Cannabidiol inhibits TGF-β1-induced epithelial-mesenchymal transition in human conjunctival epithelial cells by interrupting TGF-β/Smad signaling

Abstract

Epithelial-mesenchymal transition (EMT) plays a significant role in conjunctival fibrosis-related pathologies and has emerged as a promising therapeutic target for managing conjunctival fibrosis. Cannabidiol (CBD), a predominant non-psychoactive cannabinoid derived from the cannabis plant, has demonstrated antifibrotic effects in various extraorbital tissues. However, its influence on fibrosis-associated EMT in conjunctiva remains unexplored. Given the ubiquitous expression of cannabinoid targets in ocular tissues, including the conjunctiva, and evidence suggesting that modulation of the endocannabinoid system ameliorates ocular pathologies, this study aimed to evaluate the effects of CBD on conjunctival EMT. Cultured human conjunctival epithelial cells were stimulated with transforming growth factor-beta 1 (TGF-β1) to induce EMT. CBD treatment effectively mitigated EMT-related changes induced by TGF-β1, including increased cell elongation and migration, reduced epithelial markers (E-cadherin and zonula occludens-1, and elevated mesenchymal markers (alpha-smooth muscle actin and fibronectin) and EMT-associated transcription factor Snail. Furthermore, CBD suppressed TGF-β1-mediated Smad-2/3 phosphorylation and nuclear translocation. Treatment with a specific TGF-β/Smad pathway inhibitor (SB431542) yielded comparable results, suggesting that the inhibitory effects of CBD on EMT involve disruption of TGF-β/Smad signaling. Additionally, the EMT phenotype was associated with increased interleukin-6 (IL-6) secretion, which was also attenuated by CBD treatment. This study confirms that CBD effectively prevents EMT and EMT-associated IL-6 secretion by targeting TGF-β/Smad signaling, highlighting its therapeutic potential in mitigating conjunctival fibrosis.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-25216-9.

Introduction

Fibrosis is a dysregulated tissue repair process characterized by persistent fibroblast activation, excessive extracellular matrix (ECM) deposition, and abnormal tissue remodeling, typically resulting from severe injury or chronic inflammation¹. As a common pathological endpoint of numerous diseases, it can lead to organ dysfunction across nearly all tissues, including those of the ocular surface, such as the conjunctiva^2,3. Conjunctival fibrosis is a key factor in the pathophysiology of several ocular conditions, including ocular cicatricial pemphigoid, severe atopic keratoconjunctivitis (AKC), vernal keratoconjunctivitis (VKC), excessive scarring following glaucoma filtration surgery, pterygium, and trachoma⁴. These conditions impose a significant healthcare burden owing to their potential to progress to visual impairment or blindness⁴.

During fibrosis, cells originating from mesenchymal lineages, such as tissue-resident quiescent fibroblasts, become activated and transdifferentiate into myofibroblasts, initiating fibrotic changes⁵. Although myofibroblasts are recognized as the primary cells orchestrating fibrotic processes, growing evidence over the past two decades highlights the role of other cell types, such as epithelial cells, which undergo a process called epithelial-mesenchymal transition (EMT) and contribute to fibrogenesis⁶.

EMT is a dynamic and reversible process during which epithelial cells undergo cytoskeletal remodeling, lose cell polarity and intercellular adhesions, and acquire a motile, contractile, and invasive fibroblast-like phenotype with increased ECM secretion⁷. EMT plays a critical role in embryonic development, tumor progression, tissue regeneration, inflammation, and tissue fibrosis⁷. Upon tissue injury, epithelial cells activate the EMT program, enabling their participation in tissue repair and the restoration of tissue integrity⁸. Under physiological conditions, these cells typically revert to their epithelial state following reepithelialization and the resolution of the tissue repair response⁸. However, persistent stimuli, such as chronic inflammation, prevent this reversion, causing EMT-derived cells to persist and promote fibrosis⁹.

Transforming growth factor-beta (TGF-β) is a principal protein family responsible for initiating EMT in both in vitro and in vivo settings¹⁰. Among its isoforms, TGF-β1 is particularly recognized as a potent inducer of EMT in cultured conjunctival epithelial cells⁴. Upon activation by TGF-β, the TGF-β receptor phosphorylates Smad-2 and Smad-3, which subsequently form a complex with Smad-4. This complex translocates to the nucleus, where it recruits EMT-related transcription factors (EMT-TFs) to regulate the expression of target genes¹¹. EMT-TFs, primarily members of the Snail, Zeb, and Twist families, drive EMT by downregulating epithelial-specific proteins (e.g., E-cadherin, zonula occludens (ZO)-1, and cytokeratins) and upregulating mesenchymal-associated proteins (e.g., fibronectin, N-cadherin, and vimentin)¹².

