Dynamic cross-linked injectable hydrogel for combined oxaliplatin–resveratrol therapy in colorectal cancer peritoneal metastasis

Abstract

Peritoneal metastasis (PM) of colorectal cancer (CRC) remains a major therapeutic challenge due to the limited efficacy of current systemic and intraperitoneal treatments. To overcome these limitations, we developed an injectable, self-healing hydrogel constructed from a dual dynamic cross-linked network formed by borate ester and Schiff base bonds between phenylboronic acid-modified carboxymethyl chitosan (CMCS-PBA) and oxidized dextran (ODEX). This hydrogel was engineered for sustained intraperitoneal co-delivery of oxaliplatin (OXA) and resveratrol-loaded mesoporous silica nanoparticles (MSNs@RES). The system exhibited excellent biocompatibility and significantly inhibited cancer cell proliferation, migration, and invasion while inducing apoptosis and immunogenic cell death (ICD) in vitro. In a murine model of CRC peritoneal metastasis, the dual-drug-loaded hydrogel markedly suppressed tumor progression, reduced ascites formation, and prolonged survival. Proteomic analysis further revealed that treatment significantly modulated key signalling pathways, including HIF-1α-mediated angiogenesis, apoptosis-related cascades, and TGF-β-driven epithelial–mesenchymal transition (EMT). Collectively, this work presents a rationally designed localized delivery platform with significant potential for the treatment of colorectal cancer peritoneal metastasis.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12951-026-04791-7.

Introduction

Colorectal cancer (CRC) is one of the leading causes of cancer-related mortality worldwide, and peritoneal metastasis (PM) represents one of its most aggressive and life-threatening manifestations [1–3]. Patients diagnosed with CRC-associated PM often exhibit a median overall survival of less than six months, highlighting the limited effectiveness of current treatment strategies [4, 5].

Systemic chemotherapy remains the primary treatment for advanced CRC; however, its clinical benefit in PM is significantly constrained by poor drug penetration across the peritoneal barrier and severe off-target toxicity [6–8]. Cytoreductive surgery combined with hyperthermic intraperitoneal chemotherapy (CRS-HIPEC) has shown benefit in selected patients, but this approach is highly invasive, technically demanding, and associated with uneven drug distribution and substantial postoperative complications [9–11]. Therefore, there is an urgent need for localized intraperitoneal drug delivery systems capable of providing sustained and controlled drug release while minimizing systemic toxicity.

Injectable in situ-forming hydrogels have emerged as promising platforms for localized cancer therapy [12–14]. Compared with conventional intraperitoneal free-drug administration, hydrogel-based systems offer several advantages. First, improved local drug retention is achieved through the formation of a three-dimensional depot, which enables prolonged exposure of disseminated tumor nodules to high drug concentrations while limiting systemic circulation [15]. Second, enhanced biocompatibility arises from the hydrophilic polymer matrix, which acts as a protective interface and reduces tissue irritation caused by cytotoxic agents such as oxaliplatin [16]. Third, adaptability to the peritoneal environment is enabled by shear-thinning and self-healing properties, allowing minimally invasive injection while maintaining structural integrity under dynamic physiological conditions [17, 18].

To address these clinical challenges, we designed a dual dynamic covalently cross-linked hydrogel based on phenylboronic acid-modified carboxymethyl chitosan (CMCS-PBA) and oxidized dextran (ODEX). Carboxymethyl chitosan (CMCS) was chosen for its excellent biocompatibility, biodegradability, and abundant functional groups suitable for chemical modification. Grafting with phenylboronic acid (PBA) allows CMCS-PBA to participate in reversible boronate ester bond formation with cis-diol-containing polymers, enabling dynamic network construction. Oxidized dextran (ODEX), a biocompatible polysaccharide rich in cis-diol groups, was selected as the complementary dynamic crosslinker. The combination of CMCS-PBA and ODEX permits rapid in situ gelation, injectability, self-healing behavior, and reversible network reconstruction under physiological conditions, without the need for additional chemical crosslinkers. The network is constructed through pH-sensitive boronate ester bonds between phenylboronic acid groups and cis-diols, together with reversible Schiff base bonds formed between amino and aldehyde groups [19, 20]. This dual dynamic architecture enables rapid gelation, mechanical robustness, injectability, and efficient self-healing behavior, thereby providing a structurally adaptable platform suitable for intraperitoneal deployment. Compared with previously reported thermosensitive or single-network injectable hydrogels for intraperitoneal chemotherapy, the present dual dynamic cross-linked system provides adaptive mechanical stability and self-healing capability under the dynamic peritoneal environment. In addition, the platform enables compartmentalized co-delivery of oxaliplatin and MSN-encapsulated resveratrol, allowing sustained release of therapeutics with complementary anti-tumor activities.

While such a mechanically adaptive hydrogel can ensure sustained local drug retention, effective suppression of peritoneal metastasis additionally requires combined modulation of tumor progression pathways. To this end, we implemented a compartmentalized co-delivery strategy. Oxaliplatin (OXA), a first-line chemotherapeutic agent for CRC [21, 22], was incorporated directly into the hydrogel matrix to enable prolonged local chemotherapy. In parallel, resveratrol (RES), a natural polyphenol with chemosensitizing and anti-metastatic properties [23, 24], was encapsulated within mesoporous silica nanoparticles (MSNs) to overcome its poor aqueous solubility and limited bioavailability [25–27]. Benefiting from their large surface area and tunable pore architecture, MSNs facilitate efficient loading and controlled release of hydrophobic therapeutics [28, 29]. Importantly, RES has been shown to enhance OXA sensitivity and suppress TGF-β-driven epithelial–mesenchymal transition and angiogenesis [30, 31], suggesting potentially enhanced therapeutic efficacy when combined with sustained local OXA delivery.

