IFN-γ Facilitates Corneal Epithelial Cell Pyroptosis Through the JAK2/STAT1 Pathway in Dry Eye

Abstract

Purpose: To investigate the effect of gamma interferon (IFN-γ) on corneal epithelial pyroptosis in an experimental dry eye (DE) model and explore the underlying molecular mechanisms.

Methods: Experimental DE was established in adult wild-type (WT) C57BL/6 mice and Ifng-knockout mice on a C57BL/6 background by subcutaneous injection of scopolamine (1.5 mg/0.3 mL, three times per day) and exposure to desiccating stress. An immortalized human corneal epithelial cell line (HCE-T) was treated with IFN-γ under hyperosmolar conditions. Corneal epithelial defects, tear production, and conjunctival goblet cells were detected by fluorescein sodium staining, the phenol red cotton test, and periodic acid-Schiff staining. The mRNA expression was measured by quantitative real-time PCR. Changes in protein expression were analyzed by Western blotting and immunofluorescence staining. Cell Counting Kit-8 and lactate dehydrogenase assays and in situ TUNEL staining were used to assess cell death.

Results: The expression of IFNG and its related genes was increased in the corneas of DE mice, whereas genetic deletion of Ifng ameliorated desiccating stress-induced dry eye symptoms. We further found that IFN-γ activated the JAK2/STAT1 signaling pathway inducing corneal epithelial pyroptosis. Topical application of a STAT1 inhibitor in vivo or siRNA targeting STAT1 in vitro suppressed pyroptosis of corneal epithelial cells. In addition, the production of reactive oxygen species (ROS) was elevated in DE, and a reduction in excessive ROS release prevented pyroptosis.

Conclusions: The increase in IFN-γ participates in the pathogenesis of dry eye and promotes corneal epithelial pyroptosis by activating the JAK2/STAT1 signaling pathway. Oxidative stress might be in downstream of JAK2/STAT1, thereby contributing to pyroptosis.

Dry eye disease (DED) is a common multifactorial disease of the ocular surface that is accompanied by a series of clinical symptoms including ocular pain, light sensitivity, dryness, irritation, and fluctuating or blurry vision.^1,2 The destruction of tear film homeostasis caused by many factors can result in tear hyperosmolarity and subsequent corneal and conjunctival damage, which are critical in the initiation and development of DED.³ Chronic inflammation plays an important role in this process. Hyperosmolar stress induces inflammatory injury to the ocular surface, which in turn exacerbates the inflammatory response.^3,4 Recently, pyroptosis has attracted increasing attention in ocular surface damage because of its powerful inflammatory effect.⁵ 

Pyroptosis is a form of lytic programmed cell death characterized by cell swelling, rapid disruption of the plasma membrane and subsequent release of cytoplasmic contents and proinflammatory cytokines such as interleukin-1β (IL-1β) and IL-18.⁶ The gasdermin family bearing a membrane pore-forming activity is regarded as the pyroptosis executioner, and GSDMD, a substrate of caspase 1 (CASP1), has been thoroughly studied. First, caspase 1 undergoes autocatalytic activation within inflammasomes, which are formed by triggering of members of the nucleotide-binding domain and leucine-rich repeat containing (NLR) family, such as NLRP3. This process is often promoted by activating adaptors, such as apoptosis-associated speck-like protein containing (ASC).⁷ After being activated, caspases cleave GSDMD at the interdomain loop and release the N-terminal domain (N-GSDMD), which is then translocated to the plasma membrane to form pores and induce pyroptosis.⁸ This process is termed the inflammasome-dependent canonical pathway. In the noncanonical pathway, LPS is recognized by CASP4/5/11 rather than CASP1 in the host. Subsequently, activated CASP4/5/11 induces GSDMD cleavage.^9,10 Although pyroptosis mediated by GSDMD has been reported in DED, its upstream regulatory mechanism remains to be elucidated. 

IFN-γ, which is exclusively produced by natural killer (NK) and T helper (Th) -1 cells, is a pleiotropic cytokine that has immunomodulatory effects on various immune cells, as well as antiproliferative effects. In addition, IFN-γ has a strong association with cell death. Shao et al.¹¹ reported that IFN-γ enhanced CD8+ T-cell-mediated cytotoxic responses against keratinocytes, which is an important pathogenesis of lichen planus. In allogeneic bone marrow transplantation, a study revealed that IFN-γ could initiate a proapoptotic gene expression program and promote intestinal stem cell death.¹² Numerous studies have demonstrated the detrimental role of IFN-γ in the development of DED. A clinical trial revealed that the expression of IFN-γ was positively correlated with the severity of DED.¹³ Basic research has shown that IFN-γ can promote pathological apoptosis in the corneal and conjunctival epithelium.^14,15 Although the crosstalk between IFN-γ and apoptosis is well established, the association between IFN-γ and pyroptosis in dry eye and mechanistic details remain unclear. 

The Janus kinase (JAK)/signal transducer and activator of transcription (STAT) is a common downstream signaling pathway of IFN-γ.¹⁶ This pathway is associated with many important biological processes, such as cell proliferation, differentiation and death, and immune regulation.¹⁷ The development of some systemic autoimmune diseases, including rheumatoid arthritis and myeloproliferative neoplasms, has been attributed to an imbalance in the activation of the JAK/STAT pathway, which plays a crucial role in inflammation exacerbation.¹⁸ More recently, Karki et al.¹⁹ reported that the combination of TNF-α and IFN-γ activated the JAK/STAT1/IRF1 axis, promoting nitric oxide production and driving caspase-8/FADD-mediated necroptosis and pyroptosis, which is termed PANoptosis, in SARS-CoV-2 infection. Given the crosstalk between IFN-γ and cell death, the regulation of inflammation by the JAK/STAT pathway, and the proinflammatory property of pyroptosis, IFN-γ may be involved in corneal epithelial pyroptosis through the JAK/STAT pathway in DED. 

