Hydrogel co-delivery of 5-fluorouracil and siRNA attenuates TGF-β1-mediated MMT to prevent postoperative peritoneal metastasis and adhesion in colorectal cancer.

Jinpeng Wen: Conceptualization, Methodology, Project administration, Writing – original draft. Jiangchuan He: Methodology, Project administration. Yunhe Zheng: Methodology, Project administration. Ting Wang: Formal analysis, Methodology. Kailai Liu: Data curation, Supervision. Hanchao Zhou: Software, Investigation. Yuchen Zhang: Investigation, Validation. Yu Huang: Investigation, Validation. Chen Zhang: Investigation, Validation. Yue Zhang: Investigation, Validation. Ke Wang: Conceptualization, Funding acquisition, Writing – review & editing.

As illustrated schematically in Fig. 1 A, a free radical copolymerization of CGC, MPC and 5-FU. was used to fabricate FCGCM hydrogel under 365 nm irradiation. The hydrogel backbone, CGC macromer, was created via hydroxyl group-induced epoxy ring-opening followed by esterification reaction (as depicted in Fig. S1A–S1B). Spectroscopic characterizations confirmed the chemical modifications: FTIR analysis (Fig. S1C) exhibited characteristic absorption bands at 1715 and 1562 cm −1 , verifying the conjugation of GMA and Cit to CS (carbonyl stretching vibration and amide Ⅱ band). The 1 H NMR characterization (Fig. S1D) revealed the presence of two diagnostic resonances at 5.6 and 6.1 ppm arising from the vinyl protons of GMA. The peak at 3.1 ppm was assigned to the primary amine protons from Cit. Taken together, these spectral patterns are indicative of successful immobilization of both functional groups on CS backbone. Furthermore, compared to a single GMA grafting procedure, this modification provides more active binding sites of 5-FU without substantially affecting the structure of CS main chain [ 48 ]. Fig. 1 Preparation and characterizations of FCGCM hydrogel. (A) Schematic illustration of the preparation of FCGCM hydrogel; (B) SEM micrographs depicting the morphological characteristics of CGM and FCGCM hydrogels. Scale bar: 200 µm; (C) Quantitative analysis of EWC in CGM versus FCGCM hydrogels, mean ± SD ( n = 3), ns P > 0.05; (D) MDS results demonstrating hydrogen bonds between Cit and 5-FU over a 200-ns trajectory; (E) Cumulative 5-FU release kinetics from CGM and FCGCM hydrogels under physiological conditions, mean ± SD ( n = 5), ** P < 0.01; (F) Time-dependent in vitro degradation profiles of FCGCM hydrogel incubated in PBS versus chondroitinase-enriched solution (75 U/ml), mean ± SD ( n = 5); (G) In vivo imaging of Cy-7 labeled FCGCM hydrogel; (H) Quantitative analysis of total radiant efficiency derived from fluorescence imaging, mean ± SD ( n = 4). Fig 1 dummy alt text Preparation and characterizations of FCGCM hydrogel. (A) Schematic illustration of the preparation of FCGCM hydrogel; (B) SEM micrographs depicting the morphological characteristics of CGM and FCGCM hydrogels. Scale bar: 200 µm; (C) Quantitative analysis of EWC in CGM versus FCGCM hydrogels, mean ± SD ( n = 3), ns P > 0.05; (D) MDS results demonstrating hydrogen bonds between Cit and 5-FU over a 200-ns trajectory; (E) Cumulative 5-FU release kinetics from CGM and FCGCM hydrogels under physiological conditions, mean ± SD ( n = 5), ** P < 0.01; (F) Time-dependent in vitro degradation profiles of FCGCM hydrogel incubated in PBS versus chondroitinase-enriched solution (75 U/ml), mean ± SD ( n = 5); (G) In vivo imaging of Cy-7 labeled FCGCM hydrogel; (H) Quantitative analysis of total radiant efficiency derived from fluorescence imaging, mean ± SD ( n = 4). In order to elucidate the impact of 5-FU incorporation on hydrogel properties, the control CGM hydrogels made of CS-GMA and MPC were prepared. SEM imaging in Fig. 1 B showed that both FCGCM and CGM hydrogels had a similar three-dimensional porous structure with an average pore diameter of 100 µm. The mentioned pore size was found optimal for local drug delivery into peritoneal cavity as it ensures maximum loading of the drugs and permeation of the nutrients while preventing quick clearance from peritoneal mesothelium [ 49 , 50 ]. The complementary EDS mapping shown in Fig. S2 suggested that the 5-FU can be successfully incorporated