Honokiol Alleviates Dry Eye-Related Corneal Epithelial Damage by Activating SIRT3/PINK1-Mediated Mitophagy

Abstract

Background

Dry eye is a common ocular surface disease characterized by corneal epithelial damage, yet effective therapies targeting mitochondrial dysfunction remain limited.

Purpose

To investigate the protective effects of honokiol (HKL) against dry eye-related corneal epithelial injury and elucidate the underlying mechanisms focusing on SIRT3 and PINK1-mediated mitophagy.

Methods

A benzalkonium chloride (BAC)-induced dry eye mouse model (n=6) was established and treated with HKL eye drops. Corneal damage was assessed by slit-lamp photography, Schirmer I test, and H&E staining. SIRT3, PINK1, and mitophagy-related proteins were examined by immunofluorescence and Western blot. Human corneal epithelial cells (HCECs) were exposed to hyperosmolar stress with or without HKL (n=5). Cell viability, oxidative stress, mitochondrial morphology, and function were evaluated. PINK1 knockdown was performed to validate its role in HKL-mediated protection.

Results

In both BAC-induced mice and hyperosmolar-stressed HCECs, SIRT3 was downregulated with activation of PINK1/Parkin-mediated mitophagy, accompanied by mitochondrial dysfunction, apoptosis (increased to 36.2%, P < 0.001), and inflammation. HKL treatment restored mitochondrial dynamics balance, enhanced PINK1-dependent mitophagy, and alleviated oxidative stress, apoptosis (reduced to 17.4%, P < 0.001), and inflammation. PINK1 knockdown abrogated these protective effects, confirming PINK1 as a key mediator. Notably, HKL primarily enhanced SIRT3 activity rather than its expression, as immunofluorescence showed partial SIRT3 recovery in the corneal epithelium while Western blot showed no statistical significance. In vivo, HKL improved corneal integrity, tear secretion (from 1.8 to 3.5 mm/2min, P < 0.01), and tissue structure, consistent with in vitro findings.

Conclusion

This study demonstrates that HKL protects against dry eye-associated corneal epithelial damage potentially through a mechanism involving SIRT3 modulation and PINK1-mediated mitophagy, offering a potential therapeutic strategy for dry eye disease.

Introduction

Honokiol (HKL), a bioactive lignan derived from the bark, seed cones, and leaves of Magnolia officinalis, has been a cornerstone in traditional Chinese, Japanese, and Korean medicine for centuries.1 Traditionally, Magnolia officinalis preparations have been widely used to treat a variety of ailments, including anxiety, gastrointestinal disorders (eg, bloating, nausea, and abdominal pain), asthma, and allergic conditions.2,3 Its therapeutic applications were largely attributed to its anti-inflammatory, antioxidant, and neuroprotective properties, which have now been increasingly validated by modern pharmacological studies.4,5 Honokiol, as one of the major bioactive constituents, has accordingly emerged as a promising candidate for drug development.

As a potent and specific SIRT3 activator,6 HKL directly binds to SIRT3, enhancing its deacetylase activity, and also upregulates SIRT3 expression through transcriptional mechanisms.7 HKL exerts protective effects in cardiovascular, neurodegenerative, and metabolic disorders through SIRT3-dependent enhancement of mitochondrial function, including mitophagy promotion.8,9 Notably, honokiol exhibits favorable drug-like properties, including low molecular weight, appropriate lipophilicity, and a well-established preclinical safety profile.6 However, its potential therapeutic role in Dry eye disease (DED) has not been previously investigated.

As recently outlined by the Tear Film and Ocular Surface Society (TFOS) DEWS III Diagnostic Methodology report,10 DED is defined as a multifactorial, symptomatic disease characterized by a loss of ocular surface homeostasis, with tear film instability, hyperosmolarity, inflammation, and neurosensory abnormalities serving as core etiological factors. It impacts between 5% and 50% of people worldwide, with cases rising due to factors like aging, increased use of digital devices, and environmental influences.11–13 The clinical impact of DED extends far beyond ocular discomfort, profoundly impairing patients’ ability to perform daily activities such as reading, using digital devices, driving, and even social interactions, leading to reduced work productivity and significant psychological distress.14–16 Alarmingly, recent epidemiological data reveal that while over half of the general population experiences dry eye symptoms, only approximately one-fifth of affected individuals receive a formal diagnosis, and many suffer for years before seeking professional help.17 This substantial underdiagnosis and treatment gap contribute to disease progression. Untreated DED can lead to serious complications including persistent inflammation, corneal epithelial damage, and vision impairment.18 Furthermore, according to the TFOS DEWS III Management and Therapy Report,19 current management emphasizes a staged, etiology‑driven approach. Current therapeutic strategies, including artificial tears, anti-inflammatory agents (eg, cyclosporine and corticosteroids), and punctal occlusion, mainly provide symptomatic relief or broadly suppress inflammation.20,21 However, these conventional treatments require frequent administration, may induce local adverse effects upon long-term application, and fundamentally fail to arrest progressive mitochondrial and corneal epithelial deterioration, which serves as the core pathological driver of DED. Therefore, there is an urgent unmet clinical need for novel disease-modifying therapies targeting upstream pathological processes. With its favorable safety profile, superior ocular bioavailability, and robust mitochondrial protective effects, HKL possesses great translational potential to fill this therapeutic gap by protecting corneal epithelial cells from oxidative stress-induced cell death and maintaining ocular surface integrity.

The corneal epithelium, as the frontline of the ocular surface, is continuously exposed to environmental stressors including desiccation, UV radiation, and atmospheric oxygen. To maintain its barrier function, transparency, and rapid turnover, corneal epithelial cells require robust energy production and metabolic flexibility, rendering them heavily dependent on mitochondrial function.22,23 Mitochondria not only generate ATP through oxidative phosphorylation but also regulate redox homeostasis, calcium signaling, and apoptosis.24 In DED, hyperosmolar stress—a key pathogenic driver—induces mitochondrial reactive oxygen species (ROS) overproduction, leading to oxidative damage to lipids, proteins, and mitochondrial DNA (mtDNA). This results in mitochondrial membrane potential dissipation, impaired ATP synthesis, and activation of mitochondrial apoptotic pathways, culminating in corneal epithelial cell death and ocular surface disruption.25–27 Mitochondrial quality control mechanisms are essential for maintaining a healthy mitochondrial network, with mitophagy serving as a critical pathway that selectively eliminates damaged or depolarized mitochondria via autophagic degradation.28 The PTEN-induced putative kinase 1 (PINK1)/Parkin pathway represents the most extensively characterized mitophagy mechanism. PINK1 stabilizes on damaged mitochondria and recruits Parkin to initiate ubiquitination and autophagic clearance. This process prevents accumulation of dysfunctional organelles that generate excessive ROS and release pro-apoptotic factors.29,30 Recent evidence demonstrates that mitophagy is impaired in stressed corneal epithelial cells, leading to persistence of damaged mitochondria and exacerbation of oxidative injury in DED.31,32 Enhancing mitophagic clearance has thus emerged as a potential therapeutic strategy to restore mitochondrial homeostasis.

