Cell cycle arrest caused by p16 leads to harmful cell signals and scarring in Fuchs corneal disease

To investigate the potential contribution of lifelong stress exposure to cellular senescence in FECD, we conducted immunostaining and RT-PCR analyses on corneal specimens derived from FECD patients and compared them with healthy donor corneas. Brightfield microscopy revealed a distinct rosette-shaped morphology in FECD specimens, indicative of cellular stress, contrasting with the characteristic hexagonal morphology of normal corneal endothelium (Fig. 1a). Chronic stress in FECD is associated with progressive endothelial cell loss (ECL), leading to compromised intercellular junctions and barrier integrity, ultimately resulting in stromal edema and vision loss. Although ECL is a hallmark of FECD, the contribution of cellular senescence to disease pathogenesis remains underexplored in this aging disorder.

Molecular diagnosis, using senescence-associated beta-galactosidase (SA-β-Gal) staining, detected positive activity of beta-galactosidase (a lysosomal enzyme that becomes active in senescent cells) in FECD specimens, indicative of senescence (Fig. 1b). Immunostaining showed expression of early-senescent marker p21 (encoded by CDKN1A gene; Fig. 1c) and mature senescent marker, p16 (encoded by CDKN2A gene; Fig. 1d), which are cyclin dependent kinase inhibitors (CDKIs) and regulate cell cycle. Additionally, alpha-smooth muscle actin (α-SMA; encoded by ACTA2 gene), a fibrosis-associated cytoskeletal protein, was found to be dispersed in FECD specimens (Fig. 1e).

RT-PCR analysis revealed a significant upregulation of ZEB1 (encoding Zinc Finger E-box Binding Homeobox 1), as well as SNAI1 and SNAI2 (encoding the Snail family transcription factors Snail1 and Snail2), which are known repressors of E-cadherin and key inducers of endothelial-to-mesenchymal transition (EMT) in FECD. In addition, elevated expression of pro-fibrotic markers, including FN1 (encoding Fibronectin-1), which contributes to tissue fibrosis and scarring, was detected in FECD specimens. Components of the TGF-β signalling pathway, including TGFBIp (encoding TGF-β-induced protein) and TGFβ1, were also upregulated, along with extracellular matrix (ECM) components such as COL8A2 (encoding the α2 chain of type VIII collagen) and COL4A1 (encoding the α1 chain of type IV collagen). These findings suggest enhanced ECM deposition associated with guttae formation, a clinical hallmark of FECD. Additionally, CLU (encoding clusterin), a molecular marker linked to guttae lesions in the corneal endothelium, exhibited increased expression (Fig. 1f). These findings highlight the contribution of cellular senescence and fibrosis to FECD pathogenesis because of chronic, lifelong exposure to stress, while the multifactorial etiology of the disease facilitates its progression.

Environmental, hormonal, and age-related factors are known contributors to FECD pathogenesis; however, the role of combined stressors in disease progression remains poorly understood. To address this, we investigated the effects of individual and combined stressors on early pathogenic events and chronic stress-driven disease mechanisms.

Initially, the impact of individual stressors on senescence was assessed by subjecting HCEnC-21T cells to a single acute exposure of 4-hydroxyestradiol (4OHE₂, 20 µM) or UVA irradiation (25 J/cm²). Both treatments independently induced cellular senescence, as indicated by increased p21 expression and SA-β-Gal positivity (Supp. Fig. 1). To evaluate the consequences of concurrent stress exposure, we then treated HCEnC-21T cells with a combination of UVA and 4OHE2 under acute stress (AS) and chronic stress (CS) conditions (Fig. 2a).