In the present study, we examined the expression of IFN-γ and its related genes in the cornea in a desiccating stress-induced DE mouse model and evaluated clinical parameters of Ifng-knockout (Ifng^−/−) and wild-type (WT) mice after DE induction. Furthermore, we analyzed the changes in JAK and STAT family genes to identify the key pathway regulated by IFN-γ. Notably, we showed the association between IFN-γ and corneal epithelial pyroptosis in a DE mouse model, and these results were confirmed in vitro. 

Materials and Methods

Animal Model and Treatment

Adult C57BL/6 mice (7- to 8- weeks old; wild type) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China), and adult Ifng^−/− mice (8 weeks old) were obtained from Cyagen Biosciences Inc. (Suzhou, China). All experimental procedures strictly conformed to the Association for Research in Vision and Ophthalmology (ARVO) Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Ethics Committee of Zhongshan Ophthalmic Center (Guangzhou, China; approval ID: 2021-008). To establish the experimental dry eye model, the mice were exposed to an air draft and maintained in 30% ambient humidity with food and water available as desired. Meanwhile, the mice received a subcutaneous injection of 1.5 mg/0.3 mL scopolamine hydrobromide (Sigma-Aldrich Corp., St. Louis, MO, USA) three times per day for five days. To inhibit STAT1, 5 µL fludarabine (MedChemExpress, Shanghai, China) or the vehicle control (PBS) was injected subconjunctivally into the mice one day before the induction of dry eye. Neutralization of IFN-γ was performed in WT mice by subconjunctival injection of 10 µL rat anti-mouse IFN-γ XMG1.2 mAb (1 mg/mL, MM700; Thermo Fisher Scientific, Waltham, MA, USA) or rat isotype IgG control (A2119; Selleck Chemicals, Houston, TX, USA). The animal groups were as follows: untreated control mice maintained in a normal environment; WT mice exposed to desiccating stress; WT mice pretreated with fludarabine or PBS before exposure to desiccating stress; WT mice pretreated with anti- IFN-γ antibody or rat IgG before exposure to desiccating stress; and Ifng^−/− mice exposed to desiccating stress. 

Corneal Fluorescein Staining

Fluorescein staining was used to evaluate corneal epithelial defects. Each eye was instilled with 0.25% fluorescein sodium (Jingming, Tianjin, China), the mice were allowed to blink several times, and the eyes were rinsed with normal saline solution. Corneal fluorescein staining was visualized and captured by a slit-lamp microscope (SL-D7/DC-3/MAGENet; Topcon, Tokyo, Japan) under cobalt blue light. Corneal defects were quantified by ImageJ software (Version 1.52a; The National Institutes of Health, Bethesda, MD, USA). The covering area of corneal defects (%) = (fluorescein sodium positive area/the whole cornea) * 100%. 

Measurement of Tear Production

Tear secretion was measured by the phenol red thread (Jingming, Tianjin, China) on Day 5 after completion of the dry eye mouse model. Briefly, one phenol red thread was gently placed in the palpebral conjunctiva of the lower eyelid close to the lateral canthus for 15 seconds. The length of wetting was measured in millimeters. Eight eyes in four different mice per group were measured. Each eye was tested three times, and the average wetting length was recorded as the final length. 

Periodic Acid-Schiff (PAS) Staining

After the mice were euthanized, and the eyeballs were fixed in 4% PFA, embedded in paraffin, and then cut into paraffin sections. The sections were stained using a PAS staining kit (G1008-20ML; Servicebio, Wuhan, China) according to the manufacturer's instructions. Representative images of the superior or inferior conjunctiva were photographed by a light microscope (Eclipse 50i; Nikon, Tokyo, Japan), and all PAS-stained cells in the inferior conjunctival fornix region were counted. Three representative sections per mouse and four mice per group were used for PAS staining. 

Immunofluorescence Staining

For corneal immunofluorescence staining, paraffin sections were deparaffinized, rehydrated, and immersed in 10 mmol/L sodium citrate for antigen retrieval. The sections were rinsed in PBS for five minutes, and then blocked with 3% BSA (Sigma-Aldrich Corp.) plus 0.3% Triton X-100 (Solarbio, Beijing, China) for one hour at room temperature. The samples were incubated with anti-JAK2 (1:400, 3230s; Cell Signaling Technology, Danvers, MA, USA), anti-STAT1 (1:500, 14994s; Cell Signaling Technology), anti-NLRP3 (1:200, 15101s; Cell Signaling Technology), anti-ASC (1:100, sc-514414; Santa Cruz Biotechnology, Dallas, TX, USA), anti-cleaved-caspase1 (1:150, WH215188; ABclonal Technology, Woburn, MA, USA), and anti-N-GSDMD (1:100, 10137; Cell Signaling Technology) at 4°C overnight. After being washed with PBS three times, the samples were incubated with Alexa Fluor 488-labeled donkey anti-rabbit (1:500, 4412; Cell Signaling Technology) or anti-mouse IgG antibodies (1:500, 4408; Cell Signaling Technology) for one hour at room temperature and counterstained with 40,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich Corp.). 

For cell immunofluorescence staining, HCE-T cells on 24-chamber slides were fixed with 4% paraformaldehyde for 10 minutes. The remaining steps were performed as described for corneal immunofluorescence staining. In addition to the primary antibodies mentioned above, anti-p-JAK2 (1:200, 66245; Cell Signaling Technology), anti-p-STAT1 (1:400, 9167; Cell Signaling Technology), and anti-N-GSDMD (1:100, 36425; Cell Signaling Technology) were also used. All sections were photographed using a fluorescence microscope (DMI 8; Leica, Wetzlar, Germany). 