While Smad signaling serves as the central pathway in TGF-β-induced EMT, non-canonical TGF-β pathways, including PI3K/Akt, Rho-like GTPases, and MAPKs (Erk-1/2, p38, and JNK), also contribute to this process ¹⁰. The involvement of EMT in renal, pulmonary, liver and intestinal fibrosis has been well documented^13–16. The retina, among other ocular tissues, has been widely studied with regard to EMT^17–22. EMT in retinal pigment epithelium (RPE) has been associated with pathogenesis of subretinal fibrosis in age-related macular degeneration (AMD)²³. A growing body of research highlights the role of EMT in conjunctival fibrosis. In tissue samples from patients with pterygium, the pterygial stroma has been shown to be infiltrated by overlying epithelial cells expressing mesenchymal markers such as α-SMA and vimentin alongside the epithelial marker cytokeratin 14²⁴. Domdey et al. demonstrated increased expression of α-SMA and vimentin in cultured limbal epithelial cells following ultraviolet B (UVB) irradiation, a factor implicated in the pathogenesis of pterygium²⁵. Another study identified an upregulation of MiR-3175 in pterygial tissues and reported that conjunctival epithelial cells underwent EMT when treated with MiR-3175 in vitro²⁶. Moreover, single-cell transcriptome analysis of conjunctival tissue from a mouse model of dry eye disease identified a pro-inflammatory epithelial subtype that expressed epithelial markers (Krt15, Krt6a, and Krt14) and fibroblast markers (Col1a1 and Pdgfra)²⁷. Conjunctival epithelial cells from biopsies of patients with chronic graft-versus-host disease (cGVHD) exhibiting clinically significant dry eye displayed changes consistent with EMT, which were linked to conjunctival fibrosis²⁸. In vitro studies demonstrated that infection of conjunctival epithelial cells with Chlamydia trachomatis induced EMT-associated alterations, including upregulation of α-SMA and fibronectin and downregulation of E-cadherin, which were suggested to contribute to C. trachomatis-induced conjunctival scarring observed in trachoma²⁹. Given the involvement of EMT in conjunctival fibrosis across diverse pathological contexts, targeting EMT emerges as a promising therapeutic strategy for addressing conjunctival fibrosis.

Cannabidiol (CBD) is one of the two most abundant cannabinoids found in the cannabis sativa L. plant, the other being Δ⁹-tetrahydrocannabinol (THC)³⁰. Until recently, cannabinoid research primarily focused on THC owing to its high affinity for cannabinoid receptors³¹. However, cannabinoids influence a variety of other targets, with CBD interacting with over 65 identified molecules, including receptors, enzymes, ion channels, and transient receptor potential channels³¹. CBD has gained significant attention in recent years owing to its lack of psychotropic effects and abuse liability, unlike THC, and its apoptosis-regulating, anti-inflammatory, neuroprotective, antinociceptive, and antioxidative properties³². Beyond its action on cannabinoid receptors, CBD also targets multiple other molecules, such as TRPV1, GPR55, PPARγ, adenosine receptors, and serotonin receptors, thereby exerting pleiotropic effects³³. CBD has been shown to alleviate fibrosis in multiple organs, including the skin, heart, kidney, lung, and liver, as demonstrated by several studies that, however, did not examine EMT in their experimental models^34–44. Although CBD exhibited anti-EMT effects in studies conducted on cancer cells^45–49, to our knowledge, no research to date has explored the anti-EMT effects of cannabinoids in non-cancer contexts. Nonetheless, studies investigating the modulation of the endocannabinoid system (ECS) have demonstrated suppression of EMT. Specifically, activation of the CB2 receptor and inhibition of the CB1 receptor were shown to suppress TGF-β1-mediated EMT in lung alveolar epithelial cells and hepatic stellate cells^50,51. As EMT emerges as a promising therapeutic target against tissue fibrosis⁵², it is worthwhile to investigate the potential impact of CBD on fibrosis-related EMT.

Endogenous cannabinoids, cannabinoid receptors, and other target receptors of cannabinoids are expressed throughout the eye, including the conjunctiva and cornea, suggesting the potential importance of endocannabinoid signaling in maintaining ocular tissue homeostasis⁵³. While prior ophthalmology research has predominantly focused on the neuroprotective and intraocular pressure-lowering effects of cannabinoids, highlighting their potential as anti-glaucoma therapeutics, only a limited number of studies have explored the effects of cannabinoids on the ocular surface⁵⁴. These studies have demonstrated that cannabinoids, including CBD, mediate antinociceptive and anti-inflammatory effects during ocular surface injury⁵⁵.

Given the ubiquitous presence of cannabinoid targets in ocular tissues and the likely role of EMT in conjunctival fibrosis-related pathologies, we aimed to investigate the effects of CBD on conjunctival EMT. To achieve this, cultured human conjunctival epithelial (HCjE) cells were stimulated with TGF-β1 to induce EMT and simultaneously treated with CBD. EMT-related alterations were evaluated by analyzing cell morphology, migration, and the expression of protein markers associated with EMT. The activation of TGF-β signaling was examined through analysis of the canonical Smad pathway. Additionally, we sought to determine whether conjunctival EMT is associated with the secretion of the inflammatory cytokine interleukin-6 (IL-6) and to evaluate the impact of CBD treatment on this process. Demonstrating the anti-EMT effects of CBD may enable the targeting of fibrogenic EMT and support the development of a novel therapeutic approach for conjunctival fibrosis.

Methods and materials

Chemicals

Cannabidiol (CBD) (purity ≥ 99,5%) was kindly provided by Eubio (Vienna, Austria) as 100 mM stock solution prepared in 10% ethanol and 90% dimethyl sulfoxide (DMSO). SB431542, a TGFβ/Smad signaling pathway inhibitor, was purchased from MedChemExpress (Cat.: HY-10431) and dissolved at 5 mM in DMSO as a stock solution.

