Amphotericin B-loaded polyphenolic nanoparticles for synchronous fungal eradication and immunomodulation in treating superficial fungal infections

Abstract

Fungal infection represents an escalating global health threat, yet effective treatment remains clinically challenging due to the poor solubility, limited bioavailability, and pronounced cytotoxicity of conventional antifungal agents. To overcome these limitations, a self-assembled nanoparticle (ATE NPs) is developed by integrating amphotericin B (AmB), epigallocatechin gallate (EGCG), and tri(phenylboronic acid)-derived crosslinker pentamethyldiethylenetriamine (T-PBA) for efficient fungal eradication and accelerated tissue repair. Benefiting from the strong affinity of T-PBA towards fungal cell wall polysaccharides, ATE NPs exhibit targeted adhesion and prolonged retention at infected lesions. Within the infection-associated mildly acidic microenvironment, the cleavage of boronate ester bonds triggers the controlled release of AmB, disrupting fungal membranes and eliminating pathogens. Meanwhile, EGCG exerts potent anti-inflammatory and pro-healing effects to promote tissue regeneration. Consequently, ATE NPs exhibit robust mucoadhesion, superior antifungal activity, and remarkable wound-healing capabilities both in vitro and in vivo. In murine models of fungal keratitis and cutaneous mycosis, the multifunctional nanosystem effectively suppresses fungal proliferation, mitigates corneal inflammatory damage and accelerates healing. Furthermore, transcriptomic analysis reveals that ATE NPs induce fungal metabolic dysregulation and oxidative stress while favorably modulating host immune responses. Overall, this work provides a promising, biocompatible nanotherapeutic strategy for the integrated management of severe superficial fungal infections.

Graphical abstract

1. Introduction

Fungal infection represents an escalating global threat to public health, carrying immense clinical and economic burdens. Fungal keratitis, as an example, has become one of the most common causes of blindness worldwide, while cutaneous mycosis affects a quarter of the world population [[1], [2], [3]]. Despite the severity, developing effective antifungal strategies remains a formidable challenge. Because fungi are eukaryotic organisms, they share high structural and metabolic similarities with mammalian cells, which severely limits the availability of pathogen-specific drug targets and stifles the discovery of next-generation, resistance-defying therapies [4]. Furthermore, fungal infections are intimately intertwined with overwhelming inflammation. Therefore, dual-action therapies capable of simultaneously eradicating the pathogen and modulating the host immune response are highly desired to prevent tissue damage and accelerate delayed healing [5].

To date, clinical interventions for invasive fungal infections remain restricted to just four primary classes of antifungal antibiotics: polyenes, azoles, echinocandins, and the pyrimidine analog 5-fluorocytosine. Among these, the polyene macrocyclic antibiotic amphotericin B (AmB) remains the undisputed “gold standard” owing to its exceptional broad-spectrum efficacy and low propensity for inducing drug resistance [6]. Structurally, AmB exhibits a unique amphiphilic and amphoteric nature, allowing high affinity for fungal membrane ergosterol. It induces sterol sequestration and membrane permeabilization, which collapses cellular homeostasis and facilitates potent fungicidal action [7]. However, AmB lacks selectivity between fungal ergosterol and cholesterol in mammalian cells. This lack of specificity results in pronounced nephrotoxicity and other severe systemic adverse reactions, which heavily restrict its clinical utility [8].

On the other hand, AmB cannot effectively alleviate severe inflammation or promote tissue repair after pathogen eradication because it lacks inherent immunomodulatory and tissue-protective properties [6]. To address this limitation, natural polyphenols, such as epigallocatechin gallate (EGCG), offer a compelling therapeutic adjunct due to their potent antioxidant, anti-inflammatory, and tissue-regenerative properties [9]. Crucially, the robust radical-scavenging capacity effectively shields immune cells and host cells from oxidative stress, preserving their viability and homeostatic functions within highly inflamed microenvironments [10]. Interestingly, boronic acid derivatives can readily form dynamic boronate ester bonds with ortho- or meta-positioned dihydroxyl groups in polyphenols, while simultaneously exhibiting high affinity for cis-diol-rich polysaccharides abundant in fungal cell walls [11]. Therefore, engineering a dynamic boronate ester-crosslinked polyphenolic nanocomplex represents a promising approach to overcome the therapeutic bottlenecks of conventional AmB therapies, enabling synchronized fungal targeting, microenvironment-responsive drug delivery, and accelerated tissue repair.

In previous research, several phenylboronic acid (PBA)-containing systems have been developed for AmB delivery. PBA-modified polymeric micelles can bind the diol groups of AmB through dynamic boronate ester linkages, enabling pH-dependent drug release [12], whereas PBA-conjugated styrene-maleic acid copolymers have been used to improve AmB solubility, reduce hemolytic activity, and enhance mucin binding [13]. Multivalent PBA-containing telodendrimer prodrugs have also achieved responsive AmB delivery for systemic fungal infections [14]. Meanwhile, EGCG has been incorporated into antifungal nanomaterials, including EGCG-stabilized silver nanoparticles, while dynamic boronate-polyphenol interactions have been widely employed to construct pH-responsive capsules and antioxidant nanogels [[15], [16], [17]]. Nevertheless, these systems generally focus on AmB solubilization and toxicity reduction, or responsive polyphenol assembly as separate objectives.

Herein, building on established PBA-diol and polyphenol assembly strategies, we report a responsive self-assembled nanotherapeutic platform (ATE NPs) designed to improve fungal eradication and reduce the intrinsic cytotoxicity of AmB, with potent anti-inflammatory and tissue-regenerative capabilities. The nanoparticles were covalently crosslinked by tri(phenylboronic acid)-derived pentamethyldiethylenetriamine (T-PBA) and EGCG, enabling facile encapsulation of hydrophobic AmB with improved aqueous dispersion and stability. Meanwhile, the exposed boronic acid groups of T-PBA effectively adhere to cis-diol-rich polysaccharides from fungal cell walls, ensuring prolonged retention at infected lesions (Scheme 1). Upon exposure to the infection-associated microenvironment, which is characterized by local acidification [[18], [19], [20]] and elevated ROS levels [21,22] during fungal keratitis, the dynamic boronate ester linkages within ATE NPs undergo pH-responsive dissociation and ROS-mediated oxidative disruption [23,24]. This dual-responsive destabilization promotes the on-demand release of AmB, thereby disrupting fungal membranes and inducing fungal cell death [25], alongside the concurrent release of EGCG to suppress localized over-inflammation. The distinction of ATE NPs lies in the coordinated integration of these complementary functions within a single self-assembled nanoparticle, enabling fungal adhesion, responsive antifungal delivery, and host-tissue protection. In vitro evaluation demonstrated that ATE NPs exhibited potent fungicidal efficacy paired with favorable cytocompatibility and anti-inflammatory effects. Furthermore, ATE NPs significantly suppressed intraocular fungal proliferation in the murine fungal keratitis and cutaneous mycosis models, with attenuated inflammatory over-responses, promoted tissue regeneration, and restored corneal function. Collectively, by integrating targeted antifungal delivery, microenvironment-responsive drug release, and host immunomodulation, the ATE NPs significantly enhanced antifungal efficacy, prolonged adhesion capability, and tissue repair performance, thereby providing a promising therapeutic strategy for the effective management of superficial fungal infections.