TUNEL Staining

Corneal epithelial cell death was detected by a TUNEL FITC Apoptosis Detection Kit (Vazyme Biotechnology, Nanjing, China) according to the manufacturer's instructions. Paraffin eyeball sections were deparaffinized, rehydrated and permeabilized with proteinase K for 20 minutes at room temperature. After that, the samples were incubated with equilibration buffer for 20 minutes, and then with a TUNEL reaction mixture at 37°C for one hour protected from the light. Finally, the sections were examined under an inverted fluorescence microscope (DMI 8; Leica). The number of TUNEL-positive cells in the cornea of the whole section was counted. Three eyes from three different mice per group and three sections of each eye were used for TUNEL staining. 

Cell Culture and Treatments

The human corneal epithelial cell-transformed (HCE-T) cell line was obtained from ATCC (Manassas, VA, USA) and maintained in Dulbecco's modified Eagle medium/F12 (DMEM/F12; Gibco, Carlsbad, CA, USA) containing 5 µg/mL insulin, 10 ng/mL human epidermal growth factor (Sigma-Aldrich Corp.), 10% fetal bovine serum, and 1% penicillin/streptomycin (Thermo Fisher Scientific; HyClone, Logan, UT, USA) at 37°C in a humidified 5% CO₂ incubator. Primary human corneal epithelial cells (p-HCECs) were obtained from corneal limbal rims after corneal transplantation at Zhongshan Ophthalmic Center (Guangzhou, China) and cultured as previously reported.²⁰ 

To further explore the effect of IFN-γ on corneal epithelial exposure to desiccating stress, we established an in vitro cell model. HCE-T cells and p-HCECs were cultured in 500 mOsM hyperosmolar medium (adding 94 mM sodium chloride to the medium) supplemented with an additional 10 ng/mL recombinant human IFN-γ (Peprotech, Cranbury, NJ, USA). To neutralize IFN-γ, 10 ng/mL rat anti-mouse IFN-γ XMG1.2 mAb (MM700; Thermo Fisher Scientific) was added to the medium. Rat IgG (A2119, Selleck Chemicals) was used as an isotype control. 

RNA Interference

Small interfering RNAs (siRNAs) targeting human STAT1 were designed by Tsingke Biotechnology (Beijing, China). The sequences were as follows: 

• sense, 5ʹ-CGAACAUGACCCUAUCACATT-3ʹ, and
• antisense: 5ʹ-UGUGAUAGGGUCAUGUUCGTT-3ʹ.

A nontargeting scramble siRNA (si-NC) was used as a negative control. The siRNA was dissolved in nuclease-free water to concentration of 10 µM. To prepare the transfection mixture, siRNA and Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA, USA) were added to Opti-MEM (Invitrogen) according to the manufacturer's instructions. The mixture was then added to DMEM free of penicillin/streptomycin and used to culture cells for 24 hours. 

Cell Counting Kit-8 (CCK-8) Assay

Cell viability was measured using a CCK-8 kit (Dojindo, Shanghai, China) according to the manufacturer's instructions. In brief, a total of 1 × 10⁴ HCE-T cells in a volume of 100 µL per well were cultured in 96-well plates for 24 hours and exposed to conditioned medium for another 24 hours. Then, CCK-8 reagent (10 µL) was added to 90 µL of basic medium to generate a working solution, and 100 µL was added to each well and incubated for one hour at 37°C and 5% CO₂ condition away from the light. The absorbance at 450 nm was measured by a microplate reader (Agilent Technologies, Winooski, VT, USA). 

Lactate Dehydrogenase Assay

Cytotoxicity was tested using lactate dehydrogenase (LDH) kit (Dojindo, Shanghai, China). HCE-T cells were cultured in 96-well plates until confluent and then treated with the indicated medium for 24 hours. The supernatant was collected and mixed with the LDH working solution as indicated. After incubation at 37°C for 30 minutes in a 5% CO₂ incubator, the absorbance at 490 nm was determined by a microplate reader (Agilent Technologies). 

RNA-sequencing Analysis

Total RNA was extracted from corneal tissue using a RNeasy Plus Mini Kit (Qiagen, Valencia, CA, USA) according to the manufacturer's recommendations. A TruSeq Stranded mRNA Library Prep kit (no. 20020594; Illumina, San Diego, CA, USA) was used to construct the RNA-Seq libraries, which were sequenced on Illumina PE150 sequencers. Sequencing reads were mapped to hg19 using Star. TPM values were called using RSEM. An absolute value of log2-fold change >1, as well as P value < 0.05 was used to identify differentially expressed genes using DESeq2 analysis. 

Quantitative Real-time PCR (qRT-PCR)

Corneal tissues were collected from two eyes and pooled as one sample, and three samples were used in each group. Total RNA was extracted from corneal tissue using a RNeasy Plus Mini Kit (Qiagen), and total RNA was extracted from HCE-T cells using a PureLink RNA Isolation Kit (Invitrogen). The concentration of RNA was measured by an ND-1000 spectrophotometer (Thermo Fisher Scientific). The RNA was reverse transcribed into cDNA using a HiScript II 1st Strand cDNA Synthesis Kit (Vazyme Biotechnology). QRT-PCR was performed using SYBR Green reagents (Vazyme Biotechnology). The PCR results were analyzed by the comparative threshold cycle (Ct) method, and GAPDH was used as an endogenous reference gene. The primer sequences used in this study are listed in Supplementary Table S1. 