Brightfield microscopy revealed distinct morphological changes following stress exposure. AS induced a rosette-like cell morphology (white arrows), while CS resulted in an enlarged, flattened cellular morphology characteristic of a senescent phenotype (yellow arrow; Fig. 2b). Gene expression analysis showed that AS significantly upregulated markers associated with early EMT (SNAI1), fibrosis (ACTA2), extracellular matrix remodeling (TGFBIp), and transient senescence (CDKN1A), suggesting early senescence induction (Fig. 2c). However, CS elicited a broader and more robust upregulation of SNAI2 (a marker of sustained EMT), ACTA2 (fibrosis marker), TGFBIp, TGFβ1 (EMT markers), and the mature senescence marker CDKN2A, indicative of a more permanent senescent and fibrotic state (Fig. 2d). Notably, SNAI2, TGFβ1, and CDKN2A were exclusively elevated under CS conditions compared to AS, suggesting a shift toward a stable senescence and sustained tissue remodeling (Fig. 2e).

To further explore the link between senescence and FECD pathogenesis, we assessed DNA damage and stress response pathways. Immunostaining revealed increased levels of γH2AX, a marker of DNA double-strand breaks, in both AS (88 ± 4%) and CS (72 ± 7%) groups, compared to untreated controls (2 ± 1%) (Fig. 2f), indicating that both acute and chronic stress induces significant genotoxic stress. Expression of Tp53, a key regulator of the DNA damage response, was elevated in AS (34 ± 9%) and CS (22 ± 9%) relative to controls (1 ± 0.5%), with a higher response observed in AS (Fig. 2g). Furthermore, expression of p21, a cyclin-dependent kinase inhibitor and key downstream effector of Tp53, was elevated in AS-treated cells (36 ± 6%) compared to CS-treated cells (14 ± 4%) and controls (1 ± 1%) (Fig. 2h), indicating Tp53–p21 signalling axis in acute stress.

Interestingly, immunostaining for p16, a canonical marker of irreversible and mature senescence, revealed a significantly higher increase following CS (65 ± 9%) compared to both untreated controls (0 ± 0%) and acute stress (AS; 3 ± 3%) (Fig. 2i). This suggests that sustained stress exposure is required to trigger activation of the p16-mediated senescence pathway. Retinoblastoma protein (RB) phosphorylation is essential for E2F transcription factor release and cell cycle progression. Analysis of phosphorylated RB (pRB) levels demonstrated a reduction of pRB in AS cells and a further decrease in CS cells relative to controls (Fig. 2j), indicating progressive inhibition of cell cycle activity. The observed reduction in pRB, coupled with elevated p16 expression, supports activation of the p16–pRB axis in mediating permanent cell cycle arrest in response to chronic stress. To confirm the senescent state, cells were stained for SA-β-Gal activity. CS-treated cells showed a significantly higher proportion of SA-β-Gal positivity (60 ± 8%) compared to AS (24 ± 4%) and untreated controls (2 ± 1%) (Fig. 2k), further corroborating the onset of a stable senescent phenotype under chronic stress conditions. In addition, α-SMA, a marker of myofibroblast differentiation and fibrosis, was markedly elevated in both AS (30 ± 5.5%) and CS (66 ± 5.4%) groups compared to controls (0.3 ± 0.5%) (Fig. 2l), with significantly higher levels observed in CS. These protein-level findings are consistent with the transcriptional upregulation of EMT- and fibrosis-related genes observed via RT-PCR (Fig. 2e).

Collectively, these results suggest that while acute combined stress triggers transient senescence via the Tp53–p21 pathway, chronic exposure promotes a shift to irreversible, p16–pRB-dependent senescence. This transition is accompanied by enhanced fibrotic remodeling, indicating a critical role for persistent stress in driving pathological features characteristic of FECD.

Cellular senescence is characterized by an irreversible cell cycle exit. While we previously demonstrated that acute UVA exposure induces early senescence in CEnCs via p21-dependent G2/M arrest [12], the impact of chronic stress induced by concurrent UVA and 4OHE₂ exposure on cell cycle regulation remains unclear. To investigate this, we analysed the DNA intensity of cells after temporal induction stress using flow cytometry. AS exposure induced a higher accumulation of cells in the G2/M phase (53 ± 14%) compared to untreated controls (26 ± 2%; p < 0.01). In contrast, CS exposure promoted a shift toward G0/G1 arrest (68 ± 2%) compared to control (60 ± 2%; p < 0.01) with a concomitant reduction in G2/M (20 ± 1%) compared to AS, reflecting progression into a stable, mature senescent state (Fig. 3a).