Scheme 1.

Schematic illustration of the synthesis of ATE NPs and their proposed therapeutic mechanism against superficial fungal infections.

2. Materials and methods

2.1. Synthesis of ATE NPs

The synthesis of ATE NPs was carried out in two sequential stages. First, T-PBA was synthesized through the reaction of N,N,N′,N′,N″-pentamethyldiethylenetriamine with 4-(bromomethyl)phenylboronic acid. Briefly, N,N,N′,N′,N″-pentamethyldiethylenetriamine (0.17 g) was dissolved in N,N-dimethylformamide (DMF, 2 mL). 4-(bromomethyl)phenylboronic acid (0.90 g) was dissolved in DMF (3 mL). The two solutions were subsequently mixed and stirred at room temperature for 24 h. After completion of the reaction, the product was precipitated by the addition of ethyl acetate, collected by centrifugation, and washed three times with ethyl acetate. The resulting precipitate was then vacuum-dried to afford T-PBA as the final product.

Subsequently, ATE NPs were prepared via a one-pot self-assembly strategy using AmB, T-PBA, and EGCG. Specifically, AmB was dissolved in dimethyl sulfoxide (DMSO) to obtain a stock solution (20 mg mL⁻¹), while T-PBA, EGCG, and pentaerythritol were each dissolved in H₂O at a concentration of 20 mg mL⁻¹. For nanoparticle preparation, 300 μL of the AmB/DMSO solution was added to 12 mL of ultrapure water and sonicated for 30 s to yield a clear light-yellow solution. Subsequently, 600 μL of EGCG aqueous solution and 480 μL of T-PBA aqueous solution were added dropwise under continuous stirring. After stirring for 2.5 h, the resulting suspension was collected by centrifugation, washed thoroughly, and then ultrasonically redispersed in 10 mL of water to obtain ATE NPs.

2.2. Preparation of ATP NPs, ETP NPs, and AmB + EGCG solution

ATP NPs were prepared as follows: 300 μL of AmB/DMSO solution was added to 12 mL of ultrapure water and sonicated for 30 s. Subsequently, 600 μL of pentaerythritol aqueous solution and 480 μL of T-PBA aqueous solution were added under continuous stirring. After stirring for 2.5 h and washing three times, ATP NPs were obtained.

For the preparation of ETP NPs, 600 μL of EGCG aqueous solution was added to 12 mL of ultrapure water and mixed thoroughly, followed by the addition of 480 μL of T-PBA aqueous solution and 600 μL of pentaerythritol aqueous solution under stirring. After stirring for 2.5 h and washing three times, ETP NPs were obtained directly.

For the preparation of the AmB + EGCG solution, 300 μL of AmB/DMSO solution was first added to 12 mL of ultrapure water and sonicated for 30 s, after which 600 μL of EGCG aqueous solution was introduced under stirring. The mixture was stirred for 2.5 h to obtain the AmB + EGCG solution.

2.3. Determination of AmB content and pH-Responsive release

The AmB contents in ATE NPs and ATP NPs were determined using UV-vis spectroscopy. Briefly, 100 μL of each nanoparticle dispersion was mixed with an equal volume of 0.5 mM HCl solution to completely disassemble the nanoparticles and release the incorporated AmB. The absorbance of each sample was measured at 340 nm, and the AmB concentration was calculated using a pre-established calibration curve over the concentration range of 0.40 to 100 μg mL⁻¹.

The pH-responsive release of AmB from ATE NPs was evaluated using a dialysis method. One milliliter of ATE NPs dispersion was transferred into a dialysis bag with a molecular weight cutoff of 10 kDa and immersed in 10 mL of release medium containing 0.1% Tween 80. The release media consisted of acetate-sodium acetate buffer at pH 7.4 or pH 6.0 mixed with absolute ethanol at a volume ratio of 4:6. At predetermined time points of 100, 200, 300, 400, 500, 600, and 720 min, 100 μL of the release medium was withdrawn and immediately replaced with an equal volume of fresh medium. The amount of released AmB was quantified by measuring the absorbance at 340 nm using the corresponding calibration curve. The cumulative release percentage was calculated relative to the initial AmB content in the ATE NPs, with correction for the amount of AmB removed during repeated sampling.

2.4. Adhesion of ATE NPs on the fungal surface

First, Hoechst staining solution was added to 1 mL of C. albicans suspension (10⁷ CFU mL⁻¹, dispersed in H₂O) and incubated for 20 min. The stained fungal cells were then washed three times with physiological saline to remove residual dye, yielding Hoechst-labeled C. albicans.

Separately, 15 mg of AmB was dissolved in 300 μL of DMSO, followed by the addition of 8.5 mg of Rhodamine B (RhoB) under continuous stirring. After stirring for 24 h, ethyl acetate was added to precipitate the product. The precipitate was collected by centrifugation and washed three times with ethyl acetate to remove excess unbound RhoB. Next, the RhoB-stained AmB was used to prepare the ATE NPs and ATP NPs.

Subsequently, 200 μL of Hoechst-stained C. albicans was separately incubated with 200 μL of RhoB-labeled AmB (4 μg mL⁻¹), ATP NPs and ATE NPs (AmB equivalent: 4 μg mL⁻¹). The interactions between AmB, ATP NPs, ATE NPs and C. albicans were then characterized by confocal laser scanning microscopy.

2.5. Evaluation of In vitro antifungal activity

200 μL of C. albicans suspension (10⁷ CFU mL⁻¹, dispersed in H₂O) was mixed with 200 μL of different treatments, including H₂O (control), ETP NPs, free drugs (AmB + EGCG), ATP NPs, and ATE NPs. The concentrations of all formulations were normalized to AmB equivalents at 1, 2, 4, 8, and 16 μg mL⁻¹. The mixtures were incubated at 37 °C with shaking for 4 h. The surviving fungal cells were then quantified using the standard plate counting method. Each experimental condition was performed in triplicate.

2.6. Crystal violet staining and live/dead biofilm assay

200 μL of C. albicans suspension (10⁵ CFU mL⁻¹, dispersed in SDB) was added to a 96-well plate and incubated at 37 °C for 6 h. The culture medium was then carefully removed, and the wells were treated with SDB (control), ETP NPs, free drugs (AmB + EGCG), ATP NPs, and ATE NPs (AmB equivalent: 16 μg mL⁻¹), respectively. After 48 h of incubation, planktonic fungi were removed, and the wells were washed three times with saline. The biofilms were then stained with crystal violet solution (1 mg mL⁻¹, 200 μL per well) for 15 min, followed by three washes with saline. After air-drying, the stained biofilms were imaged. For semi-quantitative analysis, 200 μL of 95% ethanol was added to each well and incubated for 30 min to solubilize the dye. The solution was then transferred to a new 96-well plate, and the optical density at 570 nm was measured using a microplate reader. Eight replicates were performed for each group.

