Enhanced ocular delivery of tacrolimus via cationic nanoliposomes: Mitigating inflammation and rebalancing Th17/Treg responses in dry eye disease Authors: Lirui Liu, Dan Chen, Siting Sheng, Xueqi Lin, Huiling Zhao, Chujun Liu, Xiang Chen, Yang Ye, Wen Xu Published: 2026 ABSTRACT AbstractDry eye disease (DED) is a common ocular surface disorder characterized by persistent inflammation and immune imbalance, leading to visual impairment and substantial deterioration in quality of life. Although topical medications remain the mainstay of DED management, their therapeutic effectiveness is frequently limited by poor corneal penetration, low ocular bioavailability, and potential safety issues associated with long-term administration. To address this, we engineered a low-dose cationic tacrolimus nanoliposome delivery system (FK506 NLPs) designed to enhance ocular surface drug availability while targeting immune-inflammatory dysregulation. The positively charged NLPs exhibited prolonged corneal residence via mucoadhesion and improved drug permeation. In a murine model of DED, FK506 NLPs accelerated ocular surface repair, improved tear film stability, and preserved conjunctival goblet cells and lacrimal gland structure. Mechanistically, they suppressed the expression of key proinflammatory mediators in both corneal and lacrimal tissues and restored immune homeostasis by re-establishing the balance between Th17 and regulatory T cells. Importantly, FK506 NLPs also exhibited favorable in vivo biocompatibility. Taken together, these findings demonstrated that FK506 NLPs combined improved ocular delivery with coordinated anti-inflammatory and immunoregulatory effects, highlighting their potential as a promising therapeutic strategy for the management of DED.Keywords: Dry eye, Nanoparticles, Liposomes, Tacrolimus, Th17/Treg Introduction Dry eye disease (DED) is a multifactorial ocular surface disorder that arises primarily from disturbances in tear composition, volume, or dynamics, culminating in tear film instability and disruption of the ocular surface microenvironment [1]. Clinically, DED is marked by chronic ocular surface inflammation, epithelial injury, and neurosensory dysfunction, manifesting as dryness, visual fatigue, pain, burning, itching, and foreign body sensation [2]. Epidemiological studies estimate that DED affects approximately 10-20% of individuals over 40 years of age, with prevalence continuing to rise in parallel with population aging and the widespread use of contact lenses and digital devices. Together, these trends impose an escalating burden on visual performance and overall quality of life [3]. Current therapeutic strategies for DED, such as artificial tears and cholinergic agents, primarily offer short-lived symptomatic relief and fail to adequately address the underlying inflammatory pathology. To more effectively disrupt the self-perpetuating inflammatory cascade that drives DED, immunosuppressive agents have emerged as alternative treatment options [4]. Tacrolimus (FK506), a macrolide immunosuppressant with a potency 10-100 times greater than that of cyclosporine, has demonstrated considerable therapeutic promise [5]. Chronic inflammation is central to DED pathogenesis. Hyperosmolar stress induces corneal epithelial apoptosis and promotes the release of pro-inflammatory cytokines, which activate resident antigen-presenting cells (APCs), including macrophages and dendritic cells [6,7]. These APCs subsequently migrate to draining lymph nodes, where they promote the differentiation of CD4+ T cells, particularly T helper 17 (Th17) cells. The infiltration of Th17 cells into the ocular surface and lacrimal glands further compromises corneal barrier integrity, exacerbates epithelial injury, and destabilizes the tear film [7,8]. In parallel, regulatory T cells (Tregs), which are critical for maintaining immune tolerance through Foxp3-dependent transcriptional programs, progressively lose their suppressive capacity under sustained inflammatory stress, thereby intensifying Th17/Treg imbalance and driving disease progression [[9], [10], [11]]. Mechanistically, FK506 binds to FK506-binding protein 12 (FKBP12) to inhibit calcineurin-mediated dephosphorylation of nuclear factor of activated T cells (NFAT), resulting in suppression of interleukin-2 (IL-2) transcription and attenuation of T-cell activation [12,13]. Notably, Th17 cells display heightened sensitivity to FK506, underscoring its particular suitability as a targeted immunomodulatory therapy for DED-associated inflammation [14]. Nevertheless, conventional FK506 formulations are usually administered at concentrations around 1.0 mg/mL and still face major limitations. Because FK506 is highly hydrophobic and has a relatively high molecular weight, its corneal penetration is limited and it is rapidly cleared from the ocular surface. As a result, effective treatment often requires higher concentrations and prolonged dosing. This may reduce patient adherence and may also adversely affect Treg biology by suppressing Foxp3 expression and impairing Treg maturation [15]. To counteract these drawbacks, FK506 is frequently combined with adjunctive agents such as metformin, IL-2, or rapamycin to alleviate its inhibitory effects on Tregs and re-establish Th17/Treg equilibrium [[16], [17], [18]]. Intriguingly, accumulating evidence indicates that FK506 exerts a concentration-dependent, bidirectional influence on Foxp3 transcription [[19], [20], [21]]. At low intracellular levels, FK506 can facilitate nuclear translocation of NFAT and synergize with Smad3 signaling, thereby enhancing Foxp3 expression and promoting Treg differentiation [22]. Together, these observations highlight the critical importance of optimizing FK506 delivery to enable effective immunomodulation at substantially reduced doses. Recent progress in ocular pharmacotherapy has emphasized the urgent need for safer and more efficient local drug delivery strategies. The clinical utility of conventional eye drops is inherently constrained by poor ocular bioavailability, typically below 5%, owing to rapid tear turnover, blinking, and