On the basis of this material and therapeutic design rationale, we developed a dual dynamic cross-linked injectable hydrogel capable of sustained intraperitoneal co-delivery of OXA and MSNs@RES. We hypothesized that prolonged peritoneal retention, coupled with coordinated regulation of angiogenesis, epithelial–mesenchymal transition, and apoptosis-related pathways, would effectively inhibit tumor growth and metastatic dissemination. To validate this concept, we performed systematic physicochemical characterization, in vitro biological evaluation, in vivo efficacy assessment in a murine peritoneal metastasis model, and proteomic pathway analysis to elucidate the underlying molecular mechanisms. (Scheme 1).

Scheme 1.

Synthesis diagram of CPOMO hydrogel and its therapeutic mechanism in peritoneal metastasis of colorectal cancer

Materials and methods

Materials

CMCS and ODEX were purchased from Bomeibio Co., Ltd. (Qingdao, China), while NaIO₄, EDC and NHS were obtained from Aladdin Co., Ltd. (Shanghai, China). The Cell Counting Kit-8 (CCK-8), calcein-AM/PI double staining kit, and oxaliplatin (OXA) were purchased from Dalian Meilun Biotech Co., Ltd. (Dalian, China).

Cells and animals

CT26 colorectal cancer cells, HCT116 human colorectal cancer cells and L929 fibroblast cells were purchased from HySigen Biosciences Inc. (Suzhou, China). CT26 cells were cultured in RPMI 1640 medium, while HCT116 cells and L929 cells were maintained in DMEM medium; both media were supplemented with 10% fetal bovine serum (FBS), ampicillin, and streptomycin. All cells were incubated in a humidified incubator at 37 °C with 5% CO₂. RPMI 1640 and DMEM media were obtained from Beijing Solarbio Science & Technology Co., Ltd. (Beijing, China). BALB/c mice were purchased from Zhonghong Bohuan Biotechnology Co., Ltd. (Jiangxi, China) and acclimatized for 1 week prior to experiments to ensure adaptability.

Synthesis of ODEX and CMCS-PBA

ODEX was synthesized by oxidizing dextran with sodium periodate. Briefly, 1 g dextran was dissolved in 10 mL deionized water (DIW), followed by dropwise addition of NaIO₄ solution (763 mg in 8 mL DIW). The reaction was carried out in the dark for 4 h and then quenched with 1 mL ethylene glycol. The product was purified by dialysis (MWCO 3500 Da) against DIW for 3 days and subsequently lyophilized to obtain ODEX.

CMCS-PBA was prepared through EDC/NHS-mediated coupling. Briefly, 1 g CMCS was dissolved in 20 mL H₂O/DMF (v/v = 3:2). Meanwhile, PBA (0.4 g), EDC (0.46 g), and NHS (0.33 g) were dissolved in 20 mL of the same solvent mixture and activated for 30 min before being added dropwise to the CMCS solution. The reaction pH was adjusted to 5.0–5.5 and maintained for 24 h at room temperature. The resulting solution was dialyzed against 1% Na₂CO₃ solution for 2 days and then against DIW for 3 days. Finally, CMCS-PBA was obtained by lyophilization.

Synthesis and characterization of MSNs@RES

Briefly, MSNs (10 mg; XFNANO, China) and RES (10 mg; MACKLIN, China) were dispersed in anhydrous ethanol and stirred at 37 °C for 24 h. The mixture was centrifuged (10,000 rpm, 10 min), and the pellet was washed with fresh ethanol via repeated centrifugation before being dried under vacuum at 40 °C for 12 h.

The encapsulation efficiency (EE) and drug loading (DL) were determined by quantifying the unloaded RES in the supernatants using UV-Vis spectrophotometer (DR6000, Hach, Japan), based on a calibration curve. The cumulative in release profile of RES from MSNs@RES was assessed in phosphate-buffered saline (PBS; pH 7.4). The morphology of MSNs and MSNs@RES was characterized by transmission electron microscopy (TEM, JEM2100) operating at 200 kV.

The EE and DL were calculated according to the following equations:

Where is the total mass of RES added, is the mass of free RES in the supernatant, and is the mass of MSNs used.

Preparation of CMCS-PBA/ODEX Hydrogel (CPO) and MSNs@RES and OXA Loaded Hydrogel (CPOMO)

CMCS-PBA and ODEX were each dissolved in DIW to a final concentration of 4 wt%. The pH of the CMCS-PBA solution was adjusted to 8.0 using 2 M NaOH. For the blank hydrogel (CPO), equal volumes of CMCS-PBA and ODEX solutions were mixed, and gelation spontaneously occurred within ~ 20 s.

For drug-loaded hydrogels, OXA was dissolved in the ODEX solution to 2 mg/mL, and MSNs@RES were dispersed in the CMCS-PBA solution at 5 mg/mL using 10 min ultrasonic treatment followed by stirring. The dual-drug-loaded hydrogel (CPOMO) was prepared by mixing equal volumes of OXA-loaded ODEX and MSNs@RES-dispersed CMCS-PBA solutions, forming a gel within ~ 20 s. Single-drug controls were prepared similarly: CPO-OXA by mixing OXA-loaded ODEX with blank CMCS-PBA, CPOR by mixing blank ODEX with MSNs@RES-dispersed CMCS-PBA, and CPOM by mixing MSNs only in CMCS-PBA solution.

Injectable and self-healing properties of hydrogels

For the injectability test, equal volumes of the CMCS-PBA and ODEX solutions were mixed and immediately loaded into a 1 mL syringe. The mixture was extruded manually to evaluate its injectable behavior. To investigate self-healing, two CPO hydrogel samples were prepared, with one stained with Methylene Blue. Each sample was cut into two halves, and the freshly cut surfaces of a stained half and an unstained half were placed in direct contact. The macroscopic healing at the interface was visually observed and recorded after 5 min of contact.