Western Blot Analysis

Corneas from two eyes were pooled as one sample (three samples per group), and the cells to be tested were collected from 6-well plates (three samples per group). To prepare protein lysates, the samples were homogenized in cold RIPA buffer containing 1% proteinase inhibitor and 1% phosphatase inhibitor (Thermo Fisher Scientific) and then spun in a centrifuge at 13,000 rpm for 15 minutes at 4°C. The supernatant was collected, and the protein concentration was determined by BCA assays (Thermo Fisher Scientific). Equal amounts of protein were loaded on 10% SDS-PAGE gels and electronically transferred to 0.22 µm PVDF membranes (Millipore, Bedford, MA, USA). After being blocked in 5% fat-free milk for one hour at room temperature, the membranes were incubated with the following primary antibodies at 4°C overnight: anti-STAT1 (1:1000, 14994; Cell Signaling Technology), anti-JAK2 (1:1000, 3230, Cell Signaling), anti-p-STAT1 (1:1000, 9167, Cell Signaling Technology), anti-p-JAK2 (1:1000, 66245; Cell Signaling Technology), anti-NLRP3 (1:1000, 15101; Cell Signaling Technology), anti-ASC (1:500, sc-514414; Santa Cruz Biotechnology), anti-cleaved-caspase1 (1:1000, WH215188; ABclonal), anti-N-GSDMD (1:1000, 10137, 36425; Cell Signaling Technology), anti-GAPDH (1:1000, 2118; Cell Signaling Technology), and anti-β-actin (1:1000, 20536-1-AP; Proteintech, Rosemont, IL, USA). After the membranes were washed three times with TBST, HRP-conjugated goat anti-rabbit IgG (1:5000, 7074; Cell Signaling Technology) or goat anti-rabbit IgG (1:5000, 7074; Cell Signaling Technology) was applied for two hours at room temperature. The separated bands were visualized by a chemiluminescence solution (Beyotime Biotechnology, Shanghai, China). The band intensities were analyzed by ImageJ software (Version 1.52a; the National Institutes of Health). 

Reactive Oxygen Species (ROS) Assay

The production of ROS was detected by a DCFH-DA Reagent Kit (Meilun, Dalian, China). HCE-T cells (at 5 × 10⁴ cells per well) were seeded in 24-well plates for 24 hours. Subsequently, the experimental group was placed in hyperosmolar medium containing IFN-γ for 24 hours. After the medium was displaced with DCFH-DA (10 µg/mL) and DAPI (4 µg/mL) mixed in basic DMEM/F12 for 30 minutes, and the cells were washed with DMEM/F12 three times. The cells were photographed using a fluorescence microscope (DMI 8; Leica). 

Statistical Analysis

Experiments were independently performed at least three times. Student's t test was conducted for statistical comparison between two individual groups, and one- or two-way ANOVA followed by the Bonferroni's post hoc test was used for comparisons among three or more groups. Statistical analyses were performed using GraphPad Prism 8.0 software (GraphPad Software Inc., San Diego, CA, USA). The data are presented as the mean ± SD. P < 0.05 was considered as statistically significant. 

Results

Ifng Deficiency Alleviates Ocular Surface Damage in Dry Eye

To investigate changes in the IFN-γ signaling pathway in dry eye, we performed RNA sequencing and Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis on corneas collected from untreated (Ctrl) mice and mice exposed to five days of desiccating stress-treated. The sequencing results showed that 211 genes were significantly upregulated and 79 genes were downregulated in the DE group compared with the Ctrl group (Supplementary Fig. S1A). Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis revealed that the JAK-STAT signaling pathway was involved in DED (Supplementary Fig. S1B). In particular, the expression of IFNG, which is the only type II IFN, and IFNGR1 and IFNGR2 was dramatically increased in the DE group compared with the Ctrl group. Moreover, the interferon-inducible genes such as OAS gene family and GBP gene family were increased in the DE group (Fig. 1A). QRT-PCR further confirmed that IFNGR1, IFNGR2, IRF7, and MX1 levels were upregulated in response to desiccating stress (Fig. 1B). To determine the role of IFN-γ in dry eye, we subjected Ifng^−/− mice to desiccating stress conditions. Ifng knockout (KO) significantly alleviated the severity of DE. The area of corneal epithelial defects was reduced in the Ifng^−/− DE group compared with the WT DE group (Figs. 1C, 1D). Tear secretion was decreased in WT mice exposed to desiccating stress on the fifth day, whereas Ifng-KO mice showed less reduction in tear secretion (Fig. 1E). Moreover, Ifng deficiency partially reversed desiccating stress-induced goblet cell loss (Fig. 1F). These data indicate that IFN-γ signaling is activated and plays a pivotal role in the development of DED. 

Figure 1.

 

Ifng deficiency alleviates ocular surface damage in dry eye. (A) Heatmap showing the transcriptional changes in IFN-related genes in the corneas of mice after 5 days of treatment with or without desiccating stress (n = 2 mice/group). (B) IFNGR1, IFNGR2, IRF7, CXCL10, MX1, and MX2 mRNA levels in corneas after five days of desiccating stress compared with those of untreated controls (n = 3 mice/group). (C) Representative fluorescein sodium staining of samples from desiccating stress-induced dry eye mice with or without Ifng deficiency on days 1, 3, and 5 (n = 5 mice/group). (D) The percentage of the corneal defect area was calculated with ImageJ (n = 5 mice/group). (E) Phenol thread tear production test in mice (n = 4 mice/group). (F) Representative PAS staining of conjunctival goblet cells and the comparison of goblet cell density (n = 4 mice/group). Scale bar: 100 µm. Error bars: SD. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 versus untreated controls. ^#P < 0.05; ^####P < 0.0001 versus WT DE group.

IFN-γ Induces Ocular Surface Damage through the JAK2/STAT1 Signaling Pathway

The JAK/ STAT pathway has been demonstrated to be downstream of interferon. However, the specific JAK/STAT signaling factors mediated by IFN-γ in DED remain unclear. RNA sequencing analysis showed that among the JAK genes and STAT genes detected, the increase in the expression of JAK2 and STAT1 was the most pronounced in the DE group compared with the control group (Fig. 2A). qRT-PCR and immunofluorescence staining results showed that JAK2 and STAT1 were increased in the corneal tissue of DE mice (Figs. 2C, 2D). Then, we examined the activation of JAK2/STAT1 in response to IFN-γ stimulation. Western blot analysis demonstrated that the expression of phosphorylated JAK2 (p-JAK2) and p-STAT1 was significantly increased in the corneas of mice exposed to desiccating stress (Fig. 2B), which is consistent with increase in IFN-γ. 