In a separate experiment, fungal cells were cultured in confocal dishes under identical conditions and incubated at 37 °C for 48 h. After washing the biofilms three times with saline, a live/dead biofilm viability kit was used for staining according to the manufacturer's instructions. Three-dimensional images were acquired using an SP8 confocal laser scanning microscope (Leica, Wetzlar, Germany).

To evaluate the disruption of mature biofilms, pre-formed biofilms were established by culturing C. albicans in SDB for 48 h, followed by treatment with the same formulations for 24 h. The biofilm-disrupting efficacy was assessed using crystal violet staining and live/dead confocal imaging.

2.7. Measurement of intracellular ROS levels in HCECs

HCECs were seeded in a 12-well plate (10⁶ cells per well) for 24 h for adhesion. Then the media were replaced and the cells were treated with different drugs (AmB equivalent: 1.5 μg mL⁻¹) 1 h before H₂O₂ (250 μM) added. The cells were incubated for another 2 h. Subsequently, DCFH-DA (1:1000) and Hoechst 33342 (1:1000) solution in DMEM without phenol red or FBS was utilized to dye the cells for 30 min. The ROS levels were evaluated by fluorescence microscopy.

2.8. In vivo establishment and treatment of C. Albicans-infected keratitis model

Female C57BL/6 mice (8 weeks old, 16 ± 1.2 g) were chosen to evaluate in vivo treatment effect of C. albicans-infected keratitis. The mice were anesthetized with Nembutal (intraperitoneal injection, 60 mg kg⁻¹) and proparacaine hydrochloride (eye drops). The corneal epithelium layer was removed and 20 μL of C. albicans suspension (10⁷ CFU mL⁻¹) was injected onto the cornea. After infection for 24 h, the C. albicans-infected keratitis models were established and randomly divided into six groups and topically treated with 20 μL H₂O (control), ETP NPs, free drugs (AmB + EGCG), ATP NPs, ATE NPs, and AmBisome® (AmB equivalent: 12.8 μg mL⁻¹) on Day 1 and Day 2. The anterior segment images were recorded through the slit lamp image system (YZ5T, 6 6 VISION TECH Co., Ltd., China) on day 0, 1, 2, 5, 9, and 15. The area of the corneal epithelium defect was stained by 0.5% sodium fluorescein (FLS) on day 15. Then the clinical grading scale was evaluated by independent, masked observers according to five items (edema, surface regularity, neovascularization, opacity area, and opacity density) each scored from 0 to 3. Fluorescein staining was quantified by dividing the cornea into four quadrants. Each quadrant was graded on a scale of 0 to 4, where 0 indicating no staining, 1 indicating diffuse dot-like green staining, 2 indicated a green-staining area covering <1/3 of the quadrant, 3 indicated a green-staining area covering >1/3 of the quadrant, and 4 indicated a green-staining area covering >2/3 of the quadrant. The total fluorescein score was obtained by summing the scores from all four quadrants, yielding a final score ranging from 0 to 16. Non-invasive anterior segment optical coherence tomography (AS-OCT, Phoenix Micron Ⅳ, Phoenix-Micron Inc., USA) was used to monitor internal structural alterations and to quantify central corneal thickness as an indicator of inflammatory edema. In vivo confocal microscopy (Spectralis HRA MEL 90, Heidelberg, Germany) was further employed to directly visualize the stromal hyphae. To quantify the in vivo fungicidal potency, representative corneas were excised, homogenized in sterile PBS, and cultured on SDA plates for CFU counting. Body weight was measured on days 0, 1, 2, 5, 9, and 15. On day 15, the mice were euthanized and eyes were excised for further etiological and pathological analysis. Five types of major organs, including the heart, lung, liver, spleen, and kidney, were excised for pathological analysis. Ocular samples were fixed in 4% paraformaldehyde, dehydrated, embedded in paraffin, and sectioned into 4 μm slices for further analysis.

2.9. In vivo establishment and treatment of C. Albicans-infected skin wounds model

Female C57BL/6 mice (8 weeks old, 14 ± 1.4 g) were chosen to evaluate in vivo treatment effect of C. albicans-infected skin wounds. The mice were anesthetized with Nembutal (intraperitoneal injection, 60 mg kg⁻¹) to establish the skin wounds model. A full-thickness circular skin wound was prepared on the back of the mouse using the scalpel and a total of 20 μL of C. albicans suspension (10⁷ CFU mL⁻¹) was inoculated onto the wound area. After infection for 24 h, the C. albicans-infected skin wounds models were established and randomly divided into five groups and topically treated with 50 μL H₂O (control), ETP NPs, free drugs (AmB + EGCG), ATP NPs, and ATE NPs (AmB equivalent: 12.8 μg mL⁻¹) on Day 1 and Day 2. The images of the skin wounds and the change of body weight were observed on day 0, 1, 2, 4, 6, and 8. Then the mice were euthanized and skin tissues were excised for further etiological and pathological analysis with CFUs counting.

2.10. Statistical analyses

The results were presented as the mean ± standard deviation (SD). Statistical analysis was performed using the GraphPad Prism or Origin software. An independent-sample t-test was used to compare between two groups. For the comparison among (more than) three groups, comparative studies of means were carried out using one-way ANOVA analysis with Tukey's post hoc test.

3. Results

3.1. Synthesis and characterization of ATE NPs

ATE NPs were prepared via a simple self-assembly strategy. Specifically, T-PBA was first synthesized through the reaction between N,N,N′,N″,N″-pentamethyldiethylenetriamine and 4-(bromomethyl)phenylboronic acid (Fig. S1). Subsequently, EGCG was introduced into the AmB solution, followed by the dropwise addition of T-PBA solution under continuous stirring (Fig. 1a). The boronic acid moieties of T-PBA reversibly form boronate ester linkages with the 1,3-diol groups of AmB, as supported by ¹H NMR analysis (Fig. S2) and further corroborated by an Alizarin Red S competitive displacement assay (Fig. S3), and with the catechol/pyrogallol hydroxyl groups of EGCG [12,16], thereby driving the spontaneous self-assembly of the components into water-dispersible ATE NPs. Transmission electron microscopy (TEM) and high-resolution TEM (HR-TEM) images revealed that ATE NPs exhibited a uniform spherical morphology with an average diameter of approximately 100 nm (Fig. 1b and c). Owing to the presence of three positively charged units within each T-PBA molecule, ATE NPs possessed a net cationic surface (Fig. 1d), which further contributed to their good dispersion stability in aqueous solution through electrostatic repulsion. Meanwhile, the nanoparticles containing only AmB (ATP NPs) were synthesized by the self-assembly of AmB, T-PBA, and pentaerythritol (Fig. S4). The nanoparticles containing only EGCG (ETP NPs) were prepared by the self-assembly of EGCG, T-PBA, and pentaerythritol (Fig. S5). TEM imaging showed that both ATP NPs and ETP NPs exhibited spherical morphologies, with intensity-average hydrodynamic diameters of 119.03 ± 3.49 nm and 44.65 ± 0.50 nm, respectively (Fig. 1c). All three types of nanoparticles showed positive surface charges, among which ETP NPs displayed the highest zeta potential (+27.2 ± 2.8 mV), compared with ATE NPs (+16.3 ± 1.0 mV) and ATP NPs (+6.11 ± 0.62 mV) (Fig. 1d). This difference is attributed to the abundant vicinal diol groups provided by EGCG and pentaerythritol, which competitively modulate the surface shielding and charge exposure of the T-PBA core.