Sirtuin 3 (SIRT3) is the primary mitochondrial deacetylase regulating antioxidant defense, energy metabolism, and mitochondrial quality control.33 Importantly, SIRT3 promotes mitophagy by deacetylating and stabilizing PINK1, facilitating its accumulation on damaged mitochondria and subsequent Parkin recruitment.34,35 SIRT3 also activates peroxisome proliferator-activated receptor γ coactivator 1-α (PGC-1α), the master regulator of mitochondrial biogenesis, thereby coordinating the generation of new mitochondria to replace those eliminated by mitophagy.36 In various oxidative stress-related diseases, SIRT3 downregulation correlates with impaired mitophagy and mitochondrial dysfunction, suggesting SIRT3 activation as a promising therapeutic strategy.37

Given the critical role of mitochondrial dysfunction in DED pathogenesis and the capacity of SIRT3 to indirectly regulate mitophagy, we hypothesized that HKL could restore corneal epithelial homeostasis via SIRT3‑mediated enhancement of mitophagy. While previous studies have independently documented SIRT3 downregulation and impaired mitophagy in DED, to our knowledge, this study provides the first evidence exploring the potential role of SIRT3 and PINK1-mediated mitophagy as a combined therapeutic target in the corneal epithelium. By elucidating how HKL mitigates hyperosmolar-induced damage, our findings highlight the novelty of targeting mitochondrial quality control in the ocular surface. Ultimately, this study underscores the translational relevance of HKL, proposing it as a highly promising, mechanism-driven therapeutic strategy to address the current clinical limitations in DED management.

Materials and Methods

Cell Culture and Treatment Protocol

The immortalized human corneal epithelial cell (HCEC) line utilized in this study was obtained from the RIKEN Cell Bank (RCB2280; Tsukuba, Japan). Cells were maintained in Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12 (DMEM/F12; Gibco, USA) supplemented with 6% fetal bovine serum (FBS; Gibco, USA), 7µg/mL insulin (Sigma-Aldrich), 7ng/mL epithelial growth factor (Gibco, USA), and 1% penicillin-streptomycin (v/v). Cultures were incubated at 37°C in a humidified atmosphere containing 5% CO₂. HCECs between passages 3 and 5 were employed for all experimental procedures. For high osmotic pressure (HOP) stimulation, cells were cultured in medium adjusted to 500mOsm/L by adding 94mM NaCl.

HKL (Cat No. HY-N0003; MCE; USA; summary of HKL information in Supplementary Table S1) was dissolved in dimethyl sulfoxide (DMSO) to prepare a 100mM stock solution, which was then diluted with culture medium to the indicated concentrations (final DMSO concentration≤ 0.1%). To evaluate the protective effect of HKL, cells were pretreated with 10μM HKL for 2h prior to hyperosmolar exposure for indicated time periods. Control cells were maintained in normosmolar medium (312mOsm/L) with or without HKL.

Dry Eye Mouse Model

Male C57BL/6J mice, aged between 6 and 8 weeks, were procured from Beijing SPF Biotechnology Co., Ltd. (Beijing, China). All animal experimental protocols were approved by the Department of Corneal Diseases at Tianjin Eye Hospital (Approval No. ZXHK-DWLL-2025-0128) and were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The mice were maintained under controlled environmental conditions within an animal facility, with temperature regulated at 24°C±1°C and relative humidity at 50%±5%. To evaluate ocular surface clinical parameters (Schirmer I test and corneal fluorescein staining), n=6 mice per group were utilized. For subsequent ex vivo assays, including Western blot and immunofluorescence staining, n=5 mice per group were used. The sample sizes were determined based on established protocols from comparable previous studies utilizing the BAC-induced dry eye model,38,39 which demonstrate that 5 to 6 animals per group provide adequate statistical power to detect biologically meaningful differences while strictly adhering to the “Reduction” principle of animal welfare. No animals, tissues, images, or experimental replicates were excluded from the final analyses. Predefined exclusion criteria for this study included: (1) unexpected animal death prior to the experimental endpoint; (2) signs of severe ocular or systemic infection unrelated to the BAC-induced modeling; and (3) apparent technical failures during tissue collection or processing. However, none of these conditions were encountered during the study. A one-week acclimatization period was provided to allow adaptation to the housing environment. The dry eye model was established via topical administration of benzalkonium chloride (BAC).40 Animals were randomly allocated into three groups: PBS group, BAC+PBS group and BAC+HKL group. Briefly, a solution containing 5μL 0.2% BAC (Beyotime, China, Y116738) was applied to both eyes twice daily for 7 consecutive days, while control mice received 5µL of (phosphate-buffered saline) PBS. Starting 10min after BAC treatment, instill 100µM HKL into both eyes twice daily. At the end of the experiment, euthanasia was performed by intraperitoneal injection of an overdose of sodium pentobarbital (200 mg/kg, Sigma). Death was confirmed by complete cessation of heartbeat and respiration, as well as loss of pupillary reflex. This method is consistent with the euthanasia guidelines of the American Veterinary Medical Association (AVMA).

Tear Secretion Measurement and Fluorescein Staining

Tear secretion was evaluated utilizing the Schirmer I test with phenol red-impregnated cotton threads (Boruimin, China). The threads were gently placed in the outer one-third of the inferior conjunctival sac of mice for 2min. The length of the color-changed portion of the thread (from yellow to red) was measured in millimeters in accordance with the manufacturer’s guidelines to determine tear production. For evaluation of corneal epithelial integrity, 1% fluorescein sodium was instilled into the conjunctival sac, followed by manual closure of the eyelids three times. After ninety seconds, corneal epithelial damage was assessed using a slit-lamp microscope equipped with a cobalt blue filter. The cornea was systematically divided into four quadrants, each of which was individually scored. The cumulative sum of these four scores constituted the final damage grade, with a maximum possible score of 16 points.41 To minimize investigator bias, formal blinding was implemented during the in vivo clinical assessments. Specifically, the investigators performing the Schirmer I test and clinical scoring of corneal fluorescein staining (CFS) were blinded to the animal group allocations.

Hematoxylin and Eosin Staining

Following euthanasia, the mouse eyeballs were promptly excised and fixed in 4% paraformaldehyde. After a 24-hour fixation period at room temperature, the specimens were subjected to graded dehydration and subsequently embedded in paraffin. Sections were then prepared from all experimental groups and stained using hematoxylin and eosin (H&E) for histological analysis.

Immunofluorescence Staining

For the immunofluorescence analysis of frozen tissue sections, eyeballs were harvested post-euthanasia and embedded in O.C.T. compound (Sakura, USA). The embedded specimens were stored at −80°C and subsequently sectioned into 10 µm thick slices. These sections were fixed with 4% paraformaldehyde for 15 minutes, followed by blocking with 5% normal goat serum for 30min at room temperature. Thereafter, the sections were incubated overnight at 4°C with primary antibodies targeting SIRT3 (Proteintech, China, catalog no. 10099-1-AP) and PINK1 (Proteintech, China, catalog no. 23274-1-AP). After three washes with PBS, the sections were treated with appropriate Alexa Fluor-conjugated secondary antibodies, specifically Alexa Fluor 488-conjugated goat anti-rabbit (Thermo Fisher Scientific, USA) and DyLight 549-conjugated goat anti-rabbit (HuaXia Pathology, China). Subsequently, nuclei were counterstained with DAPI, and the slides were mounted using an antifade medium. Imaging was performed using a laser scanning confocal microscope (TCS SP8; Leica, Germany).

For immunofluorescence staining of cultured cells, cells were fixed with 4% paraformaldehyde and permeabilized with 0.2% Triton X-100 at room temperature for 30min. Following permeabilization, cells were blocked with 5% bovine serum albumin for one hour at room temperature and then incubated overnight at 4°C with primary antibodies against TOM20 (Proteintech, China, catalog no. 11802-1-AP) and LC3 (Santa Cruz Biotechnology, USA, catalog no. sc-398822). Nuclear staining was performed using DAPI. Quantitative analysis of fluorescence intensity was conducted utilizing ImageJ software (version 2.3.0).