multiple anatomical and physiological barriers of the ocular surface [23,24]. These factors severely limit drug residence time and transcorneal absorption [25]. In this context, nanocarrier-based delivery systems, including polymeric nanoparticles, nanoemulsions, and lipid vesicles, have emerged as promising alternatives by improving drug retention, permeability, and overall bioavailability [[26], [27], [28], [29]]. In recent years, nanozyme-based systems have also attracted increasing attention in dry eye treatment. Unlike conventional nanocarriers, nanozymes can mimic natural antioxidant enzymes and directly scavenge reactive oxygen species [30,31]. Among these platforms, cationic liposomes have garnered particular interest because their positive surface charge enables electrostatic interactions with the negatively charged mucin layer, thereby enhancing ocular surface adhesion and facilitating corneal penetration [32]. In addition, rational modulation of lipid bilayer composition allows for stable drug encapsulation and controlled release kinetics, ultimately increasing therapeutic efficacy while reducing dosing frequency [33]. Therefore, nanozyme-based systems mainly highlight intrinsic therapeutic activity, whereas cationic liposomes mainly provide efficient drug delivery and enhanced ocular surface retention. Collectively, these attributes position cationic liposomes as a leading and versatile platform in ocular drug delivery research. In this study, we engineered low-concentration (0.2 mg/mL) cationic FK506 nanoliposomes (FK506 NLPs) as a therapeutic strategy for DED aimed at simultaneously suppressing ocular surface inflammation and correcting immune dysregulation (Fig. 1). FK506 NLPs displayed a narrow and uniform particle size distribution along with sustained-release behavior, leading to a marked prolongation of ocular surface residence. Using a benzalkonium chloride (BAK)-induced murine model of DED, we demonstrated that FK506 NLPs significantly attenuated pro-inflammatory cytokine production, promoted Treg generation, and effectively restored the Th17/Treg balance. Importantly, comprehensive ophthalmic and systemic safety assessments revealed no detectable adverse effects. Taken together, these findings indicated that FK506 NLPs integrated enhanced ocular delivery with coordinated anti-inflammatory and immunoregulatory actions, underscoring their strong potential for clinical translation in the treatment of DED. Schematic illustration of the therapeutic effects of FK506 NLPs on the ocular surface, highlighting their roles in promoting corneal epithelial cell repair, restoring tear secretion, and improving goblet cell function, along with the proposed anti-inflammatory effects and restoration of Th17/Treg immune homeostasis in DED treatment. Methods Synthesis and characterization of FK506 NLPs FK506 NLPs were fabricated using a thin-film dispersion-hydration method as previously described [34]. Briefly, 5.0 mg of 1,2-dioleoyl-3-trimethylammonium-propane (DOTAP), 5.32 mg of 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE), 2.77 mg of cholesterol, 5.02 mg of 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy (polyethylene glycol)-2000] (DSPE-PEG2000), and 2.0 mg of FK506 were accurately weighed and dissolved in 6 mL of chloroform in a 100-mL round-bottom flask. The organic solvent was removed by rotary evaporation under reduced pressure at 120 rpm and 40 °C for 1 h, yielding a uniform thin lipid film on the inner wall of the flask. The resulting film was hydrated with 10 mL of phosphate-buffered saline (PBS) containing 15 glass beads at 40 °C for 1 h. Based on the amount of FK506 used and the hydration volume, the nominal FK506 concentration of the formulation was 0.2 mg/mL (2.0 mg/10 mL). The suspension was then sonicated in an ultrasonic water bath for 40 min to reduce particle size. After that, the liposomal suspension was extruded using a Mini-Extruder (Avanti Polar Lipids, Inc., Alabaster, AL, USA) through Nuclepore® polycarbonate membranes with pore sizes of 800 nm, 400 nm, and 200 nm. Each extrusion step included 11 back-and-forth passes to obtain a narrow and homogeneous size distribution. Nile Red-labeled FK506 NLPs (Nile Red NLPs) were prepared using the same protocol. All formulations were stored at 4 °C and passed through a 0.22 μm membrane filter for sterile processing before use. The mean hydrodynamic diameter, zeta potential, and polydispersity index (PDI) of FK506 NLPs and empty nanoliposomes (ENLPs) were determined using a Zetasizer Nano ZS (Malvern Panalytical, UK) at 25 °C. Prior to measurement, samples were diluted 100-fold with deionized water, and all measurements were performed in triplicate (n = 3). The microstructural morphology of the liposomes was examined by transmission electron microscopy (TEM) and cryogenic transmission electron microscopy (cryo-TEM). For TEM analysis (JEM-1230, JEOL, Japan), appropriately diluted samples were negatively stained with 2% (w/v) uranyl acetate, deposited onto copper grids, and air-dried prior to imaging at an accelerating voltage of 80 kV. For cryo-TEM (FEI Talos F200C, USA), 3-5 μL of sample was applied to hydrophilized copper grids, blotted to remove excess liquid, and rapidly vitrified in liquid ethane using an FEI Vitrobot. The vitrified grids were transferred under liquid nitrogen to a cryo-holder and imaged at 200 kV. Determination of the standard curve and in vitro drug release A standard calibration curve for FK506 was generated using a series of known concentrations prepared in methanol. In vitro release of FK506 from the NLPs was evaluated by a dynamic dialysis method employing dialysis bags with a molecular weight cut-off (MWCO) of 3500 Da. Each dialysis bag was loaded with 4 mL of FK506 NLP suspension, securely sealed, and immersed in 100 mL of artificial tears solution as the release medium. The system was maintained at 37 °C with continuous stirring at 200 rpm. At predetermined time points, 1 mL aliquots of the external medium were collected and immediately replaced with an equal volume of fresh artificial tears solution to maintain sink