into the FCGCM hydrogel and that its spatial distribution is reasonably homogeneous. The drug functionally blended within the hydrogel unlike in physically blended hydrogels where the drug is administered unevenly [ 51 ]. The quantity of the hydration capacity was assessed via PBS. Both hydrogel formulations demonstrated remarkable water retention capabilities, as evidenced by EWC values that surpassed 96% ( Fig. 1 C). After statistical analysis, EWC of CGM and FCGCM hydrogels showed no significant difference. Together, these results confirm that the structural and hygroscopic properties of hydrogel systems are retained when 5-FU is loaded. 5-FU has shown to be clinically significant and has antitumor activity against gastrointestinal tumors. It modulates the ECM and cell-interaction [ 52 ]. However, it is limited due to dose-limiting toxicity, fast clearance from the systemic circulation and lack of sufficient activity to kill tumor cells. To overcome these challenges, CGC was synthesized via covalent grafting, establishing a hydrogen bond-mediated 5-FU release mechanism within the hydrogel network. The amino and carboxyl groups of Cit can form hydrogen bonds with nitrogen and fluorine atoms, respectively, of 5-FU. As a result of such intermolecular hydrogen bonds, the molecules of 5-FU are positioned in between the molecular chains of the hydrogel. This increases resistance to the diffusion of 5-FU, thereby slowing down the initial release of the drug. Furthermore, this methodology is easily biocompatible when compared to covalent drug conjugations. The covalent conjugations usually require harsh conditions for the reaction and have negative effects on the activity of the drug as well [ 53 ]. MDS were calculated as a trajectory of 200 ns to quantify intermolecular hydrogen bonds between Cit and 5-FU ( Fig. 1 D, Movie S1). The comparative study showed that Cit&5-FU had a 2.76-fold higher occurrence of hydrogen bond formation (11,238 interactions) as compared to 5-FU self-association (4072 interactions) (Fig. S3A). Moreover, the overall mean hydrogen bond energy (−8.11 kcal/mol) of Cit&5-FU was lower than 5-FU&5-FU (−6.46 kcal/mol) (Fig. S3B), indicating Cit&5-FU binds more favorably. This is crucial for 5-FU retention within the formulation. In line with the computational predictions, 5-FU release assays demonstrated ( Fig. 1 E) that the FCGCM hydrogel attenuated the burst release of 5-FU compared to the CGM hydrogel within 24 h in vitro . The profile reflects a slow, continuous release of 5-FU and is much superior than other 5-FU loaded hydrogels reported in the literature which has >80% burst release in 24 h [ 54 , 55 ]. In addition, in SPF model (Fig. S3C), the release behavior of 5-FU from FCGCM hydrogel indicated its adaptability to dynamic postoperative changes in peritoneal microenvironment. Under the pathological conditions (acidic pH, high MMP activity), the release of 5-FU was appropriately accelerated, in line with antitumor therapeutic needs with no burst release. This verified the success of the release strategy designed based on hydrogen bond mediation. This environment-responsive behavior is due to the cooperative regulation of Cit-mediated hydrogen bonds and the MMP-susceptible CS backbone for therapeutic activity in pathological peritoneal microenvironments. In order to assess the clinical utility of FCGCM hydrogel for the prevention of peritoneal adhesions, systematic investigation of its biodegradation behavior through complementary in vitro and in vivo assays was conducted. The mass loss of the springs was investigated through in vitro degradation studies ( Fig. 1 F). With chondroitinase, full mass loss was achieved in 9 days whereas in the absence of chondroitinase, it took 15 d to lose mass completely thereby indicating the presence of an enzyme-accelerated biodegradation pathway. Using chondroitinase to mimic the in vivo enzymatic microenvironment in this experiment not only confirmed that the FCGCM hydrogel could be degraded by the CS substrate-dependent enzyme but also simulated the pathological enzymatic degradation that occurs after intraperitoneal implantation. This method reflects more accurately the true in vivo