Viscosity and plasticity characterization of hydrogels

The viscosity and plasticity of the hydrogel were qualitatively assessed by manual manipulation under ambient conditions. Viscosity was evaluated by observing adhesion to nitrile gloves during stretching and pinching. For plasticity, the hydrogel was manually molded into circular, star, and heart shapes, with shape retention documented photographically to confirm its form-stable properties.

Structural, rheological and morphological characterization of hydrogels

The ¹H NMR spectra were acquired using a nuclear magnetic resonance spectrometer (Bruker Ascend III 400 MHz) to characterize the chemical structures of OD and CMCS-PBA. FTIR spectroscopic analysis was performed on all samples over the range of 4000–500 cm⁻¹ with an FTIR spectrometer (Thermo Scientific iS5) to identify functional groups and chemical bonds.

The rheological properties of the hydrogels were characterized using a rotational rheometer (Malvern Kinexus). Time sweep and frequency sweep measurements were conducted to evaluate gelation behavior and viscoelasticity. Additional stress sweep, step-strain, and long-term time sweep tests were performed to further investigate the injectability, self-healing behavior, and rheological stability of the hydrogels under physiological conditions.

The internal morphologies of the formed CPO and CPOMO hydrogels were observed using a field emission scanning electron microscope (Hitachi S-4800).

Swelling, degradation and mechanical properties of hydrogels

The swelling and degradation properties of the hydrogel samples were evaluated according to the following procedure. Samples of uniform size were prepared, and their initial dry mass was recorded as . Each sample was then placed in PBS and maintained at 37 °C under gentle shaking at 120 rpm. At specific time points, the samples were carefully removed from the solution. Any surface-adherent liquid was blotted away using a laboratory wipe, and the swollen mass was immediately measured as . The swelling ratio was determined using the following formula:

For the degradation study, Samples were incubated in PBS at 37 °C under continuous shaking at 120 rpm. At predetermined time points, sample was retrieved and freeze-dried to obtain their dry mass (). The percentage of remaining mass was determined using the formula:

The mechanical properties of the hydrogels were evaluated through tensile and compression tests.

In vitro and in vivo evaluation of drug release

The hydrogels were incubated in PBS at 37 °C under constant shaking at 120 rpm. At predetermined time points, aliquots of the release medium were collected and replaced with an equal volume of fresh PBS. Additional release studies were also performed in PBS containing 0.5 mg/mL lysozyme at 37 °C to evaluate drug release behavior under enzyme-containing conditions. The concentration of the released drug in the supernatant was determined using a UV-Vis spectrophotometer (DR6000, Hach, Japan). The cumulative drug release was calculated as the percentage of the total amount of drug detected in the release medium relative to the initial drug loading in the hydrogel.

To assess the sustained-release capability of the hydrogel in vivo, the fluorescent dye Sulfo-Cy5.5 was employed as a model tracer to monitor its retention within the peritoneal cavity. BALB/c mice received an intraperitoneal injection of 200 µL containing either free Sulfo-Cy5.5 in PBS or Sulfo-Cy5.5 encapsulated within the hydrogel. The distribution and persistence of the fluorescent signal were tracked using an in vivo imaging system (IVIS). Quantitative analysis of fluorescence intensity was performed on the acquired images with LivingImage software (Living image 4.5.2 software AUGUST 2015) to compare the retention profiles between the two groups.

Biocompatibility assessment of hydrogels in vitro and in vivo

Sterile hydrogel extracts (CPO and CPOM) were prepared by incubating the hydrogels in complete culture medium (1 mL hydrogel/10 mL medium) at 37 °C for 24 h. L929 fibroblasts were cultured in 96-well plates and incubated with the extracts for up to 3 days. Cell viability was evaluated using a CCK-8 assay on days 1, 2, and 3, while live/dead staining (Calcein-AM/PI) was performed for qualitative assessment. Hemocompatibility was assessed by incubating hydrogel samples with diluted mouse blood and calculating the hemolysis rate from the absorbance of the supernatant at 570 nm.

For in vivo biosafety evaluation, healthy BALB/c mice received a single intraperitoneal injection of 200 µL hydrogel or PBS. Body weight and general health status were monitored for 7 days. Major organs (heart, liver, spleen, lung, and kidney) were collected after euthanasia and subjected to H&E staining for histopathological analysis. Serum was separated by centrifugation and analyzed for alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen (BUN), and creatinine (CREA) to evaluate potential hepatic and renal toxicity.

Anti-tumor effect of hydrogels in vitro

CT26 and HCT116 cells were cultured in 96-well plates and exposed to the extracts for up to 3 days. Cell viability was quantified using a CCK-8 assay at days 1, 2, and 3. Live/dead staining was performed on for qualitative assessment.

Transwell invasion and wound healing assay of hydrogels

The invasive ability of CT26 cells was evaluated using a Transwell assay. CT26 cells (1 × 10⁵ cells/mL) suspended in serum-free RPMI 1640 medium were seeded into the upper chambers (200 µL), while the lower chambers were filled with 600 µL RPMI 1640 medium containing 10% FBS and the corresponding extracts. After 24 h incubation, non-invaded cells were removed, and invaded cells were fixed, stained with 0.1% crystal violet, and imaged under an inverted microscope (Olympus, Japan).

Cell migration was assessed using a wound healing assay. CT26 cells were seeded into 6-well plates (2 × 10⁵ cells/well) and cultured to confluence. A linear scratch was created using a sterile 200 µL pipette tip, followed by PBS washing to remove detached cells. Cells were then cultured in medium containing the corresponding extracts. Images of the wound area were captured at 0, 24, and 48 h using an inverted microscope (Olympus, Japan).

Flow cytometry for apoptosis

Flow cytometry was used to detect cell apoptosis and evaluate the apoptosis-inducing ability of different group on CT26 and HCT116 cells. Cells were seeded into 6-well plates at 2 × 10⁵ cells per well, incubated for 24 h to allow adhesion, and treated with different experimental group extracts for 24 h. Cells were digested with EDTA-free trypsin-EDTA, collected by centrifugation, and washed with cold PBS. After staining with Annexin V-FITC and PI according to the kit instructions, cell apoptotic status was detected by flow cytometry (Beckman CytoFLEX S, USA) within 1 h.