Cell cycle progression is regulated by various cyclins and CDKs. The G2/M transition is facilitated by the Cyclin B1/CDK1 complex, which promotes chromosome condensation and nuclear envelope breakdown to initiate mitosis. In AS-exposed cells, Cyclin B1 expression was significantly elevated (65 ± 7% vs. 19 ± 5% in untreated controls; p < 0.001) (Fig. 3b), indicating G2/M re-entry. Conversely, CS led to increased Cyclin D1 expression (65 ± 3% vs. 46 ± 2% in controls; p < 0.01) (Fig. 3c). Despite this upregulation, nuclear translocation of Cyclin D1 was impaired under CS, leading to reduced RB phosphorylation, thereby inhibiting cell proliferation and arresting cells in the G0/G1 phase. Consequently, CDK4/6 expression was downregulated, impeding G1-S transition and promoting a permanent arrest in the G0/G1 phase (Fig. 3d).

To investigate cell cycle-specific transcriptional changes, we sorted G0/G1 and G2/M phase cells using fluorescent activated cell sorting (FACS) method (Fig. 3e). Post-AS, G2/M cells showed increased expression of SNAI1 and CDKN1A genes, suggesting early EMT and senescence initiation (Fig. 3f). In contrast, CS induced a robust upregulation of CDKN2A in G0/G1 phase (2.9-fold; p < 0.001), indicative of p16-mediated cell cycle arrest (Fig. 3g). Additionally, while ACTA2 expression was predominantly observed in G2/M cells following AS, its expression shifted to G0/G1 phase after CS, indicating a transition toward a mature senescence-associated pro-fibrotic phenotype.

To further characterize the senescence-fibrosis phenotype, we performed double-sorting (FACS II) of G0/G1 cells into α-SMA+ve and p16+ve subpopulations (Fig. 3e). FACS II analysis revealed higher proportions of α-SMA+ (94 ± 5%) and p16+ (93 ± 6%) cells within the G0/G1 populations (Fig. 3h), suggesting a strong correlation between mature senescence (p16) and fibrosis (α-SMA). Gene expression analysis using RT-PCR showed upregulation of FN1, TGFBIp, and CLU markers in α-SMA+ cells (Fig. 3i), while p16+ cells exclusively upregulated COL4A1, a component of Descemet’s membrane, suggesting that mature senescent cells may contribute to guttae formation via excessive collagen deposition (Fig. 3j).

To assess a direct correlation between p16 and α-SMA in FECD pathogenesis, we then co-stained FECD patient-derived specimens and observed their co-localization, providing in situ evidence for the senescence-fibrosis association (Fig. 3k). Furthermore, both p16+ and α-SMA+ sub-sorted cells exhibited upregulation of TGFβ1 and SNAI2, supporting a mechanism in which senescence-associated, EMT-driven fibrosis during G0/G1 arrest contributes to ECM remodelling driving FECD pathogenesis (Fig. 3l).

Senescent cells are known to secrete SASP factors that can influence healthy neighbouring cells via paracrine signalling. In the human eye, persistent secretion of SASP by mature senescent CEnCs may accumulate in the anterior chamber and activate transcriptional regulators such as NF-kB, promoting corneal endothelial dysfunction. Although pro-inflammatory cytokines and chemokines such as IL-6 and IL-8 have been detected in the aqueous humor of FECD patients, the precise role of sustained SASP in disease onset and progression remains poorly understood due to lack of robust in vitro models. To address this, we collected conditioned media (SASP) from acutely and chronically stressed CEnCs and evaluated its effects on HCEnC-21T cells and healthy donor corneas (Table 1) to delineate the functional impact of SASP in FECD progression (Fig. 4a).