Figure 2.

 

IFN-γ induces ocular surface damage through the JAK2/STAT1 signaling pathway. (A) Heatmap showing the transcriptional changes in the JAK and STAT gene families in the corneas of mice after five days of treatment with or without desiccating stress (n = 2 mice/group). (B) Representative immunoblots showing JAK2, p-JAK2, STAT1, and p-STAT1. The relative protein expression levels were calculated with ImageJ and normalized to GAPDH (n = 3 mice/group). (C) The mRNA expression levels of JAK2 and STAT1 (n = 3 mice/group). (D) Representative immunofluorescence staining of JAK2 and STAT1 (n = 3 mice/group). Scale bar: 100 µm. (E) Representative fluorescein sodium staining of samples from desiccating stress-induced dry eye mice after treatment with the STAT1 inhibitor fludarabine (FA) on days 1, 3, and 5 (n = 5 mice/group). (F) The percentage of the corneal defect area was calculated with ImageJ (n = 5 mice/group). (G) Phenol thread tear production test in mice (n = 4 mice/group). Error bars: SD. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.001 versus untreated controls. ^##P < 0.01; ^###P < 0.001; ^####P < 0.0001 versus DE group. There was no significant difference in corneal defects or tear secretion between the DE and DE+PBS groups.

STAT1, which is a transcription factor encoded by the STAT1 gene, participates in cell proliferation and death. To further clarify the role of STAT1 in DED, fludarabine (FA, 2.5 µM, 5 µL), a specific inhibitor of STAT1, was subconjunctivally injected into the DED mouse model, and PBS was used as a vehicle control (Supplementary Fig. S6A). FA attenuated ocular surface defects (Figs. 2E, 2F) and less reduction in tear secretion on the fifth day after desiccating stress treatment (Fig. 2G), whereas there was no significant difference between the PBS injection group and DE group in terms of corneal defects or tear secretion. These results suggest that STAT1 activation promotes damage to the ocular surface, whereas inhibiting STAT1 protects the cornea from desiccating stress injury to some extent. 

IFN-γ-Mediated Activation of STAT1 Facilitates Pyroptosis in the Corneas in Dry Eye

Pyroptosis, which is a proinflammatory form of programmed necrosis, was demonstrated to contribute to DED development.⁵ We also confirmed that pyroptosis participated in our DED mouse model by RNA sequencing analysis, qRT-PCR, and immunofluorescence staining (Supplementary Figs. S1C–E). Therefore we examined whether IFN-γ could regulate pyroptosis in dry eye. First, we found that knockout of Ifng downregulated the protein expression of JAK2, p-JAK2, STAT1 and p-STAT1 (Fig. 3A). The number of TUNEL-positive cells in the corneal epithelium was dramatically decreased in Ifng-KO DE mice compared with WT DE mice (Figs. 3B, 3C). To further evaluate specific changes in pyroptosis, we performed western blot analysis and immunofluorescence staining of mouse corneas. The expression levels of typical proteins involved in pyroptosis, including NLRP3, ASC, and N-GSDMD, were significantly reduced in desiccating stress-treated Ifng^−/− mice compared with WT DE mice (Figs. 3D, 3E). Additional experiments on WT DE mice treated with a neutralizing IFN-γ antibody confirmed the results obtained in Ifng^−/− mice. Consistently, the anti-IFN-γ antibody but not the rat isotype IgG decreased the expression of JAK2, STAT1 and proteins related to pyroptosis in WT DE mice (Supplementary Figs. S2A, S2B). After IFN-γ neutralization, tear secretion was increased, and corneal defects were alleviated in DE mice (Supplementary Figs. S2C-E). 

Figure 3.

 

Ifng deficiency suppresses pyroptosis in DED by downregulating of JAK2/STAT1. (A) Western blots showing the decreased expression of JAK2, p-JAK2, STAT1, and p-STAT1 in DE mice under a gene knockout background (n = 3 mice/group). (B) Representative TUNEL staining in the corneal epithelium of control mice, WT DE mice and Ifng-KO DE mice (blue, DAPI-stained nuclei; green, TUNEL-positive cells; n = 3 mice/group). Scale bar: 100 µm. (C) Quantitative analysis of the number of TUNEL-positive cells (n = 3 mice/group). (D) Western blot and quantitative analysis showing significantly reduced expression of proteins involved in pyroptosis in the corneas of Ifng-KO DE mice compared with WT DE mice. (E) Representative immunofluorescence staining of NLRP3, ASC, C-CASP1, and N-GSDMD in the corneas of control mice and WT Ifng-KO DE mice (n = 3 mice/group). Scale bar: 100 µm. Error bars: SD. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 versus untreated controls. ^##P < 0.01; ^####P < 0.0001 versus WT DE group.

To determine the pivotal role of STAT1, STAT1 inhibitor was administered to the mice. When STAT1 was specifically inhibited by pretreatment with FA, corneal epithelial cell death was significantly decreased in mice exposed to desiccating stress (Figs. 4A, 4B). Western blot analysis showed that in comparison with the vehicle control, changes in key pyroptosis proteins in DE mice treated with FA were similar to those in Ifng-KO DE mice and WT DE mice treated with an anti-IFN-γ antibody (Fig. 4C), and these results were further confirmed by immunofluorescence staining (Fig. 4D). The data were not significantly different between the DE and DE+PBS groups. In summary, these results indicate that IFN-γ is a critical cytokine promoting corneal epithelial pyroptosis through the JAK2/STAT1 signaling pathway in DED. 

Figure 4.