Evaluation for ICD effect of hydrogels

The immunofluorescence assay was performed to detect the expression of immunogenic cell death (ICD) markers (CRT and HMGB1). CT26 cells were seeded into confocal dishes at 1 × 10⁴ cells per dish, incubated for 24 h to allow adhesion, and co-cultured with different group extracts for 24 h. CT26 cells were then washed, fixed with 4% paraformaldehyde, and blocked with 5% bovine serum albumin (BSA). Cells were incubated with primary antibodies against CRT and HMGB1 at 4 °C overnight, followed by incubation with fluorescently labeled secondary antibodies at room temperature for 1 h. After DAPI nuclear staining, cells were observed and imaged under a fluorescence microscope (DP80, Olympus, Japan).

Anti-tumor effect of hydrogels in vivo

All animal experiments were approved by the Ethics Committee of Zhonghong Bohuan Biotechnology Co., Ltd (LL202509300001). A peritoneal metastasis model was established by intraperitoneal injection of 5 × 10⁵ CT26-luc cells into BALB/c mice. After tumor establishment, mice were randomly assigned to six groups (Control, CPO, OXA-RES, CPO-OXA, CPOR, and CPOMO; n = 6 per group).

One week post-injection, tumor formation was monitored by IVIS, and mice with successful tumor growth received intraperitoneal injections of 200 µL of the corresponding hydrogel or PBS (control). After one week of treatment, IVIS imaging was repeated, mice were sacrificed, and intra-abdominal tumor burden was documented. The volume of ascites in each mouse was recorded.

Proteomics analysis

Tumor tissues (50 mg) from the Control and CPOMO groups were rinsed with pre-cooled PBS and ground into powder in liquid nitrogen. Total proteins were extracted on ice for 30 min and centrifuged at 12,000 rpm for 20 min at 4 °C. Protein concentration was determined using a BCA assay kit.

For proteomic analysis, 100 µg of protein per sample was reduced with 10 mM DTT (56 °C, 30 min), alkylated with 20 mM iodoacetamide (room temperature, 30 min, dark), and digested with trypsin (protein: trypsin = 50:1, 37 °C, 16 h). Peptides were desalted using C18 columns and lyophilized.

Peptide samples were dissolved in 0.1% formic acid and analyzed by LC-MS/MS on a Q Exactive Plus coupled to an Easy-nLC 1200 system. Solvent A (0.1% formic acid in water) and solvent B (0.1% formic acid in acetonitrile) were used with a linear gradient. Data were acquired in data-dependent acquisition mode.

Protein identification and quantification were performed using MaxQuant against the UniProt mouse database. Differentially expressed proteins (DEPs) were selected with fold change > 2 or < 0.5 and P < 0.05. Functional annotation and pathway enrichment were conducted using KEGG and GO databases.

Immunostaining and histological analysis

Harvested tumor tissues were subjected to histological analysis (n = 3 per group). Tissue sections were prepared according to standard protocols and imaged using fluorescence or confocal microscopy. Quantitative analysis was performed using ImageJ software.

Tumor proliferation was evaluated by Ki-67 immunohistochemical staining, and the Ki-67-positive rate was calculated from three randomly selected high-power fields (HPFs) per section. Apoptosis was assessed by TUNEL staining and quantified as the percentage of TUNEL-positive cells relative to total DAPI-positive cells.

The expression levels of HIF-1α, VEGF, and CD31 were quantified based on mean fluorescence intensity (MFI) measured in three random HPFs per section using ImageJ software.

Western blot (WB) analysis

To further investigate the molecular mechanisms underlying the anti-tumor effects of CPOMO, Western blot analysis was performed on tumor tissues to validate proteins identified by proteomic analysis. Total proteins were extracted using RIPA buffer containing protease and phosphatase inhibitors, separated by SDS-PAGE, transferred onto PVDF membranes, and probed with antibodies against HIF-1α, VEGF, BAX, BCL2, Vimentin, E-cadherin, and N-cadherin. Protein expression was quantified using ImageJ with GAPDH as the internal control.

To assess the involvement of TGF-β signaling in EMT regulation, CT26 cells were treated with hydrogel extracts in the presence or absence of exogenous TGF-β1 (10 ng/mL), followed by Western blot analysis.

Statistical analysis

All statistical analyses were performed by Origin software. Experimental data were expressed as the mean ± standard deviation (SD). Differences between groups were evaluated using Student’s t-test or one-way analysis of variance (ANOVA). Statistical significance was defined as p < 0.05, with *p < 0.05, **p < 0.01, and ***p < 0.001. ns. no significance.

Results

Physicochemical properties of hydrogels

The successful application of an intraperitoneal drug delivery system requires rapid in situ gelation, injectability, structural adaptability, and long-term mechanical integrity. Upon mixing CMCS-PBA and ODEX aqueous solutions, gelation occurred within approximately 20 s under physiological conditions (Fig. 1A), indicating efficient formation of the dual dynamic network through borate ester and Schiff base linkages. Such rapid gelation is particularly advantageous for intraperitoneal administration, as it minimizes precursor diffusion prior to network formation and enables localized depot establishment.

Fig. 1.

Physicochemical properties of CPO hydrogel. (A) Digital images showing the gel formation process of CPO hydrogel: CMCS-PBA solution (left), ODEX solution (middle), and the formed hydrogel after mixing (right). (B) Injectability test of the hydrogel precursor, with the hydrogel shaped into the letters “NCU” post-extrusion. (C) Adhesion test and plasticity test. (D) Self-healing property of CPO hydrogel: stained and unstained hydrogel halves before (left) and after 3 min of contact (right)

The precursor solution exhibited pronounced shear-thinning behaviour and could be smoothly extruded through a syringe needle without clogging (Fig. 1B). After extrusion, the hydrogel rapidly recovered its mechanical integrity, consistent with the reversible nature of the dynamic covalent bonds. This reversible bond dissociation and reformation mechanism allows the material to withstand shear stress during injection while restoring its structure in situ, a feature essential for minimally invasive administration in the dynamic peritoneal environment.