Small Interfering RNA Transfection

Cells at approximately 50% confluence were transfected with siRNA specific to PINK1 (siPINK1) utilizing Lipofectamine 3000 for a duration of 24h. Post-transfection, the cells were pretreated with 10μM Honokiol (HKL) for 2h, followed by incubation under HOP conditions for an additional 24h. Subsequently, the cells were harvested for analysis via Western blotting, flow cytometry, and Seahorse metabolic assays. The sequences of the siRNAs employed are detailed in Supplementary Table S2.

Cell Viability Assay

Cell viability was evaluated utilizing the Cell Counting Kit-8 (CCK-8) assay. HCECs were cultured for 24h in either normal medium or hyperosmotic medium (500 mOsm/L) supplemented with varying concentrations of HKL. After the addition of the CCK-8 reagent, the cells were incubated, and the absorbance was subsequently measured at 450 nm using a microplate reader.

Real-Time Quantitative PCR Analysis

Total RNA was extracted from HCECs and mouse corneal tissues utilizing TRIzol reagent (Sigma-Aldrich) in accordance with the manufacturer’s instructions. Complementary DNA (cDNA) synthesis was conducted using the PrimeScript RT Kit (Takara). Quantitative PCR (qPCR) assays were carried out employing SYBR Premix Ex Taq (Takara) alongside gene-specific primers detailed in Supplementary Table S3. Relative mRNA expression levels were normalized to the housekeeping gene GAPDH and quantified using the 2^−ΔΔCq comparative method.

Western Blotting

Corneal tissues and HCECs were lysed using RIPA buffer (Beyotime, China) supplemented with protease inhibitors. Protein concentrations were quantified via the bicinchoninic acid (BCA) assay (Thermo Fisher Scientific, USA). Equal amounts of protein (30 μg per lane) were subjected to separation by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and subsequently transferred onto polyvinylidene difluoride (PVDF) membranes (Millipore, USA). The membranes were blocked with 5% skim milk (Biotopped, China) in PBS containing 0.1% Tween-20 for 2 hours at ambient temperature. This was followed by an overnight incubation at 4°C with primary antibodies targeting the following proteins: TOM20 (Proteintech, 11802-1-AP), PINK1 (Proteintech, 23274-1-AP), p62 (Abmart, T55546), LC3B (Abmart, T55992), TFAM (Proteintech, 22586-1-AP), Parkin (Proteintech, 14060-1-AP), SIRT3 (Proteintech, 10099-1-AP), Tim23 (Santa Cruz, sc-514463), DRP1 (CST, 8570), OPA1 (CST, 67589), Bcl2 (CST, 4223), cleaved caspase-3 (Proteintech, 87055-4-RR), Bax (CST, 5023), and β-actin (Proteintech, 20536-1-AP). Following thorough washing, membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies—either HRP-goat anti-mouse recombinant secondary antibody (Proteintech, China, RGAM001) or HRP-goat anti-rabbit recombinant secondary antibody (Proteintech, China, RGAR001)—for 1.5 hours at room temperature. Protein bands were detected using enhanced chemiluminescence reagents (Beyotime, China) and subsequently quantified by densitometric analysis employing ImageJ software.

Assessment of the Mitochondrial Membrane Potential (ΔΨm)

The mitochondrial membrane potential was evaluated utilizing the JC-1 assay kit (Dojindo, Japan). Following experimental treatments, cells were rinsed with PBS and subsequently incubated with a JC-1 working solution diluted 1:1 with complete culture medium at 37°C for 30 minutes. Thereafter, cells were washed twice with JC-1 staining buffer, overlaid with fresh complete medium, and promptly examined using fluorescence microscopy.

Flow Cytometry

Apoptotic cell death was assessed utilizing the Annexin V-FITC Apoptosis Detection Kit (Dojindo, Japan, AD10). In brief, HCECs were harvested following treatment, washed with PBS, and subsequently resuspended in binding buffer. The cells were then stained with Annexin V-FITC and propidium iodide (PI) in accordance with the manufacturer’s protocol. Quantification of apoptotic cells was performed via flow cytometry using a BD FACSAria III instrument, and the resulting data were processed and analyzed with FlowJo software (version 10.8.1).

Reactive Oxygen Species (ROS) Analysis

Intracellular levels of reactive oxygen species (ROS) were quantified utilizing a ROS Detection Kit (Beyotime, China, S0033S). Following experimental treatment, HCECs were incubated with serum-free culture medium supplemented with the fluorescent probe 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) at 37°C for 20 minutes. Subsequently, the cells were washed three times with serum-free culture medium to eliminate any unbound probe. Intracellular ROS-associated fluorescence was then immediately observed and recorded using a fluorescence microscope.

Mitochondrial ROS Measurement

Mitochondrial ROS were detected using MitoSOX Red (Beyotime, China). After treatment, HCECs were rinsed with HBSS and subsequently incubated with 5 μM MitoSOX Red at 37°C for 10 minutes in the absence of light. After three washes with PBS, fluorescence images were captured using confocal microscopy and quantitatively analyzed employing ImageJ software.

Seahorse Assay

The oxygen consumption rate (OCR) of HCECs was quantified utilizing a Seahorse XFe96 Analyzer (Agilent Technologies). Cells were plated in Seahorse XF microplates at an optimized density and incubated overnight. Prior to measurement, the cells were washed and incubated for one hour at 37 °C in a non-CO₂ incubator with unbuffered XF base medium supplemented with glucose, L-glutamine, and sodium pyruvate. OCR was assessed under basal conditions and following the sequential administration of oligomycin (1.5μM), FCCP (1μM), and a combination of rotenone and antimycin A (0.5μM each). The OCR data were normalized to cell number. Key parameters, including basal respiration, ATP-linked respiration, maximal respiration, and spare respiratory capacity, were derived from the measurements. Each assay was independently replicated a minimum of three times.

Measurement of Mitochondrial Morphology and Structure

Cells were incubated with 300 nM MitoTracker Deep Red FM (ThermoFisher, USA, M22426) for 30 min, and the staining solution was then replaced with serum-free medium. Confocal microscopy was employed to capture images of the stained cells. To comprehensively evaluate mitochondrial morphology, two distinct quantification approaches were applied to the confocal images. First, automated morphological parameters—including mean mitochondrial area, mean perimeter, form factor, and aspect ratio—were measured using the Mitochondria Analyzer plugin in ImageJ software, for which a minimum of 10 randomly selected cells per group were analyzed. Second, a detailed visual categorization was performed according to previously established criteria.42 For this highly detailed approach, five representative cells per group were randomly selected, and every single mitochondrion within these cells was visually classified into one of four distinct morphological categories (tubular, intermediate, fragmented, or swollen) by an investigator who was blinded to the experimental group allocations.

Statistical Analysis

The data are expressed as the mean±standard deviation (SD) derived from a minimum of five independent experiments, each conducted in triplicate per group. Statistical analyses and graphical visualizations were carried out using GraphPad Prism software (version 10.6.0). Differences between two groups were assessed using an unpaired, two‑tailed Student’s t‑test. For comparisons involving multiple groups, one‑way analysis of variance (ANOVA) followed by Tukey’s post hoc test was used for parametric data. A p-value less than 0.05 was regarded as indicative of statistical significance. For Seahorse assays, three independent biological replicates were conducted (n=3 per group). Given the limited sample size, permutation testing with 10,000 permutations was adopted for intergroup comparisons. Effect sizes alongside mean differences are presented to aid biological interpretation, and p-values are interpreted conservatively.

