Hierarchical drug-releasing nanocomposite hydrogel for postoperative synergistic chemo-photothermal therapy of uveal melanoma

Abstract

Postoperative recurrence and metastasis remain critical obstacles in uveal melanoma treatment. Herein, we report the synthesis of a hierarchical drug-releasing nanocomposite hydrogel (LAP NPs/ICG gel) via in situ crosslinking of chitosan, co-encapsulating indocyanine green and hyaluronic acid-based lapatinib-loaded nanoparticles (LAP NPs) for postoperative synergistic chemo-photothermal therapy of uveal melanoma. The LAP NPs/ICG gel can be injected into the tumor resection cavity and remains in vivo for more than four weeks. Under near-infrared irradiation, the hydrogel induces rapid local hyperthermia to ablate residual tumor tissue. Concurrently, progressive hydrogel degradation enables sustained release of LAP NPs, which actively target disseminated malignant lesions, enhance tumor penetration and intratumoral retention, ultimately triggering glutathione-responsive intracellular drug release for efficient tumor eradication. In vitro and in vivo studies using MuM-2B tumor models demonstrate that a single local injection of the LAP NPs/ICG gel significantly suppresses tumor growth, postoperative recurrence, and hepatic metastasis. Moreover, the hydrogel-based therapeutic system shows no detectable damage to adjacent healthy tissues or major organs, ensuring a favorable in vivo safety profile. Overall, this multimodal hydrogel system represents a promising adjunctive strategy for postoperative uveal melanoma therapy.

Graphical abstract

1. Introduction

Uveal melanoma (UM) is the most common primary intraocular malignant tumor in adults. Nearly 50% of UM cases develop distant metastases, with approximately 93% involving the liver, and the median survival following metastasis is less than 12 months [1]. Surgical resection remains the primary intervention for UM and provides substantial clinical benefit [2]. However, complete eradication of tumor tissue is highly challenging, and residual malignant cells can elevate the risk of rapid postoperative recurrence and distant metastasis [[3], [4], [5]]. Conventional postoperative adjuvant modalities, including radiotherapy and chemotherapy, exhibit limited efficacy and might induce severe side effects that irreversibly damage delicate ocular structures [[6], [7], [8]]. Therefore, the development of innovative adjuvant therapeutic strategies to safely eradicate residual microscopic tumors postoperatively has become an urgent requirement for improving the prognosis of UM patients.

Injectable hydrogel-based local photothermal therapy (PTT) has garnered increasing attention as an adjuvant strategy in postoperative cancer treatment owing to its minimally invasive nature and ability to achieve precise photothermal ablation at the surgical site. Such hydrogels can undergo in situ sol-gel transition in vivo, conformally fill irregular surgical cavities, and serve as a physical barrier against external bacterial invasion [[9], [10], [11], [12], [13]]. The exceptional optical transparency of the vitreous body enables the efficient delivery of therapeutic light to intraocular targets, conferring a distinct advantage in precise PTT for UM treatment. However, PTT is inherently limited by spatial constraints, failing to completely eradicate residual marginal lesions and remaining ineffective against circulating tumor cells released into the bloodstream during surgery [[14], [15], [16]]. Chemotherapy is an indispensable therapeutic modality for metastatic tumors that are refractory to direct eradication via surgery or radiotherapy. Co-encapsulation of chemotherapeutic agents and photosensitizers within hydrogels to design a spatiotemporally complementary photothermal-chemotherapy synergistic platform represents an ideal postoperative adjuvant strategy [[17], [18], [19], [20], [21]]. However, the hydrophilic porous matrix of hydrogels results in inefficient encapsulation of hydrophobic bioactive therapeutics and frequently causes unregulated burst release kinetics, which is undesirable for controlled therapeutic drug delivery and increases the risk of damaging surrounding healthy tissues [22,23].

The advancement of nanomedicine has opened new avenues for the development of multifunctional hydrogels as postoperative adjuvant therapies. Compared with conventional chemotherapeutic agents, nanoparticle (NP)-based delivery systems offer distinct advantages, including improved solubility of poorly water-soluble drugs, enhanced stability, and superior tumor-targeting capability [[24], [25], [26]]. However, the systemic administration of NPs is often limited by rapid reticuloendothelial system clearance and insufficient tumor accumulation, resulting in suboptimal therapeutic efficacy [27]. To address this limitation, local hydrogel-based drug depots have emerged as an effective strategy to sustain NP retention at the tumor resection site and achieve prolonged therapeutic exposure. The nanoscale dimensions and hydrophilic shell of NPs promote homogeneous distribution of hydrophobic drugs within hydrogel matrices, while the incorporation of specific ligands and stimuli-responsive functional groups into the NPs allows for precise drug release, ultimately enhancing therapeutic efficacy [28,29]. Upon degradation of the hydrogel in vivo, drug-loaded NPs are gradually released from the three-dimensional hydrogel network, which protects labile drugs from enzymatic degradation and facilitates targeted delivery, overcoming the inherent limitations of PTT in eradicating residual marginal and metastatic tumor cells [[30], [31], [32]]. To validate this synergistic strategy, indocyanine green (ICG) was selected as the photothermal agent, owing to its clinically approved status by the FDA, strong near-infrared (NIR) absorption, exceptional biocompatibility, and predictable metabolic profile, making it highly suitable for translational biomedical applications [33].

Herein, we report the development of an injectable, multimodal nanocomposite hydrogel (LAP NPs/ICG gel) based on a dynamic Schiff-base crosslinked network formed by chitosan (CS) and difunctionalized poly(ethylene glycol) (DF-PEG). This responsive hydrogel co-encapsulates glutathione (GSH)-responsive lapatinib-loaded NPs (LAP NPs) and ICG to inhibit postoperative UM recurrence and metastasis. The dynamic Schiff-base chemistry enables mild in situ gelation and injectability without any external initiators, as well as gradual hydrogel degradation for sustained NP release [34]. Lapatinib, a dual tyrosine kinase inhibitor targeting the epidermal growth factor receptor (EGFR) and human epidermal growth factor receptor 2 (HER2), was chosen for incorporation based on our earlier findings on its efficacy in suppressing tumor cell proliferation and metastasis [[35], [36], [37]]. Upon NIR laser irradiation, the LAP NPs/ICG gel rapidly generates local hyperthermia to ablate residual tumor tissues, while simultaneously enabling the sustained release of LAP NPs. These NPs actively target disseminated tumor cells and undergo intracellular GSH-triggered drug release, facilitating efficient tumor eradication (Scheme 1). Local injection of the LAP NPs/ICG gel enhances tumor penetration and intratumoral drug retention, potently suppressing UM growth, postoperative recurrence and metastasis. Moreover, the LAP NPs/ICG gel minimizes systemic side effects and spares surrounding healthy tissues, ensuring favorable biocompatibility and safety. This multimodal nanocomposite hydrogel integrates the synergistic benefits of photothermal ablation and NP-mediated targeted chemotherapy, proving to be an effective therapeutic strategy for postoperative management of UM.

