). ¹⁷ Cellular stress–induced ΔΨm dissipation leads to an aberrant MPTP opening that liberates cytochrome c (Cyt c) release to the cytosol, triggering a cascade of caspase activation that propagates apoptotic signals. Caspase-9 is activated in apoptosome composed of Apaf and Cyt c released from mitochondria. Death-receptor signaling activates initiator caspases, such as caspase-8 in death-receptor signalosome complex. Caspase-8 and -9 in turn trigger cleavage of executioner caspases including caspase-3, resulting in DNA degradation and ultimately cell death. 

Cyclosporin A (CsA), an immunomodulator, has demonstrated therapeutic effect in treating dry eye. ^9,18–22 CsA inhibited lymphocytic infiltration, reduced expressions of immune activation and inflammation markers, ^23,24 and suppressed apoptosis in the lacrimal and conjunctival epithelial cells. ^9,12,19,25 We reported that in a canine model of spontaneous DE, apoptosis was increased in conjunctival epithelial cells, whereas it decreased in T cells infiltrating at the same site of inflammation. This was reversed after the dogs were treated topically with 0.2% CsA. ⁹ The integrated effect of CsA is attributed to complex mechanisms. First, it has been well documented that CsA selectively inhibits T-cell activation through blocking Ca²⁺-activated, calcineurin-dependent nuclear factor of activated T cells (NFAT), and subsequently blocks cytokine (such as IL-2) and chemokine production. ²⁶ CsA also attenuates responses of immune as well as nonimmune cells to inflammatory stimuli via suppression of NF-κB transcriptional signaling. ^27,28 Additionally, CsA binds to intracellular cyclophylins, including its mitochondrial receptor cyclophylin D (CypD). ²⁹ CypD was first identified by the demonstration of CsA effects on mitochondrial Ca²⁺ fluxes. ³⁰ CsA inhibited MPTP opening in the brain and cardiac cells as well as models of ischemia, that took a direct or indirect part in cytoprotection. ^31,32  

It has been reported that ΔΨm was altered in Wong-Kilbourne–derived human Chang conjunctival cells after incubation with preservative containing eye drops. ³³ However, there are few systematic studies to examine a direct role of MPTP, in death/survival of the ocular surface epithelial cells and invading lymphocytes under physiologic or inflammation conditions. The unique property of CsA interaction with mitochondria allows us to investigate the underlying mechanism of CsA on MPTP activity in regulating apoptosis. We hypothesize that inflammation triggers aberrant MPTP activation. CsA induces T-cell apoptosis and preserves ocular resident epithelial cells from inflammatory apoptosis through differential modulation of MPTP activity. Using IOBA-NHC and Jurkat T-cell models, we studied the role of MPTP and effect of CsA on inflammation and immune activation-induced apoptosis of human conjunctival epithelial cells and T cells in vitro. The correlation between mitochondrial functions (MPTP and ΔΨm) and cell apoptotic/survival state, as well as underlying molecular mechanisms involving intrinsic and extrinsic apoptosis signaling pathways were also investigated. 

Materials and Methods

Materials

Reagents were purchased from the following suppliers: anti-Fas (CH11) antibody (αFas) from Millipore (Lake Placid, NY); glyceraldehyde-3-phosphate dehydrogenase (GAPDH), monoclonal anti-human antibodies to Bax, Bcl-2, caspase-3, caspase-8, caspase-9, and COX IV antibodies for flow cytometry and Western blot from Cell Signaling (Danvers, MA); anti-human antibody cytochrome C for Western blot from BD Pharmingen (San Diego, CA); APC- or PE-coupled anti-mouse IgG and peroxidase-coupled anti-mouse IgG1 or IgG2a from Life Technologies (Carlsbad, CA); broad spectrum caspase inhibitor N-(2-quinolyl)valyl-aspartyl-(2,6-difluorophenoxy)-methyl ketone (Z-VAD), caspase-3 inhibitor (Z-DEVD), caspase-8 inhibitor (Z-IETD), caspase-9 inhibitor (Z-LEHD), human recombinant TNFα, and human recombinant IFNγ were from R&D Systems (Minneapolis, MN); CsA from Allergan (Irvine, CA); Fas receptor (DX2) antibody, and the isotype control (mouse IgG1, κ) were from BD Biosciences (San Jose, CA); Fas ligand (NOK-1) antibody and the isotype control (mouse IgG1, κ) were from eBiosciences (San Diego, CA). 

Cell Culture

IOBA-NHC are spontaneously immortalized cells derived from normal human conjunctival epithelium, ³⁴ and cultured in Dulbecco's modified Eagle's medium/Ham's F12 Nutrient Mixture (DMEM/F12; Life Technologies) with supplements described previously. ³⁵ The experimental conditions used for proinflammatory or proapoptotic stimulations are specified in the figure legends. 

Jurkat cells are an immortalized line of T lymphocytes (human lymphoblastic leukemia T-cell line; American Type Culture Collection, Manassas, VA). They were cultured in RPMI 1640 containing 10% fetal bovine serum, 0.1% fungizone, 50 U/mL penicillin, and 50 μg/mL streptomycin (Life Technologies). Cells were stimulated with a combination of phorbol 12-myristate 13-acetate (PMA, 10 ng/mL; Sigma-Aldrich, St. Louis, MO) and a plate-bound anti-CD3 antibody (αCD3: 5 μg/mL; BD Biosciences) in the presence or absence of CsA for designated time periods (specified in the figure legends). 

Primary human blood CD3⁺ T cells were purchased commercially (Zen-Bio, Inc., Research Triangle Park, NC). Cells were purified from peripheral blood samples of one male and one female donor, with PI viability greater than 92%. 