Cell culture

hTERT-immortalized HCjE cells originating from healthy donors^56,57 were generously provided by Prof. Ilene Gipson (Schepens Eye Research Institute, Harvard Medical School, Boston). These cells were cultured to the exponential phase in keratinocyte-serum-free medium (Cat.:10724–011, Gibco) supplemented with bovine pituitary extract, 0.2 ng/mL EGF, 0.4 mM CaCl₂, and 100 U/mL penicillin/streptomycin (Cat.:15140122, Gibco) at 37 °C, 5% CO₂, and 95% humidity. The medium was changed every second day, and cells were trypsinized at 70% confluency using trypsin-ethylenediamine tetraacetic acid (EDTA) (0.05%), (Cat.: 25300054, Gibco). For the assays, cells from passage 6 were seeded into multiwell culture plates and stimulated for 72 h with TGF-β1 (10 ng/mL) (Cat.:100-B, R&D) once they reached 70% confluency. Concurrently, CBD (2–10 µM), SB431542 (5 µM) or vehicle (0,001% ethanol + 0,009% DMSO) was added to the media.

Cytotoxicity assay

Cells were seeded into 96-well plates and grown for 48 h before being treated with CBD at concentrations of 2, 5, 10, 20, and 50 µM for 72 h. After treatment, a 1% neutral red solution prepared in culture medium was added to each well and incubated for 3 h. Following incubation, the cells were destained using a solution of 1% glacial acetic acid and 50% ethanol. Absorbance was measured at 540 nm using a spectrophotometer (Varioskan Flash, Thermo Fisher Scientific). Relative viability was calculated by normalizing the absorbance values to the control group, which was set at 100%).

Morphological analysis

All images in the experiments were acquired using the TissueFAXS microscope system (TissueGnostics, Vienna, Austria). Brightfield images were analyzed for cell circularity using ImageJ software version 2.3. Circularity values were calculated based on the formula provided, where a value of 1 indicates a perfect circle, and a value of 0 represents an elongated polygon.

Wound healing assay

A scratch was introduced in confluent monolayers using a pipette tip. Cell debris was removed by washing with PBS, and the appropriate treatments were subsequently added. Wound areas were imaged at 0 h and 24 h post-wounding. Wound closure was quantified using ImageJ software and calculated using the following equation: 1 − (cell-free wound area at 24 h/cell-free wound area at 0 h).

Immunocytochemistry

Cells were fixed with 4% paraformaldehyde in phosphate buffered saline (PBS) for 10 min and permeabilized with 0.1% Triton X-100 in PBS for 5 min at room temperature. Cells were then blocked with 3% goat serum in PBS for 1 h at room temperature. Following blocking, cells were incubated overnight at 4 °C with primary antibodies, followed by a 2 h incubation with fluorophore-conjugated secondary antibodies at room temperature. Nuclei were stained with 4’,6-diamidino-2-phenylindole (DAPI) for 10 min. All wash steps were carried out using PBS/0.05% Tween-20. Antibodies were diluted in 1% bovine serum albumin (BSA) in PBS/0.05% Tween-20. Antibody details are provided in Supplementary Table S1. Throughout the experiment, fluorescence imaging was performed using consistent acquisition settings. Fluorescence signals were quantified using CellProfiler and ImageJ.

Western blotting

Cells were incubated on ice for 20 min with radioimmunoprecipitation assay buffer (Cat.:ab156034, Abcam) supplemented with protease inhibitors (Cat.:B14001, Selleckchem) and phosphatase inhibitors (Cat.:sc-45065, Santa Cruz). Lysates were collected by scraping and centrifuged at 16,000 × g for 20 min at 4 °C. Protein concentrations in the collected supernatants were determined using the Bradford assay. Proteins were resolved by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene fluoride (PVDF) membranes. The membranes were blocked with 5% BSA in tris-buffered saline (TBS) containing 0.05% Tween-20 for 1 h at room temperature and then probed overnight at 4 °C with primary antibodies, followed by a 1 h incubation with HRP-conjugated secondary antibodies at room temperature. All wash steps were carried out using TBS/0.05% Tween-20. Antibodies were diluted in 1% BSA in TBS/0.05% Tween-20, and details are provided in Supplementary Table S1. Proteins were visualized using chemiluminescence detection (Cat.:34580, Thermo Scientific) and imaged with the ChemiDoc™ Touch Imaging System (Cat.:1708370, Biorad). Densitometry analysis was performed using Biorad Image Lab software (Biorad, USA). Data were normalized to the loading control Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and the normalization by sum method was applied⁵⁸.

Enzyme-linked immunosorbent assay

Cell culture media were collected from multiwell plates and centrifuged at 500 × g for 10 min at 4 °C. IL-6 levels in the cell culture supernatants were measured following the protocol provided by the antibody manufacturer and normalized to cell numbers determined via DAPI staining. Details of the antibodies used are provided in Supplementary Table S1.

Statistical analysis

Data were analyzed using GraphPad Prism v10 (San Diego, CA) and are presented as mean ± standard deviation of at least three independent experiments as indicated in figure legends. One-way analysis of variance (ANOVA) was performed to compare groups, followed by Dunnett’s post-hoc test. Values of p < 0.05 were considered statistically significant. “ ∗ ” indicates a statistically significant difference compared to the TGF-β1-stimulated group (∗ p < 0.05, ∗  ∗ p < 0.01, ∗  ∗  ∗ p < 0.001, ∗  ∗  ∗  ∗ p < 0.0001), and “#” indicates a statistically significant difference compared to the control group (# p < 0.05, ## p < 0.01, (### p < 0.001; #### p < 0.0001).