Scheme 1.

Schematic illustration of the LAP NPs/ICG gel as a multimodal nanocomposite hydrogel platform for postoperative chemo-photothermal therapy. (a) Fabrication and hierarchical drug release behavior of the LAP NPs/ICG gel. (b) Therapeutic application in postoperative therapy: (1) PTT-mediated ablation of residual tumor tissue and (2) sustained release of LAP NPs from the hydrogel to eradicate disseminated tumor cells.

2. Materials and methods

2.1. Preparation of LAP-loaded NPs

LAP NPs were fabricated via the self-assembly of HA-ss-Chol conjugate, which was synthesized as reported in our previous work [36]. Briefly, HA-ss-Chol (50 mg) and LAP (20 mg) were co-dissolved in 5 mL DMSO, and the solution was slowly added dropwise into 50 mL phosphate-buffered saline (PBS, pH 7.4) under sonication to induce self-assembly. The mixture was placed in a dialysis bag (MWCO 3500 Da) and dialyzed against PBS for 48 h to remove residual solvent. Following dialysis, the dispersion was filtered through a 0.22 μm membrane to yield the final LAP NPs. Owing to the lack of intrinsic fluorescence of LAP, doxorubicin (DOX) was also encapsulated into the NPs (DOX-LAP NPs) for distribution tracking, utilizing its intrinsic fluorescence as a probe for monitoring cellular uptake, intracellular distribution and tumor penetration of NPs.

2.2. Preparation of nanocomposite hydrogel

Preparation of the nanocomposite hydrogel: ICG (1 mg) was dissolved in 1 mL of LAP NPs solution (LAP concentration: 1 mg/mL) containing 2% (v/v) acetic acid solution. Once fully dissolved, CS (0.03 g) was added under gentle stirring, followed by the dropwise addition of 50 μL DF-PEG aqueous solution (33 wt%). The mixture was immediately subjected to vortexing at 2000 rpm for approximately 30 s to induce gelation, yielding the nanocomposite hydrogel co-loaded with LAP NPs and ICG (LAP NPs/ICG gel). Control hydrogels were prepared using the same procedure with different component combinations: (i) CS gel comprising CS and DF-PEG; (ii) ICG gel comprising CS, ICG, and DF-PEG; (iii) free LAP gel comprising CS, free LAP, and DF-PEG; (iv) free DOX/ICG gel comprising CS, free DOX, ICG, and DF-PEG; and (v) DOX-LAP NPs/ICG gel comprising CS, DOX-LAP NPs, ICG, and DF-PEG.

2.3. In vitro cytotoxicity

MuM-2B cells were plated in 24-well plates at a density of 2 × 10⁴ cells per well and allowed to adhere for 12 h. Subsequently, freshly prepared hydrogel (100 μL) was added to each well (final volume: 1 mL). For the NIR-treated group, hydrogels were exposed to an 808 nm NIR laser at 1.0 W/cm² for 10 min. Following 24 h of co-incubation, the medium was replaced with serum-free RPMI-1640 medium containing the Cell Counting Kit-8 (CCK-8) reagent, followed by further incubation for 2 h. The absorbance at 450 nm was recorded using a microplate reader, and cell viability was calculated as a percentage relative to untreated controls.

2.4. In vitro anti-recurrence

To mimic postoperative tumor recurrence and the drug-depot function of the hydrogel, MuM-2B cells were seeded in 24-well plates at a density of 2 × 10⁴ cells per well and allowed to adhere for 12 h, representing the primary tumor. Subsequently, hydrogel (100 μL) was added to each well and incubated for 24 h to simulate hydrogel implantation at the tumor resection site. The supernatant, residual hydrogel, and detached cells were carefully aspirated and removed to mimic surgical tumor debulking. A fresh medium containing 1 × 10⁴ MuM-2B cells was added to each well to simulate microscopic residual tumor cells remaining after surgery. To mimic the postoperative hydrogel depot, hydrogel (60 μL) was placed into transwell inserts and transferred to the corresponding wells, allowing the sustained release of LAP NPs to continuously diffuse toward the residual tumor cells without direct contact. After the predetermined incubation period, cell viability was assessed using the CCK-8 assay.

2.5. Cellular uptake

The cellular uptake of the released NPs was evaluated using a transwell co-culture system. Briefly, MuM-2B cells (5 × 10⁴ cells per well) were seeded onto sterile round coverslips placed in 24-well plates and incubated for 12 h. Subsequently, 100 μL of the DOX-LAP NPs/ICG gel was added to transwell inserts and positioned above the wells containing the cells. At predetermined time points, the cells adhered to the coverslips were gently washed with PBS, fixed with 4% paraformaldehyde, stained with 4′,6-diamidino-2-phenylindole (DAPI), and imaged using a confocal laser scanning microscope (CLSM, Zeiss LSM 880, Germany). For flow cytometry analysis, MuM-2B cells were seeded in 24-well plates at the same density (5 × 10⁴ cells per well) and incubated for 12 h. Hydrogel (100 μL) was then placed into transwell inserts, which were transferred into the corresponding wells. At selected time points, cells were harvested via trypsinization, washed with PBS, and resuspended in 0.2 mL PBS. Fluorescence intensity was analyzed using a flow cytometer (Beckman Coulter Cytomics FC-500) to quantify cellular uptake.

2.6. In vitro MuM-2B 3D tumor spheroid efficacy evaluation

MuM-2B cells (2 × 10³ cells per well) were seeded in ultra-low attachment 96-well plates (Labselect, 11548) with 100 μL of culture medium and incubated for 7 days to allow spheroid formation. Mature spheroids were then transferred into 24-well plates containing fresh medium, followed by the placement of transwell inserts loaded with hydrogel (100 μL) above each well. At 24, 48, and 72 h after hydrogel treatment, spheroids were imaged, and their diameters were quantified using ImageJ software. Subsequently, spheroids were enzymatically dissociated with trypsin, and the resulting single-cell suspensions were collected and resuspended in 500 μL PBS for viability assessment using the CCK-8 assay.

2.7. Penetration in tumor spheroids

A 3D multicellular tumor spheroid model was established and treated with either free DOX or DOX-LAP NPs at an equivalent DOX concentration of 10 μg/mL. Following 6 h of incubation, spheroids were thoroughly washed with PBS and imaged via Z-stack scanning using CLSM. The 3D fluorescence distribution of DOX within the spheroids was reconstructed using ImageJ software.