In Situ Detection of MPTP Opening

MPTP opening was determined using the calcein-cobalt (Co²⁺) technique that tracks the fluorescent dye calcein AM (CalAM) trapped in the mitochondria. ¹⁶ Co²⁺ is a metal channel selective fluorescence-quencher that does not cross mitochondrial membrane. MPTP opening was determined using a transition pore assay kit (MitoProbe; Life Technologies) based on the manufacturer's instructions with minor modifications. Hanks' balanced salt solution was replaced by phosphate-buffered salt solution (PBS with Ca²⁺, Mg²⁺) to provide sufficient phosphate (Pi) concentration essential for MPTP response. ³⁶ Calcein AM fluorescence was quantified using a flow cytometer (BD Biosciences). 

Measurement of Change in Mitochondrial Membrane Potential (ΔΨm)

ΔΨm was determined using a JC-1 assay reagent (MitoProbe; Life Technologies). JC-1 is a mitochondrial-sensitive dye exhibiting potential-dependent accumulation in the mitochondria. It is relatively independent of mitochondrial mass as compared with other mitochondrial dye such as rhodamine-123. ³⁷ Cells were cultured in the presence or absence of stimuli for a designated period of time (specified in the figure legends), and prepared at 1 × 10⁶ cells/mL in PBS. They were then incubated for 30 minutes with JC-1 (2 μM). JC-1 fluorescence was quantified by flow cytometry. Mitochondrial membrane depolarization (↓ΔΨm) is indicated by a switch to a decrease in JC-1 red fluorescence accompanied with an increase in JC-1 green fluorescence. 

Determination of Fas/FasL Expressions, Caspase-3, -8, and -9 Activities, and Apoptosis

Cells were cultured to 70% confluency and stimulated with TNFα (25 ng/mL) and IFNγ (500 U/mL) in the presence or absence of a general caspase inhibitor (1–100 μM, Z-VAD-FMK), caspase-3 inhibitor (100 μM, Z-DEVD), caspase-8 inhibitor (100 μM, Z-IETD), or caspase-9 inhibitor (100 μM, Z-LEHD), for 24, 48, and 72 hours with or without CsA (10 μM). Cells were harvested to determine the level of Fas and FasL expressions, the activities of caspase-3, -8, and -9, as well as apoptosis using flow cytometry. Early or late apoptotic cells were identified via Annexin V-FITC/PI staining (BD Biosciences). 

Western Blot Analysis on Apoptosis Mediators

Bcl-2, Bax, Cyt c, and the active forms of caspases were analyzed by Western blot. IOBA-NHC were cultured in the media without supplements and stimulated with TNFα (25 ng/mL) and IFNγ (500 U/mL), or etoposide (100 μg/mL), for designated time periods. Cells (2 × 10⁷) were lysed for 5 minutes in 50 μL cold buffer containing 250 mM sucrose, 1 mM EDTA, 0.05% digitonin, 25 mM Tris-HCl (pH 7.5), 1 mM DTT, and 0.1 mM PMSF. The cytosol supernatant was used for detection of caspases-3, -8, -9, and Cyt c. The mitochondrial fraction was used to evaluate Bcl-2 and Bax. Proteins (10 μg) were separated on NuPage 4–12% Bis-Tris gels (Invitrogen, Grand Island, NY), and blotted using primary antibodies for Cyt c, Bcl-2, Bax, or anti-human cleaved caspase-3, -8, or -9 (1 μg/mL). Secondary horseradish peroxidase–linked antibody (1:2000) was detected by chemiluminescence. GAPDH and COX IV (1:1000) were used as controls. 

IL-2 ELISA

Jurkat T cells were stimulated with PMA (10 ng/mL) and αCD3 (5 μg/mL) in the presence or absence of CsA (1 nM–10 μM) for 24 hours. CsA was prepared in molecular-grade ethanol (0.02% EtOH). Supernatants were collected from Jurkat culture and frozen at −80°C until the time of assay. IL-2 sandwich ELISA was performed according to the manufacturer's instruction (BD Biosciences). 

Statistics

Statistically significant differences between experimental groups were determined by one-way ANOVA or Student's t-test, where P ≤ 0.05 was considered significant. 

Results

Ca²⁺ Overload–Induced MPTP Opening and Effect of CsA

Intracellular calcium (iCa²⁺) is essential for cell signaling and amplification of apoptosis. Under physiologic conditions, iCa²⁺ elevation results in a transient increase in mitochondrial membrane permeability, allowing Ca²⁺ and small proteins to reach mitochondrial matrix to maintain ΔΨm. Excessive iCa²⁺ in response to endogenous or environmental insults, however, results in a persistent, sometimes irreversible MPTP opening, leading to cell death.  

MPTP is sensitive to Ca²⁺ change. To establish a positive control to study the role of inflammation and immune activation in MPTP opening state, we first examined the impact of Ca²⁺ overload on MPTP in IOBA-NHC and T cells. MPTP opening was induced by ionomycin in normal untreated IOBA-NHC (Fig. 1), as well as resting (unstimulated) Jurkat T cells and primary human blood CD3⁺ T cells (Fig. 2). The level of MPTP induction was ionomycin dose–dependent in both cell types. This was illustrated by a correspondent decrease in CalAM retained in the mitochondria upon cell exposure to CoCl₂ (Figs. 1A, 2A). This Ca²⁺-induced MPTP activation was inhibited by CsA (P < 0.01) in IOBA-NHC (Fig. 1B). At 10 μM, CsA completely prevented ionomycin-induced MPTP opening in IOBA-NHC (Figs. 1B, 1C). Mitochondrial CalAM in IOBA-NHC cells exposed to ionomycin (Fig. 1C, blue peak) was fully restored in the presence of CsA (green peak). In contrast, CsA had a minimal effect on ionomycin-induced MPTP opening in Jurkat T cells at all concentrations studied (1–500 nM) (Fig. 2A). At higher concentrations, CsA (1–10 μM) did not prevent CalAM loss in Jurkat T cells (Figs. 2B, 2C, P > 0.05). CsA (10 μM) slightly inhibited CalAM loss in primary human blood CD3⁺ T cells (Fig. 2D, P < 0.05). Cell viability was not significantly affected by transient treatment of ionomycin and/or CsA at all concentrations tested. Based on dose–response studies, 10 nM ionomycin was the optimal concentration determined for both cell types and, therefore, was used as a control in all subsequent experiments. Additionally, an increased Ca²⁺ concentration in IOBA-NHC cells with ionomycin was also confirmed by fluorometric imaging plate reader calcium assay (data not shown) (FLIPR Calcium Assay; Molecular Devices, Sunnyvale, CA). The findings demonstrated a direct impact of Ca²⁺ flux on the MPTP openings in IOBA-NHC and T cells under the normal condition. 