Results

CBD inhibits EMT-related alterations induced by TGF-β1

The cytotoxicity of CBD (0–50 µM) on HCjE cells was evaluated using the neutral red assay (Fig. 1A). After 72 h of treatment, significant cytotoxicity was observed only at 50 µM. As a result, subsequent experiments were performed using non-toxic concentrations of 2, 5 and 10 µM CBD. First, we assessed the morphological hallmarks of EMT and quantified cell morphology using the circularity index, where a value of 1 represents a perfect circle, and 0 indicates an elongated polygon (Fig. 1B). TGF-β1 stimulation led to the loss of tight intercellular contacts and the formation of elongated spindle-shaped cells, characteristic of mesenchymal cells, resulting in a reduced circularity index compared to the control. Concurrent treatment with CBD preserved cell–cell contacts, and the typical polygonal morphology of epithelial cells and in most cases reduced the number of cells exhibiting a fusiform morphology and increased the circularity index. This effect was dose-dependent. To further investigate cell migration, a wound healing assay was conducted (Fig. 1C). Compared to the control, TGF-β1 significantly facilitated cell migration, resulting in almost complete wound closure 24 h post-wounding. The addition of CBD to the culture medium alongside TGF-β1 inhibited wound closure in a concentration-dependent manner. Since 2 µM CBD did not show a significant effect on cell morphology and migration, further experiments to assess EMT at the protein level were conducted using 5 and 10 µM CBD. Given that EMT impacts intercellular contacts, the cytoskeleton, and the ECM, we analyzed the cell–cell adhesion protein E-cadherin and the tight junction protein ZO-1, which are markers of the epithelial phenotype, as well as the ECM protein fibronectin, the cytoskeletal protein α-SMA, and the EMT-TF Snail, which are associated with the mesenchymal state. These analyses were performed using immunocytochemistry (Fig. 1D) and western blotting (Fig. 1E). Compared to the control, TGF-β1 downregulated ZO-1 and E-cadherin while upregulating fibronectin and α-SMA, and induced Snail expression, as demonstrated by immunocytochemistry and western blot assays. These results indicate that TGF-β1 activates EMT in HCjE cells. When the cells were treated simultaneously with CBD alongside TGF-β1, they exhibited increased expression of E-cadherin and ZO-1, along with decreased expression of fibronectin and α-SMA, approaching the levels observed in control cells. Additionally, TGF-β1-induced Snail expression was repressed by CBD treatment. These effects were observed in a dose-dependent manner. CBD effectively counteracted EMT-related alterations induced by TGF-β1 in HCjE cells at the protein level.

Fig. 1.

CBD inhibits EMT-related alterations induced by TGF-β1 in HCjE cells. HCjE cells were stimulated for 72 h with TGF-β1 (10 ng/mL) to induce EMT. TGF-β1-stimulated cells were concomitantly treated with CBD (2–10 μM). (A) Cell viability was determined using the neutral red assay (n = 3). (B) Morphological changes in the cells were assessed, and the circularity index was calculated using the formula 4π x (area)/(circumference)²; a value of 1 indicates a perfect circle, while 0 indicates an elongated polygon (n = 5). (C) Cell migration was assessed via a wound healing assay (n = 3). Time points indicate hours after wounding. Data are expressed as the percentage of wound closure, calculated as [1—(cell-free wound area at 24 h / cell-free wound area at 0 h)]. Cyan lines indicate the wound area. (D) Immunocytochemistry assays demonstrated changes in epithelial markers E-cadherin and ZO-1, the mesenchymal marker fibronectin, and the EMT-related transcription factor Snail (n = 3). (E) Changes in E-cadherin, ZO-1, Snail, and the mesenchymal marker α-SMA were analyzed by western blotting. GAPDH served as the loading control. Scale bar = 50 µm. All data are presented as mean ± SD. “n” indicates number of independent experiments. Statistical significance is indicated as * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 compared to TGF-β1; # p < 0.05, ### p < 0.001, #### p < 0.0001 compared to control. CBD, cannabidiol; EMT, epithelial-mesenchymal transition; TGF-β1, transforming growth factor-beta 1; HCjE, human conjunctival epithelial; ZO, zonula occludens; α-SMA, alpha-smooth muscle actin; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

CBD suppresses TGF-β1-activated Smad signaling

TGF-β/Smad signaling is one of the major intracellular signaling pathways promoting both EMT and fibrosis⁵⁹. We aimed to investigate whether the EMT-inhibiting effects of CBD in HCjE cells involve TGF-β/Smad signaling. Smad-2 and Smad-3 are core transcription factors that regulate TGF-β/Smad signaling. Upon phosphorylation, they form a complex with Smad-4 and translocate to the nucleus to regulate target gene expression⁵⁹. Immunocytochemistry assays showed that stimulation of HCjE cells with TGF-β1 promoted the nuclear translocation of Smad-2/3, which was inhibited by CBD in a dose-dependent manner (Fig. 2A). Similarly, TGF-β1-induced phosphorylation of Smad-2/3 was decreased upon concomitant treatment with CBD, as demonstrated by western blotting (Fig. 2B).

Fig. 2.