Fig. 1.

Synthesis and characterization of ATE NPs. (a) Preparation of ATE NPs by AmB, T-PBA, and EGCG. Created in BioRender. (2026) https://BioRender.com/6ox3epv. (b) TEM images of ATE NPs. Inset: HR-TEM image of ATE NPs. Scale bar: 200 nm for TEM, and 50 nm for HR-TEM. (c) Intensity-average hydrodynamic diameters of ATE NPs, ATP NPs, and ETP NPs. (d) Zeta potential and PDI of ATE NPs, ATP NPs, and ETP NPs. (e) TEM mapping analysis of ATE NPs with different colored dots representing C, N, O, and B. Scale bar: 500 nm. (f) FT-IR spectra of ATE NPs, AmB, EGCG, and T-PBA. (g) UV-vis absorption spectra of ATE NPs, AmB, and EGCG. (h) Scavenging rates of ABTS after exposure to ATE NPs at various concentrations. (i) Scavenging rates of DPPH after exposure to ATE NPs at various concentrations. (j) Release curves of AmB from ATE NPs over time in HAc-NaAc buffer at different pH.

To verify the successful formation of ATE NPs, energy-dispersive X-ray spectroscopy (EDS) mapping was performed. As shown in Fig. 1e, the mapping images revealed a homogeneous distribution of C, O, N, and B elements throughout the nanoparticle. Furthermore, Fourier-transform infrared (FT-IR) spectroscopy was performed to confirm the successful incorporation of AmB. As shown in Fig. 1f, AmB, EGCG, and T-PBA exhibited typical O-H stretching vibrations in the 3350-3420 cm⁻¹ range. Notably, ATE NPs displayed a broadened and slightly red-shifted O-H band at 3389 cm⁻¹, suggesting the formation of an extensive hydrogen-bonding network. Crucially, a new absorption band emerged at 1011 cm⁻¹, which was assigned to B-O-C stretching vibrations. Together with the noticeable shifts in the aromatic ring (1610 cm⁻¹) and B-O bands (1315-1388 cm⁻¹), these spectral results clearly validated the formation of boronate ester linkages. These intermolecular interactions were further corroborated by UV-vis spectroscopy (Fig. 1g). The characteristic absorption peak of AmB at 337 nm was red-shifted to 340 nm in ATE NPs, accompanied by a significant increase in absorbance intensity and pronounced spectral broadening. These spectral changes further confirmed the strong intermolecular associations, thereby confirming the successful encapsulation of AmB. Moreover, X-ray diffraction (XRD) patterns of both ATE NPs and ATP NPs lacked the characteristic crystalline peaks observed in AmB, indicating the amorphous state of these nanoparticles (Fig. S6).

To further quantitatively evaluate the antioxidant activity of ATE NPs, 2,2′-Azino-bis (3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) and 1,1-Diphenyl-2-picrylhydrazyl (DPPH) radical scavenging assays were performed. As shown in Fig. 1h, ATE NPs exhibited pronounced ABTS radical scavenging capability, with a concentration-dependent effect due to the abundant phenolic hydroxyl groups of the incorporated EGCG. When the concentration of ATE NPs reached 120 μg mL⁻¹, the ABTS radical scavenging efficiency reached 54.22% ± 1.15%. This potent antioxidant performance was further corroborated by the DPPH assay, which demonstrated a radical scavenging rate exceeding 62.78% ± 0.48% at a relatively low concentration of ATE NPs of 32 μg mL⁻¹ (Fig. 1i).

The in vitro stimuli-responsive drug release behavior of ATE NPs was investigated in the mildly acidic microenvironment of fungal infections. Under simulated normal physiological conditions (pH 7.4), the nanoplatform exhibited a slow and sustained drug release. However, upon exposure to a mildly acidic pathological environment (pH 6.0), the release kinetics accelerated significantly, achieving more than 90% drug release efficiency over 8 h (Fig. 1j). This responsive behavior is fundamentally attributed to the rapid hydrolysis of the pH-sensitive boronate ester bonds under acidic conditions. The cleavage of these dynamic linkages drives the structural disassembly of the nanocarriers, thereby facilitating the rapid release of the encapsulated AmB. Furthermore, ATE NPs remained visually well dispersed without obvious sedimentation during 7 days of storage in H₂O (Fig. S7).

3.2. In vitro adhesion performance and antifungal activity of ATE NPs

In this research, we further investigated the interactions between ATE NPs and C. albicans. As shown in Fig. 2a, ATE NPs effectively adhered to C. albicans. Colocalization analysis of Rhodamine B (RhoB)-labeled AmB and Hoechst-stained C. albicans confirmed that ATE NPs accumulated extensively on the fungal surface, generating strong red fluorescence signals. In contrast, only weak red fluorescence was observed in C. albicans treated with ATP NPs, while almost no detectable red fluorescence was found in the free AmB-treated group. This is because the T-PBA of ATE NPs can react with β-glucans and mannans on the fungal cell wall to form boronate ester bonds, thereby enabling strong adhesion to the fungal surface. Moreover, the superior adhesion performance of ATE NPs is also attributed to the strong interactions between the EGCG of ATE NPs and C. albicans [26].

Fig. 2.

In vitro adhesion performance and antifungal activity. (a) Fluorescent images of C. albicans after incubation with AmB, ATP NPs, and ATE NPs (RhoB labeled AmB, and C. albicans was stained by Hoechst). Scale bar: 40 μm. (b) The in vitro fungicidal evaluation. C. albicans was incubated with different concentrations of ETP NPs, AmB + EGCG, ATP NPs and ATE NPs (1, 2, 4, 8 μg mL⁻¹, AmB equivalent). (c) Live (green)/dead (red) staining images of C. albicans after incubation with H₂O (control), ETP NPs, AmB + EGCG, ATP NPs and ATE NPs. The images were presented as mixed channels of NO1 (green) and PI (red). Scale bar: 80 μm. (d) SEM images of C. albicans after incubation with H₂O (control), ETP NPs, AmB + EGCG, ATP NPs and ATE NPs. Scale bar: 2 μm. (e) TEM images of C. albicans sections after incubation with H₂O (control), ETP NPs, AmB + EGCG, ATP NPs and ATE NPs. Scale bar: 500 nm. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

To further evaluate the antifungal activity of ATE NPs, the MICs of free AmB and ATE NPs against C. albicans were determined using the CLSI M27 broth microdilution method. The MIC of ATE NPs was 0.16 μg mL⁻¹ (AmB equivalent), whereas that of free AmB was 1.28 μg mL⁻¹, indicating enhanced antifungal activity of AmB after incorporation into ATE NPs. The in vitro antifungal activity of ATE NPs was further evaluated using a standard colony-forming unit (CFU) assay. As shown in Fig. 2b, ATE NPs exhibited superior antifungal efficacy against C. albicans compared with the free-drug group. At an AmB-equivalent concentration of 2 μg mL⁻¹, ATE NPs achieved more than 99% fungal eradication within 4 h, and the antifungal effect of ATE NPs displayed an obvious concentration-dependent behavior. In contrast, ETP NPs did not show any significant antifungal activity throughout the tested concentration range, indicating that EGCG, T-PBA, and pentaerythritol alone do not possess intrinsic fungicidal properties. Representative CFU images further visually confirmed the potent broad-spectrum antifungal activity of ATE NPs. As shown in Fig. S8, both C. albicans and C. neoformans exhibited markedly reduced colony growth following treatment with ATE NPs compared with all other groups. Consistently, live/dead fluorescence staining assays demonstrated that ATE NPs achieved the most effective fungal eradication among all experimental groups (Fig. 2c).