 

STAT1 inhibitor suppresses pyroptosis in dry eye. (A) Representative TUNEL staining in the corneal epithelium of control mice, DE mice and DE mice treated with FA (blue, DAPI-stained nuclei; green, TUNEL-positive cells; n = 3 mice/group). Scale bar: 100 µm. (B) Quantitative analysis of the number of TUNEL-positive cells (n = 3 mice/group). (C) Western blot analysis showing change in the expression of pyroptosis-related proteins after STAT1 inhibition in DE mice (n = 3 mice/group). (D) Representative immunofluorescence staining of NLRP3, ASC, C-CASP1, and N-GSDMD in the corneal epithelium of control mice, DE mice and DE mice treated with FA (n = 3 mice/group). Scale bar: 100 µm. Error bars: SD. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 versus untreated controls. ^#P < 0.05; ^##P < 0.01; ^####P < 0.0001 versus DE group. There was no significant difference between the DE and DE+PBS groups in these data.

IFN-γ Enhances HCE-T Cell Pyroptosis Under Hypertonic Stress through the JAK2/STAT1 Pathway In Vitro

To explore the direct effects of IFN-γ on corneal epithelial cells in DED, HCE-T cells were cultured in 500 mOsM hyperosmolar medium in vitro to imitate desiccating stress, and exogenous IFN-γ (10 ng/mL) was added concurrently. HCE-T cells were treated for six, 12, and 24 hours, and then the cells were collected at three time points for mRNA expression analysis. The data showed that pyroptosis-related genes were increased most significantly at 24 h (Fig. 5A). Then, the LDH release assay, cell viability analysis and protein expression analysis were all based on this time point. IFN-γ treatment significantly increased LDH release (Fig. 5B) and decreased cell viability under hypertonic stress (Fig. 5C). As expected, compared with those in the control group, NLRP3, ASC, C-CASP1 and N-GSDMD levels were increased in HCE-T cells treated with IFN-γ under hyperosmolar conditions (Fig. 5D), which was confirmed by immunofluorescence staining (Fig. 5E). The administration of IFN-γ alone in the absence of hyperosmolarity had no effect on HCE-T viability or pyroptosis, and IFN-γ plus hyperosmolarity significantly promoted the expression of genes related to pyroptosis in HCE-T cells compared with HCE-T cells treated with hyperosmolar medium alone (Supplementary Figs. S3A, 3B). To further determine the direct effects of IFN-γ on potentiating pyroptosis under hypertonic conditions, we added an anti-IFN-γ antibody to the HCE-T cells or p-HCECs culture medium. The results showed that the viability was increased and pyroptosis was inhibited in HCE-T cells and p-HCECs after the neutralization of IFN-γ compared with the IgG isotype control in the presence of hyperosmolarity plus IFN-γ. The data showed no significant difference between the Hy + IFN-γ and Hy + IFN-γ + IgG groups (Supplementary Figs. S4A–C, Supplementary Figs. S5A, 5B). 

Figure 5.

 

IFN-γ promotes HCE-T cell pyroptosis under hypertonic stress. (A) The mRNA levels of NLRP3, AIM2, ASC, CASP1, and GSDMD in control and IFN-γ plus hyperosmolarity-treated HCE-T cells (n = 3/group). (B) LDH assay showing increased release of LDH in HCE-T cells treated with IFN-γ plus hyperosmolarity (n = 5/group). (C) Decreased viability in HCE-T cells treated with IFN-γ plus hyperosmolarity, as measured by the CCK-8 kit (n = 5/group). (D) Representative immunoblots showing NLRP3, ASC, C-CASP1, and N-GSDMD. Relative protein expression levels were normalized to GAPDH (n = 3/group). (E) Representative immunofluorescence staining of NLRP3, ASC, C-CASP1, and N-GSDMD in control and IFN-γ plus hyperosmolarity-treated HCE-T cells (n = 3/group). Scale bar: 100 µm. Error bars: SD. *P < 0.05; **P < 0.01; ***P < 0.001; ***P < 0.0001 versus untreated controls.

The mRNA levels of JAK2 and STAT1 were elevated in parallel with the increase in pyroptosis-related genes in HCE-T cells in vitro (Fig. 6A). Activated forms of JAK2 and STAT1 were measured by western blot analysis and immunofluorescence staining. As shown in Figures 6B and 6C, the protein expression levels of p-STAT1 and p-JAK2, as well as STAT1 and JAK2, were significantly upregulated when HCE-T cells were exposed to hyperosmolar medium containing IFN-γ. After IFN-γ was neutralized, the expression of JAK2 and STAT1 was partially reduced compared with that in the IgG isotype control group, confirming the direct regulatory effect of IFN-γ on JAK2/STAT1 (Supplementary Figs. S4D, S4E). To determine the effects of STAT1 on corneas observed in vivo, we transfected HCE-T cells with siRNA targeting STAT1. When STAT1 was effectively knocked down, the levels of molecules related to pyroptosis were significantly decreased in HCE-T cells under hyperosmolar conditions (Figs. 6D, 6E). Taken together, these results demonstrate that IFN-γ strengthens pyroptosis in HCE-T cells under hypertonic stress in a JAK2/STAT1-dependent manner. 

Figure 6.

 

Inhibiting STAT1 suppresses pyroptosis in vitro. (A) The mRNA expression of JAK2 and STAT1 in control and IFN-γ plus hyperosmolarity-treated HCE-T cells (n = 3/group). (B) Western blot and quantitative analysis of proteins related to pyroptosis in HCE-T cells (n = 3/group). (C) Representative immunofluorescence staining of JAK2, p-JAK2, STAT1 and p-STAT1 in HCE-T cells (n = 3/group). (D) The mRNA expression of pyroptosis-related genes in si-NC and si-STAT1 HCE-T cells with or without IFN-γ plus hyperosmolarity treatment (n = 3/group). (E) Western blot and quantitative analysis of STAT1, p-STAT1, NLRP3, ASC, C-CASP1, and N-GSDMD in HCE-T cells. *P < 0.05; **P < 0.01; ***P < 0.001; ***P < 0.0001 versus controls. ^#P < 0.05; ^###P < 0.001; ^####P < 0.0001 versus Hy + IFN-γ + NC group.