2.8. In vitro anti-metastasis effect

MuM-2B cells were pretreated for 24 h with 10 μL of various hydrogel formulations in a transwell co-culture system using RPMI-1640 medium containing 0.1% fetal bovine serum (FBS). The pretreated cells were then collected, resuspended, and seeded into the upper chambers of 24-well transwell inserts at a density of 5 × 10⁴ cells per well, while the lower chambers contained RPMI-1640 medium with 10% FBS as a chemoattractant. Non-migratory cells on the upper membrane were gently removed after 24 h using a cotton swab. Migrated cells on the lower surface were fixed with 4% paraformaldehyde and stained with 0.1% crystal violet. Following imaging under a light microscope, the stained cells were eluted with 33% acetic acid, and absorbance at 570 nm was measured to quantify cell migration.

2.9. In vitro anti-invasion effect

The invasion assay was performed following the same procedure as the migration assay, with the exception that transwell membranes were pre-coated with 60 μL of matrigel prior to seeding cells into the upper chambers.

2.10. Biodistribution and tumor penetration

The biodistribution of LAP NPs/ICG gel was evaluated in a MuM-2B subcutaneous tumor model. Briefly, 4 × 10⁶ MuM-2B cells were subcutaneously inoculated into the right lower back of each mouse. The tumor volume (V) was monitored using a vernier caliper to measure length (L) and width (W), and calculated according to the following equation: V = 0.5 × L × W². When tumor volume reached approximately 100 mm³, the free DOX/ICG gel and DOX-LAP NPs/ICG gel were peritumorally injected at an equivalent DOX dose of 10 mg/kg. At 48 h post-injection, mice were sacrificed, and tumors along with major organs (heart, liver, spleen, lung, and kidney) were excised for ex vivo imaging. Tumor tissues were subsequently sectioned and stained with DAPI for analysis via CLSM.

2.11. In vivo anti-tumor efficacy

When subcutaneous tumors reached approximately 100 mm³, tumor-bearing mice were randomly divided into 7 groups (8 mice per group): (i) untreated control, (ii) free LAP gel, (iii) LAP NPs gel, (iv) ICG gel + NIR, (v) LAP NPs/ICG gel, (vi) LAP NPs/ICG gel + NIR, and (vii) ICG gel + NIR + LAP NPs. Each mouse received a peritumoral injection of 100 μL of the designated hydrogel formulation. For groups that received NIR irradiation, mice were exposed to an 808 nm laser (1.0 W/cm²) for 120 s at two-day intervals, for a total of five sessions. In the ICG gel + NIR + LAP NPs group, LAP NPs were administered intravenously at 5 mg/kg immediately after each NIR irradiation. The first day of treatment was designated as day 0. The tumor size and body weight were monitored every two days. Mice were humanely euthanized if tumor volume exceeded 1500 mm³ or upon signs of health deterioration. On day 27, all mice were sacrificed, and tumors along with major organs were harvested for histological analysis.

2.12. In vivo anti-recurrence assessments

When subcutaneous tumors reached approximately 150 mm³, mice underwent partial tumor resection, during which the overlying and adjacent skin tissues were carefully preserved. Mice were then randomly assigned to five groups (8 mice per group): (i) untreated control, (ii) free LAP gel, (iii) LAP NPs gel, (iv) ICG gel + NIR, and (v) LAP NPs/ICG gel + NIR. Each mouse received a 100 μL injection of the corresponding hydrogel formulation into the surgical site. For NIR-treated groups, mice were irradiated using an 808 nm laser (1.0 W/cm²) for 120 s at two-day intervals, for a total of five sessions. Tumor recurrence and body weight were monitored throughout the study. The first detection time of a subcutaneous nodule was recorded as the local recurrence time. On day 27, all mice were sacrificed, and recurrent tumors were collected for histological analysis.

2.13. In vivo safety assessment

Healthy female BALB/c mice were randomly allocated to five groups (five mice per group) and administered the same drug formulations as those used in anti-recurrence experiments. Clinical behavior and body weight were monitored throughout the study. On day 27, blood samples were collected for biochemical analysis: lactate dehydrogenase (LDH), creatine kinase (CK), aspartate transaminase (AST), alanine aminotransferase (ALT), blood urea nitrogen (BUN), creatinine (CREA), total cholesterol (TC), triglyceride (TG), and glucose (GLU). The mice were then sacrificed, and major organs as well as the skin adjacent to the irradiation site were harvested for histological examination. To evaluate the potential impact of photothermal treatment with the LAP NPs/ICG gel on ocular tissues, 5 μL of the hydrogel was injected into the mouse eyeballs through the sclera using a Hamilton syringe. The temperature variation at the ocular site was then monitored under 808 nm NIR irradiation at 1.0 W/cm² for 120 s. After treatment, the eyeballs were harvested for histological and immunohistochemical analyses, including H&E staining, TUNEL assay, and CD68 immunofluorescence staining, to assess tissue morphology, apoptosis, and macrophage infiltration.

2.14. Statistical analysis

All experiments were conducted in triplicate or with more specimens. Data are presented as the mean ± standard deviation. Comparisons between two groups were conducted using Student's t-test, and one-way analysis of variance was used for comparisons involving more than two groups. Statistical significance is denoted using asterisks: *p < 0.05, **p < 0.01, and ***p < 0.001.

3. Results

3.1. Preparation and characterization of the LAP NPs/ICG gel

The GSH-responsive amphiphilic polymer HA-ss-Chol was synthesized according to our earlier report and fully characterized (Figs. S1 and S2) [36]. Spherical LAP NPs with a diameter of 54.8 nm were prepared via the self-assembly of HA-ss-Chol in an aqueous solution (Fig. S3). LAP was encapsulated into the hydrophobic core, achieving a drug loading capacity of 14.7 wt%. LAP NPs exhibited negligible changes in particle sizes and drug loading retention in serum-containing PBS (Fig. S4), with no observable sedimentation, indicating their favorable stability under physiological conditions and reduced risk of premature drug leakage. The crosslinker DF-PEG was synthesized through esterification between PEG and 4-formylbenzoic acid (Fig. S5). The appearance of distinctive proton resonances at 7.97 and 8.23 ppm in the ¹H NMR spectrum (Fig. S6), indicates aromatic protons, confirming the successful synthesis of DF-PEG. CS, exhibiting excellent biocompatibility and biodegradability [38], was used as the matrix to encapsulate LAP NPs and ICG via a simple blending method, with DF-PEG serving as the crosslinker to fabricate a synergistic therapeutic system (Fig. 1a). The resulting functional hydrogel, LAP NPs/ICG gel, underwent gelation at room temperature, as visually confirmed by the vial inversion method (Fig. 1b). It also exhibited favorable injectability, which is readily extrudable through a 22G syringe needle and could form a predefined pattern spelling “GEL” upon injection (Fig. 1c). The LAP NPs/ICG gel and rhodamine-stained hydrogel (pink) spontaneously fused into a single hydrogel at room temperature without external intervention, which could be lifted using tweezers (Fig. 1d). The freeze-dried LAP NPs/ICG gel exhibited loading contents of 2.4 wt% and 2.6 wt% of LAP and ICG, respectively. Scanning electron microscopy (SEM) images of the LAP NPs/ICG gel (Fig. 1e) revealed a honeycomb-like three-dimensional porous network with pore sizes in the range of 30–80 μm, which is favorable for drug encapsulation and controlled release [39]. Compared with the CS gel, the LAP NPs/ICG gel displayed noticeably smaller pore sizes, likely owing to the increased crosslinking density arising from hydrogen bonding between CS and HA [40].