application scenario of the hydrogel as a peritoneal drug delivery system. Using a covalent conjugation strategy the near-infrared fluorescent dye Cy7 was attached to the hydrogel network for in vivo tracking. Real-time fluorescence imaging showed very strong peritoneal localization after implantation which then decreased over time ( Fig. 1 G). Based on the quantitative analysis complete signal extinction was observed by day 10 post operatively ( Fig. 1 H). Thus, confirmation of the complete degradation of the hydrogel was made. T1-weighted MRI images of the subcutaneously injected hydrogels (Fig. S4) indicated the signal loss of the hydrogels. The hydrogels demonstrated nearly complete degradation by Day 10, which was consistent with the fluorescence tracking and in vitro degradation studies. The degradation time is well matched with the critical period for postoperative peritoneal adhesion formation (7–10 d) and early tumor cell seeding [ 56 ]. Together, these biodegradation characteristics of the FCGCM hydrogel, it is highly suitable for applications in the peritoneal cavity. The peritoneum and intra-abdominal visceral organs experience gliding motion which applies cyclic compressive loads on the hydrogels. Thus, the self-healing capacity and customized mechanical properties of hydrogels for intraperitoneal applications are key design criteria [ 57 ]. In line with our molecular engineering strategy, the FCGCM hydrogel network, exhibits autonomous repair capability after mechanical damage. The repair process occurs through the dynamic rearrangement of hydrogen bonding networks and electric charge interactions at the fracture sites that ultimately help the damaged hydrogel to reassemble back ( Fig. 2 A). Fig. 2 Rheological properties of FCGCM hydrogel. (A) Schematic representation of the self-healing process in FCGCM hydrogels; (B) Strain amplitude sweeps (0.1%−1000% strain, 1 Hz); (C) Time-dependent oscillatory shear measurements (1% strain, 1 Hz); (D) Thixotropic recovery evaluated through alternating low-high strain cycles (1% → 500% → 1%, 1 Hz); (E, F) Thermal responsive behavior assessed via time sweeps at 45 °C for FCGCM (E) and CGM (F) precursor solutions; (G, H) Urea-mediated hydrogen bonds disruption studies comparing FCGCM (G) and CGM (H) hydrogel networks under oscillatory shear (1% strain, 1 Hz). Fig 2 dummy alt text Rheological properties of FCGCM hydrogel. (A) Schematic representation of the self-healing process in FCGCM hydrogels; (B) Strain amplitude sweeps (0.1%−1000% strain, 1 Hz); (C) Time-dependent oscillatory shear measurements (1% strain, 1 Hz); (D) Thixotropic recovery evaluated through alternating low-high strain cycles (1% → 500% → 1%, 1 Hz); (E, F) Thermal responsive behavior assessed via time sweeps at 45 °C for FCGCM (E) and CGM (F) precursor solutions; (G, H) Urea-mediated hydrogen bonds disruption studies comparing FCGCM (G) and CGM (H) hydrogel networks under oscillatory shear (1% strain, 1 Hz). Oscillatory rheometry was utilized to progressively characterize the viscoelastic properties of hydrogels. Crossover points between the storage modulus (G’) and loss modulus (G’’) were observed for both the formulations under strain amplitude sweeps (0.1%−1000% strain, 1 Hz) ( Fig. 2 B). When the strain was low, G’ was greater than G’’, which means that the elastic behavior in the shear modulus was dominating. G’’ was greater than G’ when the strain was critical and the behavior was dominating. It was observed that FCGCM presented one crossover strain similar to that of the CGM. The results suggest that the incorporation of 5-FU did not alter the original hydrogel mechanical structure as drug loading generally alters the mechanical properties of the polymer hydrogel [ 58 ]. The time sweep tests indicated excellent structural stability, as both of the hydrogels have a constant G' and G" value throughout 200 s ( Fig. 2 C). The complementary steady-shear assays (Fig. S5) illustrated the pseudoplastic behavior, revealing a decrease in viscosity with increasing shear rate which confirms the requisite shear-thinning properties that ensure injectability. Thixotropic cycling tests were performed to examine self-recovery capability ( Fig. 