The hydrogel demonstrated good adhesion to multiple substrates and could be moulded into various shapes without fracture (Fig. 1C). Such plasticity facilitates close contact with irregular peritoneal tumor nodules, which is expected to enhance local drug retention and therapeutic efficiency. Moreover, separated hydrogel segments fused seamlessly within 3 min (Fig. 1D), confirming effective self-healing. In the context of peritoneal implantation, this property is critical, as mechanical disruption caused by organ movement could otherwise compromise depot integrity and drug release control.

Collectively, these physicochemical characteristics establish the hydrogel as a structurally robust and functionally adaptable platform suitable for intraperitoneal drug delivery.

Characterization of MSNs@RES and physicochemical verification of hydrogel components

Transmission electron microscopy revealed that MSNs possessed a uniform spherical morphology with ordered mesoporous channels, which were preserved after RES loading (Fig. 2A). The encapsulation efficiency (85.1%) and drug loading capacity (28.3%) indicate effective incorporation of RES into the nanoporous matrix. Sustained RES release over 60 h without burst behavior (Fig. 2B) confirms controlled diffusion from the mesopores, which is crucial for overcoming the rapid degradation and poor bioavailability associated with free RES.

Fig. 2.

Comprehensive characterization of MSNs@RES and CMCS-PBA/ODEX hydrogel components. (A) Transmission electron microscopy (TEM) micrographs of pristine MSNs (left) and resveratrol-loaded MSNs (MSNs@RES, right). Scale bar: 500 nm. (B) In vitro cumulative release profile of resveratrol (RES) from MSNs@RES in PBS at 37 °C. (C) ¹H nuclear magnetic resonance (¹H NMR) spectra of CMCS-PBA (left) and ODEX (right). (D) FTIR spectra of CMCS, CMCS-PBA, DEX, ODEX, and blank CPO hydrogel. (E) Rheological behaviors of CPO and CPOMO hydrogels: time sweep (left), frequency sweep (middle), and complex viscosity (right). (F) Scanning electron microscopy (SEM) images of freeze-dried CPO (left) and CPOMO (right) hydrogels. Scale bar: 200 μm. (G) Compressive stress-strain curves of CPO and CPOMO hydrogels. Data are presented as mean±standard deviation (SD, n = 3)

¹H NMR spectra provided clear evidence for successful oxidation of dextran and grafting of phenylboronic acid onto CMCS (Fig. 2C). Specifically, ODEX exhibited a characteristic aldehyde proton signal at 9.0-9.8 ppm, consistent with previous reports of dextran oxidation [32]. In CMCS-PBA, new aromatic proton signals appeared at 7.2–7.8 ppm, corresponding to the phenyl protons of grafted PBA, in agreement with literature precedent for PBA-modified chitosan derivatives [33]. These observations together confirmed the intended chemical modifications.

FTIR spectra further supported the formation of the hydrogel network (Fig. 2D). The aldehyde absorption peak in ODEX and the aromatic ring vibrations in CMCS-PBA were clearly observed, while the emergence of a C = N stretching vibration around 1630–1650 cm⁻¹ in the hydrogel indicated successful Schiff base formation. Additionally, the boronic acid-related bands shifted, suggesting establishment of boronate ester interactions, in line with previously reported dynamic cross-linking behavior in similar systems [34]. Collectively, the NMR and FTIR data provide strong evidence for the successful construction of the dual dynamic cross-linked hydrogel.

Rheological measurements showed that both CPO and CPOMO hydrogels exhibited dominant elastic behavior (G′ > G″) across the tested frequency range (Fig. 2E), indicating stable network formation. Incorporation of MSNs@RES slightly increased the storage modulus and complex viscosity, with G′ reaching approximately 400 Pa and G″ around 45 Pa, suggesting nanoparticle-mediated reinforcement of the polymer matrix. This mechanical enhancement may contribute to improved structural stability during prolonged intraperitoneal retention.

Additional rheological tests were performed to further evaluate the practical applicability of the hydrogel. Stress sweep measurements demonstrated a stress-responsive behavior, in which the hydrogel network underwent structural disruption under elevated stress, supporting its injectability. Step-strain tests showed rapid recovery of the storage modulus to ~ 95% of its original value after repeated high-strain deformation, indicating excellent self-healing capability. Furthermore, extended time sweep measurements at 37 °C revealed stable viscoelastic properties over 1 h, with G′ ≈ 460 Pa and G″ ≈ 20 Pa throughout the testing period, confirming the long-term rheological stability of the hydrogel under physiologically relevant conditions (Figure S1).

SEM imaging revealed an interconnected porous architecture in both CPO and CPOMO hydrogels (Fig. 2F), favorable for drug diffusion and mass transport. MSNs were uniformly distributed within the CPOMO matrix without visible aggregation. Swelling studies showed rapid water uptake reaching over 150% within 12 h (Figure S2), which supports diffusion-based drug release while maintaining structural stability. Approximately 50% of the hydrogel mass remained after 14 days, suggesting sustained presence within the peritoneal cavity during the therapeutic window.