conditions. The collected samples were measured by UV-Vis spectrophotometry at 210 nm, and FK506 concentrations were determined according to the standard calibration curve. Retention time of FK506 NLPs on the ocular surface In vivo fluorescence imaging was performed to evaluate the residence time of FK506 NLPs on the ocular surface. Mice were anesthetized by intraperitoneal injection of sodium pentobarbital (200 mg/kg), after which 5 μL of Nile Red NLPs or an equivalent concentration of free Nile Red was topically instilled onto the corneal surface. Fluorescence signals were recorded at predefined time points (0, 0.5, 3, 5, 10, 15, and 25 min) using an in vivo imaging system (AniView100, BioLutec, China) with excitation and emission wavelengths set at 535 and 600 nm, respectively. For quantitative analysis, regions of interest (ROIs) were defined using the automatic analysis function of the AniView system. The same ROI was copied from a template image to other images at the same position, and background subtraction was performed based on the selected reference image. Establishment, treatment, and clinical evaluation of the in vivo DED model Experimental DED was induced by topical administration of BAK (0.2% w/v in PBS, 10 μL per eye) to both eyes twice daily (9:00 a.m. and 9:00 p.m.) for seven consecutive days. Following model establishment, mice were randomly allocated to four treatment groups and received topical treatment twice daily for 3 days with one of the following formulations: 0.9% saline (control), free FK506 solution (FK506-Free, 0.2 mg/mL), a commercial FK506 hydrate ophthalmic solution (Talymus®, 1 mg/mL), or FK506 NLPs (FK506-equivalent concentration 0.2 mg/mL). All treatments were administered at a volume of 10 μL per eye and at identical time points to ensure intergroup consistency. Therapeutic efficacy was assessed on days 0, 1, 2, and 3 after DED induction by evaluating tear secretion, tear film break-up time (TBUT), and corneal fluorescein staining. Tear secretion was quantified using the phenol red thread test, in which a phenol red-impregnated cotton thread was placed in the lower conjunctival fornix (approximately one-third of the distance from the lateral canthus) for 60 s, and the length of the wetted portion (mm) was measured. TBUT was determined following the instillation of 2 μL of 1% fluorescein sodium into the conjunctival sac. Under cobalt blue illumination using a slit-lamp imaging system, the interval between eyelid opening and the appearance of the first dark spot on the corneal surface was recorded, and the mean value of three measurements per eye was calculated. For corneal fluorescein staining, an additional 2 μL of 1% fluorescein sodium was applied, and the eyelids were gently closed three to four times to ensure uniform dye distribution. Corneal images were captured under cobalt blue illumination with a slit-lamp imaging system, and staining severity was graded according to a standardized 0-4 scale: 0, no staining; 1, slight punctate staining (<30 spots); 2, punctate staining (>30 spots) without diffuse staining; 3, severe diffuse staining without plaque formation; and 4, severe diffuse staining with dense plaque formation [35]. RNA isolation and quantitative real-time PCR (RT-qPCR) Total RNA was isolated from corneal and lacrimal gland tissues using the FlaPure Animal Tissue Total RNA Extraction Kit (GenePure, Beijing, China) in accordance with the manufacturer's instructions. RNA quantity and purity were determined with a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA). Only samples meeting quality criteria (A260/280 ratio >1.8 and RNA concentration ≥100 ng/μL) were used for subsequent analyses. First-strand cDNA synthesis was performed using PremixScript™ RT Master Mix (Perfect Real Time, Takara, Shiga, Japan) under the following conditions: 37 °C for 15 min, 85 °C for 5 s, and holding at 4 °C. RT-qPCR was carried out with TB Green Premix Ex Taq™ II (Takara, Shiga, Japan) using a thermal cycling program consisting of an initial denaturation at 95 °C for 30 s, followed by 40 cycles of 95 °C for 5 s and 60 °C for 30 s. Each experimental group comprised three biological replicates, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) served as the internal reference gene. Relative mRNA expression levels were calculated using the 2−ΔΔCt method. Primer sequences used in this study are provided in Table S1. Immunofluorescence staining Fixed tissues were embedded in optimal cutting temperature (OCT) compound and cryosectioned into 7 μm-thick sections. The sections were permeabilized with 0.5% Triton X-100 (Solarbio, Beijing, China) for 20 min, followed by three washes with PBS. Nonspecific binding was blocked by incubating the sections with 10% goat serum (Solarbio, Beijing, China) for 1 h at room temperature. Sections were then incubated overnight at 4 °C with the following primary antibodies: rabbit anti-cytokeratin 10 (1:200; ab76318, Abcam), rabbit anti-IL-1β (1:400; ab9722, Abcam), rabbit anti-MMP-9 (1:200; 10375-2-AP, Proteintech Group), rabbit anti-CD4 (1:100; 19068-1-AP, Proteintech Group), mouse anti-IL-17 (1:50; sc-374218, Santa Cruz Biotechnology), and mouse anti-Foxp3 (1:50; sc-53876, Santa Cruz Biotechnology). After washing with PBS, sections were incubated for 1 h at room temperature with the corresponding fluorescent secondary antibodies, including Cy3-conjugated goat anti-rabbit IgG (H + L) (1:200; A0516, Beyotime) and FITC-conjugated goat anti-mouse IgG (H + L) (1:200; A0568, Beyotime). Nuclei were counterstained with DAPI, and sections were mounted using an antifade mounting medium. Images were acquired with a confocal laser-scanning microscope, and fluorescence signals were analyzed using ZEN 2 software. Quantification of positively stained areas was performed with ImageJ. Flow cytometry Single-cell suspensions were prepared from cervical lymph nodes of neonatal mice. Red blood cells were lysed by incubation with 2 mL of Lysing Buffer (Cat.