Results

BAC-Induced Corneal Damage and Alterations in Mitophagy and SIRT3 Expression in Mouse Cornea

Corneal epithelial integrity was assessed by fluorescein staining. Unlike PBS controls with smooth, unstained corneas, BAC-treated mice exhibited dense punctate fluorescein staining with confluent areas on day 7, indicating compromised epithelial barrier function (Figure 1A). CFS scores were significantly elevated in the BAC group, reflecting persistent corneal injury (Figure 1B, n=6). Schirmer I test also showed markedly reduced tear secretion in BAC-exposed mice (Figure 1F, n=6). H&E staining revealed 4–5 layers of well-organized epithelial cells in the PBS group, whereas BAC-treated corneas displayed epithelial thinning, disorganized cellular architecture, and edema (Figure 1E), consistent with fluorescein staining.

Figure 1.

BAC-induced corneal damage and alterations in mitophagy and SIRT3 expression in mouse cornea. (A) Representative corneal fluorescein staining images of PBS and 0.2% BAC-treated mice at day 7. (B) Quantitative analysis of CFS scores (n=6). (C) Western blot analysis of SIRT3 protein expression in mouse cornea. (D) Densitometric quantification of SIRT3 protein levels (n=5). (E) H&E staining of corneal sections. (F) Tear secretion measured by Schirmer I test (n=6). (G) Quantitative analysis of SIRT3 immunofluorescence intensity (n=5). (H) Immunofluorescence staining of SIRT3 in mouse cornea. (I) Western blot analysis of mitophagy-related proteins (PINK1, Parkin, LC3, p62, Tom20, Tim23). (J–O) Densitometric quantification of PINK1, Parkin, LC3, p62, Tom20, and Tim23 protein levels (n=5). Data are presented as the mean ± SD. ns, not significant; * P < 0.05, ** P < 0.01, *** P < 0.001 versus PBS group. Exact p-values for all statistical comparisons are detailed in Supplementary Table S4.

Western blot analysis (n=5) showed that the LC3 II/LC3 I ratio was significantly increased in the BAC group, along with elevated PINK1, Parkin and p62 expression. No significant changes were observed in Tom20, or Tim23 levels, indicating BAC specifically initiated mitophagy (Figures 1I–O). SIRT3 protein expression (n=5) was downregulated in corneal tissues of BAC-treated mice (Figures 1C and D), and immunofluorescence confirmed reduced SIRT3 signal in the corneal epithelium after BAC treatment (Figure 1H and G).

In summary, 0.2% BAC successfully induced corneal epithelial injury in mice, accompanied by mitophagy activation and SIRT3 downregulation, suggesting a potential role for SIRT3 in BAC-induced mitophagy and corneal damage, though the underlying mechanisms require further investigation.

HOP Induces SIRT3 Downregulation, Dynamic Changes in Mitophagy-Related Proteins, Mitochondrial Dysfunction, and Morphological Damage in HCECs

HCECs were cultured under normal osmotic pressure or HOP, and assessments were performed at 6h and 24h to evaluate SIRT3 expression, mitophagy levels, mitochondrial respiratory function, and morphology (n=5). As shown in Figures 2A and E, SIRT3 protein expression decreased in a time-dependent manner following HOP. Key mitophagy-related proteins displayed distinct temporal patterns: PINK1 expression increased at 6h but declined by 24h, remaining above control levels (Figure 2A and C). Parkin expression showed sustained accumulation from 6h to 24h (Figure 2A and D). The LC3 II/LC3 I ratio was unchanged at 6h but significantly elevated at 24h (Figure 2A and F). p62 and Tom20 levels were reduced at both time points (Figure 2A, B, G), indicating that HOP induces mitophagy with complex dynamic regulation.

Figure 2.

HOP induces SIRT3 downregulation, dynamic changes in mitophagy-related proteins, mitochondrial dysfunction, and morphological damage in HCECs. (A) Western blot analysis of SIRT3 and mitophagy-related proteins (PINK1, Parkin, LC3, p62, Tom20) in HCECs treated with hyperosmolar medium for 6h and 24h. (B–G) Densitometric quantification of p62, PINK1, Parkin, SIRT3, LC3, and Tom20 (n=5). (H) Representative MitoTracker staining images showing mitochondrial morphology in HCECs treated with Normal or hyperosmolar medium for 6h,24h. White boxes mark the regions magnified in the lower panels. (M–O) Densitometric quantification of mean branch lenth, mean perimeter, mean branch diameter (n=10). Mitochondrial morphology was quantified (P) according to the criteria detailed in “Material and Methods”. (I) Seahorse XF analysis of OCR in HCECs treated with Normal or hyperosmolar medium for 6h,24h. Quantitative analysis of basal respiration (J), maximal respiration (K), spare respiratory capacity (L) and ATP production (Q) (n=3). Data are presented as the mean ± SD. ns, not significant; * P < 0.05, ** P < 0.01, *** P < 0.001 versus Ctrl group; # P < 0.05, ## P < 0.01, ### P < 0.001 versus HOP 6h group; @@ P < 0.01 versus fragmented of Ctrl group; andandand P < 0.001 versus swollen of Ctrl group.

Mitochondrial function was assessed by Seahorse XF analysis. Compared with controls, the HOP 6h group showed no significant changes in basal respiration, ATP-linked respiration, or maximal respiratory capacity, whereas all these parameters were significantly reduced in the HOP 24h group (Figure 2I, n=3). Quantitative analysis revealed a trend toward decreased basal respiration, maximal respiratory capacity, spare respiratory capacity and ATP production in the HOP 24h group compared to the control group (Figure 2J, K, L, Q, n=3). Permutation test p-values were >0.05, likely due to the limited statistical power associated with the small sample size. These results indicate that mitochondrial respiratory function remains compensatory under short-term hyperosmolar exposure but becomes severely impaired after prolonged exposure, supporting the use of the 24h time point for subsequent HKL treatment.

Mitochondrial morphology was examined by MitoTracker staining and confocal microscopy. Control HCECs displayed filamentous, reticular mitochondria (Figure 2H). In the HOP 6h group, mitochondria showed shortening and fragmentation with disrupted networks. The HOP 24h group exhibited more severe disruption, with punctate, fragmented aggregates and swollen, vacuolated mitochondria. Quantitative analysis (n=10) confirmed significantly increased fragmentation and decreased branch length and perimeter in HOP groups, with more pronounced changes at 24h than at 6h (Figure 2M–P).

In summary, HOP directly downregulates SIRT3 in HCECs and induces differential expression of mitophagy-related proteins, including an initial rise and subsequent decline of PINK1, sustained Parkin accumulation, and late-stage LC3 II elevation, accompanied by progressive impairment of mitochondrial respiratory function and gradual disruption of mitochondrial morphology.

HKL Alleviates HOP-Induced Oxidative Stress, Mitochondrial Damage, and Morphological Disruption in HCECs

To identify the optimal concentration of HKL for mitigating HOP-induced cellular injury, cell viability was evaluated via CCK-8 assays. Under normal osmotic pressure, HKL concentrations up to 20μM did not adversely affect cell viability (Figure 3A). Cells were pretreated with varying concentrations of HKL (2, 5, and 10μM) for 2h, followed by exposure to HOP for 24h. Cell viability demonstrated a concentration-dependent improvement, reaching a maximum at 10μM before declining at higher concentrations (Figure 3B). Notably, 10μM HKL significantly enhanced cell viability under hyperosmolar conditions without exhibiting cytotoxic effects under normal osmotic pressure. Consequently, 10μM HKL was selected for subsequent mechanistic investigations.

Figure 3.