Mechanical testing demonstrated adequate compressive strength (≈ 35 kPa for CPOMO, Fig. 2G), sufficient to resist mechanical stress in vivo. Together, these data confirm that the hydrogel-nanoparticle composite maintains structural integrity while providing a permissive architecture for controlled drug delivery.

pH-responsive drug release and in vivo retention

The tumor microenvironment of peritoneal metastasis is mildly acidic, which can be exploited for selective drug release. RES and OXA were quantified at their characteristic UV absorption wavelengths of 306 nm and 210 nm, respectively, using independently established calibration curves (Figure S3). CPOMO exhibited accelerated release of both drugs under acidic conditions (pH 6.5) compared with physiological pH 7.4 (Fig. 3A, B). After 8 days, cumulative OXA release reached 86.8% at pH 6.5 versus 67.9% at pH 7.4, while RES displayed a similar release trend. This pH-responsive behavior is likely attributable to partial destabilization of borate ester bonds under acidic conditions, resulting in loosening of the hydrogel network and enhanced drug diffusion. Additional release studies were further performed in PBS containing 0.5 mg/mL lysozyme, a concentration comparable to the physiological level of lysozyme in human body cavity [14]. A moderately accelerated release behavior was observed under enzymatic conditions, possibly due to partial loosening of the dynamic hydrogel network (Figure S4).

Fig. 3.

In vitro pH-responsive drug release and in vivo retention behavior of CPOMO hydrogel. (A) Cumulative release kinetics of OXA from CPOMO in PBS at pH 6.5 and 7.4 (37 °C). (B) Cumulative release kinetics of RES from CPOMO in PBS at pH 6.5 and 7.4 (37 °C). (C) Quantitative analysis of fluorescence intensity from in vivo imaging of BALB/c mice. (D) In vivo fluorescence images of mice intraperitoneally injected with Sulfo-Cy5.5-loaded PBS (Cy5.5@PBS) or Sulfo-Cy5.5-loaded CPOMO (Cy5.5@CPOMO) at different time points. Data are presented as mean ± SD (n = 3)

Importantly, neither drug exhibited significant burst release, which is critical to avoid acute toxicity. The sustained release profile supports prolonged tumor exposure while In vivo imaging further demonstrated prolonged peritoneal retention of hydrogel-encapsulated Cy5.5 compared to rapid clearance of free dye (Fig. 3C, D). The persistent fluorescence signal over 3 days confirms formation of a stable intraperitoneal depot, supporting sustained local drug delivery.

Biocompatibility of hydrogels

Biocompatibility is essential for clinical translation. L929 fibroblast viability remained above 95% after exposure to hydrogel extracts (Fig. 4A, B), indicating negligible cytotoxicity. Hemolysis rates were below 5% (Fig. 4C), meeting established safety standards for biomaterials.

Fig. 4.

In vitro and in vivo biocompatibility assessment of hydrogels. (A) Live/dead staining images of L929 fibroblast cells treated with hydrogel extracts for 3 days. Scale bar: 100 μm. (B) Cell viability of L929 cells treated with hydrogel extracts for 1, 2, and 3 days, detected by Cell Counting Kit-8 (CCK-8) assay. (C) Hemolysis rate of hydrogels. (D) Hematoxylin and eosin (H&E) staining images of major organs (heart, liver, spleen, lung, kidney) from mice injected with hydrogels. Scale bar: 100 μm. (E) Serum biochemical analysis of ALT, AST, BUN, and CREA. Data are presented as mean ± SD (n = 3)

In vivo histological analysis of major organs revealed no significant tissue damage or inflammatory infiltration (Fig. 4D). Furthermore, serum biochemical analysis showed no significant abnormalities in ALT, AST, BUN, or CREA levels among the different treatment groups (Fig. 4E), indicating the absence of obvious hepatic or renal toxicity. These results further support the favorable biosafety of the hydrogel system and its suitability for intraperitoneal administration.

In vitro anti-tumor and anti-metastatic activity

Treatment with CPOMO significantly reduced CT26 cell viability compared with free drug or single-drug hydrogel controls (Fig. 5A, D). Similar anti-proliferative effects were further confirmed in HCT116 cells, in which CPOMO treatment also showed the strongest inhibitory activity among all groups (Figure S5). The observed reduction exceeded the additive effects of individual treatments, suggesting combination interaction between OXA and RES. RES is known to sensitize CRC cells to platinum-based chemotherapy by modulating apoptotic and EMT-related pathways, which likely contributes to the enhanced efficacy observed.

Fig. 5.

In vitro anti-tumor activities of CPOMO hydrogel against CT26 cells. (A) Live/dead staining images of CT26 cells treated with different hydrogel extracts for 24 h. Scale bar: 100 μm. (B) Transwell assay images of CT26 cells. Scale bar: 100 μm. (C) Wound healing assay images of CT26 cells at 0, 24, and 48 h post-scratching. Scale bar: 100 μm. (D) CCK-8 assay results of CT26 cells treated with different hydrogel extracts. (E) Quantitative analysis of wound closure rate at 24 and 48 h. (F) Quantitative analysis of invaded CT26 cells per field. (G) Flow cytometry dot plots showing apoptotic status of CT26 cells. (H) Immunofluorescence images of immunogenic cell death (ICD) markers: CRT, red and high mobility group protein HMGB1, green. Scale bar: 50 μm. (I) Quantitative analysis of apoptosis rate. (J, K) Quantitative analysis of nuclear HMGB1 and CRT-positive cell ratio. Data are presented as mean ± SD. (n = 3)

Transwell and wound-healing assays demonstrated substantial suppression of migration and invasion in the CPOMO group (Fig. 5B, C, E, F). Given that peritoneal metastasis relies on cellular dissemination and mesothelial invasion, inhibition of these processes is particularly relevant.

Flow cytometry revealed a markedly increased apoptotic rate in CPOMO-treated cells (Fig. 5G, I). Consistently, flow cytometry also demonstrated increased apoptosis induction in HCT116 cells following CPOMO treatment compared with free drugs and single-drug formulations (Figure S6). Furthermore, immunofluorescence analysis demonstrated enhanced calreticulin (CRT) exposure and HMGB1 release (Fig. 5H, J, K), indicating induction of immunogenic cell death. The ability to induce ICD-related responses suggests that CPOMO may not only exert direct cytotoxicity but also potentially contribute to immune-related anti-tumor effects, which warrants further investigation.