CBD suppresses TGF-β1-activated Smad signaling in HCjE cells. HCjE cells were stimulated for 72 h with TGF-β1 (10 ng/mL) to induce EMT. TGF-β1-stimulated cells were concomitantly treated with CBD (2–10 μM). (A) Immunocytochemistry assays demonstrated nuclear translocation of Smad-2/3 (n = 3). (B) Phosphorylation of Smad-2/3 was determined by western blotting. Scale bar = 50 µm. All data are presented as mean ± SD. “n” indicates number of independent experiments. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 compared to TGF-β1; ## p < 0.01, #### p < 0.0001 compared to control. CBD, cannabidiol; TGF-β1, transforming growth factor-beta 1; HCjE, human conjunctival epithelial; EMT, epithelial-mesenchymal transition.

Inhibition of Smad signaling prevents EMT induced by TGF-β1

We aimed to examine whether the inhibition of TGF-β/Smad signaling in HCjE cells hinders the EMT process. SB431542, a specific inhibitor of TGF-β/Smad signaling, was employed for this purpose. Compared to TGF-β-stimulated cells, concomitant treatment with SB431542 alongside TGF-β maintained the epithelial phenotype of the cells and increased the circularity index (Fig. 3A) and inhibited wound closure (Fig. 3B). The decreased levels of E-cadherin and ZO-1, as well as the increased expression of fibronectin and Snail induced by TGF-β, were restored to levels similar to those of the control when the cells were treated with SB431542 (Fig. 3C and 3D). The inhibition of Smad-2/3 nuclear translocation (Fig. 3E) and Smad-2/3 phosphorylation (Fig. 3F) by SB431542 was also confirmed.

Fig. 3.

Inhibition of Smad signaling prevents EMT induced by TGF-β1 in HCjE cells. HCjE cells were stimulated for 72 h with TGF-β1 (10 ng/mL) to induce EMT. TGF-β1-stimulated cells were concomitantly treated with the TGF-β/Smad inhibitor SB431542 (5 μM). (A) Morphological changes in the cells were assessed, and the circularity index was calculated using the formula 4π x (area)/(circumference)²; a value of 1 indicates a perfect circle, while 0 indicates an elongated polygon (n = 5). (B) cell migration was assessed via a wound healing assay (n = 3). Time points indicate hours after wounding. Data are expressed as the percentage of wound closure, calculated as [1—(cell-free wound area at 24 h / cell-free wound area at 0 h)]. Cyan lines indicate the wound area. (C) Immunocytochemistry assays demonstrated changes in epithelial markers E-cadherin and ZO-1, the mesenchymal marker fibronectin, and the EMT-related transcription factor Snail (n = 3). (D) Changes in E-cadherin, ZO-1, and Snail were analyzed by western blotting. GAPDH served as the loading control. (E) Immunocytochemistry assays demonstrated nuclear translocation of Smad-2/3 (n = 3). (F) Phosphorylation of Smad-2/3 was determined by western blotting. Scale bar = 50 µm. All data are presented as mean ± SD. “n” indicates number of independent experiments. Statistical significance is indicated as * p < 0.05, *** p < 0.001, **** p < 0.0001 compared to TGF-β1; # p < 0.05, ## p < 0.01, ### p < 0.001, #### p < 0.0001 compared to control. EMT, epithelial-mesenchymal transition; TGF-β1, transforming growth factor-beta 1; HCjE, human conjunctival epithelial; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; ZO, zonula occludens.

CBD prevents increased IL-6 secretion associated with EMT phenotype

Secretion of inflammatory mediators by EMT-cells has been proposed as a mechanism through which they activate nearby fibroblasts and exacerbate fibrosis^60–62. IL-6 has been shown to activate fibroblasts ⁶³. In previous sections, we confirmed the occurrence of EMT in HCjE cells in response to TGF-β. As a next step, we aimed to uncover whether EMT is linked with increased IL-6 secretion and whether this can be inhibited by CBD treatment. Results revealed that TGF-β promoted IL-6 secretion, while treatment with CBD decreased IL-6 levels in a dose-dependent manner (Fig. 4). The concurrent elevation of IL-6 secretion and EMT activation suggests that cells in the mesenchymal state acquire an inflammatory secretome that produces IL-6, which can be prevented by CBD.

Fig. 4.

CBD prevents increased IL-6 secretion associated with EMT phenotype. HCjE cells were stimulated for 72 h with TGF-β1 (10 ng/mL) to induce EMT. TGF-β1-stimulated cells were concomitantly treated with the TGF-β/Smad inhibitor SB431542 (5 μM) or CBD (5–10 μM). IL-6 levels in cell culture supernatants were measured via ELISA (n = 3). “n” indicates number of independent experiments. Results were normalized to cell number. All data are presented as mean ± SD. Statistical significance is indicated as **** p < 0.0001 compared to TGF-β1; p < 0.0001 compared to control. CBD, cannabidiol; IL-6, interleukin-6; EMT, epithelial-mesenchymal transition; TGF-β1, transforming growth factor-beta 1; ELISA, enzyme-linked immunosorbent assay.