HKL alleviates HOP-induced oxidative stress, mitochondrial damage, and morphological disruption in HCECs. (A) Cell viability of HCECs treated with indicated concentrations of HKL under normosmolar conditions for 24h (n=5). (B) Cell viability of HCECs pretreated with indicated concentrations of HKL for 2h followed by hyperosmolar stimulation for 24h (n=5). (C) Western blot analysis of mitochondrial dynamics-related proteins OPA1 and DRP1. (D) Representative MitoTracker staining images showing mitochondrial morphology in HCECs treated with Normal or hyperosmolar medium for 24h, with or without HKL treated. White boxes mark the regions magnified in the lower panels. Mitochondrial morphology was quantified (G) according to the criteria detailed in “Methods”. (E and F) Densitometric quantification of OPA1 and DRP1 (n=5). (H–J) Densitometric quantification of mean branch lenth, mean perimeter, mean branch diameter (n=10). (K) Representative images of intracellular ROS detection by DCFH-DA staining. (L) Quantitative analysis of intracellular ROS fluorescence intensity (n=5). (M) mtROS was detected by MitoSOX staining in HCECs, and was visualized by confocal microscopy. (N) Quantitative analysis of MitoSOX fluorescence intensity (n=5). Data are presented as the mean ± SD. ns, not significant; * P < 0.05, ** P < 0.01, *** P < 0.001 versus Ctrl group; # P < 0.05, ## P < 0.01, ### P < 0.001 versus HOP group; @@ P < 0.01, @@@ P < 0.001versus 0μM group; andandand P < 0.001 versus swollen of Ctrl group.

To investigate the mechanism by which HKL preserves mitochondrial morphology, the expression of the fusion protein OPA1 and the fission protein DRP1 was examined (n=5). As shown in Figure 3C, HOP treatment downregulated OPA1 and upregulated DRP1, indicating an imbalance favoring excessive fission. HKL reversed these changes, restoring OPA1 expression and suppressing DRP1 expression (Figure 3E and F). These results suggest that HKL restores mitochondrial dynamics balance by upregulating fusion and downregulating fission proteins, thereby counteracting HOP-induced mitochondrial damage.

Morphological analysis (n=10) of mitochondria revealed a marked reduction in fragmentation and swelling in the HOP+HKL group, accompanied by partial restoration of filamentous mitochondrial structures (Figure 3D). Quantitative assessment further confirmed that mitochondrial swelling and fragmentation were significantly diminished in the HOP+HKL group compared to the HOP group (Figure 3G, n=5), with concomitant increases in mean branch length and mean perimeter (Figure 3H and I) and a decrease in mean branch diameter (Figure 3J). These findings demonstrate that HKL effectively preserves mitochondrial morphology against disruption caused by HOP.

To investigate whether HKL exerts protective effects through antioxidant mechanisms, intracellular total ROS and mitochondrial superoxide (mtROS) were assessed using DCFH-DA and MitoSOX Red, respectively. The HOP group showed increased green fluorescence (total ROS) and red fluorescence (mtROS), indicating HOP-induced accumulation of both total and mitochondrial ROS (Figure 3K and M, n=5). Quantitative analysis confirmed significant elevations in both ROS levels compared to controls (Figure 3L and N, n=5). HKL treatment markedly reduced both fluorescence signals, demonstrating effective suppression of oxidative stress, particularly mtROS generation.

In conclusion, HKL alleviates hyperosmolar-induced oxidative damage, restores mitochondrial dynamics balance, preserves mitochondrial network integrity, and enhances cell viability in HCECs, suggesting its cytoprotective effects are primarily mediated through antioxidant activity targeting mitochondria.

PINK1 Knockdown Attenuates HKL-Induced Protection Against HOP-Induced Mitophagy and Mitochondrial Dysfunction in HCECs

To evaluate the role of PINK1 in HKL-mediated mitochondrial protection, PINK1 was knocked down in HCECs using siRNA. Knockdown efficiency was confirmed by Western blot and qPCR. Compared with the negative control (NC) group, PINK1 siRNA3 significantly reduced PINK1 protein levels (Figure 4A and C) and mRNA expression to less than 30% of control levels (Figure 4B), confirming effective knockdown for subsequent experiments.

Figure 4.

PINK1 knockdown attenuates HKL-induced protection against HOP-induced mitophagy and mitochondrial dysfunction in HCECs. (A) Western blot validation of PINK1 knockdown efficiency using PINK1 siRNAs. (B) qRT-PCR analysis of PINK1 mRNA levels. (C) Quantification of PINK1 protein levels. HCECs were transfected with negative control (NC) or PINK1 siRNA (siPINK1) for 24 h, followed by HKL (10μM) treatment 2h and then hyperosmolar stimulation for 24h. (D) Western blot analysis of mitophagy-related proteins (PINK1, Parkin, LC3, p62, Tom20) in HCECs under indicated conditions. (E–I) Quantification of relative protein levels (PINK1, Parkin, LC3, p62, Tom20) (n=5). Representative immunofluorescence images (J) and colocalization analysis (K–N) of LC3 (green) and Tom20 (red) to label mitophagosomes, reflecting mitophagys. The white line in (J) denotes the X‑axis distance adopted for colocalization analysis in (K-N). (O) Representative JC-1 staining images showing mitochondrial membrane potential (ΔΨm). (P) Quantitative analysis of JC-1 red/green fluorescence ratio (n=5). Data are presented as the mean ± SD. ns, not significant; * P < 0.05, *** P < 0.001 versus NC group; ## P < 0.01, ### P < 0.001 versus NC HOP+HKL group.

The involvement of PINK1 in HKL-regulated mitophagy was assessed by examining key mitophagy-related proteins. As shown in Figure 4D, HOP treatment for 24 h upregulated PINK1, Parkin, and the LC3 II/LC3 I ratio (Figure 4E–H, n=5), consistent with Figure 2. HKL further enhanced these changes and decreased p62 levels, indicating increased mitophagic flux. However, following PINK1 knockdown, the HKL-induced upregulation of Parkin and LC3 II/LC3 I was significantly attenuated (Figure 4F and G, n=5), p62 and Tom20 expression was restored (Figure 4H and I, n=5). These results indicate that HKL promotes mitophagy in a PINK1-dependent manner, and that PINK1 silencing abolishes this effect.

Mitophagy activation was further assessed by examining the colocalization of LC3, an autophagosome marker, with Tom20, a marker of the outer mitochondrial membrane. As shown in Figure 4J, minimal colocalization between LC3 (green) and Tom20 (red) was observed in the NC group. Hyperosmolar treatment markedly increased the number of colocalized puncta, indicative of autophagosomal engulfment of damaged mitochondria. HKL treatment further intensified this colocalization signal, supporting its role in promoting mitophagic flux. However, following PINK1 knockdown, the colocalization between LC3 and Tom20 remained weak despite HKL administration. Quantitative fluorescence analysis revealed a reduced overlap of red and green intensity profiles in the siPINK1 HOP+HKL group relative to the HKL-treated group. Collectively, these organelle-level data substantiate the critical role of PINK1 in mediating HKL-induced mitophagy.

The loss of mitochondrial membrane potential (ΔΨm) serves as a critical marker of mitochondrial dysfunction. JC-1 staining was used to evaluate ΔΨm across groups. As shown in Figure 4O, NC cells displayed strong red fluorescence (JC-1 aggregates) with a high red/green ratio, indicating normal ΔΨm. Following HOP exposure, green fluorescence (JC-1 monomers) increased and the red/green ratio decreased, indicating depolarization (Figure 4P, n=5). HKL treatment reversed these changes, restoring red fluorescence and increasing the red/green ratio, suggesting protection of mitochondrial function. However, this effect was markedly attenuated following PINK1 knockdown: the siPINK1 HOP+HKL group showed increased green fluorescence and a reduced red/green ratio compared with the HOP+HKL group, similar to the HOP group (Figure 4P). These findings indicate that PINK1 deficiency impairs HKL-mediated preservation of ΔΨm, highlighting the critical role of PINK1 in mitochondrial protection.