Exposure of HCEnC-21T cells to acute SASP induced an EMT-like- fibroblastic phenotype, whereas chronic SASP exposure resulted in enlarged, senescent-like cells characteristic of mature senescence (Fig. 4b). SA-β-Gal staining demonstrated a significantly higher proportion of senescent cells in cultures exposed to chronic SASP (14 ± 3%) compared to acute SASP (2 ± 1.5%) and controls (1 ± 0.6%; p < 0.01) (Fig. 4c). Additionally, chronic SASP increased the proportion of cells arrested in the G0/G1 phase (60 ± 3%) relative to acute SASP (50 ± 1%) and untreated controls (50 ± 9%), consistent with the onset of cell cycle arrest associated with mature senescence (Fig. 4d).

To validate these findings in human tissue, we exposed normal cadaveric donor corneas to SASP-conditioned media. Chronic SASP exposure significantly increased Sa-β-Gal positive cells (20 ± 7%) compared to acute SASP (2 ± 2%) and control tissues (0.2 ± 0.4%; p < 0.01) (Fig. 4e), supporting the in vitro findings of senescence induced by chronic SASP exposure.

Gene expression profiling of SASP-exposed tissues revealed increased expression of EMT marker (SNAI2), along with ACTA2, TGFBIp, TGFβ1, and CDKN1A genes following acute SASP treatment (Fig. 4f). In contrast, chronic SASP induced elevated levels of SNAI2, ACTA2, TGFBIp, COL4A1, CDKN1A, and the mature senescence marker (CDKN2A) (Fig. 4g). Notably, chronic SASP exposure resulted in a more pronounced upregulation of fibrotic (ACTA2), ECM (TGFBIp, COL8A2, COL4A1), and senescence-related (CDKN1A and CDKN2A) gene expression compared to acute SASP (Fig. 4h), suggesting a progressive shift from a pro-EMT to a pro-ECM-altering state.

To identify specific SASP components responsible for these effects, we profiled cytokines and chemokines present in the conditioned media using a multiplex assay. Acute SASP was enriched in pro-inflammatory factors including IL-6, MIP-1B and IFN-γ, along with elevated levels of TGFBIp (detected using ELISA), a known ECM remodelling mediator. In addition to these markers, chronic SASP uniquely contained increased levels of pro-fibrotic cytokines (IL-17, MCP-1) and pro-senescent factors (IL-8, and TNF-α) compared to untreated controls (Fig. 4i; Table 2). These findings demonstrate that SASP derived from chronically stressed cells not only reinforce senescence but also promotes fibrotic responses in neighboring healthy cells through a self-sustaining feedback mechanism.

IL-17A (IL-17 ligand) promotes EMT by inducing TGFβ production [16] and activating fibroblasts, while IL-8, which binds to CXCR1/2 receptors, induces DNA damage–mediated senescence. Given the presence of these cytokines in chronic SASP and their known roles in inflammation and fibrosis, we investigated the therapeutic potential of targeting IL-17 and IL-8 pathways in FECD.

Gene expression analysis of FECD patient-derived specimens revealed a significant upregulation of IL-17A (209-fold, p < 0.001) and CXCR2 (23-fold, p < 0.01), compared to healthy controls (Fig. 5a), indicating activation of these cytokine pathways in FECD pathogenesis. Further, double sorting of α-SMA+ cells from the G0/G1 phase revealed an 870-fold increase in CXCR2 expression (p < 0.0001) (Fig. 5b). Similarly, G0/G1 sorted p16+ cells exhibited elevated expression of CXCR2 (86-fold; p < 0.01) and IL-17A (2092-fold; p < 0.0001) (Fig. 5c), suggesting that these pro-senescent and pro-fibrotic cytokine pathways are strongly enriched in senescent and fibrotic cell populations (Fig. 5d).