Figure 1

 

CsA inhibits ionomycin-induced MPTP opening in IOBA-NHC. (A) IOBA-NHC cells were resuspended to 1 × 10⁶ cells/mL in PBS (with Ca²⁺, Mg²⁺). Cells were pretreated with CsA (10, 100 nM) for 10 minutes at 37°C, followed by additions of CalAM (2 μM), CoCl₂ (400 μM), and ionomycin (1–500 nM). Cells were incubated for 15 minutes in the dark and resuspended in PBS. CalAM fluorescence was measured using flow cytometry. MPTP opening was induced in the normal IOBA-NHC cells by ionomycin in a dose-dependent manner. CsA at 100 nM concentration results in a trend of inhibition (P > 0.05) in MPTP opening in the cells treated with 1 to 25 nM ionomycin. (B) IOBA-NHC were stimulated with ionomycin at 10 nM (within the range that cellular response was detectable) in the presence or absence of CsA (1–10 μM). CsA at 5 or 10 μM dose significantly inhibited MPTP opening. *P < 0.01 (comparison: +CsA versus –CsA), **P < 0.001 (cells only versus cells + ionomycin). (C) A flow cytometry histogram illustrated the shift of CalAM peaks under various conditions representing the opening status of MPTP. The purple peak represents cells stained with CalAM alone. The red peak represents CalAM retained in the mitochondria of IOBA-NHC (cells treated with CalAM+ CoCl₂). The green peak demonstrates a complete CsA (10 μM) blockage of MPTP opening induced by ionomycin (10 nM, blue peak). Data are presented as mean fluorescence intensity (MFI) of three independent samples + SEM.

Figure 2

 

CsA exhibits a minimal effect on ionomycin-induced MPTP opening in T cells. (A) Jurkat cells were prepared for MPTP assay the same way as IOBA-NHC (details in Fig. 1 legend). MPTP opening was induced in resting Jurkats by ionomycin in a dose-dependent manner. CsA (10, 100 nM) had no effect on ionomycin-induced loss of mitochondrial CalAM. (B) Cells were stimulated with ionomycin (10 nM) in the presence or absence of CsA at higher concentrations (1, 5, 10 μM), which had no significant effect on MPTP opening (P > 0.05). **P < 0.005 (cells only versus cells + ionomycin). (C) A histogram shows an overlap between the blue (+iono, −CsA) and green (+iono, +CsA) peaks, indicating that ionomycin-induced MPTP activation in Jurkats was not affected by CsA. (D) Primary human blood CD3⁺ T cells were stimulated the same way as Jurkat, with ionomycin (10 nM) in the presence or absence of CsA, (1, 5, 10 μM). MPTP opening was not affected by CsA at 1 or 5 μM. A slight MPTP recovery was detected in the primary T cells treated with 10 μM CsA, *P < 0.05 (comparison: +CsA versus −CsA), **P < 0.005 (cells only versus cells + ionomycin). Data are presented as MFI of three independent samples + SEM.

Inflammation-Induced MPTP Opening, Mitochondrial Membrane Dissipation (↓ΔΨm), Apoptosis, and Effect of CsA

Proinflammatory cytokines TNFα and IFNγ were increased in tears as well as conjunctival epithelia of DE patients. Conjunctival epithelial cells including IOBA-NHC produce a plethora of cytokines and chemokines in response to TNFα and/or IFNγ stimulation. ³⁵ Desiccating stress-induced conjunctival epithelial cell apoptosis was intensified by exogenous IFNγ. ¹⁴ In the next set of experiments, we investigated if inflammation has a direct role on the opening state of MPTP and, thus, may impair mitochondrial functions leading to IOBA-NHC cell apoptosis. The stimulatory condition used was described previously. ³⁵ MPTP was evaluated in IOBA-NHC cells exposed to TNFα (25 ng/mL) or IFNγ (500 U/mL) alone, or their combination. Ionomycin-induced MPTP opening was used as a benchmark. TNFα or IFNγ alone at 24 hours did not elicit MPTP activation shown by the unchanged mitochondrial CalAM (Fig. 3A). At 48 hours, MPTP opening was instigated slightly by TNFα (CalAM: 13%↓), and significantly by IFNγ (41%↓) (Fig. 3B). Marked MPTP opening occurred after TNFα and IFNγ simultaneous treatment for 24 hours (Fig. 3A), as shown by the reduced mitochondrial CalAM intensity, which was further decreased (76%↓) after 48 hours (Fig. 3B). Treatment with CsA (10 μM) significantly blocked MPTP opening by TNFα, IFNγ, or their synergistic stimulation (Fig. 3C). Interestingly, CsA treatment led to a higher accumulation of mitochondrial CalAM than the cells-only control (see the Discussion section for potential explanation). The TNFα + IFNγ–induced MPTP activation was found concomitant with mitochondrial membrane depolarization (Fig. 3D). The ΔΨm decrease in IOBA-NHC treated with TNFα + IFNγ (48 hours) was demonstrated by a marked increase in JC-1 green fluorescence, which corresponded with a decrease in the JC-1 red fluorescence (Fig. 3D). Pre- or cotreatment with CsA completely prevented the TNFα and IFNγ–instigated ΔΨm loss in IOBA-NHC (Fig. 3E). We then further examined if the same inflammatory stimulation eliciting MPTP activation and ΔΨm loss will lead to IOBA-NHC apoptosis. As we reported previously, TNFα or IFNγ alone (for up to 72 hours) did not cause significant IOBA-NHC cell death. ³⁵ In the current study, apoptosis was induced slightly by TNFα + IFNγ at 24 hours, and significantly after 48 hours (Fig. 4A). Over 30% of IOBA-NHC cells became Annexin V⁺PI⁺ by 48 hours, the same time when significant MPTP activation and ΔΨm loss took place. Pre- or co-CsA treatment completely prevented or blocked IOBA-NHC cell death induced by TNFα + IFNγ (Fig. 4B). However, CsA protection was lacking when given 24 hours post TNFα + IFNγ treatment. IOBA-NHC cells remained viable in 41.5 μM CsA (1:10 dilution of Restasis; Allergan, Inc.) for at least 72 hours, in the absence of TNFα and IFNγ (Fig. 4C). 