Scanning electron microscopy (SEM) images provided further insight into C. albicans morphology post-treatment. Among all groups, the ATE NPs-treated group showed the most pronounced morphological collapse and surface wrinkling (Fig. 2d). Furthermore, bio-transmission electron microscopy (Bio-TEM) was employed to compare the overall morphology of C. albicans after different treatments. The control and ETP NPs treated fungal cells retained relatively regular morphologies, whereas the AmB + EGCG and ATP NPs treated cells exhibited deformation and shrinkage (Fig. 2e). These qualitative observations were consistent with the SEM and live/dead staining results.

3.3. In vitro antibiofilm effect of ATE NPs

Fungal biofilm infections pose a severe and growing threat in clinical practice. Surrounded and encased by a protective self-produced exopolymeric matrix, biofilm provides a protective barrier to physical and chemical agents that could potentially injure or kill fungi [27]. Therefore, preventing the formation of biofilm in the early stage and destroying the mature biofilm in the late stage are two valid antibiofilm methods. Consequently, the antibiofilm effect of ATE NPs on C. albicans was verified for both preventing and destroying the biofilms.

As shown in Fig. 3a, the biofilms in the control groups maintained relatively intact structures with strong fluorescence signals. In contrast, biofilm formation was inhibited to varying extents and the biofilm architecture was partially disrupted after treatment with the other formulations, among which the ATE NPs-treated group exhibited the sparsest biofilm structure. Quantitative analysis of fluorescence intensity and biofilm thickness further demonstrated the results. After treatment with ATE NPs, the green fluorescence intensity showed a 51.83-fold decrease, and the biofilm thickness was reduced from 30.98 ± 2.26 μm to 9.19 ± 0.60 μm. Notably, compared with the control group, the ETP NPs-treated group also showed a certain reduction in green fluorescence intensity and biofilm thickness. This phenomenon might be attributed to the reported ability of EGCG to interfere with ergosterol biosynthesis in C. albicans and inhibit its transition from the yeast form to the hyphal form, thereby suppressing biofilm maturation [[28], [29], [30]] (Fig. 3b and c). Crystal violet staining further confirmed these observations with the biofilm biomass decreased by 77.44% after ATE NPs treated. (Fig. 3d and e). After dispersing the treated biofilms by repeated pipetting with physiological saline, the residual fungal burden was quantified by CFU counting and representative plates, showing that ATE NPs inhibited 99.98% of fungal growth (Fig. 3f and g). Taken together, these results suggest that ATE NPs are highly effective in preventing C. albicans biofilm formation.

Fig. 3.

In vitro antibiofilm effect. (a) 3D live (green)/dead (red) staining images of unmatured C. albicans biofilms after different treatments. The images were presented as mixed channels of NO1 (green) and PI (red). Scale bar: 80 μm. (b,c) Quantification of green fluorescence intensity and thickness of unmatured biofilms after different treatments. (d,e) Quantification and images of unmatured biofilms stained by crystal violet after different treatments. (f,g) Quantification and images of the live fungi in unformed biofilms on SDA plates after different treatments. (h) 3D live (green)/dead (red) staining images of matured C. albicans biofilms after different treatments. The images were presented as mixed channels of NO1 (green) and PI (red). Scale bar: 80 μm. (i,j) Quantification of red/green fluorescence ratio and thickness of matured biofilms after different treatments. (k,l) Quantification and images of matured biofilms stained by crystal violet after different treatments. (m,n) Quantification and images of the live fungi in formed biofilms on SDA plates after different treatments. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

To evaluate the eradication of mature biofilms, 2-day-old C. albicans biofilms were exposed to various treatments, followed by measurements such as live/dead staining, crystal violet staining, and CFU enumeration. In the live/dead staining analysis, ATE NPs induced notably more fungal cell death, as evidenced by markedly increased red fluorescence compared with other treatments, and the biofilm thickness was reduced from 26.43 ± 2.13 μm to 12.31 ± 0.60 μm compared with the control group (Fig. 3h–j). Similarly, the crystal violet staining showed that over 79.54% of the biomass in the mature biofilm was eradicated (Fig. 3k and l), with a 99.9% reduction in C. albicans after treatment with ATE NPs (Fig. 3m and n), which was superior to both other treatments. Taken together, these results highlight the exceptional biofilm eradication capability of ATE NPs.

3.4. In vitro anti-inflammatory activity of ATE NPs

Given the limited biosafety of AmB, the biocompatibility of ATE NPs was first evaluated using a cell counting kit-8 assay prior to in vitro anti-inflammatory assessment. As shown in Fig. S9a, when the AmB-equivalent concentration reached 8 μg mL⁻¹, the cell viability in the AmB group dropped below 60%, whereas the ATE NPs-treated group maintained cell viability above 80%, indicating that ATE NPs significantly alleviated the cytotoxicity of AmB. Furthermore, T-PBA exhibited negligible cytotoxicity toward HCECs over the tested concentration range of 1-32 μg mL⁻¹, with cell viability remaining above 85% and no statistically significant differences observed among the different concentrations (Fig. S9b).

To evaluate the therapeutic potential of ATE NPs in fungal infection-associated inflammation, we first explored the elimination of intracellular reactive oxygen species (ROS) caused by inflammation. Human Corneal Endothelial Cells (HCECs) were treated with H₂O₂ and the generated intracellular ROS was analyzed using DCFH-DA, combined with confocal laser scanning microscopy (Fig. 4a). HCECs treated with H₂O₂ showed a significant accumulation of intracellular ROS. Treatment with ATP NPs reduced 61.51% of the intracellular ROS, which was attributed to the ROS-scavenging capacity of boric acid [31], while the other three groups achieved an intracellular ROS scavenging efficiency of over 90%, and the intracellular ROS content in the ATE NPs and ETP NPs groups was lower than that in the AmB + EGCG group. These results demonstrate that EGCG has a better ROS scavenging ability than dynamic boronate, and the synergistic effect of EGCG and the boronate bonds can further enhance ROS clearance (Fig. 4b).

Fig. 4.

In vitro anti-inflammatory activity evaluation. (a,b) Images and quantification of fluorescence intensity of DCFH staining. Scale bar: 200 μm. (c-f) Representative blots of IL-10, TNF-α, IL-1β, and IL-6 are shown along with quantitative analysis. (g) Expression of TNF-α, IL-1β, IL-6 and IL-10 measured by Western blot in RAW 264.7 cells.