Fig. 1.

Preparation and characterization of LAP NPs/ICG gel. (a) Schematic illustration of the fabrication process and drug release mechanism of LAP NPs/ICG gel. Photographs demonstrating (b) gelation behavior, (c) injectability, and (d) self-healing ability of the LAP NPs/ICG gel. (e) Representative SEM images of the freeze-dried hydrogel. (f) Frequency sweep and (g) strain sweep of the LAP NPs/ICG gel at 25 °C. (h) Temperature changes of hydrogels containing various concentrations of ICG following 808 nm laser irradiation (1 W/cm², 10 min). Mean ± s.d., n = 3. (i) Temperature elevation profiles of CS gel and LAP NPs/ICG gel under 808 nm laser irradiation (1 W/cm²). (j) Photothermal stability of the LAP NPs/ICG gel under repeated irradiation cycles (10 min ON, 1 W/cm²; 20 min OFF). (k) TEM image and DLS analysis (inset) of LAP NPs released from the gel matrix. (l) In vitro GSH-triggered LAP release profile of the released LAP NPs. Mean ± s.d., n = 3. (m) In vitro LAP release profile of the LAP NPs/ICG gel. Mean ± s.d., n = 3. (n) In vivo degradation profile of the LAP NPs/ICG gel following subcutaneous implantation in mice. Mean ± s.d., n = 3.

The rheological behavior of the LAP NPs/ICG gel was investigated using frequency (Fig. 1f) and strain (Fig. 1g) sweep measurements. The variation in the loss modulus (G″) during the frequency sweep is characteristic of hydrogels cross-linked via dynamic interactions rather than permanent covalent bonds [17], which is consistent with the self-healing behavior (Fig. 1d), owing to the dynamic nature of Schiff-base. Such dynamic cross-linking underlies the self-healing ability of the LAP NPs/ICG gel, enabling bonds to reform after cleavage. At a fixed angular frequency of 1 rad/s, the storage modulus (G′) fell below G″ at a strain of approximately 400%, reflecting the fracture of Schiff-base bonds and concomitant restoration of liquid-like behavior. This mechanical adaptability enables the hydrogel to be injected after gelation, which conforms to irregular surgical cavities and facilitates localized drug delivery while supporting tissue integration.

The photothermal properties of the LAP NPs/ICG gel were systematically assessed under 808 nm NIR irradiation (1 W/cm²). Fig. 1h shows that the increase in the temperature depends on the ICG concentration, following 10 min irradiation, with the 1 mg/mL formulation reaching an equilibrium temperature of 36.3 °C. This temperature was significantly higher than those achieved at lower concentrations (22 °C for 0.2 mg/mL and 28.7 °C for 0.5 mg/mL) and comparable with 38.5 °C observed for 2 mg/mL. Consequently, the 1 mg/mL ICG formulation was identified as the optimal concentration for subsequent experiments. The optimized LAP NPs/ICG gel demonstrated rapid photothermal conversion, with a temperature increase of 15.5 °C achieved within 120 s irradiation, reaching 36.8 °C after 10 min (Fig. 1i). The hydrogel exhibited a laser power-dependent temperature increase (Fig. S7). Under irradiation at 1 W/cm², the temperature of the LAP NPs/ICG gel increased to 47.8 °C within 5 min, exceeding the 45 °C threshold required to elicit a photothermal effect sufficient for tumor ablation [41]. Following five cycles of NIR irradiation (1 W/cm², 10 min ON/20 min OFF), the thermal profile remained nearly identical to that of the first cycle, with a stable temperature increase, plateauing at approximately 36 °C (Fig. 1j). This demonstrates the outstanding photothermal stability of the LAP NPs/ICG gel.

Considering that HA-coated NPs can enhance specific targeting capability toward UM cells [35], it is essential to evaluate whether intact LAP NPs can be effectively released from the hydrogel matrix. LAP NPs/ICG gels were incubated in PBS (pH 6.5) for 48 h, and the released nanostructures in the degradation medium were characterized via transmission electron microscopy (TEM). TEM images revealed abundant nanospherical structures with an average particle size of 60.5 nm (Fig. 1k), matching those of the LAP NPs originally embedded in the hydrogel. The GSH-responsive drug release of the NPs was further investigated using a dialysis method (Fig. 1l). In PBS containing 10 mM GSH, NPs exhibited a rapid release profile, with 67.9% of LAP released within 24 h and a cumulative release of 87.9% after 60 h. Notably, the structure of NPs collapsed after drug release (Fig. S8). In contrast, the final release in GSH-free PBS was only 24.1%, confirming that the LAP NPs/ICG gel liberates encapsulated LAP NPs during degradation, while retaining their intrinsic GSH-triggered drug release capability. As hyaluronidase (HAase) is abundantly expressed in tumor cells [42], we investigated the drug release behavior of NPs in PBS containing varying concentrations of HAase (Fig. S9). The results indicated that HAase efficiently accelerated drug release by enzymatically degrading the HA backbone.

Subsequently, the drug release behavior of the LAP NPs/ICG gel was assessed using a dialysis method. As shown in Fig. 1m, the LAP concentration in the dialysate gradually increased during the first 24 h, reaching a cumulative release of 16.8%. Upon GSH addition, the release rate was markedly accelerated, achieving 93% cumulative release after 60 h, whereas the GSH-free system maintained a slow release profile, with only 23.6% cumulative release. This release pattern could be attributed to Schiff-base cleavage under mildly acidic conditions (pH 6.5), inducing gel matrix degradation and subsequent liberation of LAP NPs, which then undergo GSH-facilitated drug release, resulting in a hierarchical release profile. The in vivo biodegradation of LAP NPs/ICG gel was evaluated through subcutaneous implantation in BALB/c nude mice (Fig. 1n). The hydrogel displayed a nearly linear degradation profile, with 40.3% of the initial mass remaining after four weeks. The intratumoral release behavior of LAP NPs/ICG gel was further evaluated by measuring intratumoral LAP concentrations after peritumoral administration. As shown in Fig. S10, sustained LAP retention in tumor tissues was observed for at least 21 days, indicating the prolonged local drug depot capability of the hydrogel. Moreover, NIR irradiation did not alter the sustained release behavior of the hydrogel, supporting its potential for long-acting local drug delivery. Thus, the LAP NPs/ICG gel injection into the tumor site can rapidly elevate the local temperature under laser irradiation and sustainably release LAP over an extended period, synergistically enhancing the anti-tumor efficacy.