Figure 3

 

CsA inhibits TNFα and/or IFNγ-induced MPTP opening, and loss of mitochondrial membrane potential (ΔΨm) in IOBA-NHC. Cells were stimulated with TNFα (25 ng/mL) and/or IFNγ (500 U/mL) for 24 and 48 hours in the presence or absence of CsA (10 μM). Cells were resuspended to 1 × 10⁶ cells/mL PBS, and subjected to MPTP assay or JC-1 staining. Ionomycin-induced MPTP opening was used as a reference control. (A) MPTP opening was not elicited by TNFα or IFNγ alone at 24 hours but the combined treatment of TNFα + IFNγ (*P < 0.01). (B) MPTP opening was slightly increased by TNFα (P > 0.05), but significantly by IFNγ alone by 48 hours. MPTP activation was further enhanced by TNFα + IFNγ at 48 hours (*P < 0.005, **P < 0.0001). (C) Treatment with CsA extensively blocked MPTP activation as a result of TNFα, IFNγ, or TNFα+IFNγ stimulation. *P < 0.001 (comparison: +CsA versus −CsA). CsA treatment led to restoration of, and even higher accumulation of, Calcein AM dye than the cell only control. (D) ΔΨm was decreased in IOBA-NHC exposed to TNFα + IFNγ for 24 hours, and continued to drop by 48 hours. An increase in the green fluorescence accompanied by a decrease in red fluorescence represents mitochondrial depolarization or a loss in ΔΨm. (E) CsA pre- (for 24 hours) or cotreatment (in the presence of TNFα + IFNγ for 48 hours) inhibited ΔΨm loss. Data are representative of three independent experiments.

Figure 4

 

CsA protects IOBA-NHC from TNFα and IFNγ-induced apoptosis. IOBA-NHC were grown in six-well plates and cultured in the presence or absence of TNFα (25 ng/mL) and IFNγ (500 U/mL) for 24 or 48 hours. Cells were subjected to Annexin V/PI apoptosis assay by flow cytometry. (A) Percentage (%) apoptotic cells (Annexin V⁺PI⁺ in upper right quadrant: 5.5%) was greater than the untreated control (2.5%) at 24 hours, and continued to increase by 48 hours of exposure to TNFα + IFNγ (30%). (B) CsA pre- (for 24 hours) or cotreatment (in the presence of TNFα + IFNγ for 48 hours) significantly inhibited TNFα and IFNγ-induced IOBA-NHC apoptosis. *P < 0.001 (comparison: +CsA versus −CsA). (C) IOBA-NHC cells remained viable (Annexin V⁺PI⁺ cells: <6%) following incubation with CsA (1–10 μM, or 41.5 μM 1/10 dilution from Restasis [0.05% cyclosporine ophthalmic emulsion]) for 72 hours. Data are representative of three independent experiments.

T cells escape from programmed cell death in human and animal models of DE. Several T-cell subtypes were found increased in the cell number and activities in DE patients. ²³ To determine the role of MPTP activation in T-cell fate and the effect of CsA, Jurkats were stimulated with PMA and αCD3. The level of mitochondrial CalAM was increased at 24 hours, and subsided with time in both unstimulated (Fig. 5A) and activated (Fig. 5B) Jurkat cells. The enhanced mitochondrial CalAM at 24 hours indicates an initial increase in T-cell proliferative activities upon culturing/stimulation. Ionomycin elicited MPTP opening in both unstimulated and PMA/αCD3-activated Jurkat cells. Addition of CsA to unstimulated or activated T cells (without ionomycin) led to an accumulation of CalAM, more than cells alone (see the Discussion section for potential explanation). However, in the presence of ionomycin, CsA was not able to block MPTP opening in either unstimulated (Fig. 5A) or activated (Fig. 5B) Jurkat cells. In unstimulated T cells, ΔΨm remained constant through 72 hours (Fig. 5C, left panel). However, ΔΨm was continuously decreased in the PMA + αCD3–activated T cells over a period of 72 hours (Fig. 5C, middle panel). This activation-induced ΔΨm loss in Jurkat was exacerbated by CsA (10 μM) in a time-dependent fashion (Fig. 5C, right panel). 

Figure 5

 

Effect of CsA on MPTP, ΔΨm, and apoptosis in resting or activated T cells. Jurkat T cells were stimulated with PMA (10 ng/mL) and αCD3 (5 μg/mL) in the presence or absence of CsA (10 μM) for 24, 48, or 72 hours. Cells were prepared for MPTP, JC-1, and Annexin V/PI assays the same way as described. (A) In the unstimulated T cells, mitochondrial CalAM was increased initially (24 hours), but returned to the baseline by 72 hours. In the absence of ionomycin, CsA addition caused a further increase of mitochondrial CalAM. MPTP opening was induced by ionomycin, which was not affected by CsA. (B) In PMA + αCD3 stimulated T cells, mitochondrial CalAM was greater than that of unstimulated cells, but also subsided with time. Ionomycin-induced MPTP opening in activated cells was not affected by CsA. (C) PMA and αCD3 induced a time-dependent ΔΨm loss in activated Jurkat T cells, which was further enhanced by CsA. (D) CsA dose-dependently induced apoptotic cell death in the unstimulated T cells. *P < 0.001(−/+CsA). (E) CsA dose-dependently inhibited IL-2 production by activated Jurkats, P < 0.005 (−/+ CsA). (F, G) PMA + αCD3 induced a time-dependent induction of apoptosis in activated Jurkat T cells [F] and primary human blood CD3⁺ T cells [G]. This was exacerbated in the presence of CsA [F, G]. Data are representative of two or three independent experiments.