Then we further investigated the therapeutic potential of ATE NPs in fungal infection-associated inflammation. In a lipopolysaccharide (LPS)-induced RAW 264.7 macrophage inflammatory model, the expression levels of key anti-inflammatory cytokine interleukin-10 (IL-10) and pro-inflammatory cytokines (tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), and interleukin-6 (IL-6)) were assessed using quantitative polymerase chain reaction (q-PCR) and western blot analysis (Fig. 4c–g). LPS stimulation markedly upregulated TNF-α, IL-1β, and IL-6, while significantly downregulating IL-10 expression in the control group (Fig. 4g). After treatment, compared with other groups, the ATE NPs group significantly suppressed the expression of TNF-α, IL-1β, and IL-6, while markedly upregulating IL-10 expression, demonstrating potent anti-inflammatory effects (Fig. 4c–f).

Collectively, these results demonstrate that ATE NPs exert strong protective anti-inflammatory effects by effectively suppressing pro-inflammatory cytokine production and eliminating intracellular ROS, which further highlights their great potential for the treatment of inflammation associated with fungal infections.

3.5. In vivo therapeutic efficacy of ATE NPs in treating C. albicans-infected keratitis

Based on the remarkable antifungal and antibiofilm effects, we further investigated the therapeutic efficacy of ATE NPs on keratitis infected with C. albicans (Fig. 5a). Due to tear circulation, the retention ability of drugs on the ocular surface directly affects therapeutic efficacy. Therefore, we first evaluated the ocular surface retention of ATE NPs (Fig. 5b and c). Compared with free RhoB, RhoB-ATE NPs showed better retention ability at all time points. At 4 h post-administration, almost no ocular surface residue of free RhoB was observed, whereas RhoB-ATE NPs still retained approximately 12.6% at 8 h and 5.6% at 16 h. After 24 h incubation of C. albicans suspension on cornea, the keratitis model was successfully established as the appearance of corneal edema and turbidity. The mice were then randomly divided into six groups and treated with H₂O (control), ETP NPs, AmB + EGCG, ATP NPs, ATE NPs, and AmBisome®, a clinically used liposomal AmB formulation, at the same AmB-equivalent concentration on Day 1 and Day 2. The treatment effects were recorded by the slit lamp system. As shown in Fig. 5d and e, the corneas in all groups exhibited edema and exudation, accompanied by a small amount of purulent discharge on day 0 (clinical grading scale: Control 4.8, ETP NPs 5.2, AmB + EGCG 5.2, ATP NPs 4.6, ATE NPs 5.0, and AmBisome® 4.4). By day 2, a marked aggravation of corneal inflammation and edema was observed in each group with a significant inflammatory response in the eyelids (clinical grading scale: Control 10.8, ETP NPs 11.4, AmB + EGCG 9.2, ATP NPs 8.2, ATE NPs 6.4, and AmBisome® 7.4). From day 5 to day 9, keratitis continued to deteriorate in the control, ETP NPs, and AmB + EGCG groups, with severe corneal edema, opacity, and neovascularization (clinical grading scale on day 9: Control 13.8, ETP NPs 13.6, and AmB + EGCG 12). Conversely, the ATP NPs, ATE NPs, and AmBisome® groups experienced gradual alleviation of corneal edema and effective suppression of the inflammatory response (clinical grading scale on day 9: ATP NPs 9.2, ATE NPs 7, and AmBisome® 9). On day 15, the control, ETP NPs, and AmB + EGCG groups exhibited further progression of keratitis, accompanied by corneal hemorrhage and extensive neovascularization (clinical grading scale: Control 14, ETP NPs 14, and AmB + EGCG 12). The AmBisome® group showed partial alleviation of the corneal lesions, with a clinical grading score of 8. In comparison, the ATE NPs treatment group achieved the most favorable therapeutic effects, with restoration of corneal transparency, and effective control of the inflammatory response (clinical grading scale: ATE NPs 6). Besides, the intraocular pressure (IOP) was measured during the treatment (Fig. 5f). The IOP gradually increased to 16, and 14.2 mmHg in the control, and ETP NPs groups because of the inflammation and edema while the IOP increased and then returned to 13 mmHg in the AmB + EGCG group. In the ATP NPs and ATE NPs groups, the IOP slightly increased and then returned to 11.6 mmHg and 9.8 mmHg, indicating the validity and biosafety of the treatment.

Fig. 5.

The in vivo antifungal ability of ATE NPs on a C. albicans-induced keratitis model. (a) Schematic diagram of murine keratitis infection with C. albicans and treatment protocol with different formulations. (b) Real-time ocular fluorescence imaging of RhoB-ATE NPs and RhoB at different time points. (c) Fluorescent intensity decay curves of RhoB-ATE NPs and RhoB in the eye area over time. (d) Slit lamp photos of eyes after different treatments on the 0th, 1st, 2nd, 5th, 9th, and 15th day. (e) Ocular clinical scores after different treatments. (f) Measurement of IOP.

The antifungal and anti-inflammatory efficacy of ATE NPs treatment was further evaluated through ophthalmic examinations. On day 15, fluorescein sodium staining revealed the corneal epithelial defects (Fig. 6a and b). There was no corneal epithelial defect in the healthy group while the control, ETP NPs, and AmB + EGCG groups exhibited obvious corneal epithelial defects. The AmBisome® group retained residual corneal epithelial defects, and the ATP NPs group exhibited mild corneal epithelial defects, while the ATE NPs group experienced almost complete corneal healing and an absence of sodium fluorescein staining. Due to the inflammation and edema, the corneal thickness could increase in keratitis and the anterior segment images were recorded by optical coherence tomography (Fig. 6c and d). On day 4, the corneal thickness was 185.9, 173.8, 172.3, 165.4, and 164.0 μm in the control, ETP NPs, AmB + EGCG, ATP NPs, and AmBisome® groups while it was 93.6 μm in the ATE NPs group, demonstrating the successful recovery of the keratitis, with subsequent normalization of the corneal structure. The in vivo laser scanning confocal microscopy was applied to verify fungal organisms and hyphae on day 4 (Fig. 6e and f). A lot of hyphae were detected in the control, and ETP NPs groups while the hyphae were reduced in the AmB + EGCG, ATP NPs, and AmBisome® groups. The ATE NPs group showed a marked reduction of hyphae, demonstrating the excellent antifungal properties and the ability to disrupt fungal physiology. Fungal plate counting was applied to verify the antifungal effects (Fig. 6g and h). Consistent with the results of in vivo laser scanning confocal microscopy, the CFUs counting of AmB + EGCG, ATP NPs, and AmBisome® groups reduced and the ATE NPs group achieved the best antifungal effect compared with control group. Furthermore, the anti-inflammatory and corneal healing repair treatment effect was evaluated through Hematoxylin-Eosin (H&E) staining, immunohistochemical (IHC) staining, and Picrosirius red staining (Fig. 6i–k, S10). In the H&E staining, the cornea exhibited significant edema, irregular corneal epithelium, and inflammatory cell infiltration in the control, ETP NPs, AmB + EGCG, and ATP NPs groups. In the ATE NPs group, the corneal edema was alleviated. and corneal thickness returned to normal, with a well-organized corneal structure. Inflammation-related IHC staining, including IL-1β and IL-6 was applied to reveal the anti-inflammatory effect. In the control, and ETP NPs groups, IL-1β, and IL-6 were both highly expressed while they were reduced in the AmB + EGCG, and ATP NPs groups. IL-1β, and IL-6 were significantly down-regulated in the ATE NPs group, which verified the alleviation of the inflammatory reaction. The deposition of newly formed collagen fiber was evaluated through Picrosirius red staining. In the control, ETP NPs, and AmB + EGCG groups, collagen fibers were disorganized and sparsely distributed in the epithelium and stromal layer. However, there was a large amount of dense and well-aligned collagen fibers in the ATE NPs group, especially under the epithelium layer, which demonstrated effective collagen deposition and tissue regenerative repair in cornea.