3.2. In vitro anti-tumor efficacy of the LAP NPs/ICG gel

Favorable biocompatibility is a prerequisite for biological applications. The cytocompatibility of the hydrogel components was first evaluated in both MuM-2B UM cells and ARPE-19 retinal pigment epithelial cells using the CCK-8 assay (Fig. S11). The ICG gel exhibited high biocompatibility, with cell viability exceeding 80% following 24 h incubation even at concentrations up to 500 μg/mL, confirming the safety of CS and ICG within the tested range. In contrast, LAP NPs and free LAP displayed concentration-dependent cytotoxicity. Notably, at a LAP concentration of 10 μg/mL, LAP NPs induced significantly lower viability in MuM-2B cells compared with that in ARPE-19 cells, whereas free LAP caused similar cytotoxicity in both cell lines. Western blot analysis confirmed markedly higher CD44 expression in MuM-2B cells than that in ARPE-19 cells (Fig. S12), while flow cytometry further demonstrated significantly enhanced cellular internalization of LAP NPs for MuM-2B cells (Fig. S13). These findings indicate that the HA-enriched surface of LAP NPs can mediate active targeting toward CD44-overexpressing MuM-2B cells, enhancing intracellular drug accumulation and the selective cytotoxicity of LAP NPs [35,43,44]. Subsequently, various hydrogel formulations were co-incubated with MuM-2B cells for 24 h. As shown in Fig. 2a, cells treated with the ICG gel maintained 93.9% viability, whereas treatment with LAP NPs/ICG gel reduced cell viability to 58.1%. Under 808 nm laser irradiation, cytotoxicity was further enhanced, with cell viability declining sharply to 19.2% in the LAP NPs/ICG gel group. Western blot analysis revealed that laser irradiation enhanced the inhibition of EGFR phosphorylation (Fig. S14), which could contribute to the enhanced antitumor efficacy of the combined treatment. To simulate postoperative residual tumor conditions, after 1 day of incubation with the different hydrogels, the original medium was replaced with fresh medium containing suspended MuM-2B cells, and hydrogel-loaded transwell inserts were transferred to the wells (Fig. 2b). In the CS and ICG gel groups, tumor cell viability increased markedly on day 2. In contrast, the LAP gel and LAP NPs gel suppressed tumor cell proliferation, with viability declining to 61% and 41.8%, respectively, on day 4.

Fig. 2.

In vitro anti-tumor efficacy of the LAP NPs/ICG gel. (a) Relative viability of MuM-2B cells following incubation for 24 h with different treatments. Mean ± s.d., n = 5. (b) Regrowth of MuM-2B cells post-reseeding on day 1 following various treatments. Mean ± s.d., n = 5. (c) CLSM images and (d) corresponding mean fluorescence intensity (MFI) of MuM-2B cells treated with DOX-LAP NPs/ICG gel at different time points. Mean ± s.d., n = 3. (e) Flow cytometry analysis and (f) quantitative MFI analysis of MuM-2B cells after DOX-LAP NPs/ICG gel treatment at the indicated time points. Mean ± s.d., n = 3. (g) Morphological changes (scale bar: 200 μm), (h) spheroid diameter variations, and (i) relative cell viability of MuM-2B MCSs after exposure to different hydrogels for 72 h. Mean ± s.d., n = 3. (j) Representative images of crystal violet-stained MuM-2B cells in migration and invasion assays. Quantitative analysis of (k) cell migration and (l) invasion rates under different treatments. Mean ± s.d., n = 3.

The cellular uptake of LAP NPs released from the hydrogel matrix was investigated in MuM-2B cells. To facilitate visualization and tracking, doxorubicin (DOX) was encapsulated into the hydrophobic core of LAP NPs as a fluorescent probe to form DOX-LAP NPs with a size of 64.9 nm (Fig. S15). The cellular uptake was examined in a transwell culture system, with the DOX-LAP NPs/ICG gel placed in the upper chamber and MuM-2B cells seeded on coverslips at the bottom of the basal chamber. CLSM results (Fig. 2c and d) revealed efficient, time-dependent intracellular accumulation of the released DOX-LAP NPs. Quantitative flow cytometry (Fig. 2e and f) further confirmed a progressive increase in intracellular fluorescence over time, demonstrating that the LAP NPs/ICG gel enables degradation-mediated, sustained release of NPs into tumor cells, ensuring prolonged retention and continuous drug delivery.

Three-dimensional multicellular spheroids (MCSs) of MuM-2B cells were established as tumor-mimicking models and treated with various hydrogel formulations. The hydrogels were loaded into the upper chambers of transwell inserts, allowing released drugs to diffuse into the lower chamber and exert cytotoxicity in the MCSs. Morphological analysis (Fig. 2g) and quantification of spheroid diameters (Fig. 2h) showed that the control-, CS gel-, and ICG gel-treated MCSs maintained compact architectures with minimal size changes after 72 h. In contrast, free LAP gel- or LAP NPs gel-treated MCSs exhibited partial peripheral disintegration accompanied by a moderate increase in diameter. LAP NPs/ICG gel-treated MCSs displayed the most prominent structural disruption and loosening of the spheroid architecture, with an increase in the apparent spheroid diameter relative to the initial size. Cell viability assessment (Fig. 2i) revealed that all drug formulations induced varying degrees of cytotoxicity toward MCSs, with the LAP NPs/ICG gel achieving the most potent effect, reducing viability to 47.2%, significantly lower than that in the other groups.

Cell migration is a key driver of UM progression and metastasis that undermines therapeutic efficacy and patient survival [45]. Transwell assays were conducted using MuM-2B cells to evaluate the anti-metastatic potential of different hydrogel formulations. As shown in Fig. 2j, control-, CS gel-, and ICG gel-treated cells exhibited pronounced and aggressive migratory behavior. In comparison, exposure to the free LAP gel and LAP NPs/ICG gel markedly inhibited cell migration, with migration rates of 53.2% and 34.3%, respectively (Fig. 2k). The superior inhibitory effect of the LAP NPs/ICG gel compared with that of the free LAP gel is likely attributed to enhanced cellular uptake mediated by the NP delivery system. Invasion assays consistently revealed the strongest suppression by the LAP NPs/ICG gel, with an invasion rate of only 26.3% (Fig. 2l). Collectively, the LAP NPs/ICG gel efficiently inhibits MuM-2B cell proliferation, regrowth, migration, and invasion through sustained drug release, highlighting its considerable potential for localized therapy against UM.