CsA dose-dependently promoted apoptosis of resting/unstimulated Jurkat cells (Fig. 5D), contrary to CsA protection of IOBA-NHC cells. Activation of Jurkat cells by PMA + αCD3 elicited a time-dependent cell death (Fig. 5F, middle panel) compared with the unstimulated cells (Fig. 5F, left panel). The activation-induced Jurkat cell death was further enhanced by CsA (10 μM) (Fig. 5F, right panel). This finding from Jurkat cells (Fig. 5F) was confirmed in primary human blood CD3⁺ T cells (Fig. 5G). The primary human blood T cells appeared to be more sensitive than Jurkat T cells to PMA + αCD3 stimulation and CsA treatment. 

Additionally, CsA completely abolished PMA + αCD3–induced IL-2 production at subnanomolar concentrations (1 nM) (Fig. 5E). 

Role of Fas/FasL in MPTP Activation and Effect of CsA

Fas expression was increased in the conjunctival and lacrimal biopsies from dry eye dogs. ⁹ Fas was upregulated in the conjunctival epithelium by exogenous IFNγ in mice under desiccating stress. ¹⁴ To determine if Fas and/or FasL are expressed by IOBA-NHC, and if MPTP is a direct target of Fas/FasL-induced apoptosis in the presence or absence of inflammation, IOBA-NHC cells were exposed to an agonistic αFas antibody with or without IFNγ and/or TNFα stimulation. Fas (Fig. 6A), and to a lesser extent FasL (Fig. 6B), were found constitutively expressed by IOBA-NHC. Fas was increased at 24 and 48 hours by IFNγ or TNFα + IFNγ, but not TNFα alone. FasL was induced by IFNγ (48 hours) or TNFα + IFNγ (24 and 48 hours), but again, not significantly by TNFα alone. Such Fas or FasL overexpression was attenuated by CsA (Figs. 6A, 6B), though posttreatment of CsA did not significantly inhibit FasL induction by TNFα + IFNγ (Fig. 6B). The FACS plots for FasL were illustrated in Figure 6C with the isotype control to confirm the low level, but significant expressions of FasL by IOBA-NHC cells, as well as the inhibitory effect of CsA upon inflammatory stimulation. αFas alone (15 μg/mL for 72 hours) failed to induce substantial MPTP opening in IOBA-NHC (data not shown). Addition of IFNγ markedly increased MPTP opening by 24 hours. The extent of MPTP opening by αFas + IFNγ was comparable to that by ionomycin. CsA abrogated αFas + IFNγ–induced MPTP activation (Fig. 6D). 

Figure 6

 

IFNγ upregulates Fas/FasL expression and enhances Fas-induced MPTP activation in IOBA-NHC, and the inhibitory effect of CsA. IOBA-NHC cells were stimulated with TNFα (25 ng/mL), IFNγ (500 U/mL) or TNFα + IFNγ for 24 or 48 hours in the presence or absence of CsA (10 μM). (A) Fas was increased by IFNγ or TNFα + IFNγ stimulation at 24 and 48 hours, which was suppressed by CsA pretreatment (24 hours before TNFα + IFNγ), or cotreatment (24 or 48 hours with TNFα + IFNγ). Fas upregulation at 48 hours was also inhibited by CsA posttreatment (24 hours after TNFα + IFNγ). **P < 0.001 (cells only versus IFNγ or +TNFα + IFNγ), *P < 0.005 (−/+ CsA). (B) FasL expression was increased by IFNγ or TNFα + IFNγ (**P < 0.001: Cells only versus IFNγ or +TNFα + IFNγ), which was inhibited by CsA pre- or co-, but not posttreatment. **P < 0.001(−/+CsA) (C) TNFα and IFNγ-induced FasL upregulation and the effect of CsA (at 24 hours' time point) are illustrated by the FasL signal peaks in the FACS plots (blue = unstimulated cells; red = cells activated with TNFα+IFNγ; green = activated cells treated with CsA; filled gray = isotype control). (D) IOBA-NHC were stimulated with agonistic αFas antibody (15 μg/mL), IFNγ (500 U/mL) or their combination for 24 hours in the presence or absence of CsA (10 μM). MPTP opening was induced slightly by αFas or IFNγ alone, and markedly by αFas and IFNγ joined treatment. CsA significantly protected IOBA-NHC from αFas and IFNγ-induced MPTP opening. *P < 0.005 (Cells only versus +αFasR + IFNγ); **P < 0.001 (−CsA versus +CsA). Data are representative of three independent experiments.

For T cells, activation-induced FasL upregulation and FasL–Fas interaction controls immune responses by eliminating T cells expressing self-reactive T-cell receptors (TCRs). ³⁸ Fas is abundantly expressed while FasL is present at a low level in resting T cells, but increased upon stimulation. ^39,40 This was confirmed in our study. Fas (Fig. 7A) and FasL (Fig. 7B) were both increased in activated Jurkat cells. CsA slightly enhanced Fas upregulation (72 hours; Figs. 7B, 7C, left), and significantly augmented FasL increase in activated Jurkat cells (Figs. 7B, 7C, right), supporting that NFAT plays a functional role in FasL expression of T cells. 