Figure 4 demonstrates that PINK1 knockdown significantly compromises HKL-induced mitophagy activation, as reflected by reduced protein expression and LC3/Tom20 colocalization, as well as impaired mitochondrial functional protection. Collectively, these results confirm that HKL promotes mitophagy at least partially through a PINK1-dependent pathway, thereby protecting HCECs from HOP-induced mitochondrial damage.

PINK1 Knockdown Attenuates HKL-Induced Protection Against HOP-Induced Apoptosis and Inflammation in HCECs

Having established the pivotal role of PINK1 in HKL-regulated mitophagy, we next examined the downstream effects on cell fate and inflammation. Apoptosis was assessed by flow cytometry with Annexin V/PI staining. The NC group showed few apoptotic cells (Figure 5A). HOP treatment increased early and late apoptosis, raising the total apoptosis rate to 36.2%±0.8%, confirming induction of apoptosis (Figure 5B). HKL significantly reduced this rate to 17.4%±0.3%. However, this anti-apoptotic effect was markedly diminished in the siPINK1 HOP+HKL group, where the apoptosis rate rebounded to 27.7%±1.6%, indicating that PINK1 deficiency impairs HKL-mediated cytoprotection (Figure 5B, n=5).

Figure 5.

PINK1 knockdown attenuates HKL-induced protection against HOP-induced apoptosis and inflammation in HCECs. HCECs were transfected with NC or siPINK1 for 24 h, followed by HKL (10μM) treatment 2h and then hyperosmolar stimulation for 24h. (A) Representative flow cytometry images of Annexin V/PI staining. (B) Quantitative analysis of total apoptosis rate (n=5). (C–E) qRT-PCR analysis of IL-1β, IL-6, and TNF-α mRNA levels (n=5). (F) Western blot analysis of apoptosis-related proteins (Bax, Bcl-2, cleaved caspase-3). (G–I) Densitometric quantification of Bax, Bcl-2, and cleaved caspase-3 (n=5). Data are presented as the mean ± SD. ns, not significant; * P < 0.05, ** P < 0.01,*** P < 0.001 versus NC group; # P < 0.05, ### P < 0.001 versus NC HOP+HKL group.

At the molecular level, HOP treatment upregulated the pro-apoptotic protein Bax, downregulated the anti-apoptotic protein Bcl-2, and increased cleaved caspase-3 levels (Figure 5F–I, n=5). HKL reversed these changes, effects that were largely abrogated by PINK1 knockdown, aligning with the flow cytometry results.

Given the association between corneal epithelial damage and inflammation, we examined whether HKL modulates inflammation through PINK1. HOP significantly increased IL-1β, IL-6, and TNF-α mRNA levels (3- to 5-fold) (Figure 5C–E, n=5). HKL suppressed these increases, but PINK1 knockdown markedly attenuated this anti-inflammatory effect, with cytokine levels comparable to those in the HOP group. Together, these results demonstrate that PINK1 is essential for HKL’s anti-inflammatory action and that HKL mitigates HOP-induced inflammation through a PINK1-dependent mitophagy pathway.

HKL Alleviates BAC-Induced Corneal Damage and Mitochondrial Dysfunction and Enhances PINK1-Mediated Mitophagy in a Mouse Dry Eye Model

To evaluate the therapeutic efficacy of HKL in dry eye-associated corneal damage and its underlying mechanisms in vivo, a BAC-induced mouse dry eye model was established and treated with HKL. As shown in Figure 6A, the BAC+HKL group exhibited markedly reduced fluorescein staining and improved corneal transparency compared with the BAC+PBS group. Quantitative analysis confirmed significantly lower CFS scores in the BAC+HKL group (Figure 6B, n=6). Tear secretion was also significantly increased in the BAC+HKL group (Figure 6C, n=6). H&E staining revealed restored epithelial thickness and orderly cellular arrangement in the BAC+HKL group, indicating preservation of corneal structural integrity (Figure 6D).

Figure 6.

HKL alleviates BAC-induced corneal damage and mitochondrial dysfunction and enhances PINK1-mediated mitophagy in a mouse dry eye model. Mice are treated with PBS, 0.2% BAC+PBS or 0.2% BAC+HKL (100μM) at day 7. (A and B) Representative corneal fluorescein staining images and quantitative analysis of CFS scores (n=6). (C) Tear secretion measured by Schirmer I test (n=6). (D) H&E staining of corneal sections. (E) Representative immunofluorescence images of SIRT3 (red) in corneal epithelium. (F) Representative immunofluorescence images of PINK1 (green) in corneal epithelium. (G) Western blot analysis of SIRT3, mitochondrial dynamics-related proteins (OPA1, DRP1), and mitophagy-related proteins (PINK1, Parkin, LC3, p62, Tom20) in cornea. Quantitative analysis of SIRT3 (H) and PINK1 (I) immunofluorescence intensity (n=5). (J-Q) Densitometric quantification of OPA1, DRP1, PINK1, Parkin, SIRT3, p62, LC3, and Tom20 (n=5). Data are presented as the mean ± SD. ns, not significant; * P < 0.05, ** P < 0.01,*** P < 0.001 versus PBS group; # P < 0.05, ## P < 0.01, ### P < 0.001 versus PBS+BAC group.

Based on in vitro findings of SIRT3 downregulation and the critical role of PINK1, immunofluorescence staining was performed to assess SIRT3 and PINK1 expression in corneal tissues. Strong SIRT3 (red) fluorescence was observed in the control corneal epithelium, which was markedly reduced in the BAC+PBS group. The BAC+HKL group showed partial restoration of SIRT3 fluorescence, with significantly higher mean intensity than the BAC+PBS group (Figure 6E and H, n=5). PINK1 (green) fluorescence was weak in the PBS group, increased in the BAC+PBS group, and further enhanced in the BAC+HKL group, indicating that HKL upregulates PINK1 expression and potentiates mitophagy signaling in vivo (Figure 6F and I, n=5).

Western blot analysis (n=5) showed that SIRT3 protein expression was significantly reduced in the BAC+PBS group, with a non-significant upward trend in the BAC+HKL group (Figure 6G and N). BAC+PBS group exhibited OPA1 downregulation and DRP1 upregulation, indicating excessive mitochondrial fission (Figure 6J and K). Mitophagy markers (PINK1, Parkin, LC3 II/LC3 I) were elevated in the BAC+PBS group compared with the PBS group, and further increased in the BAC+HKL group, accompanied by downregulation of p62 and Tom20, as well as restoration of OPA1 and reduction of DRP1 (Figures 6L–Q). These findings indicate that HKL promotes clearance of damaged mitochondria and protects corneal epithelial cells by enhancing PINK1-mediated mitophagy and modulating mitochondrial dynamics.

Overall, Figure 6 demonstrates that HKL mitigates corneal epithelial damage, improves tear secretion, and preserves corneal tissue structure in a BAC-induced dry eye model. These effects are associated with upregulation of PINK1-mediated mitophagy and regulation of mitochondrial dynamics, supporting that HKL protects corneal epithelial cells from hyperosmolar-induced mitochondrial dysfunction and apoptosis via the PINK1-mediated mitophagy pathway, offering a promising therapeutic strategy for dry eye disease. In summary, HKL protects corneal epithelial cells from dry eye-induced injury by activating SIRT3/PINK1-mediated mitophagy (Figure 7, Created with BioGDP.com43).

Figure 7.