3.3. Biodistribution and tumor penetration

The biodistribution of drugs following peritumoral injection of hydrogels was evaluated in MuM-2B tumor-bearing mice. After 48 h of administration, major organs and tumors were excised for ex vivo fluorescence imaging. Fig. 3a shows red fluorescence in the heart, liver, kidney, and tumor in DOX-LAP NPs/ICG gel- and free DOX/ICG gel-treated mice. Remarkably, DOX-LAP NPs/ICG gel-treated mice exhibited markedly stronger intratumoral fluorescence, with 1.8-fold higher average intensity than free DOX/ICG gel-treated mice (Fig. 3b). Significantly lower fluorescence was observed in the kidneys of DOX-LAP NPs/ICG gel-treated mice, suggesting that hydrogel-mediated sustained release of NPs enhances intratumoral drug retention while reducing renal clearance compared with the free drug formulation.

Fig. 3.

Biodistribution and tumor penetration. (a) Ex vivo fluorescence images of tumors and major organs harvested from MuM-2B tumor-bearing mice at 48 h after peritumoral injection of the DOX-LAP NPs/ICG gel or free DOX/ICG gel. (b) Quantitative analysis of MFI in excised tumors and major organs. Mean ± s.d., n = 3. (c) MFI of tumor sections and (d) representative CLSM images showing intratumoral fluorescence distribution at 48 h post-injection, red: DOX; blue: DAPI-stained nuclei. Mean ± s.d., n = 3. (e) CLSM images, (f) 3D fluorescence reconstruction, and (g) quantitative MFI analysis of MuM-2B multicellular spheroids treated with DOX-LAP NPs or free DOX. Mean ± s.d., n = 3.

To further assess intratumoral drug penetration, DAPI-stained tumor sections were examined via CLSM, and fluorescence intensity profiles were quantified using ImageJ software (Fig. 3c and d). Tumors from DOX-LAP NPs/ICG gel- and free DOX/ICG gel-treated mice displayed DOX-associated red signals. However, the DOX-LAP NPs/ICG gel-treated mice demonstrated highly intense and uniformly distributed fluorescence across the entire tumor tissue, with an average signal 1.6-fold higher than that of the free DOX/ICG gel-treated mice. This enhanced intratumoral accumulation is likely due to NP-facilitated cellular internalization and improved tissue penetration [46]. To further validate this, DOX-LAP NPs and free DOX were co-incubated with MuM-2B MCSs, and their penetration capabilities were assessed via Z-stack CLSM scanning. Bright DOX fluorescence appeared throughout the spheroid interior of DOX-LAP NPs-treated MCSs, penetrating up to 90 μm from the spheroid surface (Fig. 3e and f), whereas fluorescence from free DOX was primarily confined to the periphery. Quantitative analysis confirmed significantly higher fluorescence intensity in NP-treated cells than in free DOX-treated cells (Fig. 3g), underscoring the superior tumor penetration capability of NPs, which facilitates deep intratumoral drug distribution.

3.4. In vivo anti-tumor efficacy

The anti-tumor efficacy of various hydrogel formulations was evaluated in a subcutaneous MuM-2B tumor-bearing mouse model following peritumoral injection. For groups receiving NIR treatment (ICG gel + NIR group and LAP NPs/ICG gel + NIR group), 808 nm laser irradiation (1 W/cm², 120 s/session) was applied to the tumor site on days 0, 3, 6, 9, and 12 (Fig. 4a). In the ICG gel + NIR + LAP NPs group, LAP NPs were intravenously administered at a dose of 5 mg/kg immediately after each NIR irradiation. The LAP NPs/ICG gel exhibited efficient photothermal conversion, with the local temperature rapidly increasing to 46.3 °C within 60 s, which continuously increased throughout the irradiation period, ultimately reaching 51.1 °C by 120 s (Fig. 4b). The tumor volume and body weight were monitored during the treatment period to evaluate therapeutic efficacy and systemic toxicity.

Fig. 4.

In vivo anti-tumor efficacy. (a) Experimental schedule for in vivo anti-tumor treatment. (b) Temperature changes at the local site of the LAP NPs/ICG gel under 808 nm laser irradiation (1 W/cm²). (c) Relative tumor volume growth curves during the treatment period. (d) Tumor inhibition rates calculated from tumor volumes. (e) Individual tumor growth trajectories of MuM-2B tumors in different treatment groups. (f) Photographs of tumors collected on day 27. (g) Tumor weights in various treatment groups. (h) Survival curves of MuM-2B tumor-bearing mice following various treatments. (i) Relative body weight changes of tumor-bearing mice. (j) Proliferation rates, (k) apoptosis rates, and (l) histological analyses (H&E, Ki67 and TUNEL staining) of MuM-2B solid tumors after treatment. Cell nuclei were stained with DAPI (blue), Ki-67 signals are shown in red, and TUNEL signals in green. Mean ± s.d., n = 3.

All hydrogel-treated groups exhibited varying degrees of tumor growth inhibition (Fig. 4c). Compared with the free LAP gel group, the LAP NPs gel group showed markedly slower tumor growth, attributable to the prolonged intratumoral drug retention and enhanced tumor penetration of the released NPs. The therapeutic outcome was further amplified by the combination of PTT and chemotherapy. Mice treated with LAP NPs/ICG gel + NIR showed substantially reduced tumor volumes compared with those treated with either LAP NPs/ICG gel or ICG gel + NIR alone, achieving a tumor inhibition rate of 93.0% (Fig. 4d), with complete tumor regression in three mice at the end of treatment (Fig. 4e). Notably, in ICG gel + NIR + LAP NPs-treated mice, LAP NPs were systemically administered via tail vein injection to synergize with PTT. These mice exhibited a tumor inhibition rate of 80.8%, which was significantly inferior to that of LAP NPs/ICG gel + NIR-treated mice. Moreover, repeated intravenous injections were required, highlighting the superior therapeutic advantage of hydrogel-mediated sustained drug release over systemic administration for localized tumor treatment. The enhanced anti-tumor efficacy of LAP NPs/ICG gel + NIR was further confirmed via tumor imaging (Fig. 4f), tumor weight measurements (Fig. 4g), and the prolonged survival of treated mice (Fig. 4h). Meanwhile, mice in all treatment groups maintained stable body weights throughout the study (Fig. 4i).

Liver tissues harvested at the end of the experiment were subjected to hematoxylin and eosin (H&E) staining to evaluate UM metastasis in mice (Fig. S16). The results revealed varying degrees of tumor cell infiltration in the liver across different groups, with metastatic burden assessed using the proportion of the liver area occupied by tumor cells (Fig. S17). Compared with the free LAP gel group, the LAP NPs gel group showed a marked reduction in metastasis rate, highlighting the substantial advantages of LAP NPs-mediated targeted therapy in suppressing distant metastasis of UM. The LAP NPs/ICG gel + NIR group exhibited the lowest hepatic metastasis, with a metastasis rate of 0.1%, significantly lower than those of the other groups, underscoring the synergistic efficacy of localized PTT and sustained targeted chemotherapy against UM metastases. Additionally, no significant histopathological abnormalities were observed in the heart, spleen, lung, or kidney tissues (Fig. S18), indicating the safety of the local delivery strategy at the tested drug dosage.