Figure 7

 

Effect of CsA on Fas and FasL expression in the resting and activated Jurkat T cells. Jurkats were stimulated with PMA (10 ng/mL) and αCD3 (5 μg/mL) in the presence or absence of CsA (10 μM) for 24, 48, or 72 hours. Surface expressions of Fas (A, C, left) and FasL (B, C, right) were determined using flow cytometry. Fas and FasL were both upregulated (48 and 72 hours) in the activated T cells compared with the resting cells (*P < 0.00001: Cells only versus activated cells). CsA treatment further enhanced FasL upregulation (**P < 0.0003: −CsA versus +CsA), but had a minimal effect on Fas expression. Data are representative of two independent experiments.

Inflammation-Induced Mitochondrial Translocation of Bax, Cytochrome c Release, Caspase Activation, and Effect of CsA

Caspase activation is a central event in apoptosis signaling. Since the majority of the dying cells were positive for AnnexinV and PI, representing cells at a late stage apoptosis or necrosis, we examined if inflammation-induced IOBA-NHC death was via caspase-dependent apoptosis. We first evaluated if inflammation-induced IOBA-NHC death can be blocked by caspase inhibitors. AnnexinV⁺PI⁺ cells (%) were decreased by a pan caspase inhibitor Z-VAD in a dose-dependent manner (Fig. 8A). Z-VAD at 100 μM completely blocked TNFα and IFNγ-triggered IOBA-NHC death at all time points, suggesting that such inflammatory cell death is primarily via apoptosis. Similar studies were conducted using Z-DEVD, Z-IETD, or Z-LEHD, the inhibitors specific for caspase-3 (involved in both intrinsic and extrinsic apoptosis), caspase-8 (extrinsic), or caspase-9 (intrinsic), respectively. IOBA-NHC cell death was partially attenuated by all three inhibitors. Inhibition by a single caspase inhibitor was less than that of the pan inhibitor Z-VAD (Fig. 8B). Apoptosis was decreased by 24%, 63%, or 35% in the presence of Z-DEVD, Z-IETD, or Z-LEHD at 100 μM at 72 hours, respectively. The results suggest that IOBA-NHC cells proceeded through both intrinsic and extrinsic apoptotic pathways upon TNFα + IFNγ stimulation. 

Figure 8

 

CsA protects IOBA-NHC cells from inflammation-induced apoptosis via inhibition of caspase activities. IOBA-NHC were stimulated with TNFα (25 ng/mL) and IFNγ (500 U/mL) for 24, 48, or 72 hours in the presence or absence of various caspase inhibitors and CsA (10 μM). (A) TNFα and IFNγ-induced cell apoptosis was inhibited by the pan caspase inhibitor Z-VAD in a dose-dependent manner (1, 10, and 100 μM). This inflammation-induced IOBA-NHC apoptosis was completely blocked by Z-VAD at 100 μM (*P < 0.00001), and (B) partially inhibited by the specific caspase-3, -8, or -9 inhibitor (Z-DEVD, Z-IETD, or Z-LEHD, respectively; **P < 0.001) at the same concentration (100 μM). (C) Western blot revealed a marked increase of the cleaved (activated form) caspase-3, -8, and -9 in the cytosol protein fraction of IOBA-NHC cells stimulated with TNFα and IFNγ. (D–F) Pre- or cotreatment with CsA attenuated TNFα and IFNγ-induced activation of caspase-3, -8, and -9 in IOBA-NHC cells (*P < 0.05, **P < 0.005). Data are representative of three independent experiments.

To elucidate the mechanism of CsA protection of IOBA-NHC from inflammation-induced apoptosis in addition to blocking MPTP, we further investigated CsA effect on caspase activation. TNFα + IFNγ stimulation (48 hours) resulted in a marked increase in the active (cleaved form) caspase-3, -8, and -9 (Fig. 8C). CsA significantly attenuated inflammation-induced activation of all three caspases in pre- or cotreated IOBA-NHC cells (Figs. 8D–F). 

It was reported that IL-2 release by activated Jurkat cells requires caspase activities. ⁴¹ We have shown in the current report that CsA inhibited IL-2 release by activated Jurkats (Fig. 5F). To determine the role of caspases in activation-induced Jurkat apoptosis and CsA exacerbation of T-cell death, we studied the ability of caspase inhibitors in mitigating apoptosis in activated Jurkat cells in the presence of CsA. PMA and αCD3-induced Jurkat cell death was completely abolished by Z-VAD (pan), and to a lesser extent by the individual caspase inhibitors (100 μM) (Figs. 9A–C). The inhibitory capacity of caspase inhibitors on activated Jurkat apoptosis was largely hindered by CsA. A time-dependent further induction of T-cell apoptosis was induced by CsA in the presence of caspase-8 or -9 inhibitors. In the presence of CsA, caspase-8 or -9 inhibitor induced more apoptosis at 48 and 72 hours as compared with that of caspase-3 inhibitor. This is likely a result of a time-dependent upregulation of caspase-8 and -9 activities by CsA. Indeed, the intracellular activities of caspases-8 and -9 were further increased in the activated Jurkats in the presence of CsA over 24, 48, and 72 hours (Figs. 9E, 9F). The intracellular activity of caspase-3 was also enhanced by CsA (Fig. 9D), but the level of induction subsided over time. This may explain the relatively less induced apoptosis level in Jurkat cells treated with caspase-3 inhibitor in the presence of CsA. Nevertheless, our finding suggests that the manifestation of T-cell death may be attributed to caspase activation directly induced by CsA. 

Figure 9

 

CsA exacerbates activation-induced Jurkat T cell death via enhancing caspase activities. Jurkats were stimulated with PMA (10 ng/mL) and αCD3 (5 μg/mL) for 24, 48, or 72 hours in the presence or absence of caspase inhibitors (100 μM) and CsA (10 μM). (A–C) PMA and αCD3 induced a time-dependent Jurkat apoptosis, which was mitigated by pan or specific caspase-3, -8, or -9 inhibitors to various extents. Apoptotic T cells (%) was significantly increased in the presence of CsA. (D, E) The activities of caspase-3, -8, and -9 were increased in activated Jurkats, and further enhanced in the presence of CsA. *P < 0.005: activated cells versus activated cells + caspase inhibitor; **P < 0.0005: caspase Inhibitor versus caspase inhibitor + CsA. Data are representative of three independent experiments.