Schematic diagram illustrating the protective mechanism of Honokiol against dry eye-induced corneal epithelial damage. BAC/HOP induces SIRT3 downregulation, leading to mitochondrial dynamics imbalance, mitochondrial fragmentation and dysfunction. This triggers compensatory PINK1-mediated mitophagy, which is insufficient to clear all damaged mitochondria, resulting in apoptosis, inflammation, and corneal epithelial damage. HKL treatment partially restores SIRT3, enhances PINK1-dependent mitophagic flux, and restores mitochondrial dynamics balance, promoting clearance of damaged mitochondria and recovery of mitochondrial function. This inhibits apoptosis and inflammation, ultimately protecting corneal epithelial integrity. PINK1 knockdown abolishes these protective effects, confirming its essential role in HKL-mediated protection. Red arrows indicate inhibitory effects, green arrows indicate activating effects, and black curved arrows indicate the developmental process. Created with BioGDP.com.

Discussion

In the present investigation, we successfully utilized a BAC-induced murine model alongside a HOP in vitro model to explore the critical role of mitochondrial quality control in the pathogenesis of DED. Both BAC exposure and HOP consistently compromised corneal epithelial barrier integrity and induced a significant downregulation of SIRT3, which uniquely coincided with the concurrent activation of PINK1/Parkin-mediated mitophagy. While recent investigations have begun to implicate SIRT3 in DED pathogenesis—primarily through FOXO1-mediated classical autophagy and the mtROS-TXNIP-NLRP3 signaling axis44–46—its direct involvement in the specific clearance of damaged mitochondria remained largely unexplored. Our findings provide the initial macroscopic observation that SIRT3 downregulation occurs in tandem with canonical mitophagy activation in the corneal epithelium, thereby expanding the functional scope of SIRT3 from general cellular autophagy to the highly specialized domain of mitochondrial network homeostasis.

A pivotal finding of our study is the multi-phasic and highly temporally dynamic nature of the mitophagic response under HOP. Through detailed time-course analyses in HCECs, we observed strikingly asynchronous expression patterns among key mitophagy-related proteins. The early upregulation of PINK1 at 6h, followed by a subsequent decline at 24h, strongly contrasts with the progressive accumulation of Parkin and the delayed increase in the LC3 II/LC3 I ratio at 24h. This temporal discrepancy is not merely an observational artifact but likely reflects the sequential biological progression of mitophagy, encompassing initial damage recognition, subsequent signal amplification, and ultimate autophagic degradation.47 Furthermore, our in vivo data revealed a lack of significant alterations in the expression levels of p62, Tom20, or Tim23 despite robust PINK1/Parkin upregulation. The stability of p62, an autophagy substrate typically degraded upon flux activation,48 suggests that BAC-induced mitophagy in the living organism may remain arrested at an early initiation stage without achieving complete autophagic flux. This phenomenon could be attributed to complex regulatory negative feedback loops or compensatory mechanisms operating within the intricate in vivo microenvironment.

Crucially, this early activation of mitophagy appears to function as a transient, compensatory adaptive mechanism rather than a sustained protective response. Our Seahorse metabolic analysis provided functional evidence for this biphasic response: while mitochondrial respiration was adequately maintained after 6h of HOP, essential functional parameters—including basal respiration, maximal respiratory capacity, spare respiratory capacity, and ATP production—exhibited a distinct declining trajectory by 24h. This metabolic exhaustion indicates that under prolonged HOP, the continuous accumulation of mitochondrial damage eventually overwhelms the innate cellular mitophagic capacity. Consequently, this overwhelmed state results in pronounced mitochondrial fragmentation, bioenergetic failure, and the structural deterioration of the corneal epithelium observed at later time points.

These temporal and functional dynamics established a robust mechanistic rationale for introducing HKL as a targeted therapeutic intervention aimed at bolstering mitochondrial quality control. Our systematic evaluation demonstrates that HKL preserves mitochondrial network homeostasis through highly multifaceted mechanisms. Primarily, HKL effectively counteracts the HOP-induced imbalance in mitochondrial dynamics by restoring the expression of the inner membrane fusion protein OPA1 while simultaneously suppressing the fission mediator DRP1.49–51 The coordinated action of these GTPases is essential; OPA1 facilitates inner membrane fusion and cristae remodeling, whereas DRP1 drives mitochondrial division. By pharmacological reestablishing this fusion-fission equilibrium, HKL prevents the progressive fragmentation of the mitochondrial network, thereby minimizing the pool of severely damaged mitochondria that require autophagic removal and enhancing overall mitochondrial quality control.

Beyond dynamic remodeling, HKL’s most fundamental protective mechanism lies in its ability to enhance the specific clearance of dysfunctional mitochondria via the PINK1 pathway. PINK1 functions as the master cellular sensor for mitochondrial depolarization.47 Our functional validation experiments revealed that following PINK1 knockdown, the HKL-induced upregulation of Parkin and LC3 II, as well as the degradation of p62 and Tom20, were significantly attenuated. Moreover, the subsequent failure of HKL to preserve mitochondrial membrane potential in the absence of PINK1 confirms that its cytoprotective efficacy is fundamentally dependent on an intact PINK1-mediated autophagic flux.52 This bidirectional linkage highlights a positive feedback loop where HKL enhances autophagic clearance to restore membrane potential, an effect that is completely uncoupled when PINK1 is silenced.

The preservation of mitochondrial integrity by HKL directly dictates the cellular fate by profoundly modulating downstream apoptotic and inflammatory signaling cascades. Dysfunctional mitochondria serve as the central hub for intrinsic apoptosis; increased outer membrane permeability leads to the pathological release of pro-apoptotic factors, such as cytochrome c, into the cytosol.52,53 By clearing these compromised organelles, HKL stabilizes the Bcl-2/Bax ratio and successfully interrupts the downstream caspase-3 apoptotic cascade.54,55 Concurrently, the efficient mitophagic clearance of damaged mitochondria drastically reduces the generation of mtROS and prevents the cytoplasmic leakage of damage-associated molecular patterns (DAMPs), such as oxidized mitochondrial DNA.56 Because these specific DAMPs are potent endogenous triggers for the NLRP3 inflammasome and subsequent IL-1β maturation 63, their proactive clearance provides a compelling mechanistic explanation for the broad suppression of multiple inflammatory cytokines (IL-1β, IL-6, TNF-α) observed following HKL treatment. Ultimately, the attenuation of local corneal inflammation by HKL likely disrupts the vicious cycle of ocular surface damage, facilitating the macroscopic clinical improvements observed in our in vivo model, including the restoration of normal epithelial structure and the enhancement of tear secretion.57

In this study, SIRT3 detection results differed between in vivo and in vitro assays. Immunofluorescence showed partial restoration of SIRT3 signal in the corneal epithelium after HKL treatment, whereas Western blot of whole corneal homogenates showed only a non-significant recovery trend. This discrepancy may firstly be due to methodological sensitivity, as immunofluorescence specifically detects the epithelial layer while Western blot captures mixed signals from all corneal layers. If HKL primarily modulates SIRT3 in the epithelium, the signal may be diluted by stromal and endothelial contributions in Western blot, obscuring statistical significance. Secondly, the distinction between regulation of protein activity versus expression should be considered. Previous studies have demonstrated that HKL can enhance SIRT3 enzymatic activity via direct binding,6 independent of changes in protein expression levels. Consequently, the lack of significant changes in SIRT3 expression by Western blot is consistent with a mechanism centered on SIRT3 activation, while immunofluorescence signal recovery may reflect enhanced protein stability or conformational changes. These complementary approaches collectively support HKL-mediated regulation of SIRT3 and PINK1/Parkin axis. In vivo, HKL upregulated PINK1 fluorescence and protein expression, increased Parkin and the LC3 II/LC3 I ratio, and showed downward trends in p62 and Tom20, confirming the cross-model reproducibility of our in vitro findings.