# 555899, BD Biosciences) on ice for 5 min, and the reaction was quenched by the addition of PBS. Cells were then centrifuged, washed, and enumerated using a Countess® II automated cell counter. For ex vivo stimulation, 1 mL of the single-cell suspension was incubated with 2 μL of Leukocyte Activation Cocktail containing BD GolgiPlug (Cat.# 550583, BD Biosciences) at 37 °C in a humidified atmosphere with 5% CO2 for 4-6 h. Following stimulation, cells were washed twice with PBS and prepared for subsequent staining procedures. Cell viability was assessed by staining with 1 μL of Fixable Viability Stain 780 (Cat.# 565388, BD Biosciences) for 5 min at 37 °C, followed by two washes with Stain Buffer (FBS) (Cat.# 554656, BD Biosciences). Surface marker staining was then performed by incubating cells in 100 μL of Stain Buffer containing fluorochrome-conjugated antibodies against CD3ε (PE-Cy7 hamster anti-mouse CD3ε, Cat.# 552774, clone 145-2C11), CD4 (FITC rat anti-mouse CD4, Cat.# 553046, clone RM4-5), and CD25 (BV421 rat anti-mouse CD25, Cat.# 564370, clone 3C7) for 20-30 min at 4 °C in the dark. After washing, cells were fixed and permeabilized using freshly prepared reagents from the Transcription Factor Buffer Set (Fix/Perm Buffer, Cat.# 562574, BD Biosciences) for 40-50 min at 2-8 °C in the dark. Intracellular cytokine and transcription factor staining was subsequently carried out by resuspending the fixed cells in 100 μL of Perm/Wash Buffer and incubating them with Alexa Fluor 647-conjugated rat anti-mouse IL-17A (Cat.# 560184, clone TC11-18H10, BD Biosciences) and PE-conjugated rat anti-mouse Foxp3 (Cat.# 12-5773-82, clone FJK-16S, Invitrogen) for 40-50 min at 2-8 °C in the dark. Cells were then washed and resuspended in 0.35 mL of Stain Buffer for acquisition on a BD LSRFortessa flow cytometer. Flow cytometric data were analyzed and quantified using FlowJo software (version 10.8.1). Statistical analysis All quantitative data were expressed as mean ± standard deviation (SD). Statistical analyses were performed using GraphPad Prism version 9.5 and SPSS version 27. Comparisons between two groups were conducted using two-tailed unpaired Student's t-tests. In contrast, multiple-group comparisons were analyzed by one-way analysis of variance (ANOVA) followed by appropriate post hoc tests. Statistical significance was defined as follows: ns, not significant; ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001; and ∗∗∗∗P < 0.0001. Results and discussion Characterization of FK506 NLPs FK506 NLPs were successfully fabricated as described above, and the overall preparation workflow is schematically illustrated in Fig. 2a. The morphological features of the resulting nanoparticles were examined by TEM and cryo-TEM (Fig. 2b). Both imaging modalities consistently revealed that FK506 NLPs possessed a spherical to near-spherical morphology with smooth surfaces and well-defined boundaries, indicative of structurally intact and uniform nanoliposomes. Preparation and characterization of FK506 NLPs. (a) Schematic illustration of the FK506 NLP synthesis process. (b) TEM and cryo-TEM images showing the morphology of FK506 NLPs. Scale bars: 200 nm (left) and 100 nm (right). (c) Hydrodynamic size distribution based on particle density, and (d) zeta potential measurements comparing FK506 NLPs with ENLPs. n = 3. (e) In vitro release profile of FK506 from FK506 NLPs over time, demonstrating initial burst release followed by sustained drug retention. n = 3. (f) FTIR spectra depicting structural characteristics and intermolecular interactions of FK506 NLPs, ENLPs, and FK506-Free. (g) Quantitative analysis of relative fluorescence intensity at each time point post-application of Nile Red NLPs, or free dye, illustrating enhanced ocular surface retention. n = 3. (h) Representative real-time in vivo fluorescence images of mouse eyes after topical administration. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) Dynamic light scattering (DLS) analysis further characterized the physicochemical properties of the formulations (Fig. 2c). The mean hydrodynamic diameters of FK506 NLPs and ENLPs were 197.0 ± 0.55 nm and 186.5 ± 1.31 nm, respectively, with corresponding PDI values of 0.22 ± 0.01 and 0.14 ± 0.01. These results indicated a relatively narrow size distribution and good homogeneity. As shown in Fig. 2d, FK506 NLPs exhibited a positive zeta potential of +28.30 ± 0.32 mV, slightly higher than that of ENLPs (+24.34 ± 0.71 mV), reflecting enhanced colloidal stability and dispersibility attributable to the cationic lipid components. Furthermore, no obvious change in particle size or zeta potential of FK506 NLPs was observed over 72 h, indicating good stability (Fig. S1). The in vitro release behavior of FK506 from FK506 NLPs was assessed using a dynamic dialysis method (n = 3), and the resulting release profile is presented in Fig. 2e. The release kinetics displayed a characteristic biphasic pattern, consisting of an initial burst release followed by a prolonged sustained-release phase over 60 h, consistent with the expected behavior of controlled-release nanocarriers. Approximately 26.16% of the encapsulated drug was released within the first 12 h, after which drug liberation proceeded at a markedly slower rate, reaching a cumulative release of 68.64% by the end of the observation period. Fourier-transform infrared (FTIR) spectroscopy was employed to further elucidate the chemical structure and intermolecular interactions among FK506-Free, ENLPs, and FK506 NLPs over a wavenumber range of 4000-500 cm−1 (Fig. 2f). FK506-Free exhibited characteristic absorption bands, including a broad O–H stretching vibration at 3431.8 cm−1, C–H stretching vibrations at 2929.5 cm−1 and 2857.3 cm−1, and a pronounced C=O stretching band at 1731.1 cm−1, consistent with its macrolide lactone structure. ENLPs displayed typical lipid-associated peaks, with O–H stretching at 3432.1 cm−1, C–H stretching at 2925.4 cm−1 and 2857.3 cm−1, and a C=O stretching vibration at 1737.6 cm−1, confirming the presence of ester linkages within the lipid bilayer and the absence of FK506-related signals. In the FK506 NLPs spectrum, the O–H stretching band shifted to 3436.5 cm−1, suggesting the formation of new hydrogen-bonding interactions