The following experiments were carried out to determine if caspase activation was a result of imbalance between pro- and antiapoptotic protein expressions (Bax versus Bcl-2), and Cyt c release from compromised mitochondria due to inflammation. Cytosol or mitochondrial proteins were analyzed for Cyt c, Bax, and Bcl-2. IOBA-NHC were stimulated with TNFα and IFNγ, or etoposide (a potent apoptosis inducer). A small amount of cytosol Cyt c was observed in the untreated IOBA-NHC. A marked increase in cytosol Cyt c was detected at 48 hours (Fig. 10, top panel). The amount of inflammation-induced Cyt c release was comparable to that by etoposide. GAPDH levels stayed constant regardless of treatment conditions. In the mitochondrial fraction, Bax was significantly increased (approximately 2-fold), whereas Bcl-2 decreased (approximately 1.5-fold) in TNFα + IFNγ stimulated cells based on the normalized protein contents using a mitochondrial protein Cox IV (Fig. 10, bottom panel). 

Figure 10

 

Cytochrome c release and mitochondrial translocation of Bax in inflammation-stimulated IOBA-NHC. Western blot was performed on cytosol or mitochondrial proteins from IOBA-NHC treated with TNFα (25 ng/mL) and IFNγ (500 U/mL), or etoposide (100 μg/mL ) for 4 or 48 hours. GAPDH and COX IV were used as housekeeping controls for cytosolic and mitochondrial proteins, respectively. Data are representative of two independent experiments.

Discussion

We have examined the role of MPTP and relevant apoptosis pathways in inflammation-induced apoptosis, and mechanisms underlying regulatory effects of CsA on apoptosis in human conjunctival epithelial cells versus T cells. Using in vitro models, we demonstrated for the first time that TNFα and IFNγ induced MPTP opening and ΔΨm loss, along with increased apoptosis in IOBA-NHC cells. This was accompanied by upregulation of Fas/FasL, Bax, and caspase activation. CsA protected IOBA-NHC cells from inflammation-induced cell death by blocking intrinsic (mitochondria-mediated) and extrinsic (death-receptor–initiated) apoptosis pathways. CsA preserved IOBA-NHC through stabilization of MPTP and maintaining mitochondrial membrane potential, as well as inhibition of inflammation-induced Fas/FasL expression, and caspase activation. In contrast, in activated Jurkat and/or primary human blood T cells, MPTP opening and cell apoptosis were increased, along with upregulation of Fas/FasL and caspase activities. CsA had a minimal effect on MPTP state in T cells, but further promoted membrane potential loss, exacerbated T-cell apoptosis, and further increased Fas/FasL expression and caspase activation. 

Apoptosis can be initiated at the entry of plasma membrane upon death-receptor ligation (extrinsic) or at mitochondria (intrinsic), depending on types of stimuli, sensitivity of targeted cells, propensity of local players to receptor activation, MPTP induction, as well as involvement of pro/antiapoptotic proteins and caspase activities. IFNγ is a T_H1 cytokine produced mainly by activated T cells and natural killer (NK) cells, and plays a pivotal role in immune regulation and inflammation, usually in synergy with other cytokines, such as IFNγ and TNFα. IFNγ was indicated in corneal allograft rejection, Sjögren's syndrome, and dry eye. ⁴² IFNγ-induced apoptosis was demonstrated in vitro in the Wong-Kilbourne derivative of Chang conjunctival cells via preferential STAT1 activation, antagonizing the prosurvival effect of NF-κB. ¹³ Mice lacking IFNγ (B6γKO) showed resistance to desiccating stress-induced conjunctival cell apoptosis, which can be elicited by exogenous IFNγ. ¹⁴ We reported that IOBA-NHC produced a plethora of cytokines and chemokines in response to IFNγ and/or TNFα. Exposure of IOBA-NHC to TNFα or IFNγ alone for 24, 48, and 72 hours did not induce early cell apoptosis, but did result in <5% PI⁺ population at 48–72 hours. ³⁵ In the current study, IFNγ or TNFα alone had minimal effects on MPTP until 48 hours. A synergistic treatment of IFNγ and TNFα resulted in an earlier (at 24 hours) and greater onset of MPTP activation. The results indicate that, first, IFNγ, not TNFα, is the primary MPTP inducer in IOBA-NHC, suggesting that MPTP alterations occur very early during apoptosis, even before phosphatidylserine (PS) redistribution to outer plasma membranes detectable by annexin V staining. Second, the pro/antiapoptosis balance of TNFα signaling was potentiated by IFNγ and tipped off to trigger TNFα death-receptor–mediated pathway. Therefore TNFα-induced proapoptotic activity has counterbalanced the antiapoptotic activity rendered via NF-κB activation (which is the dominant pathway when TNFα is present alone). We found NF-κB nuclear translocation as early as 30 minutes after IOBA-NHC cells exposed to TNFα (unpublished observations). Moreover, IFNγ-elicited MPTP activation and subsequent IOBA-NHC apoptosis were accelerated and amplified by TNFα addition. This may be attributed to an activation of both extrinsic and intrinsic pathways via proapoptotic Bcl-2 family members that link extrinsic and intrinsic pathways by inducing mitochondrial translocation of Bax, a major proapoptotic protein. It was reported that death-receptor–induced caspase-8 activation cleaved Bid and, subsequently, attacked mitochondria and liberated cytochrome c, prompting mitochondria-mediated apoptosis. ^43,44 In IOBA-NHC, mitochondrial Bax was increased by 2-fold after cells were exposed to TNFα + IFNγ. This was correlated with Cyt c release and activation of both caspase-8 (extrinsic) and caspase-9 (intrinsic). Pre- and cotreatment with CsA significantly inhibited caspase activation. Bax and Bcl-2 levels did not appear to be directly affected by CsA. It is also interesting to note that, in TNFα and IFNγ-stimulated IOBA-NHC cells, CsA treatment led not only restoration to, but also to more accumulation of CalAM dye than the control (cells-only). This may be attributed to CsA binding to its mitochondrial receptor CypD on MPTP, so that the open/shut flickering of mitochondrial pores are blocked by CsA binding, resulting in a retention of CalAM. This was also observed in Jurkat cells in the absence of ionomycin. However, this accumulation appears to be a transient event in that CalAM level subsided over time. Overall, our findings suggest that CsA attenuation of inflammatory apoptosis of IOBA-NHC is likely a result of its blockade of MPTP opening and inhibition of caspase activation than its direct inhibition of specific proapoptotic proteins. 