Fig. 6.

Ophthalmic pathological characterization of the C. albicans-induced keratitis model. (a) Sodium fluorescein staining of corneas with different treatments. (b) Quantification of fluorescein staining scores in (a). (c) OCT imaging photos after different treatments. Scale bar: 500 μm. (d) Statistical results of corneal thickness from OCT images after different treatments. (e,f) In vivo laser scanning confocal microscope images of fungal hyphae and quantification analysis. Scale bar: 100 μm. (g,h) Images of fungal colonies on SDA plates and quantification analysis of the residual rate. (i) H&E staining, IL-1β staining, IL-6 staining, and Sirius Red staining images of ocular sections after different treatments. Scale bar: 200 μm. (j) Relative expression levels of the inflammatory factor IL-6 after different treatments. (k) Relative expression levels of Sirius Red after different treatments. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

To further investigate potential tissue toxicity, comprehensive histopathological examinations of the retina and major organs (heart, liver, spleen, lung, and kidney) were performed using H&E staining (Fig. S11). Consistent with the biochemical findings, none of the groups (including the control, ETP NPs, AmB + EGCG, ATP NPs and ATE NPs groups) exhibited noticeable pathological lesions, tissue necrosis, structural disruption, or abnormal inflammatory infiltration after treatment. The retinal microstructure also maintained its normal morphological integrity without any signs of intraocular toxicity. In parallel, body weight monitoring in both the C. albicans-infected keratitis mice and the infected skin wound mice showed a stable and gradual increase throughout the treatment period across all groups (including the control, ETP NPs, AmB + EGCG, ATP NPs and ATE NPs groups), indicating the absence of overt systemic distress or treatment-related toxicity (Fig. S12).

In summary, ATE NPs treatment can effectively treat fungal keratitis and restore corneal transparency by eradicating pathogenic fungi and controlling inflammatory response. Notably, ATE NPs achieved more effective fungal clearance and corneal recovery than AmBisome®, demonstrating therapeutic advantages over the clinically used liposomal AmB formulation.

3.6. In vivo therapeutic efficacy of ATE NPs in treating C. albicans-infected skin wounds

The in vivo treatment effect of ATE NPs was then evaluated in the C. albicans-infected skin wound model (Fig. 7a). The C. albicans suspension was dropped onto the murine skin wounds. The skin wound model was successfully established after the typical symptoms, including redness, swelling, and exudation after 24 h. The mice were then treated with AmB + EGCG, ATP NPs, ETP NPs, and ATE NPs on Day 1 and Day 2, respectively. The healing status and photos of skin wounds were recorded on day 0, 1, 2, 4, 6, and 8 (Fig. 7b–d). On day 0, skin wound infection models were successfully established in all groups, with the presence of irregular wounds and exudate. On day 1, the wound area remained stable with slight decrease in all groups and the area was 91.33%, 101.92%, 91.73%, 92.07% and 90.43% in the control, ETP NPs, AmB + EGCG, ATP NPs and ATE NPs groups. During day 2 to day 4, slow wound healing, with bleeding, exudate, and tissue edema were observed in the control, ETP NPs, AmB + EGCG, and ATP NPs groups while the wound area gradually decreased, accompanied by reduced edema in the ATE NPs group. On day 8, wound area decreased to 38.74%, 43.23%, 29.62%, and 32.18% in the control, ETP NPs, AmB + EGCG, and ATP NPs groups and the infected skin wounds were not completely healed. Compared with other groups, the ATE NPs treatment achieved the best therapeutic effect and the wound area recovered to 18.23% and the wounds exhibited nearly complete re-epithelialization, which confirmed the excellent in vivo antifungal and anti-inflammatory effects.

Fig. 7.

The in vivo antifungal ability of ATE NPs on a murine C. albicans-induced wound infection model in mice. (a) The scheme illustrated the construction of the murine skin fungal infection model and the treatment protocol with different formulations. (b) Photos of mice wounds after different treatments on the 0th, 1st, 2nd, 4th, 6th, and 8th day. (c) Representative wound contour maps at different time points. (d) Statistical results of wound areas after different treatments. (e) Quantification analysis of the residual rate. (f) H&E staining, IL-6 staining, IL-1β staining, and Sirius Red staining images of wound tissues after different treatments. Scale bar: 1 mm. (g) Relative expression levels of the inflammatory factor IL-6 after different treatments. (h) Relative expression levels of Sirius Red after different treatments. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

After the treatment, the tissues were collected and the quantitative analysis of CFUs was performed (Fig. 7e, S13). In the control and ETP NPs groups, a lot of CFUs were formed while the number of CFUs decreased in the AmB + EGCG and ATP NPs groups, but the fungi were not completely eradicated. The ATE NPs group achieved the best antifungal effect, indicating the excellent antifungal efficacy. The anti-inflammatory and wound healing effects were further evaluated through histopathological analysis including H&E staining, IHC staining, and Picrosirius red staining (Fig. 7f–h, S14). In H&E staining, the skin tissues did not recover completely, with visible tissue gaps, substantial exudation, inflammatory cell infiltration, and disorganized tissue structure in the control group. In the ETP NPs, AmB + EGCG, and ATP NPs groups, the tissues exhibited less exudation and inflammatory cell infiltration while the tissue structure was also disrupted. However, it exhibited a dramatic reduction in exudation and inflammatory cell infiltration after ATE NPs treatment, with the formation of a well-formed epithelium, continuous keratinocyte layer, indicating the excellent healing effects of the skin. Then the Picrosirius red staining was used for the newly formed collagen fiber in the wound area. There was a small quantity of collagen fiber in the central repaired tissue in the control, ETP NPs, and AmB + EGCG groups while there was more collagen fiber in the ATP NPs group. However, there was a large amount of well-aligned collagen fibers in the ATE NPs group, approximately 2.2-fold higher than those in the control group, demonstrating effective collagen deposition and tissue regenerative repair in the skin wounds. The cytokine expression levels of IL-1β and IL-6 were assessed by IHC. The cytokines were highly expressed in the control and ETP NPs groups. In the AmB + EGCG, ATP NPs, and ATE NPs groups, the expression notably decreased after the treatment which indicated the significant inhibition of the inflammatory reaction of infection. In conclusion, the prominent antifungal and anti-inflammatory effects of ATE NPs treatment were verified in C. albicans-infected skin wounds with potential to promote healing.