during drug encapsulation. The persistence of C–H stretching peaks at 2925.4 cm−1 and 2857.3 cm−1 indicated preservation of the lipid aliphatic chain structure. In comparison, the C=O stretching band shifted to 1738.5 cm−1, further confirming successful incorporation of FK506 into the lipid nanocarriers and strong non-covalent interactions, likely hydrogen bonding, between the drug and lipid components. Ocular surface retention of FK506 NLPs To evaluate the ocular surface retention behavior of FK506 NLPs, in vivo fluorescence imaging was performed in anesthetized mice (Fig. 2g and h). Eye drops containing either Nile Red NLPs or free Nile Red were topically instilled onto the ocular surface, and fluorescence signals were continuously monitored using a small-animal imaging system. Signal intensity was visualized using a red-to-blue pseudocolor scale to reflect relative fluorescence levels over time. At baseline (0 s), comparable fluorescence intensities were observed in both groups. However, as early as 5 min post-instillation, the residual fluorescence signal in the Nile Red NLPs group was markedly higher than that in the free Nile Red group. In mice receiving free Nile Red, fluorescence intensity declined rapidly between 5 and 25 min and was nearly undetectable by 25 min. In contrast, fluorescence signals in the Nile Red NLPs group remained largely stable throughout the observation period. A slight increase was even observed at 25 min. This may reflect progressive accumulation mediated by electrostatic interactions with ocular surface mucins. ROI analysis (Fig. 2g) supported this result. Fluorescence intensity in the Nile Red NLPs group was approximately 1.4-fold and 2.0-fold higher than that in the free dye group at 15 and 25 min, respectively. Collectively, these results indicated that FK506 NLPs substantially prolonged ocular surface residence compared with FK506-Free eye drops. This enhanced retention was likely attributable to stable electrostatic interactions between the positively charged nanoliposomes and negatively charged mucins on the ocular surface, which extended residence time, facilitated corneal permeation, and ultimately improved local drug bioavailability [36]. Biocompatibility evaluation of FK506 NLPs While cationic liposomes offer significant advantages for ocular drug delivery, their permanent or pH-dependent positive charge may pose potential safety risks, including disruption of cellular and nuclear membranes, increased reactive oxygen species (ROS) generation, and DNA damage [37]. Therefore, a thorough evaluation of FK506 NLPs' biocompatibility is essential. The Cell Counting Kit-8 (CCK-8) assay and Live/Dead staining demonstrated that FK506 NLPs showed no significant cytotoxicity toward human corneal epithelial cells (HCECs) at concentrations up to 20 μg/mL (Figs. S2 and S3). As shown by the CCK-8 assay (Fig. S2), cell viability in the FK506 NLP-treated groups remained above 90%. Consistently, Live/Dead staining (Fig. S3) revealed strong green fluorescence, indicating viable cells, with almost no red fluorescence, suggesting minimal cell death after FK506 NLP treatment. Mice received topical administration of 10 μL of 0.9% saline (w/v), FK506-Free (0.2 mg/mL), Talymus® (1 mg/mL), or FK506 NLPs (FK506-equivalent concentration 0.2 mg/mL) twice daily for 3 days. Slit-lamp examination under diffuse illumination (Fig. S4) revealed smooth, transparent corneal surfaces in all groups, with no evidence of congestion, epithelial defects, inflammation, or neovascularization. Corneal fluorescein staining showed no punctate or patchy green fluorescence, indicating intact epithelial integrity. Histological analysis with hematoxylin and eosin (H&E) staining confirmed normal corneal architecture, including well-organized epithelial, stromal, and endothelial layers, and no inflammatory infiltration or structural abnormalities. Anterior segment optical coherence tomography (AS-OCT) further demonstrated normal corneal morphology and stable central thickness across all groups, with no signs of edema or deformation. Systemic safety was evaluated by H&E staining of major organs, including the heart, liver, spleen, lungs, and kidneys, which revealed no histopathological alterations (Fig. S5). Moreover, body weight remained stable over the 4-day observation period, with no significant differences among treatment groups (Fig. 3f). These results collectively indicated that FK506 NLPs exhibited excellent ocular and systemic biocompatibility, supporting their safe use in therapeutic applications. Therapeutic effects of FK506 NLPs on ocular surface injury in DED. (a) Schematic illustration of DED mouse model establishment, treatment regimens, and subsequent evaluation timeline. (b) Representative fluorescein staining images of corneal surfaces across different treatment groups over three consecutive days. (c) Clinical scoring of corneal damage based on a standardized 0-4 grading scale. (d) Tear secretion measurements using the phenol red thread test, and (e) TBUT recorded over 3 days. (f) Daily body weight monitoring of DED mice subjected to various treatments. n = 8. ∗∗P < 0.01; ∗∗∗P < 0.001. (g) AS-OCT images of the cornea and (h) corresponding quantitative analysis of central corneal thickness. Blue arrows indicate detached corneal epithelium. Scale bar = 50 μm. n = 5. ∗∗∗∗P < 0.0001. (i) H&E-stained corneal sections and (j) quantification of corneal epithelial thickness. Blue arrows indicate detached corneal epithelium. Scale bar = 50 μm. n = 5. ns, P > 0.05; ∗∗∗P < 0.001; ∗∗∗∗P < 0.0001. (k) PAS-stained images depicting conjunctival structure and goblet cell counts. Orange triangles indicate goblet cells. Scale bar = 50 μm. (l) H&E-stained histological images of the lacrimal glands. Green triangles indicate structural abnormalities accompanied by lymphocyte infiltration. Scale bar = 50 μm. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) Therapeutic effects of FK506 NLPs on ocular surface injury in DED BAK is widely employed to model DED because it reproducibly induces ocular surface epithelial injury and impairs lacrimal gland function, recapitulating hallmark features of DED such as ocular surface inflammation, decreased tear secretion, and corneal epithelial damage [38]. Notably, adoptive transfer of CD4+ T cells from BAK-exposed donors into T cell-deficient nude mice is sufficient to induce a comparable DED phenotype, accompanied by conjunctival inflammatory infiltration and elevated IFN-γ and IL-17 levels in draining lymph nodes [39]. These findings highlight the central role of CD4+ T cells in BAK-induced pathology and validate this model as a robust platform for investigating immune mechanisms and evaluating therapeutic interventions. Accordingly, mice were treated with 0.2% BAK to induce DED and were randomly assigned to four groups: Saline, FK506-Free, Talymus®, and FK506 NLPs (Fig. 3a). Fluorescein sodium staining under cobalt blue light was used to assess corneal epithelial integrity, with scores assigned on a standardized 0-4 scale. As shown in Fig. 3b and c, Saline-treated mice exhibited severe corneal damage, characterized by diffuse punctate or confluent staining with scores exceeding 10. FK506-Free and Talymus® treatments produced partial improvement, yet residual green fluorescence was evident. In contrast, FK506 NLPs markedly reduced staining, with scores declining to 4.4 by day 3, indicating substantially enhanced epithelial recovery. Tear secretion, measured by the phenol red thread test (Fig. 3d), was slightly improved in the FK506-Free group (1.68 ± 0.19 mm to 2.08 ± 0.13 mm) and moderately enhanced in the Talymus® group (1.66 ± 0.23 mm to 2.67 ± 0.42 mm). FK506 NLPs treatment produced a pronounced increase in tear volume, from 2.06 ± 0.16 mm to 3.90 ± 0.33 mm, indicating superior stimulation of lacrimal secretion. Tear film stability, evaluated by TBUT (Fig. 3e), increased modestly in FK506-Free (0.56 ± 0.06 s to 0.99 ± 0.56 s) and Talymus® groups (0.36 ± 0.04 s to 1.27 ± 0.11 s), whereas FK506 NLPs substantially prolonged TBUT from 0.29 ± 0.04 s to 1.80 ± 0.10 s, reflecting improved tear film stability and reduced evaporation. Corneal sensitivity was further evaluated using the modified Cochet-Bonnet filament as described in the methods section [40]. The BAK group showed a marked reduction in corneal sensitivity threshold compared with the Normal group (0.25 ± 0.42 cm vs. 2.17 ± 0.52 cm, P < 0.0001), indicating impaired corneal sensory function after DED induction. Treatment with FK506 NLPs significantly increased the sensitivity threshold in BAK-treated mice (1.67 ± 0.52 cm vs. 0.25 ± 0.42 cm, P = 0.0001) (Fig. S6). In vivo confocal microscopy was further used to examine microstructural changes in the corneal stroma and endothelium (Fig. S7). Compared with the normal groups, the BAK group showed evident abnormalities in both layers. In the stroma, multiple activated dendritiform cells were observed, together with thin, disorganized corneal nerve fibers and increased branching, indicating marked stromal disturbance under DED conditions. In the endothelial layer, visualization was partially limited by corneal opacity, but the observable endothelial cells showed loss of normal morphology. After treatment with FK506 NLPs, these stromal and endothelial changes were visibly alleviated, and corneal microstructure appeared closer to that of the normal controls. After 3 days of treatment, AS-OCT and H&E staining were performed to evaluate corneal structure and cellular morphology (Fig. 3g–i, and S8). Saline-treated eyes exhibited significant epithelial thinning, structural deformation, and hyperreflective punctate infiltration. FK506 NLPs preserved epithelial integrity, restored cell alignment, mitigated stromal disruption, and reduced inflammatory infiltration. Quantitative analysis (Fig. 3h–j) revealed that corneal epithelial thickness (31.74 ± 2.44 μm) and central corneal thickness (220.60 ± 6.76 μm) in the FK506 NLPs group were comparable to those of healthy controls. Periodic Acid-Schiff (PAS) staining was employed to assess conjunctival goblet cells (Fig. 3k), which contribute to tear film stability via mucin secretion. Saline-treated mice displayed a marked reduction in goblet cell number and irregular morphology. FK506-Free and Talymus® partially restored cell density and shape, whereas FK506 NLPs significantly improved goblet cell structure, yielding plump, oval-shaped cells with well-defined borders and numbers approaching healthy levels. Lacrimal gland architecture and weight were evaluated to gauge secretory capacity (Fig. 3l and S9). Healthy controls exhibited uniform acinar size, compact cellular arrangement, and well-preserved glandular architecture. Saline-treated glands showed pronounced vacuolar degeneration, disrupted acinar alignment, and lymphocyte infiltration. FK506 NLPs markedly restored glandular structure, with uniform acinar size, minimal vacuolation, reduced inflammatory infiltration, and restored weight (17.64 ± 0.62 mg). Collectively, these results demonstrated that FK506 NLPs effectively promoted ocular surface repair, restored tear secretion and stability, preserved goblet cell density, and protected lacrimal gland structure, indicating superior therapeutic efficacy over conventional FK506 formulations in the treatment of DED. Therapeutic effects of FK506 NLPs on ocular surface inflammation in DED Inflammation is a central driver of DED pathogenesis. Matrix metalloproteinase-9 (MMP-9) is a critical biomarker of DED, specifically upregulated under hyperosmotic stress, where it disrupts tight junctions in corneal epithelial cells and compromises barrier integrity [40,41]. IL-1β, an upstream mediator of the inflammatory cascade, activates the NF-κB pathway and promotes the release of downstream pro-inflammatory cytokines, including TNF-α, IL-6, and MMP-9 [42]. Building on the previously demonstrated therapeutic effects of FK506 NLPs on ocular surface repair, we further evaluated their anti-inflammatory efficacy by measuring IL-1β and MMP-9 expression in both corneal and lacrimal gland tissues. qRT-PCR analysis of corneal tissue (Fig. 4a and b) showed that FK506 NLPs significantly suppressed IL-1β and MMP-9 expression