Fas/FasL-regulated apoptosis is essential for maintaining immune privilege of the eye. In humans and mice, FasL is not only upregulated in activated T cells to induce self-apoptosis, it is constitutively expressed in ocular tissues to mitigate damage from aberrant immune response. Fas/CD95 was also detected in the eye, and upregulated in the conjunctiva and retina in DE and uveitis. ^8,14,45 In IOBA-NHC, Fas was constitutively expressed, and enhanced by IFNγ. MPTP was not affected by Fas receptor ligation via soluble FasL or by an agonist αFas antibody alone for 72 hours. Addition of IFNγ markedly induced MPTP opening. IFNγ-enhanced Fas and FasL upregulation and MPTP activation in IOBA-NHC were significantly abrogated by CsA treatment. This result suggests that in IOBA-NHC—one of the main resident epithelial cell types on the ocular surface (OS)—IFNγ sensitizes Fas-mediated death-receptor signaling via overexpression of Fas, whose normal expression is required to maintain OS tissue homeostasis. Whether IFNγ and αFas instigated IOBA-NHC cell death is regulated only via extrinsic pathway, or involves cross-talk with the mitochondrial pathway, is currently under investigation. 

The main mechanism of CsA inhibition of T-cell activation is by blocking calcineurin-dependent NFAT transcriptional activation. Activation-induced cell death is an essential process for T-cell development and function, which can be affected by the nature of the initial stimulating events. ⁴⁶ The concentration-dependent CsA modulation of T-cell fate has been demonstrated in several studies. It was shown that Jurkat or PBMC activated with OKT3 (anti-CD3) and restimulated after 6 days with PMA and ionomycin underwent apoptosis, which was inhibited by CsA (approximately 83 nM) after 72 hours. ⁴⁷ CsA at the concentration attained in rodents (10 μM) elicited T-cell apoptosis by 24 hours via release of preformed TGFβ1, independent of calcineurin inhibition. ⁴⁸ Additionally, the sensitivity of T cells to CsA depends on activation of other transcriptional pathways such as MAPK and NF-κB. Jurkat cells expressing active NFAT alone are sensitive to CsA, whereas coexpression of a constitutively active NFAT and MEKK1 renders T cells' resistance to CsA (calcineurin-independent). ⁴⁹ Activated T-cell death appeared to be prevented by a survival signal afforded by NF-κB. ⁵⁰ In the current study, Jurkat activation and IL-2 production induced by PMA/αCD3 was inhibited by subnanomolar CsA (>1 nM). At 5 to 10 μM, CsA promoted resting T-cell death, and exacerbated activated T-cell death. At all concentrations tested (10 nM–10 μM), CsA was not able to prevent MPTP opening. Additionally, it was reported that IL-2 production in Jurkat was caspase-mediated. ⁴¹ Our study has revealed that CsA hindered the effect of caspase inhibitors, thus facilitating activation-induced T-cell death, suggesting a calcineurin-dependent mechanism of CsA action, rather than cytotoxicity. Nevertheless, CsA at a broad concentration range used in the current studies exerted no protection for MPTP opening in Jurkat, and had a minimal effect on primary human blood CD3⁺ T cells. This may be attributed to a lower mitochondrial content in T lymphocytes compared with epithelial cells. Cells of the immune system have few mitochondria in general. ⁵¹ Additionally, different OS inflammatory disorders are characterized by different types of activated T cells. The outcome of the immune activation state in the ocular tissue microenvironment is likely a result of the accumulative response from various T-cell subtypes, as well as their communications with ocular resident cells. CsA has been used for treating dry eye, allergic conjunctivitis, and rheumatoid arthritis (RA), where T_H1, T_H2, or T_H17 exhibits a prominent role in the disease pathogenesis, respectively. However, CsA action on various T-cell subsets including effector as well as suppressive T cells has not been fully studied. It is to be determined if CsA differentially regulates MPTP response under different inflammatory conditions for T_H1, T_H2, and T_H17, as well as for suppressive T regulatory cells. It is also imperative to note that most of the experiments in the current report were conducted using the immortalized cell lines. Although IOBA-NHC cells resemble some crucial characteristics of human conjunctival epithelium, ³⁴ caution should be taken when extrapolating in vitro findings to presume the MPTP status in OS pathology in humans. Therefore, our models remain an experimental tool that helps to serve as a basis for studying mechanism pathways and conducting in vivo studies. Experiments are being performed to establish an epithelial and T-cell coculture to determine MPTP response during their interactions under normal and inflammatory conditions. Mitochondrial involvement in animal models of ocular inflammation such as dry eye will also be investigated. 

Taken together, our findings suggest that the differential effect of CsA on conjunctival epithelial cells versus T cells in MPTP modulation may contribute to its therapeutic efficacy in treating human and animal models of ocular inflammation such as dry eye disease. 

Acknowledgments

Disclosure: J. Gao, Allergan, Inc. (F, E, R); R. Sana, Allergan, Inc. (F, E); V. Calder, Allergan, Inc. (F, C); M. Calonge, Allergan, Inc. (F, C); W. Lee, Allergan, Inc. (F, E); L.A. Wheeler, Allergan, Inc. (F, E, R, S); M.E. Stern, Allergan, Inc. (F, E, R, S)