3.7. Evaluation of ocular and systemic biosafety

Given the delicate and complex biological structure of the eye, assessing the biocompatibility of ATE NPs is essential for the clinical translation potential. To this end, the in vivo ocular and systemic biosafety of ATE NPs was comprehensively assessed through a series of clinical, microstructural, hematological, biochemical, and histopathological examinations.

The in vivo ocular tolerance of ATE NPs was systematically evaluated following topical administration. Healthy female C57BL/6 mice were administered 20 μL of ATE (12.8 μg mL⁻¹ AmB equivalent, 20 μL) or AmB NPs (12.8 μg mL⁻¹, 20 μL) topically once daily for 3 consecutive days, while the ATE NPs-7 days group received the same dose once daily for 7 consecutive days. The slit-lamp examination in Fig. S15 showed that the corneas in both the ATE NPs-treated and ATE NPs-7 days groups remained clear and transparent, without detectable conjunctival hyperemia, inflammation, corneal opacity, or surface irritation. Consistently, the clinical grading score of the ATE NPs group was 0.33 ± 0.58, showing no statistically significant difference compared with the untreated normal group (0.00). No significant increase in the clinical grading score was observed after 7 consecutive days of ATE NPs administration. Sodium fluorescein staining further demonstrated negligible fluorescence signals on the corneal surfaces of both the Normal, ATE NPs, and ATE NPs-7 days groups, suggesting that corneal epithelial integrity remained intact following repeated ATE NP administration. In contrast, the AmB-treated group showed more evident ocular surface irritation, with a higher fluorescein staining score of 1.67. The optical coherence tomography imaging further confirmed the structural safety of ATE NPs, showing no corneal swelling or edema after treatment. The central corneal thickness in the ATE NPs group was 74.84 μm, which was comparable to that of the normal group at 78.70 μm. Similarly, no significant alteration in corneal thickness was detected in the ATE NPs-7 days group. At the microscopic level, in vivo laser scanning confocal microscopy provided detailed visualization of all corneal layers, including the corneal endothelium, stroma and epithelium. As depicted in Fig. S16, both the Normal and the Normal + ATE NPs groups exhibited intact, well-organized layers with normal morphological characteristics. Quantitative analysis of the corneal endothelial cells further confirmed no statistically significant loss in endothelial cell density, with 1543.6 cells mm⁻² in the ATE NPs group versus 1570.6 cells mm⁻² in the normal group. These findings demonstrate that ATE NPs did not induce any detectable cellular toxicity in corneal tissues and exhibited less corneal irritation compared with conventional AmB treatment, even after 7 consecutive days of administration.

Beyond ocular tolerance, the systemic biosafety of ATE NPs was evaluated by hematological and biochemical analyses. Routine blood examination showed that key hematological parameters, including white blood cells (WBC), red blood cells (RBC), platelets (PLT), hemoglobin (HGB), and hematocrit (HCT), remained comparable among the control, ATE NPs-treated, and ATE NPs-7 days groups, indicating that ATE NPs did not induce systemic inflammation, hematological disturbance, or blood incompatibility (Fig. S17). Other erythrocyte- and platelet-related indices, including red blood cell distribution width (RDW), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), mean platelet volume (MPV), and platelet distribution width (PDW), were likewise unchanged, further supporting the absence of hematological toxicity. Serum biochemical analysis further demonstrated that liver function markers, alanine aminotransferase (ALT) and aspartate aminotransferase (AST), as well as kidney function markers, BUN and creatinine, remained within normal ranges after 3 or 7 days of ATE NPs administration, suggesting that systemic exposure to ATE NPs at the therapeutic dose did not result in detectable hepatotoxicity or nephrotoxicity (Fig. S18).

Collectively, these results demonstrate that ATE NPs possess excellent ocular and systemic biocompatibility, with no detectable ocular irritation, corneal structural damage, hematological abnormality, organ toxicity, or adverse systemic effects after repeated administration for up to 7 consecutive days, thereby underscoring their strong potential for safe ophthalmic application and further clinical translation.

4. Discussion and conclusion

As a global health threat, fungal infections continue to pose a formidable clinical challenge owing to limited drug permeability, poor bioavailability, and the escalating emergence of antifungal resistance. Although AmB is regarded as the “gold standard” antifungal agent, its poor aqueous solubility and severe nephrotoxicity substantially limit its clinical utility. In this study, we successfully developed a novel self-assembled nanoparticle formulation, ATE NPs, by rationally integrating AmB and EGCG. This platform not only significantly improves the local delivery efficiency and biosafety of AmB, but also achieves synergistic therapeutic outcomes, including potent antifungal activity, strong anti-inflammatory effects, and enhanced wound healing, thereby providing an innovative strategy for the treatment of fungal infections.

The specific binding of T-PBA to β-glucans and mannans in the fungal cell wall endows ATE NPs with excellent adhesive properties, markedly prolonging drug retention at the infected site. Moreover, the intrinsic antioxidant and anti-inflammatory properties of EGCG play a pivotal role in the treatment of fungal keratitis. ATE NPs effectively scavenged ROS, thereby alleviating infection-induced oxidative stress, which is critical for protecting corneal tissue and reducing inflammatory injury. In addition, EGCG promotes adaptive immune-mediated tissue repair and regeneration, accelerating wound healing.

This adhesion-enhanced delivery strategy, together with the superior fungicidal mechanisms of ATE NPs, enables substantially improved antifungal efficacy both in vitro and in vivo compared with free AmB and control groups. In a C. albicans-induced keratitis model, ATE NPs demonstrated significantly superior therapeutic efficacy compared with all other treatment groups. Clinical scoring, fluorescence imaging, and histopathological analyses consistently showed that ATE NPs markedly reduced fungal burden, alleviated inflammatory cell infiltration, preserved corneal structural integrity, and accelerated corneal epithelial regeneration. These findings provide strong evidence supporting the potential clinical translation of ATE NPs as a safe and effective antifungal therapeutic.

In summary, this study successfully constructed an intelligent, stimulus-responsive multifunctional nanoparticle system, in which the synergistic integration of structural design and multimodal mechanisms enables efficient antifungal activity, potent anti-inflammatory effects, and accelerated tissue repair. This innovative platform provides valuable insights for the development of advanced nanotherapeutics for fungal infectious diseases.

CRediT authorship contribution statement

Yongnan Chen: Writing – review & editing, Writing – original draft, Validation, Methodology, Investigation, Conceptualization. Fuyao Chen: Writing – original draft, Methodology. Shiya Wang: Methodology. Wenbin Dai: Writing – review & editing, Validation, Supervision. Qiao Jin: Writing – review & editing, Supervision, Project administration. Fan Jia: Writing – review & editing, Supervision, Project administration, Funding acquisition. Yang Ye: Writing – review & editing, Writing – original draft, Supervision, Funding acquisition, Conceptualization.

Ethics approval and consent to participate

All the experiments on animals were accredited by the Institutional Ethics Committee and followed the requirements for the care and use of laboratory animals of the Second Affiliated Hospital, School of Medicine, Zhejiang University (Approval No. 2026-075).

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.

Appendix A. Supplementary data

The following is the Supplementary data to this article.

Data availability

Data will be made available on request.