compared with Saline-treated controls (P < 0.0001). FK506 NLPs exhibited stronger inhibition than FK506-Free for IL-1β (P < 0.001) and MMP-9 (P < 0.05), and more effectively reduced MMP-9 relative to Talymus® (P < 0.01). In lacrimal glands (Fig. 4c and d), both cytokines were substantially elevated in the Saline group. Treatment with FK506 NLPs markedly downregulated their expression (P < 0.0001), demonstrating greater efficacy than FK506-Free (P < 0.01 for both) and Talymus® (P < 0.001 for MMP-9). Therapeutic effects of FK506 NLPs on ocular surface inflammation in DED. (a-d) Relative mRNA expression levels of IL-1β and MMP-9 in the cornea (a, b) and lacrimal glands (c, d) of DED mice following different treatments. n = 3. ns, P > 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001; ∗∗∗∗P < 0.0001. Immunofluorescence images showing (e) IL-1β (green), (g) MMP-9 (green), and (i) CK-10 (red) in corneal sections, with nuclei counterstained with DAPI (blue). Scale bar = 50 μm. Quantitative analyses of fluorescence intensity for (f) IL-1β, (h) MMP-9, and (j) CK-10. n = 5. ns, P > 0.05; ∗∗∗∗P < 0.0001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) Immunofluorescence staining corroborated these transcriptional findings (Fig. 4e–g). IL-1β and MMP-9 signals were markedly increased in Saline-treated DED eyes, moderately reduced by FK506-Free and Talymus®, and substantially suppressed by FK506 NLPs. Quantitative analysis (Fig. 4f–h) revealed that FK506 NLPs decreased IL-1β by 82.3% and MMP-9 by 57.5% relative to the Saline group, highlighting their potent anti-inflammatory activity in mitigating corneal inflammation and restoring ocular surface homeostasis. Chronic ocular surface inflammation in DED promotes structural alterations, including epithelial keratinization and squamous metaplasia [43]. Cytokeratin-10 (CK-10), an intermediate filament protein expressed in differentiated squamous epithelium, serves as a key biomarker for keratinization [44]. Under healthy conditions, CK-10 expression is minimal, but it is markedly upregulated in response to DED-induced epithelial stress. Immunofluorescence analysis (Fig. 4i and j) revealed pronounced CK-10 expression in the Saline-treated group (72.28 ± 0.18 a.u.), significantly exceeding all treatment groups (P < 0.0001). FK506-Free and Talymus® partially reduced CK-10 expression (P < 0.01 vs. Saline), whereas FK506 NLPs achieved the lowest levels (14.36 ± 2.32 a.u.), not significantly different from healthy controls. These findings demonstrated that FK506 NLPs effectively reversed DED-induced corneal squamous metaplasia and restored epithelial homeostasis, complementing their anti-inflammatory and tissue-protective effects. FK506 NLPs restore Th17/Treg homeostasis in DED Immune dysregulation is a key driver of chronic DED progression [45]. Th17 cells exacerbate corneal tissue damage by secreting pro-inflammatory cytokines, including IL-17 and IL-23, whereas Tregs maintain immune homeostasis via anti-inflammatory mediators such as IL-10 and TGF-β. Moreover, experimental evidence suggests that Th17 responses can be relatively resistant to Treg-mediated suppression in DED. An imbalance between Th17 and Treg populations, coupled with sustained local inflammation, perpetuates ocular surface damage and disease progression [47]. To investigate the immunomodulatory effects of FK506 NLPs, flow cytometry was conducted on cervical lymph nodes following three days of treatment (Fig. 5a–d). In the Saline-treated DED group, the proportion of CD4+ IL-17A+ Th17 cells was markedly elevated, reaching 3.95 ± 0.09%, compared with 1.61 ± 0.02% in healthy controls. Administration of FK506-Free and Talymus® partially mitigated this increase, reducing Th17 cell proportions to 2.59 ± 0.06% and 1.82 ± 0.10%, respectively. FK506 NLPs, however, exhibited the most pronounced suppressive effect, lowering Th17 cell frequency to 1.77 ± 0.07%, which was significantly decreased relative to both the Saline and FK506-Free groups (P < 0.0001), and was comparable to that observed with Talymus®(Fig. 5a and b). FK506 NLPs modulate Th17/Treg balance and cytokine expression in DED. (a-d) Representative flow cytometry plots and corresponding quantitative analysis of CD4+ IL-17A+ Th17 cells and CD4+ CD25+ Foxp3+ Treg cells in the cervical lymph nodes. n = 3. ns, P > 0.05; ∗∗∗P < 0.001; ∗∗∗∗P < 0.0001. (e-h) Relative mRNA expression levels of IL-17A (e), RORγt (f), TGF-β (g), and Foxp3 (h) in the corneas of DED mice following different treatments. n = 3. (i-l) Relative mRNA expression levels of IL-17A (i), RORγt (j), TGF-β (k), and Foxp3 (l) in the lacrimal glands of DED mice after the indicated treatments. n = 3. ns, P > 0.05; ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001; ∗∗∗∗P < 0.0001. Concurrently, the population of CD4+ CD25+ Foxp3+ Treg cells was assessed (Fig. 5c and d). The Saline, FK506-Free, and Talymus® groups exhibited Treg proportions of 3.55 ± 0.06%, 4.06 ± 0.12%, and 4.56 ± 0.20%, respectively. FK506 NLPs significantly augmented the Treg population to 5.92 ± 0.18%, representing a statistically significant increase over all other treatment groups (P < 0.001), thereby highlighting its enhanced capacity to promote immunoregulatory responses. To further elucidate the impact of FK506 NLPs on Th17/Treg balance, qRT-PCR analyses were performed on corneal, lacrimal gland, and lymph node tissues. In the cornea (Fig. 5e–h and S10a-b), BAK-induced DED resulted in pronounced upregulation of Th17-associated markers, including IL-17, IL-23, and the transcription factor RORγt, which were only moderately reduced by FK506-Free or Talymus® treatments. Treatment with FK506 NLPs, however, substantially suppressed these Th17-associated markers, restoring them to near-control levels, while simultaneously enhancing Treg-associated markers (IL-10, TGF-β, and Foxp3) that were significantly downregulated in the Saline group. Comparable trends were observed in lacrimal glands (Fig. 5i–l and S10c-d) and cervical lymph nodes (Fig. S11a–f), where FK506 NLPs effectively attenuated Th17-associated expression and upregulated Treg-related