Tumor sections were processed for H&E, Ki67, and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining to investigate the anti-tumor mechanism of the hydrogels (Fig. 4j–l). H&E staining of the control group revealed densely packed, highly pleomorphic nuclei with minimal necrosis, whereas hydrogel-treated groups showed varying degrees of cell damage. The LAP NPs/ICG gel + NIR group exhibited the most extensive tumor destruction, characterized by pronounced cell shrinkage and necrotic areas (Fig. 4l). Ki67 staining corroborated the superior anti-tumor efficacy of LAP NPs/ICG gel + NIR, showing a dramatic reduction in proliferative red-fluorescent signals compared with all other groups, with the lowest proliferation rate of 2.0% (Fig. 4j). TUNEL staining revealed the highest level of apoptotic cells in the LAP NPs/ICG gel + NIR group, with an apoptosis rate of 38.8% (Fig. 4k). These histological results demonstrate that LAP NPs/ICG gel + NIR effectively induces tumor cell apoptosis and necrosis while suppressing proliferation, validating its superior therapeutic efficacy in vivo.

3.5. Inhibition of postoperative tumor recurrence and metastasis

For tumors of considerable sizes, pharmacological monotherapy often proves inadequate, necessitating surgical resection followed by adjuvant drug therapy to achieve optimal clinical outcomes [47,48]. Surgical cytoreduction and delayed initiation of adjuvant therapy frequently result in locoregional recurrence and distant metastasis. The demonstrated tumor-suppressive efficacy of LAP NPs/ICG gel + NIR suggests its considerable potential for mitigating postoperative relapse in UM cases. To test this, a MuM-2B subcutaneous tumor model was established, and tumors were allowed to reach approximately 150 mm³. Around 90% of each tumor was surgically resected to simulate residual disease following incomplete resection. Mice were randomly assigned to five groups, receiving 100 μL treatments at the resection site: PBS (Control), free LAP gel + NIR, LAP NPs gel, ICG gel + NIR, or LAP NPs/ICG gel + NIR. In groups undergoing photothermal treatment, the resection site was exposed to an 808 nm laser (1 W/cm², 120 s/session) on days 0, 3, 6, 9, and 12 (Fig. 5a). Tumor recurrence, regrowth kinetics, and body weight were continuously tracked throughout the study.

Fig. 5.

Prevention of postoperative tumor recurrence and metastasis in MuM-2B tumors. (a) Experimental schedule for in vivo anti-recurrence treatment. (b) Local recurrence rates in each treatment group. (c) Growth curves of recurrent tumors. (d) Photographs of excised recurrent tumors on day 27. (e) Individual tumor progression trajectories in each group. (f) Weights of recurrent tumors. (g) Body weight changes of mice during treatment. (h) Proliferation rates and (i) histological analyses (H&E and Ki67 staining) of recurrent tumors post-treatment. Mean ± s.d., n = 3. (j) Representative H&E staining images of liver tissues from each group.

Tumor recurrence in the control group was first observed on day 9 and reached 100% recurrence on day 18 (Fig. 5b). Hydrogel-based treatments markedly delayed recurrence: the free LAP gel and ICG gel + NIR-treated groups reached complete recurrence on days 23 and 24, respectively, and the LAP NPs gel group showed an 87.5% recurrence rate at the endpoint. In contrast, the LAP NPs/ICG gel + NIR group demonstrated the strongest suppression of tumor relapse, with a recurrence rate of 25%. Tumor growth curves and post-treatment photographs (Fig. 5c–e) revealed that hydrogel-based synergistic photothermal-chemotherapy effectively suppressed recurrent tumor progression, with tumors in the LAP NPs/ICG gel + NIR group exhibiting significantly slower growth than ICG gel + NIR and LAP NPs gel groups. Consistent results were obtained in tumor weights (Fig. 5f), with LAP NPs/ICG gel + NIR yielding the lowest average tumor mass (0.025 g), substantially lower than those of LAP NPs gel (0.17 g) and ICG gel + NIR (0.23 g). Moreover, no significant body weight loss occurred after treatment (Fig. 5g), indicating minimal systemic toxicity. Histological analyses further substantiated the anti-recurrence effect (Fig. 5h and i). H&E staining revealed varying degrees of cell damage among the treatment groups, with LAP NPs/ICG gel + NIR exhibiting the highest tumor cell necrosis. Ki-67 immunofluorescence analysis revealed a significantly suppressed cell proliferation index of 4.4% in the LAP NPs/ICG gel + NIR-treated group, notably lower than those in LAP NPs gel (8.6%) and ICG gel + NIR (18.7%) groups.

The suppression of hepatic metastasis following LAP NPs/ICG gel + NIR treatment was evaluated through H&E staining of liver tissues (Fig. 5j). Compared with the metastasis rate in the control group, the metastasis rates in the free LAP gel, LAP NPs gel, ICG gel + NIR, and LAP NPs/ICG gel + NIR groups were significantly reduced, with the LAP NPs/ICG gel + NIR-treated group exhibiting the lowest metastatic rate of 0.12% (Fig. S19). Consequently, postoperative tumor recurrence and metastatic progression were effectively restrained by this nanocomposite hydrogel-based synergistic treatment.

3.6. In vivo safety evaluation

To evaluate in vivo safety, healthy BALB/c mice were subcutaneously injected with the hydrogel formulations and exposed to NIR irradiation under the same conditions as in the anti-tumor study (Fig. 6a). Body weights of all mice remained stable and steadily increased throughout the treatment period (Fig. 6b). At the end of treatment, serum samples were analyzed for key biochemical indicators: LDH, CK, AST, ALT, BUN, CREA, GLU, TC, and TG. No significant differences were detected in serum biochemical parameters between the hydrogel-treated groups and healthy controls (Fig. 6c–j, Fig. S20), suggesting that these hydrogel formulations did not cause detectable acute or chronic toxicity in the heart, liver, or kidneys (Table S1). H&E analysis of the major organs (heart, liver, spleen, lung, and kidney) confirmed the absence of discernible tissue damage across all groups (Fig. 6k). Skin tissues adjacent to the hydrogel injection sites were collected on day 27 for H&E, TUNEL, and CD68 immunofluorescence staining (Fig. 6l). The skin tissues in all groups maintained an intact morphology, with no significant apoptosis or abnormal macrophage infiltration (Fig. S21). To further evaluate the potential impact of photothermal treatment on ocular tissues, LAP NPs/ICG gel was injected into the eyeballs through the sclera, followed by NIR irradiation. Histological and immunohistochemical analyses revealed that the eyes treated with LAP NPs/ICG gel + NIR maintained normal ocular morphology, comparable to that of the untreated control group (Fig. S22). No thermal necrosis, abnormal infiltration of CD68-positive macrophages, or elevated numbers of TUNEL-positive apoptotic cells were observed in the retina, choroid, or sclera. Collectively, these results demonstrate excellent in vivo safety of the hydrogels and support their potential for clinical translation. Nevertheless, this study primarily focused on the short-term postoperative application of the hydrogel as a transient drug reservoir for UM, while the long-term degradation behavior, metabolic fate of degradation products, and potential effects on visual function were not investigated. These parameters would need to be systematically evaluated in future studies if long-term intraocular implantation is envisioned.