Oxidative Stress Is Downstream of STAT1 and Promotes Pyroptosis in Dry Eye

Numerous studies have demonstrated that ROS production can activate the NLRP3 inflammasome, which is a key pivot of pyroptosis.^21–23 In our study, we found that oxidative genes were increased in corneas of mice with dry eye (Supplementary Fig. S6B). A similar result was also observed in HCE-T treated with IFN-γ under hypertonic conditions (Fig. 7A). The ROS detection kit further validated that the levels of oxidative stress were increased in the experimental group (Fig. 7B). Furthermore, we treated HCE-T cells with or without N-acetylcysteine (NAC), a ROS scavenger, to explore the relationship between oxidative stress and pyroptosis. The use of NAC had no effect on the JAK2/STAT1 signaling pathway, whereas inhibiting STAT1 reduced oxidative gene levels (Fig. 7C, Supplementary Fig. S6B). These results indicate that oxidative stress may be downstream of JAK2/STAT1. After NAC treatment of HCE-T cells under hypertonic stress, the key components of the NLRP3 inflammasome, including NLRP3, ASC and C-CASP1, were reduced. Correspondingly, the expression of N-GSDMD was also decreased (Fig. 7D). This finding suggests that the presence of oxidative stress contributes to corneal epithelial cell pyroptosis in DED. 

Figure 7.

 

Oxidative stress downstream of JAK2/STAT1 facilitates pyroptosis in DED. (A) The mRNA expression of genes associated with oxidative stress in HCE-T cells (n = 3/group). (B) ROS production was measured by a DCFH-DA Reagent Kit (n = 3/group). (C) Representative immunoblots showing JAK2, p-JAK2, STAT1, and p-STAT1 in HCE-T cells with or without NAC treatment. Relative protein expression levels were to GAPDH (n = 3/group). (D) Western blot and quantitative analysis of NLRP3, ASC, C-CASP1, and N-GSDMD in HCE-T cells treated with or without NAC (n = 3/group). Scale bar: 100 µm. Error bars: SD. **P < 0.01; ***P < 0.001; ****P < 0.0001 versus controls. ^###P < 0.001; ^####P < 0.0001 versus Hy + IFN-γ group.

Discussion

Globally, millions of people suffer from DED, which affects quality of life and causes economic burdens, and dry eye has gained increasing attention in the field of ocular diseases. DED is now considered to be a chronic ocular surface inflammatory disease.³ Extensive evidence indicates that the expression of proinflammatory cytokines such as interleukin (IL)-6, IL-1β, IFN-γ, and TNF-α is increased on the ocular surface in experimental DE models or in patients with DED.^24–26 A well-recognized pathologic cause of ocular surface inflammation is aqueous deficiency due to lacrimal gland dysfunction and goblet cell loss.³ Type I IFN and type II IFN have been reported to be involved in the inflammatory damage of DE. In the CD25-KO model of SS-associated DE, high expression levels of IFN-γ and IFN-γR were critical for glandular apoptosis, whereas knockout of Ifng reduced lacrimal gland destruction and secretory dysfunction.²⁷ IFN-γ can also induce unfolded protein response in goblet cells leading to mucin deficiency in the CD25-KO mice,²⁸ whereas studies on type I IFN in the conjunctiva are limited.²⁹ B6.Aec1Aec2 mice, another SS-like disorder model, showed the detrimental role of IFN-α. The mice deficient in Ifnar1 were protected from mononuclear cell infiltration within the lacrimal glands.³⁰ In our study, although several genes related to type I IFN were increased, type II IFN genes still dominate. Moreover, the expression of IFNG was upregulated, and IFNA was not detected by RNA-seq. We hypothesize that the inflammatory activation in the desiccating stress model is weaker than that in SS-associated DE model, and another reason is that our test samples were corneas rather than lacrimal glands. Therefore IFN-γ may play a more important role in DE. Correspondingly, we identified that the mRNA levels of IFN-γ-responsive genes such as IRF7, MX1, and CXCL10 were increased in the cornea after the mice were exposed to desiccating stress. Knockout of Ifng decreased corneal defects, enhanced aqueous tear secretion and reduced the loss of goblet cells. More importantly, for the first time, we found that IFN-γ could promote corneal epithelial pyroptosis in DE. We further demonstrated the importance of JAK2/STAT1 signaling in this process, which was supported by the evidence that Ifng KO downregulated JAK2/STAT1 and that inhibiting STAT1 decreased the molecular makers of pryroptosis. Furthermore, our data showed increased levels of oxidative stress, which may be a response to the activation of JAK2/STAT1, were involved in pyroptosis. 

Pyroptosis is a type of proinflammatory cell death that has been reported to be involved in a variety of ocular diseases, such as keratitis, DED, cataracts, glaucoma, and uveitis.³¹ To date, among the signaling pathways associated with pyroptosis, NLRP3 inflammasome-dependent cleavage of GSDMD has been well studied in DED.^32,33 Its upstream regulatory signal, however, still needs to be further clarified. IFN-γ, which is the only member of the type II interferon class, contributes to the development of DED.³⁴ In addition to a wide range of proinflammatory effects via regulatory roles of immune cells, numerous studies have linked IFN-γ to different forms of cell death. It was reported that IFN-γ could induce tumor cell ferroptosis by stimulating the enzyme acyl-CoA synthetase long-chain family member 4 (ACSL4),³⁵ whereas an in vitro model of vitiligo using the B6 cell line showed that IFN-γ induced pyroptosis by upregulating GSDMD, CASP1, and CASP8.³⁶ In the present study, we provide the first evidence that IFN-γ can facilitate pyroptosis in DE. Knockout of Ifng significantly suppressed the expression of NLRP3, ASC, and N-GSDMD in the corneas of DE mice. To exclude the possibility of pyroptosis suppression due to the influence of the genetic background on the inadequate development of dry eye, we also performed experiments on WT DE mice using a neutralizing IFN-γ antibody, which further validated the direct effects of IFN-γ on pyroptosis. These results are consistent with a finding that IFN-γ can prime the expression of NLRP3 and ASC in thyroid cells, thereby strengthening pyroptosis in autoimmune thyroiditis.³⁷ Somewhat contradictorily, Vandanmagsar et al.³⁸ reported that an increase in adipose tissue expression of NLRP3 increased IFN-γ expression in obese mice. In fact, NLRP3 can lead to CASP1 activation and subsequent IL-18 secretion. One of the functions of IL-18 is to induce IFN-γ production.³⁹ The interaction between IFN-γ and NLRP3 forms a positive feedback loop to promote pyroptosis and inflammatory aggravation. 