To evaluate the functional consequences of cytokine inhibition, HCEnC-21T cells were exposed to chronic SASP in the presence or absence of IL-17A and CXCR2 antagonists (Fig. 5e). Morphological assessment showed improved cellular integrity and reduced senescent morphology following IL-17A and CXCR2 blockade (Fig. 5f). This was accompanied by significant downregulation of key markers associated with fibrosis and senescence such as FN1 (Fig. 5g), ACTA2 (Fig. 5h), CDKN1A (Fig. 5i), and CDKN2A (Fig. 5j). SA-β-Gal staining confirmed a substantial reduction in the proportion of senescent cells after blocking IL-17A (5 ± 3%) and CXCR2 (3 ± 1%) compared to untreated chronic SASP-exposed cells 12 ± 4% (p < 0.05) (Fig. 5k). Moreover, the proportion of α-SMA+ve cells decreased significantly following treatment with IL-17A (21 ± 9%) and CXCR2 (23 ± 12%) antagonists relative to chronic SASP conditions (63 ± 7%) (p < 0.01) (Fig. 5l). These results demonstrate that IL-17 and IL-8 signaling are key drivers of senescence and fibrosis in FECD. Pharmacological blockade of IL-17A and CXCR2 not only restores cell homeostasis but also effectively mitigates the dual pathogenic features of FECD, senescence and fibrosis, highlighting their therapeutic potential.

Several therapies have targeted the selective clearance of senescent cells to alleviate SASP overload in the tissue. Our findings indicate that mature senescence (via p16) is a key driver of fibrosis in chronic stress-induced FECD. To evaluate the therapeutic potential of senescent cell clearance in FECD, we investigated the efficacy of a senolytic cocktail, Dasatinib and Quercetin (D + Q), in a UVA-induced FECD mouse model. Female mice were subjected to corneal UVA exposure 500 J/cm² UVA [3], followed by systemic administration of DQ on day 1 and at weeks 1, 2 and 3. Animals were sacrificed at week 4, and the proportion of senescence and fibrosis markers were assessed by whole-mount immunostaining and RT-PCR (Fig. 6a).

Corneal clarity depends on the transparency which is maintained by CEnC density. In vivo confocal imaging using Heidelberg Retinal Tomography (HRT) demonstrated that UVA disrupted the normal hexagonal CEnC mosaic, accompanied by progressive cell enlargement by week 4 (Fig. 6b). DQ treatment preserved the regular endothelial cell morphology and significantly restored endothelial cell density by 21% relative to UVA-only treated eyes at week 4 (p < 0.01). Confocal microscopy using tight junctional protein (ZO-1) showed extensive multinucleation and nuclear size heterogeneity post-UVA, both of which were reduced with DQ treatment (Fig. 6c), consistent with preserved endothelial structure observed by HRT images.

Next, to determine whether DQ-mediated clearance of senescent cells attenuates fibrosis, we performed whole-mount immunostaining of mice corneas. DQ treatment significantly reduced senescence-associated heterochromatic foci, visualized by the presence of di-methylation of Lys9 on histone H3 (H3K9me3) (10.1 ± 4.5%) compared to UVA irradiation (29.4 ± 2.5%; p < 0.01) (Fig. 6c). The proportion of p16+ve senescent cells was also significantly decreased after DQ treatment (18.7 ± 2.3%) compared to UVA irradiation (31.9 ± 4.5%; p < 0.01) (Fig. 6d). Notably, the number of α-SMA+ve cells (9 ± 1.4%) were significantly reduced by DQ treatment relative to UVA-only controls (14.4 ± 1.3%; p < 0.01) (Fig. 6e), demonstrating the anti-fibrotic effects of senescent cell elimination. Additionally, DQ treatment led to significant downregulation of senescent- (CDKN1A, CDKN2A, CXCR2), fibrotic- (ACTA2, and IL17A) and ECM-associated (COL4A1, TGFBIp, TGFβ1) markers, indicating a broad suppression of the fibrotic and senescent transcriptomic programs. Pharmacological clearance of p16+ve senescent cells using DQ mitigates UVA-induced endothelial senescence and fibrosis, thereby preserving corneal structure and function and holds a promise for slowing FECD progression.