In vivo therapeutic efficacy in a peritoneal metastasis model

In the CT26-luc peritoneal metastasis model, CPOMO significantly suppressed tumor burden as evidenced by reduced bioluminescence intensity (Fig. 6B, D, F). The reduction in tumor signal correlated with decreased intra-abdominal tumor nodules and ascites formation (Fig. 6E–H). Ascites reduction is clinically meaningful, as malignant ascites contributes substantially to morbidity in PM patients.

Fig. 6.

In vivo therapeutic efficacy of CPOMO hydrogel in BALB/c mice model of colorectal cancer peritoneal metastasis. (A) Schematic illustration of the animal experiment protocol. (B) In vivo bioluminescence images of mice in different groups at day 7 and 14. (C) Survival curves of mice (n = 6 per group). (D, F) Quantitative analysis of bioluminescence intensity at day 7 and 14. (E) Digital photographs of intra-abdominal tumors after mouse sacrifice. (G) Digital photographs of collected ascites from each group. (H) Quantitative analysis of ascites volume. Data are presented as mean ± SD. (n = 6)

Importantly, CPOMO markedly improved survival, with 50% of treated mice surviving beyond 60 days (Fig. 6C). In contrast, free drugs and single-drug hydrogels provided limited survival benefit. These findings demonstrate that sustained, localized, and combined drug delivery translates into meaningful therapeutic improvement in vivo.

Molecular mechanisms underlying therapeutic efficacy

Principal component analysis (PCA) indicated that the selected tumor tissues showed good agreement and could be further analyzed (Figure S7). Proteomic analysis identified 159 differentially expressed proteins between control and CPOMO-treated tumors, enriched in angiogenesis, apoptosis, and EMT-related pathways (Fig. 7A–C). These findings were validated by Western blot analysis.

Fig. 7.

Proteomic analysis and Western blot validation of the anti-tumor mechanisms of CPOMO hydrogel. (A) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of differentially expressed proteins (DEPs) between Control and CPOMO groups. (B) Gene Ontology (GO) function enrichment analysis of DEPs. (C) Volcano plot of DEPs (red dots: up-regulated proteins, blue dots: down-regulated proteins, black dots: non-significant proteins). (D, F) WB bands and quantitative analysis of HIF-1α-VEGF pathway-related proteins. (E, G) WB bands and quantitative analysis of epithelial-mesenchymal transition. (H, J) WB bands and quantitative analysis of apoptotic pathway-related proteins (BAX, BCL2). (I, K) WB bands and quantitative analysis of epithelial-mesenchymal transition (EMT)-related proteins (Vimentin, N-cadherin, E-cadherin) and TGF-β1 rescue assay results. Data are presented as mean ± SD. (n = 3)

CPOMO treatment downregulated HIF-1α and VEGF expression (Fig. 7D, F), consistent with reduced angiogenesis. Immunofluorescence staining further confirmed decreased CD31-positive microvessel density (Fig. 8A–F). Inhibition of tumor angiogenesis may restrict nutrient supply, thereby contributing to suppression of tumor progression.

Fig. 8.

Immunofluorescence and immunohistochemical validation of key signalling pathways regulated by CPOMO hydrogel. (A-C) Immunofluorescence images of HIF-1α (A), VEGF (B), and CD31 (C) in tumor tissues. Scale bar: 100 μm. (D-F) Quantitative analysis of mean fluorescence intensity (MFI) of HIF-1α, VEGF, and CD31. (G) Ki-67 immunohistochemical images of tumor tissues (proliferation marker). Scale bar: 50 μm. (H) TUNEL immunofluorescence images of tumor tissues (apoptosis marker. Scale bar: 100 μm. (I) Quantitative analysis of Ki-67 positive rate. (J) Quantitative analysis of apoptotic index. Data are presented as mean ± SD. (n = 3)

Concurrently, CPOMO upregulated pro-apoptotic BAX and downregulated anti-apoptotic BCL2 (Fig. 7H, J), suggesting activation of intrinsic apoptotic pathways. EMT markers were also significantly modulated, with decreased Vimentin and N-cadherin and increased E-cadherin expression (Fig. 7E, G). Rescue experiments using exogenous TGF-β1 partially reversed these changes, suggesting a potential involvement of TGF-β pathway inhibition in EMT suppression (Fig. 7I, K).

Ki-67 and TUNEL staining further corroborated reduced proliferation and enhanced apoptosis in vivo (Fig. 8G–J). Taken together, these findings suggest that CPOMO may exert multi-level therapeutic effects associated with suppressing angiogenesis, inhibition of EMT- related metastasis, and promoting apoptosis.

Discussion

Peritoneal metastasis represents one of the most challenging clinical scenarios in colorectal cancer management, with limited therapeutic options and dismal prognosis [35]. Injectable dynamic hydrogels have been widely explored for biomedical applications. In this context, the hydrogel platform developed in this study builds upon previously reported systems and further extends their potential for colorectal cancer therapy through an optimized design and functional integration. These features collectively support its application as a localized drug delivery system for peritoneal metastasis treatment. In this study, we developed a dual dynamic covalently cross-linked injectable hydrogel (CPOMO) for sustained intraperitoneal co-delivery of oxaliplatin and resveratrol-loaded mesoporous silica nanoparticles, and systematically evaluated its therapeutic efficacy and underlying mechanisms in a murine model of CRC peritoneal metastasis. The results demonstrate that this localized delivery platform not only suppressed tumor progression and reduced ascites formation but also significantly prolonged survival, with mechanistic insights revealing coordinated modulation of angiogenesis, apoptosis, and epithelial–mesenchymal transition pathways.