Discussion

EMT is a form of cell plasticity that enables epithelial cells to shift along the epithelial-mesenchymal spectrum⁶⁴. Although its role in cancer cell invasion and dissemination has been well acknowledged over the years, EMT triggered by the persistent inflammatory tissue microenvironment is increasingly recognized as an important contributor to fibrosis^65,66. Lineage tracing experiments have demonstrated that epithelial cells undergoing in vivo EMT do not transdifferentiate into fibroblasts or myofibroblasts, as previously thought; instead, they undergo a partial EMT⁶⁷. Although targeting EMT has been shown to be a promising approach against fibrosis^68–75, further research is needed to elucidate the exact mechanisms by which the cells in an intermediate EMT state participate in fibrosis. Several studies have indicated that the paracrine secretion of inflammatory and profibrotic mediators by EMT-cells activates local fibroblasts^60–62. As mentioned earlier, CBD has been reported to exert antifibrotic effects in extraorbital tissues. However, its impact on fibrosis-associated EMT remains unknown. Given that multiple molecular targets of cannabinoids are prevalent in ocular tissues and modulation of the ECS improves various eye conditions^31,76, further research is required to explore the therapeutic potential of CBD in the field of ophthalmology. A major promoter of fibrosis and EMT is TGF-β, which becomes activated in response to tissue injury and inflammation⁷⁷. In our study, we found that stimulation of HCjE cells with TGF-β1 promoted IL-6 secretion and activated the EMT program, as evidenced by the acquisition of a spindle-shaped appearance, increased cell migration, and upregulation of EMT markers. Both elevated IL-6 secretion and EMT-related alterations were significantly prevented by CBD treatment through suppression of the TGF-β-Smad-Snail axis.The mechanism by which CBD inhibits TGF-β1–induced EMT is illustrated in Fig. 5.

Fig. 5.

Mechanism by which CBD disrupts TGF-β/Smad–driven EMT. CBD inhibits TGF-β–induced EMT in HCjE cells by interfering with the TGF-β/Smad signaling pathway. Upon TGF-β stimulation, receptor-mediated phosphorylation of Smad-2 and Smad-3 occurs, enabling their association with Smad-4 and subsequent nuclear translocation. Inside the nucleus, this Smad complex activates transcription of EMT-inducing factors such as Snail, leading to downregulation of epithelial markers (E-cadherin, ZO-1) and upregulation of mesenchymal markers (fibronectin, α-SMA). CBD attenuates this signaling cascade by reducing Smad-2/3 phosphorylation and preventing efficient nuclear accumulation of the Smad complex. Consequently, Snail-driven transcription is suppressed, blocking EMT progression and preserving epithelial phenotype. By the same mechanism, CBD also abrogates proinflammatory IL-6 secretion found to be associated with the EMT state. CBD, cannabidiol; EMT, epithelial-mesenchymal transition; TGF-β1, transforming growth factor-beta 1; HCjE, human conjunctival epithelial; ZO, zonula occludens; α-SMA, alpha-smooth muscle actin; IL-6, interleukin-6.

TGF-β has a critical function in neuronal and vascular development, as well as in maintaining epithelial integrity and tissue homeostasis in the eye⁷⁸. Dysregulation of TGF-β has been associated with a number of eye conditions, including trabecular meshwork fibrosis, reduced aqueous outflow, and elevated intraocular pressure in glaucoma^79–81; pericyte-myofibroblast transition linked to subretinal fibrosis and choroidal neovascularization in AMD^82,83; and EMT of lens epithelial cells in anterior subcapsular cataract and posterior capsule opacification following cataract surgery^84,85. TGF-β signaling has also been reported to be involved in ocular surface pathologies. Impeding TGF-β signaling in CD4⁺ T cells led to an improvement in dry eye disease in mice exposed to desiccating stress⁸⁶. TGF-β expression was higher in pterygium compared with healthy conjunctiva and was secreted to a greater extent by cultured pterygial fibroblasts compared with normal conjunctival fibroblasts^87,88. Ohtomo et al. demonstrated TGF-β1-positive eosinophils along with TGF-β1-related proteins in the giant papillae of patients with VKC and revealed that TGF-β1, in synergy with IL-4 and IL-13, increased eotaxin production by cultured conjunctival and corneal fibroblasts, highlighting the role of TGF-β in tissue eosinophilia linked with VKC⁸⁹. Asano-Kato et al. reported that TGF-β1, compared with other cytokines they investigated, stimulates vascular endothelial growth factor production from conjunctival fibroblasts to the greatest degree, implicating its role in neovascularization in papillary hypertrophy in AKC and VKC⁹⁰. Infection with C. trachomatis provoked upregulation of TGF-β1 and TGF-β2 in addition to EMT-related changes in cultured conjunctival epithelial cells²⁹. Overexpression of TGF-β was associated with bleb failure resulting from excessive conjunctival scarring following trabeculetomy in rats⁹¹. Furthermore, Benito et al. found that while both cultured human conjunctival and corneal epithelial cells secreted TGF-β1 and TGF-β2 under basal conditions, conjunctival cells predominantly secreted TGF-β1, whereas corneal cells secreted more TGF-β2⁹². The study also reported that stimulation of these cells with TGF-β1 and TGF-β2 activated Smad-2 and promoted the secretion of cytokines GM-CSF, IL-6, IL-8, and IL-10, demonstrating the significance of TGF-β/Smad signaling in regulating inflammatory processes in ocular surface epithelial cells. TGF-β1 stimulation of cultured corneal epithelial cells has been shown to induce senescence, a state of permanent proliferative arrest associated with aging and age-related diseases⁹³. Senescent cells of the RPE have been shown to secrete IL-6, IL-8, and TGF-β, thereby providing an inflammatory tissue microenvironment that initiates EMT and fibrosis⁹⁴. The participation of the Smad-dependent TGF-β signaling pathway in the pathogenesis of many of the aforementioned conditions has been demonstrated by previous research^{29,83,95–97}. The current study showed that TGF-β activated EMT and increased IL-6 secretion. Both of these changes were prevented when TGF-β/Smad signaling was inhibited with SB431542. This concordance points to an association between the EMT phenotype and IL-6 secretion, reinforcing the notion that epithelial cells acquire inflammatory characteristics by undergoing EMT. Furthermore, inhibition of EMT by SB431542 identifies TGF-β/Smad signaling as a potential therapeutic target against conjunctival EMT.