The study followed the tenets of the Declaration of Helsinki and was approved by the institutional review board of the Massachusetts Eye and Ear, Boston, MA, USA (protocol code 2019P003394 and 09/11/2022 of approval). Normal human donor corneas from cadavers were procured from Eversight eye bank (Ann Arbor, MI) and the de-identified FECD specimens were obtained from Price Vision group (Indianapolis, IN, USA) after written consent from the donor’s next-of-kin or the patient for tissues to be used for research. The donor characteristics are listed in Table 3. Association for research in vision and ophthalmology (ARVO) guidelines for the use of animals in ocular research was followed in this study. C57BL/6 J wild-type female mice (7-15 weeks old) were obtained from Charles River (Boston, MA, USA) or were bred in the animal facility of Schepens Eye Research Institute, Mass Eye and Ear (Boston, MA, USA) and were not randomized for this study.

The cells and tissues exposed to the stressors or SASP were stained for SA-β-Gal using senescence histochemical staining kit (Sigma-Aldrich, Burlington, MA, USA) according to the manufacturer’s protocol. The images were acquired using EVOS XL core microscope (Invitrogen, Waltham, MA, USA) and percentage of SA-β-Gal positive cells were manually counted and reported as percentage of the total population.

After each stress, cells were collected using trypsin digestion, washed with PBS and fixed in 70% ethanol for 30 min on ice. After washing, the cells were treated with 100 µg/ml RNaseA and stained with 50 µg/ml propidium iodide (PI). Cell cycle data was acquired using BD LSR II flow cytometer (BD Biosciences, Temecula, CA, USA) with at least 10,000 events per sample. Single cells were gated by measuring forward and side scatter and cell doublets were excluded. Quantification was carried out using FlowJo cell cycle analysis (v10.6.2, FlowJo, LLC).

For Fluorescence-Activated Cell Sorting (FACS) based single cell sorting experiments, G0/G1 and G2/M phase cells were sorted using BD FACS Aria III (BD Biosciences, Temecula, CA, USA) and collected in PBS on ice. These cells were immediately centrifuged at 850 g for 10 min at 4 °C and resuspended in Trizol (Invitrogen) for RT-PCR analysis.

For double cell sorting, the cells obtained from CS were fixed in 70% ethanol and treated with 50 µg/ml RNaseA followed by PI staining (50 µg/ml). The G0/G1 population was sorted in PBS and immediately placed on ice. The sorted cells were immunostained (as per the immunostaining protocol given below) using α-SMA, p16 and p21 antibodies and their respective secondary antibodies (Supp. Table). The immunostained sorted cells were subjected to a second round of cell sorting with untreated samples as controls. The G0/G1 sorted cells were further sorted into α-SMA+ve and α-SMA-ve, p16+ve and p16-ve, and p21+ve and p21-ve populations separately. The sorted cells were immediately placed on ice, centrifuged and dissolved in Trizol for RT-PCR analysis, as explained in the next paragraph. 1

RNA was extracted from cultured CEnCs or tissues using Trizol or Trizol LS (Invitrogen) respectively, purified by washing in absolute propanol and 75% ethanol sequentially, and collected in nuclease free water to measure the RNA quality and quantity using NanoDrop (Thermofisher, Waltham, MA, USA). Complementary DNA (cDNA) was synthesized by iScript (Bio- Rad, Hercules, CA, USA) in a T100 Thermal cycler (Bio-Rad) at 25 °C for 5 min, 46 °C for 20 min, and 95 °C for 1 min. RT-PCR was performed using TaqMan gene expression assays (ThermoFisher) using Eppendorf Realplex2 epgradient S mastercycler (Eppendorf, Enfield, CT, USA). Results obtained were normalized to GAPDH internal control, and the relative gene expression was quantified as 2^{^ΔΔ(-CT)}.