Fig. 6.

In vivo safety profile of LAP NPs/ICG gel. (a) Experimental schedule for in vivo safety assessment of the hydrogel formulations. (b) Body weight changes of mice in each treatment group. (c–j) Serum levels of LDH, CK, AST, ALT, BUN, CREA, TC and TG in mice following different treatments. Mean ± s.d., n = 3. (k) Representative H&E stained images of major organs collected post-treatment. (l) H&E staining, CD68, and TUNEL immunofluorescence images of skin tissues adjacent to the irradiation site in healthy mice. Cell nuclei are stained with DAPI (blue), CD68 signals are shown in red, and TUNEL signals in green.

4. Discussion

To date, preventing postoperative recurrence and fatal hepatic metastasis remains a formidable challenge in UM management. Although several therapeutic strategies have been clinically and experimentally explored, their therapeutic outcomes remain suboptimal. Plaque brachytherapy currently represents the clinical gold standard for eye-preserving treatment; however, its long-term efficacy is limited by radiation-associated complications, including radiation retinopathy, optic neuropathy, and irreversible vision loss, particularly in lesions located near the fovea or optic disc [1,49]. Similarly, transpupillary thermotherapy (TTT) is restricted by insufficient tissue penetration, with therapeutic effects confined to the irradiated region, making complete eradication of marginal lesions and disseminated tumor cells difficult [50]. Apart from physical modalities, pharmacological interventions also face inherent limitations. Intravitreal chemotherapy is hindered by rapid vitreous clearance, requiring repeated invasive injections that increase the risk of vitreous hemorrhage and retinal detachment [51]. Systemic therapies, including tebentafusp and molecular targeted inhibitors, have exhibited limited efficacy in controlling residual intraocular tumor cells owing to restricted drug delivery to ocular tissues, while high systemic exposure might result in off-target toxicities [2,52].

In this context, hydrogel-based adjuvant strategies have emerged as promising approaches for local postoperative tumor control via thermal ablation. However, heterogeneous light distribution and the limited penetration depth of NIR irradiation prevent complete tumor eradication. The localized nature of hyperthermia further limits its ability to eliminate disseminated tumor cells and occult micrometastases, reducing its effectiveness in preventing distant metastasis [20,53]. Hydrogels directly encapsulating free chemotherapeutic agents frequently suffer from burst release and insufficient tissue penetration, resulting in suboptimal therapeutic outcomes [23].

To overcome these limitations, we developed a hierarchical drug-release platform based on an injectable LAP NPs/ICG gel integrating photothermal ablation with NP-mediated chemotherapy. The hydrogel functions as a long-acting depot for sustained release of LAP NPs, followed by intracellular GSH-triggered drug release after NP internalization. This multistage delivery regime enhances tumor penetration and intracellular accumulation while minimizing premature drug leakage. Meanwhile, the hydrogel exhibits excellent injectability and prolonged local retention exceeding four weeks, allowing conformal filling of irregular post-resection cavities and reducing the need for repeated dosing. Compared with conventional modalities, the combination of localized hyperthermia and HA-mediated NP delivery improves therapeutic precision and promotes drug accumulation within residual tumor tissues, significantly inhibiting postoperative recurrence and hepatic metastatic burden in the MuM-2B model, while maintaining excellent biocompatibility and safety.

Despite these promising results, the subcutaneous MuM-2B tumor model exhibits inherent limitations. Unlike the subcutaneous compartment, ocular tissues possess complex barriers including the blood-retinal barrier, blood-aqueous barrier, vitreous humor dynamics, and continuous intraocular pressure [[54], [55], [56]]. These features cannot be fully reproduced in subcutaneous models, particularly considering intraocular pharmacokinetics, clearance behavior, and the specialized immune microenvironment of UM, which might limit direct translation of dosing and therapeutic parameters to clinical ocular settings. Nevertheless, the subcutaneous model remains a widely accepted and essential platform for evaluating in vivo biocompatibility, responsive drug release, and the chemo-photothermal synergistic efficacy of novel biomaterials [[57], [58], [59]].

5. Conclusions

In summary, we developed a hierarchical drug-releasing nanocomposite hydrogel (LAP NPs/ICG gel) incorporating GSH-responsive LAP-loaded NPs and ICG for postoperative UM therapy. This hydrogel exhibited excellent injectability, self-healing capacity, and stable photothermal performance. Upon 808 nm NIR irradiation, the LAP NPs/ICG gel induced rapid local hyperthermia for photothermal ablation, while simultaneously enabling the sustained release of LAP NPs (60.5 nm), which actively targeted disseminated UM cells and released LAP intracellularly, synergistically inhibiting tumor cell proliferation, regrowth, and migration through combined photothermal and chemotherapeutic effects. Peritumoral injection of LAP NPs/ICG gel enhanced tumor penetration and prolonged intratumoral drug retention. A single administration combined with NIR irradiation efficiently suppressed tumor growth, with a 93% tumor inhibition rate and markedly prolonged survival. In a MuM-2B model with partially resected tumors, this treatment reduced tumor recurrence to 25% and aggressively impeded the progression of recurrent tumors. It mitigated the risk of postoperative hepatic metastasis, with a metastatic rate of 0.12%, a significant reduction relative to other treatment groups. The LAP NPs/ICG gel caused no detectable damage to adjacent healthy tissues or major organs, ensuring excellent in vivo biocompatibility and safety. Collectively, this multimodal hydrogel delivery system integrates the synergistic benefits of PTT and NP-mediated targeted chemotherapy, providing a promising strategy for the localized postoperative treatment of ocular malignancies.

CRediT authorship contribution statement

Zhihao Guo: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Writing – original draft. Ruohua Zhu: Formal analysis, Funding acquisition, Investigation, Methodology, Software, Supervision, Validation, Visualization. Cailing Wei: Methodology, Supervision, Validation, Visualization. Bingrui Wang: Formal analysis, Methodology. Yunda Lv: Methodology. Jiangcheng Tan: Methodology, Supervision. Linyun Xiu: Formal analysis, Methodology. Ji-Liang Li: Conceptualization, Formal analysis, Funding acquisition, Resources, 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.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

Data will be made available on request.