Given that IFN-γ can recruit immune cells to the ocular surface exacerbating inflammatory damage and GSDMD is also expressed in immune cells, it is difficult to determine whether IFN-γ can directly promote pyroptosis in corneal epithelial cells in vivo. Previously, Chen et al.⁵ demonstrated that SDHCECs exposed to hyperosmolarity, which is a dry eye in vitro model, undergo GSDMD-mediated pyroptosis. Similarly, our results also showed that the expression of NLRP3 and CASP1 was increased in HCE-T cells under hyperosmolar stress. The administration of exogenous IFN-γ alone had no effects on cell viability and pyroptosis; however, IFN-γ plus hyperosmolarity significantly reduced cell viability and increased NLRP3 inflammasome components and GSDMD. This pyroptosis-potentiating effect of IFN-γ was partly reversed after IFN-γ was neutralized. More recently, Zhou et al.⁴⁰ revealed that IFN-γ promoted GSDMB expression in several cell lines and that the interdomain of GSDMB was cleaved by NK cell-delivered granzyme A, resulting in pyroptotic killing. However, the expression of GSDMB was not detected by RNA-seq in the corneas of mice in this study. Therefore GSDMB may not be the main executioner of pyroptosis in DE mice. Overall, GSDMD-mediated pyroptosis is directly enhanced by IFN-γ under ocular surface hyperosmolar conditions and is an important pathogenic mechanism in DED. 

The JAK/STAT pathway sustains inflammatory events in autoimmune disorders such as rheumatoid arthritis, thus making it an emerging target in inflammation.^18,41 There are four JAK kinases (JAK1–3 and tyrosine kinase 2 [TYK2]) and seven STAT proteins (STAT1–4, 5A, 5B and 6) in mammals. JAK kinases are phosphorylated in response to cytokine stimulation, which is followed by the phosphorylation of STAT proteins, which induce expression of caspases and death receptors.^17,42 This process can mediate cell death and amplify downstream inflammation. IFN-γ has been reported to be an important initiator of the JAK/STAT pathway. In our study, we found among the JAK and STAT gene families, the increases in JAK2 and STAT1 were most obvious in DE mice. Knockout or neutralization of IFN-γ significantly decreased the expression levels of JAK2 and STAT1. These results were further confirmed by in vitro experiments, which suggested that IFN-γ primarily triggers activation of the JAK2/STAT1 signaling pathway in DE. To assess the role of JAK2/STAT1 in pyroptosis, STAT1 was specifically inhibited in vivo and in vitro. We discovered that inhibiting STAT1 suppressed NLRP3 inflammasome-induced pyroptosis in corneal epithelial cells. Consistent with our results, Wang et al.⁴³ found that microglial pyroptosis in spinal cord injury was inhibited when the JAK2/STAT1 pathway was suppressed by Toll-like receptor 4 deficiency. Nevertheless, another report showed that the upregulation of JAK2 in macrophages activated the AIM2 inflammasome enhancing IL-1β secretion, which increased AKT and ERK signaling and thereby promoted proliferation and pyroptosis in macrophages.⁴⁴ Because AIM2 can also be regulated by IFN-γ through JAK/STAT signaling, the role of AIM2 in DED needs to be addressed in future investigations. Here, we primarily demonstrated that IFN-γ facilitated corneal epithelial cell pyroptosis in DE by activating the NLRP3 inflammasome through the JAK2/STAT1 pathway. 

The corneal epithelium serves as the outermost layer of the eye and confronts a variety of environmental stresses that cause the production of ROS. Our previous study showed that excessive ROS release can lead to impaired autophagy and promote cell death in DED.⁴⁵ Although an increase in ROS could directly cause STAT1 signaling activation,⁴⁶ in our study, scavenging ROS had no effect on expression of p-JAK2 or p-STAT1, and inhibiting STAT1 reduced oxidative gene expression, thereby situating STAT1 upstream of ROS. Zheng et al.⁴⁷ demonstrated that an increase in ROS generation triggered NLRP3 inflammasome activation leading to IL-1β secretion in a murine DE model. Our in vitro experimental results further verified that reducing ROS levels with NAC could suppress corneal epithelial cell pyroptosis. Taken together, these results suggest that STAT1-induced overproduction of ROS may facilitate pyroptosis in DED. 

In conclusion, we demonstrate that IFN-γ promotes corneal epithelial cell pyroptosis by activating the JAK2/STAT1 pathway in DED and suggest that targeting IFN-γ or its downstream signaling may be a potential therapeutic approach to control DED. 

Acknowledgments

Supported by the National Natural Science Foundation of China to Jin Yuan (No. 82171015), National Natural Science Foundation of China to Bowen Wang (No. 82101083), and Guangdong Basic and Applied Basic Research to Bowen Wang (2022A1515010445). The funding organization played no role in the design or execution of this research. 

Disclosure: X. Yang, None; X. Zuo, None; H. Zeng, None; K. Liao, None; D. He, None; B. Wang, None; J. Yuan, None