CEnC-21T cells, human tissues, and whole mount mice corneas were fixed in 4% paraformaldehyde (PFA) overnight at 4 °C, permeabilized with 0.1% triton x-100 and blocked with 2% bovine serum albumin (BSA) in PBS. Primary antibodies (Supp. Table) were diluted in 1% BSA and incubated overnight at 4°C. The cells were incubated with secondary antibodies for 1-hour at RT. Washing was performed with PBS between each step, and the samples were mounted with vecta-shield (Vector Laboratories, Newark, CA, USA) containing DAPI. The images were acquired under Leica DMi8 microscope or Leica SP8 confocal microscope (Leica, Deerfield, IL, USA). Whole mount mice corneas were imaged using confocal microscopy with Z-stacks used to create a maximum projection image (DM6000S with LCS 1.3.1 software, Leica). 1

The SASP from AS and CS was collected as mentioned above and used to detect TGFBI (BIGH3) using ELISA method (Thermofisher, Waltham, MA, USA) as per the manufacturer’s protocol.

The SASP factors collected from AS and CS was detected using Bio-Plex Pro Human Cytokines 17-plex assay (Bio-Rad, Hercules, CA, USA) and the plate was read using a MAGPIX® (Luminex corporation, Austin, TX, USA) machine as recommended by the manufacturer.

The expression of IL17A and CXCR2 was assessed in the FECD patient specimens and subsequently in sorted α-SMA+ve and p16+ve cells from the G0/G1 phase (as described above) by RT-PCR analysis. HCEnC-21T immortalized cells were exposed to SASP obtained from chronic stress. Concurrently, healthy cells were treated with 10 µM secukinumab (interleukin-17A ligand inhibitor, Selleckchem, Houston, TX, USA), a monoclonal antibody, combined with chronic-SASP to neutralize the effects of IL17A. In a separate experiment, cells were exposed to a combination of 0.2 µM SB225002 (a CXCR2 antagonist, Medchem express, Monmouth Junction, NJ, USA) along with chronic SASP to evaluate the inhibitory effect of CXCR2 antagonist. After 48 h of culture with SASP and the respective IL-antagonists, cells underwent SA-β-Gal staining, α-SMA staining, and RT-PCR analysis.

In the in vivo study involving C57BL/6 J mice, the left eye was designated as the untreated control, while the right eye underwent UVA irradiation using LED light to deliver a dosage of 500 J/cm² on day 1, followed by 4 weeks of recovery period as described earlier [3]. Central corneal thickness (µm) and the endothelial cell density (cells/mm²) was measured before irradiation and at 1, 2 and 4 weeks post-UVA exposure [3]. A single dose of combined Dasatinib (1 mg/Kg) and Quercetin (10 mg/Kg) was administered intraperitoneally on days 1, 7, 14 and 21 after UVA exposure [79]. The vehicle used for the treatment consisted of 5% Tween-80 + 5% polyethylene glycol, 90% 1X PBS. Mouse eyeballs were enucleated, rinsed with PBS, and transported on ice for corneal excision on day 28. The corneas were either fixed in 4% PFA for immunostaining or stored in Trizol for subsequent RT-PCR analysis.

Statistical analysis was carried out using Graphpad Prism v10.0.0 (Graphpad Software Inc, CA, USA) using an unpaired, two-tailed, non-parametric test with Welch’s correction (95% CI), one-way analysis of variance (ANOVA) with Tukey’s post-hoc test, and two-way ANOVA with Sidak’s post-hoc test to determine statistical significance deemed at p < 0.05. All the data were analyzed by blinded investigators, which were confirmed by an unblinded investigator. All the data are reported as mean ± SEM.

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.