The intraocular pressure-lowering properties of cannabinoids against ocular hypertension and their neuroprotective effects against retinal excitotoxicity in glaucoma have been relatively well documented^31,98. Favorable effects of cannabinoids on diabetic retinopathy and retinal ischemia, achieved by mitigating neurotoxicity, oxidative stress, and inflammation have also been demonstrated in previous studies⁵⁴. Most investigations into the effects of cannabinoids on eye tissues have primarily focused on THC, CBD, or cannabinoid receptor modulators. However, only a few studies have explored the effects of cannabinoids on ocular surface conditions. In a mouse model of dry eye disease, desiccating stress was shown to upregulate CB1 and CB2 receptors in the cornea and conjunctiva⁹⁹. In the same study, topical cannabinoid treatment with THC decreased IL-1β levels and the CD4⁺/CD8⁺ ratio in the cornea and conjunctiva. Another study showed that topical administration of CBD, the CBD derivative HU-308, and Δ⁸THC to the eyes reduced pain scores and corneal neutrophil infiltration in mice subjected to chemical cauterization of the corneal epithelium¹⁰⁰. The inflammation-reducing properties of ECS modulation have also been demonstrated in experimentally induced uveitis. In mouse models of endotoxin-induced uveitis, topical application of CB2 agonists decreased leukocyte-endothelial adhesion and neutrophil infiltration in the iris, cornea, and ciliary body, as well as reduced NF-κB and AP-1 activity, along with inflammatory mediators TNF-α, IL-1β, IL-6, CCL5, and CXCL2^101–103. Our results showed that CBD counteracted TGF-β1-mediated conjunctival EMT and associated IL-6 secretion. This positions CBD as a potential candidate against EMT and local inflammatory responses associated with the EMT phenotype.

In dry eye disease, tear film instability and hyperosmolarity often result from immune-mediated inflammation, which triggers an inflammatory secretome in the epithelial cells of the ocular surface¹⁰⁴. Mucin secretion by the ocular surface epithelium becomes disrupted when the cells undergo EMT, a process implicated in dry eye disease associated with cGVHD¹⁰⁴. Corneal abrasions resulting from severe dry eye disease can progress to corneal ulcers and sight loss¹⁰⁵. While a variety of pharmacological treatments exist, non-adherence due to low tolerability is frequent¹⁰⁶. Ocular neuropathic pain is a debilitating condition that is challenging to control¹⁰⁷. Topical pharmacotherapies often lack the desired efficacy and are coupled with deleterious side effects¹⁰⁸. Non-steroidal anti-inflammatory drugs and topical anesthetics may impede wound healing, whereas systemic analgesia with opioids can cause respiratory depression^108,109.

Outside ophthalmology, systemic antifibrotics are approved for pulmonary fibrosis. Nintedanib is a receptor tyrosine kinase inhibitor of PDGF, FGF, and VEGF receptors¹¹⁰. Pirfenidone suppresses synthesis of profibrotic mediators including TGF-β1, TNF-α, PDGF, and IL-1β¹¹¹. Both agents have demonstrated inhibition of epithelial–mesenchymal transition in experimental models^112–114. However, their role in ocular fibrosis remains largely unknown. In pulmonary fibrosis, their systemic use is limited by modest efficacy and frequent treatment discontinuation due to adverse effects^115,116.

Current antifibrotic strategies in ophthalmology remain constrained by limited efficacy and significant safety concerns. Antimetabolites such as mitomycin C, an alkylating agent that crosslinks DNA and inhibits fibroblast proliferation, and 5-fluorouracil (5-FU), a pyrimidine analog that blocks thymidylate synthase and disrupts DNA synthesis, are applied locally after trabeculectomy to suppress subconjunctival scarring¹¹⁷. Despite their ability to modulate wound healing, both have a narrow therapeutic window. Mitomycin C often fails to prevent bleb failure and may induce conjunctival epithelium breakdown, hypotony, and endophthalmitis¹¹⁸. 5-FU is associated with corneal epithelial toxicity and persistent epithelial defects¹¹⁹. Anti-angiogenic therapy with bevacizumab, which inhibits VEGF-driven angiogenesis and fibroblast activation, has shown comparable efficacy to mitomycin C in improving trabeculectomy success¹²⁰. However, its use is complicated by an increased risk of filtering bleb leakage¹²⁰. Targeted inhibition of TGF-β signaling, such as the TGF-β2 neutralizing antibody CAT-152 (lerdelimumab), failed in a phase III trial to improve surgical outcomes compared with placebo¹²¹. Tranilast, an antiallergy drug that inhibits histamine release, also modulates TGF-β and MAPK signaling, mechanisms associated with therapeutic effects in keloids, hypertrophic scars, scleroderma, and cardiac fibrosis¹²². Previous studies have reported favorable outcomes when Tranilast was used as an adjunct in glaucoma filtration surgery¹²³, and anti-EMT activity has also been demonstrated in preclinical models¹²⁴. Systemic exposure, however, has been associated with hepatotoxicity¹²⁵.