markers, demonstrating robust immunomodulatory effects across multiple tissues. Immunofluorescence analysis further corroborated these observations. Dual-label staining localized CD4+ IL-17+ and CD4+ Foxp3+ cells within the cornea and lacrimal glands. In the cornea (Fig. 6a–c), DED induction in Saline-treated mice significantly increased CD4 and IL-17 fluorescence (76.95 ± 3.87 a.u. and 65.30 ± 6.05 a.u., respectively). FK506 NLPs treatment markedly suppressed both markers (CD4: 21.8 ± 2.92 a.u.; IL-17: 21.78 ± 2.17 a.u.), approaching levels observed in healthy controls. Quantitative analysis revealed reductions of 55.15% for CD4+ cells and 43.52% for IL-17+ cells. CD4+ Foxp3+ expression (Fig. 6d–f) was also favorably modulated; FK506 NLPs significantly reduced CD4 levels while restoring Foxp3 expression to 25.87 ± 3.56 a.u., corresponding to a 25.9% decrease in CD4 and an 11.1% increase in Foxp3 relative to Saline-treated mice. FK506 NLPs modulate Th17/Treg balance and fluorescence intensity in DED. (a) Representative immunofluorescence images of corneal tissue stained for CD4 (red) and IL-17A (green). (b, c) Quantitative analysis of fluorescence intensities for CD4 and IL-17A in the cornea. (d) Representative immunofluorescence images of the cornea stained for CD4 (red) and Foxp3 (green). (e, f) Corresponding quantitative analysis of CD4 and Foxp3 fluorescence intensities in the cornea. (g) Representative immunofluorescence images of the lacrimal gland stained for CD4 (red) and IL-17A (green). (h, i) Quantitative analysis of CD4 and IL-17A fluorescence intensities in the lacrimal gland. (j) Representative immunofluorescence images of the lacrimal gland stained for CD4 (red) and Foxp3 (green). (k, l) Corresponding quantitative analysis of CD4 and Foxp3 fluorescence intensities in the lacrimal gland. Nuclei were counterstained with DAPI (blue). Merged images are shown at the bottom of each panel. Scale bars = 50 μm. n = 10. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) In the lacrimal glands (Fig. 6g–l), both CD4 and IL-17 fluorescence were markedly elevated in the Saline group and only modestly reduced by FK506-Free or Talymus®. By contrast, FK506 NLPs significantly attenuated these pro-inflammatory signals, reducing CD4 and IL-17 fluorescence by 38.11% and 24.4%, respectively. CD4+ Foxp3+ analysis revealed a pronounced decrease in CD4 and an increase in Foxp3 following FK506 NLP administration, with a 41.2% reduction in CD4 and a 14.1% rise in Foxp3 compared with Saline-treated mice. FK506 may regulate Foxp3 transcription in a concentration-dependent manner via NFAT-associated mechanisms [20]. Therefore, the prominent Treg-associated signals observed with FK506 NLPs are more likely related to improved local delivery and sustained low-level exposure, which may maintain effective immunomodulation while reducing large local concentration fluctuations. Under this condition, FK506 may suppress Th17-associated inflammatory responses without causing excessive local exposure that could be unfavorable for Treg stability and function. At the same time, we cannot exclude that avoiding chronic high local FK506 exposure may also help reduce potential suppression of Treg biology. We hope that this study will be followed by further work to identify the cellular and molecular determinants of the Th17/Treg shift induced by FK506 NLPs. Collectively, these results demonstrated that FK506 NLPs effectively restored Th17/Treg homeostasis in DED by suppressing pro-inflammatory CD4+ IL-17+ Th17 cells while simultaneously enhancing immunoregulatory CD4+ Foxp3+ Treg populations across both ocular surface and glandular tissues, underscoring their potent and systemic immunomodulatory potential for therapeutic intervention. Conclusion In conclusion, this study successfully engineered FK506 NLPs that markedly enhanced ocular surface adhesion via stable electrostatic interactions with mucins, thereby significantly prolonging drug retention time on the ocular surface. Our findings demonstrated that FK506 NLPs not only facilitated effective repair of ocular surface damage in DED but also robustly suppressed the IL-1β/MMP-9-mediated inflammatory cascade and restored Th17/Treg immune homeostasis by attenuating the expression of key pro-inflammatory cytokines. Compared with conventional FK506 formulations, this nanocarrier system exhibited superior therapeutic durability, enhanced bioavailability, and excellent biocompatibility. Collectively, these results provided novel mechanistic insights and a strong technological foundation for nanocarrier-based immunomodulatory interventions, underscoring their considerable research significance and promising translational potential for the clinical management of DED. Funding This work was supported by the National Natural Science Foundation of China (Grant No. 82271042). CRediT authorship contribution statement Lirui Liu: Conceptualization, Data curation, Formal analysis, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing. Dan Chen: Conceptualization, Methodology, Validation, Writing – original draft. Siting Sheng: Methodology. Xueqi Lin: Methodology. Huiling Zhao: Methodology. Chujun Liu: Visualization. Xiang Chen: Methodology. Yang Ye: Formal analysis, Methodology, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing. Wen Xu: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing – review & editing. Declaration of competing interest 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. Acknowledgements Not applicable. Footnotes This work was supported by the National Natural Science Foundation of China (Grant No. 82271042). Supplementary data to this article can be found online at https://doi.org/10.1016/j.mtbio.2026.103248. Contributor Information Yang Ye, Email: yeyang94@zju.edu.cn. Wen Xu, Email: xuwen2003@zju.edu.cn. Supplementary data The following is the Supplementary data to this article: Multimedia component 1 Data availability Data will be made available on request. References Associated Data Supplementary Materials Multimedia component 1 Data Availability Statement Data will be made available on request.