An important observation in this study that merits further investigation is that, although HKL significantly enhanced PINK1-mediated mitophagy and prior research has established the downregulation of SIRT3 under hyperosmolar stress, HKL treatment did not consistently increase SIRT3 protein expression levels in either HCECs (Figure S1) or mouse corneal tissues (Figure 6G and N). This finding contrasts with previous studies reporting that HKL upregulates SIRT3 transcription via PGC-1α-dependent positive feedback mechanism,7 suggesting that HKL’s mode of action in corneal cells may be cell type-specific or dependent on the cellular stress context. While our extensive data establish a clear protective role for HKL via PINK1-mediated mitophagy, the specific regulatory relationship among HKL, SIRT3, and PINK1 requires nuanced interpretation. Given that HKL is known to directly enhance SIRT3 enzymatic activity independent of changes in protein abundance, we hypothesize that HKL may predominantly modulate the SIRT3/PINK1 axis at the level of enzymatic activation rather than transcriptional upregulation. However, it is explicitly acknowledged that the proposed direct link between SIRT3 and PINK1 remains largely inferential in this study. Because specific evaluations of SIRT3 enzymatic activity and functional validations of the HKL-SIRT3 physical interaction were not performed, this signaling axis represents an intriguing working hypothesis. The data presented in Figure 5 indirectly support this interpretation: if HKL had no regulatory effect on SIRT3, the marked phenotypic reversal observed following PINK1 knockdown would be unlikely. It is precisely because HKL activates SIRT3 to promote PINK1-mediated mitophagy that PINK1 knockdown serves as a critical interruption point in this signaling pathway. Thus, the findings in Figure 5 substantiate the functional linkage of the SIRT3/PINK1 axis, even if this connection is primarily mediated through modulation of SIRT3 activity. The marked phenotypic reversal observed following PINK1 knockdown provides compelling functional evidence supporting this crosstalk, laying a solid foundation for future targeted biochemical investigations. Several potential explanations may underlie the lack of HKL-induced upregulation of SIRT3 in mouse corneal tissues and HCECs. First, the transcriptional upregulation of SIRT3 via PGC-1α induced by HKL may require longer treatment durations or specific cellular conditions. Second, the expression levels or activity of PGC-1α may be inherently low in HCECs, thereby limiting activation of this positive feedback loop. Third, HOP itself may suppress PGC-1α activity or function through certain mechanisms, counteracting the upregulatory effects of HKL. These hypotheses warrant further exploration in future research.

DED is a prevalent ocular surface disorder, with current therapeutic approaches predominantly centered on the use of artificial tears and anti-inflammatory agents.19,58 These treatments, however, do not specifically target the underlying pathological mechanisms associated with mitochondrial dysfunction. HKL, a natural compound, demonstrates a favorable safety profile alongside well-defined mitochondrial protective properties, thereby representing a promising novel therapeutic candidate for dry eye management. This study presents several lines of evidence supporting the clinical potential of HKL: (1) a proposed mechanism whereby HKL confers protection via the inferred SIRT3/PINK1-mediated mitophagy pathway; (2) demonstrated in vivo efficacy, as HKL ameliorates multiple key parameters in a BAC-induced dry eye model; and (3) practical administration through topical ophthalmic formulations, which minimizes systemic adverse effects.

The innovative contributions of this research are highlighted by the following: (1) the first establishment of a mechanistic link between SIRT3 and PINK1-dependent mitophagy within a dry eye model; (2) the initial report that HKL safeguards corneal epithelial cells through PINK1-mediated mitophagy; (3) the proposal of a regulatory model in which HKL activates SIRT3 via direct binding rather than by upregulating its expression, offering novel insights into the action mechanisms of natural compounds; and (4) the construction of a comprehensive evidence framework spanning molecular mechanisms to in vivo therapeutic efficacy, thereby providing robust experimental support for HKL as a candidate treatment for DED.

This study has certain limitations. First, the regulatory relationship between SIRT3 and PINK1 remains unclear and requires further assessment of PINK1 acetylation, SIRT3 activity, and HKL-SIRT3 binding. Second, the effect of HKL on lacrimal gland function has not been directly evaluated, and whether the improvement in tear secretion is due to restoration of the corneal-lacrimal gland reflex arc remains to be determined. Finally, a methodological limitation is the absence of an a priori sample size calculation. Although our sample sizes were guided by established literature protocols, certain in vitro analyses, such as the Seahorse metabolic assays (n=3 biological replicates), were conducted with relatively small sample sizes. While these metabolic findings are strongly corroborated by our orthogonal assessments of mitochondrial function (eg, ROS production and apoptosis), the small sample size may limit statistical power. Future studies utilizing larger cohorts and predefined power analyses will be beneficial to further validate these parameters.

Conclusion

This study demonstrates that HKL protects against dry eye-associated corneal epithelial damage via PINK1-mediated mitophagy, functioning through an inferred SIRT3/PINK1 regulatory axis. In both BAC-induced dry eye mice and hyperosmolar-stressed human corneal epithelial cells, SIRT3 was downregulated and mitophagy activated, leading to mitochondrial dysfunction, apoptosis, and inflammation. HKL alleviated oxidative stress, restored mitochondrial dynamics, and enhanced PINK1-dependent mitophagy. PINK1 knockdown confirmed its essential role in HKL’s protection. Notably, we hypothesize that HKL may primarily enhance SIRT3 activity rather than its expression; however, as SIRT3 activity was not directly measured, this proposed mechanistic link requires future direct experimental validation. In vivo, HKL improved corneal integrity and tear secretion, consistent with in vitro findings. Collectively, this study highlights the promising therapeutic potential of HKL in treating DED.

Funding Statement

Supported by grants from the National Natural Science Foundation of China (82171024), the Open Project of Institute of Optometry and Vision Science in Nankai University (NKSGY202401), The Science & Technology Development Fund of Tianjin Education Commission for Higher Education (2025ZXZD023), Tianjin Key Medical Discipline Construction (TJYXZDXK-3-004A-3) and Tianjin Key Medical Discipline (Specialty) Construction Project (TJYXZDXK-016A).

Abbreviations

BAC, benzalkonium chloride; BAX, BCL-2-associated X protein; Bcl-2, B-cell lymphoma 2; CCK8, cell counting kit-8; DED, dry eye disease; DMSO, dimethyl sulfoxide; DRP1, dynamin-related protein 1; HCECs, human corneal epithelial cells; HKL, Honokiol; HOP, high osmotic pressure; H&E staining, hematoxylin and eosin staining; LC3, microtubule-associated protein 1 light chain 3; mtDNA, mitochondrial DNA; mtROS, mitochondrial superoxide; NC, negative control; OCR, oxygen consumption rate; OPA1, optic atrophy 1; PINK1, PTEN-induced putative kinase 1; p62, Sequestosome-1; ROS, reactive oxygen species; SIRT3, sirtuin 3; TFAM, mitochondrial transcription factor A; TIM23, translocase of inner mitochondrial membrane 23 homolog; TNF-α, Tumor Necrosis Factor-alpha; TOM20, translocase of outer mitochondrial membrane 20; ΔΨm, mitochondrial membrane potential.

Data Sharing Statement

The published article includes all datasets generated or analyzed during this study.

Author Contributions

Yulei Huang: Conceptualization, Data Curation, Formal Analysis, Methodology, Writing – Original Draft, Writing – Review & Editing. Xindan Cao: Data Curation, Methodology, Writing – Original Draft. Xinlin Yan: Data Curation, Investigation, Methodology. Jiaqi Lin: Methodology, Validation. Zhengyu Cao: Data Curation, Software. Xiafei Chen: Validation. Jiahao Zhu: Investigation. Yi Guan: Visualization. Guozhen Chen: Validation. Xuan Li: Conceptualization, Data Curation, Funding Acquisition, Project Administration, Resources, Supervision, Writing – Review & Editing. All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.