In this context, CBD emerges as a particularly attractive candidate. In conjunctival epithelial cells, our study confirmed the inhibitory effect of SB-431542 on TGF-β–induced EMT, while previous investigations demonstrated similar anti-EMT activity along with antifibrotic effects in other preclinical models^126,127. Similarly, CBD inhibits TGF-β/Smad signaling and EMT, as demonstrated in this study. However, SB-431542 lacks translational safety data and has been linked to adverse effects in animal studies, including progression of skin papillomas to squamous cell carcinoma in mice¹²⁸. In contrast, CBD is supported by clinical use and is characterized by well-established pharmacokinetic and toxicity data, with documented human exposure across multiple contexts (oral, transdermal, inhaled) and a consistently favorable safety profile¹²⁹. Moreover, its pleiotropic activity, encompassing anti-inflammatory, antioxidant, and antifibrotic properties, may confer broader therapeutic benefits in complex diseases in which EMT contributes to pathogenesis. Taken together, these findings underscore the need for novel, safer, and more targeted antifibrotic approaches in ophthalmology, positioning CBD as a promising candidate for future clinical translation in ocular fibrosis.

Ocular permeation of CBD remains a major challenge for topical delivery because of its high lipophilicity and rapid precorneal clearance, and studies have investigated strategies such as nanoformulations and CBD analogs to improve bioavailability¹³⁰. The conjunctiva, as an accessible ocular surface tissue, may be less affected by these limitations; however, no study to date has reported conjunctival pharmacokinetics following topical CBD application. Topical CBD formulations have been evaluated in vivo at 0.4–1.6% w/v (≈13–50 mM) and shown to reduce ocular surface inflammation in mice¹³¹, while a 1% CBD formulation (~ 32 mM) lowered intraocular pressure and produced measurable anterior tissue exposure in rabbits^130,132. Topical ocular drug delivery is highly limited, with generally less than 5% of the instilled dose retained in the eye because of precorneal clearance and nasolacrimal drainage¹³³. Data from topical cyclosporine A, another highly lipophilic compound that requires emulsified formulations for ocular delivery, show that after topical application in rabbits only ~ 0.1% of the applied dose was recovered in conjunctival tissue¹³⁴. These findings indicate that the proportion of drug penetrating conjunctival epithelial cells is markedly lower than the millimolar concentrations present in dosing solutions. On this basis, exposing conjunctival cells to 2–10 µM CBD in vitro represents a conservative and physiologically relevant approximation of achievable tissue levels after topical administration.

Overall, as a compound with diverse properties, CBD may improve ocular surface pathologies resulting from inflammation and fibrosis through regulation of EMT and the associated inflammatory secretome, while also exerting neuroprotective and antinociceptive effects.

Conclusion

Our study revealed the anti-EMT effects of CBD in conjunctival epithelial cells, mediated through inhibition of the TGF-β-Smad-Snail axis. We further identified an association between EMT state and IL-6 secretion which was mitigated by CBD. Given the wide range of cellular targets affected by CBD, it is important to further dissect the upstream mechanisms modulating TGF-β/Smad signaling to clarify the molecular basis of its anti-EMT activity. Moreover, conjunctival fibrosis develops through complex in vivo processes involving multiple cell types, paracrine signaling, and extracellular matrix remodeling. Since our study is limited to an in vitro epithelial cell model, it does not recapitulate these multicellular interactions or the stromal and tissue remodeling components of fibrosis. Therefore, validation in animal models will be necessary to assess the translational relevance of our findings. Future studies should also investigate tissue-scale EMT dynamics and epithelial–fibroblast interactions to better evaluate the therapeutic potential of CBD in conjunctival fibrosis–related pathologies.

Abbreviations

AKC

Atopic keratoconjunctivitis

CB1

Cannabinoid receptor type 1

CB2

Cannabinoid receptor type 2

CBD

Cannabidiol

cGVHD

Chronic graft-versus-host disease

E-cadherin

Epithelial cadherin

ECM

Extracellular matrix

ECS

Endocannabinoid system

EGF

Epidermal growth factor

EMT

Epithelial–mesenchymal transition

GAPDH

Glyceraldehyde-3-phosphate dehydrogenase

HCjE

Human conjunctival epithelial cells

TGF-β

Transforming growth factor-beta

THC

Δ⁹-Tetrahydrocannabinol

VKC

Vernal keratoconjunctivitis

ZO-1

Zonula occludens-1

α-SMA

Alpha-smooth muscle actin

Author contributions

TBA supervised the project. AB and EMA designed the project. AB wrote the manuscript. AB performed cell culture, immunocytochemistry and western blot experiments. AB, AE and JP performed ELISA experiments. AB performed data analysis and prepared figures. EMA assisted in drafting the manuscript and contributed to scientific discussions.

Data availability

Data supporting this study are included within the article and/or supporting materials. Further inquiries should be directed to the corresponding author.

Declarations

Competing interests

The authors declare no competing interests.