Importantly, the conceptual advance of this study lies in integrating dynamically adaptive hydrogel engineering with sustained combination therapy for colorectal cancer peritoneal metastasis. The rational design of the hydrogel system addresses several key limitations of conventional intraperitoneal chemotherapy. The dual dynamic cross-linking network—comprising borate ester and Schiff base bonds—enables rapid in situ gelation, injectability, and self-healing properties. These features are particularly advantageous for intraperitoneal application, where the hydrogel must withstand continuous mechanical agitation while maintaining structural integrity. The observed shear-thinning behavior and rapid mechanical recovery indicate its suitability for minimally invasive administration and stable depot formation within the peritoneal cavity.

The incorporation of MSNs@RES addresses the longstanding challenge of delivering hydrophobic chemosensitizers like resveratrol with poor bioavailability. MSNs provided efficient RES encapsulation and sustained release, while incorporation into the hydrogel matrix further prolonged drug release. The observed pH-responsive release behavior may facilitate preferential drug release within the acidic tumor microenvironment. This feature may enhance therapeutic selectivity while reducing systemic exposure, although direct evidence of tumor-specific release in vivo warrants further investigation using imaging-guided pharmacokinetic studies.

The therapeutic superiority of CPOMO over free drugs and single-drug hydrogels supports the concept of enhanced combination therapy. Resveratrol has been widely reported to enhance the responsiveness of tumor cells to platinum-based chemotherapy [36]. The observation that CPOMO significantly enhanced apoptosis compared to single-drug treatments is consistent with this chemosensitizing effect. Furthermore, the induction of immunogenic cell death markers (CRT exposure and HMGB1 release) suggests the potential involvement of anti-tumor immune responses. ICD is increasingly recognized as a critical determinant of long-term therapeutic success, as it has been reported to promote dendritic cell maturation and cytotoxic T lymphocyte infiltration [37]. Whether the observed ICD-related responses translate into adaptive immune memory and protection against tumor rechallenge remains to be determined in future studies employing immunocompetent models with longer follow-up.

Proteomic analysis provided broad insights into the molecular changes associated with CPOMO treatment. The significant enrichment of differentially expressed proteins in angiogenesis-related pathways (HIF-1α, VEGF) is consistent with the reported anti-angiogenic properties of resveratrol and the known role of oxaliplatin in disrupting tumor vascularization [38]. Immunofluorescence validation further supported reduced HIF-1α and VEGF expression accompanied by decreased microvessel density, suggesting suppression of tumor angiogenesis.

Concurrently, modulation of EMT markers—including downregulation of mesenchymal markers (Vimentin, N-cadherin) and upregulation of epithelial marker E-cadherin—suggested partial reversal of the mesenchymal phenotype associated with metastatic dissemination. TGF-β signaling is recognized as an important regulator of EMT in CRC [39], and the partial reversal of CPOMO-induced EMT changes by exogenous TGF-β1 suggested the potential involvement of TGF-β-related signaling in this process. The dual inhibition of angiogenesis and EMT by CPOMO may therefore contribute to suppression of peritoneal metastasis progression.

Despite the promising therapeutic efficacy and favorable biosafety demonstrated in the present study, several limitations should be acknowledged. First, although the injectable hydrogel exhibited desirable physicochemical properties and sustained drug release behavior, its degradation kinetics and long-term retention under the complex intraperitoneal environment remain to be further investigated. In addition, the long-term biodistribution and clearance profile of mesoporous silica nanoparticles following hydrogel degradation were not evaluated in the current study. Furthermore, while no obvious systemic toxicity was observed, the potential inflammatory responses associated with repeated intraperitoneal administration require additional assessment. Future studies addressing these aspects, together with validation in larger and clinically relevant models, will further facilitate the translational development of this localized drug delivery platform for colorectal cancer peritoneal metastasis.

Conclusion

We developed a dual dynamic covalently cross-linked injectable hydrogel (CPOMO) for sustained intraperitoneal co-delivery of oxaliplatin and resveratrol-loaded mesoporous silica nanoparticles to treat colorectal cancer peritoneal metastasis. The hydrogel rapidly formed in situ and exhibited injectability, self-healing capability, mechanical stability, and pH-responsive drug release, enabling prolonged retention and tumor-targeted delivery within the peritoneal cavity.

CPOMO demonstrated excellent biocompatibility and superior therapeutic efficacy compared to free drugs and single-drug hydrogels. It effectively suppressed tumor growth, reduced ascites formation, and significantly prolonged survival in a murine peritoneal metastasis model. Mechanistic investigations suggested that CPOMO treatment was associated with inhibition of HIF-1α/VEGF-related angiogenesis and TGF-β-related epithelial–mesenchymal transition, along with promotion of apoptosis, supporting a multi-pathway therapeutic profile.

Overall, this study presents a mechanically adaptive and biologically combined localized delivery platform with strong potential for improving therapeutic outcomes in colorectal cancer peritoneal metastasis.

Author contributions

**Haonan Huang: ** Writing – original draft, Visualization, Resources, Methodology, Data curation, Conceptualization. **Haoyu Huang: ** Writing – review & editing, Visualization, Methodology, Investigation, Validation. **Cheng Zuo: ** Writing – original draft, Visualization, Validation. **Jing Yang: ** Writing – original draft, Validation, Validation, Software. **Huizi Li** and **Fan Jiang: ** Resources, Methodology, Formal analysis. **Ruibing Li** and **Lipeng Zhang: ** Supervision, Resources, Methodology. **Siyu Rao: ** Resources, Project administration. **Fanrong AI: ** Writing – review & editing, Resources, Project administration, Funding acquisition. **Zhen Zong: ** Supervision, Resources, Methodology, Conceptualization. **Shengxun Mao: ** Writing – review & editing, Supervision, Resources, Methodology, Conceptualization, Funding acquisition.

Funding

This work was supported by the internal funding project of Nanchang University (No. 2023efyBo5), the Science Fund for Distinguished Young Scholars of Jiangxi Province (No. 20224ACB214005); the Jiangxi Province Key Research and Development Plan (No. 20232BBE50017).

Data availability

Data will be made available on request.

Declarations

Competing interests

The authors declare no competing interests.