2 D). With a stress of 500% to 1%, the FCGCM hydrogel exhibited a rapid recovery of modulus, reaching 89.6% recovery in 4 s and complete restoration of its original viscoelastic properties in 16 s. Despite evidence of shear-thinning, self-recovery and basic mechanical characteristics sufficient to hold its structure together for short-term post-operative use this may be important as its in vivo degradation occurs within 10-day and so too would its critical therapeutic window its crosslinking, which is controlled by hydrogen bonding, may lead to diminished modulus recovery upon repeated cyclic strain and environmental perturbations [ [59] , [60] , [61] , [62] ]. Rheological analysis at 45 °C was undertaken using complementary time-dependent protocols to demonstrate the gelation kinetics of the FCGCM and CGM hydrogels. As shown in Fig. 2 E–2F, both formulations exhibited distinct viscoelastic phase transitions: during the initial 160 s, predominant viscous behavior (G'' > G') indicated that the precursor solutions were in a quasi-liquid state, followed by a sharp modulus inversion (G' > G'') at approximately 164 s. In order to assess how non-covalent interactions impact integrity mechanistically, selective bond disruption assays were performed. Hydrogels were subjected to incubation involving NaCl (an electrostatic screening agent) or urea (a hydrogen bond disruptor). According to the time-sweep assay ( Fig. 2 G–2H and S6), hydrogels treated with NaCl or urea have lower modulus than pristine hydrogels, suggesting that complementary hydrogen bonds and electrostatic attractions are responsible for the mechanical robustness of hydrogels. This confirms the dual crosslinking of the FCGCM hydrogel, where reversible bonds (hydrogen and electrostatic) coexist with irreversible covalent networks that allows self-healing, which are the key features for minimally invasive peritoneal application. Amino-functionalized mesoporous silica nanoparticles were synthesized by direct grafting of APTES onto MSNs to deliver TGF-β1 siRNA efficiently to inhibit MMT [ 46 ]. The AMSNs thus produced spontaneously adsorb anionic siRNA and form stable complexes through electrostatic interaction ( Fig. 3 A). The N 2 adsorption-desorption isotherms were used to characterize the mesostructural properties of MSNs and AMSNs. Assessment via the BET as well as BJH patterns ( Fig. 3 B and S7) resulted in Type IV isotherms for both nanoparticles, confirming uniform mesoporous structures. The average pore diameter of the MSNs (5.67 nm) reduced to 4.71 nm in AMSNs, confirming APTES functionalization. The pore size produced is suitable for siRNA loading because the hydrodynamic diameter of siRNA is 3–5 nm [ 63 ]. Thermogravimetric analysis (TGA) of MSNs and AMSNs presented losses of weight of 18.54% and 19.89%, respectively ( Fig. 3 C). Moreover, using the TGA results, the density of APTES functional group in AMSNs was estimated to be 1.35 wt%. The emergence of a characteristic N1s peak due to X-ray photoelectron spectroscopy analysis (Fig. S8) validated surface modification. Fig. 3 Preparation and characterizations of AMSNs/siRNA complexes. (A) Schematic illustrating the formation of AMSNs/siRNA complexes; (B) Pore size distribution profiles; (C) TGA curves; (D) Zeta potential and particle size measurements of MSNs, AMSNs and AMSNs/siRNA complexes prepared at nitrogen-to-phosphate (N/P) ratios of 1:1, 2:1, 4:1, 8:1 and 16:1, mean ± SD, n = 3; (E) TEM images of MSNs, AMSNs and AMSNs/siRNA complexes, scale bar: 100 nm; (F) AMSNs protect siRNA from degradation by rnase. Fig 3 dummy alt text Preparation and characterizations of AMSNs/siRNA complexes. (A) Schematic illustrating the formation of AMSNs/siRNA complexes; (B) Pore size distribution profiles; (C) TGA curves; (D) Zeta potential and particle size measurements of MSNs, AMSNs and AMSNs/siRNA complexes prepared at nitrogen-to-phosphate (N/P) ratios of 1:1, 2:1, 4:1, 8:1 and 16:1, mean ± SD, n = 3; (E) TEM images of MSNs, AMSNs and AMSNs/siRNA complexes, scale bar: 100 nm; (F) AMSNs protect siRNA from degradation by rnase. As shown in Fig. 3 D, the condensation of siRNA into AMSNs complexes with hydrodynamic diameters ranging from 160 nm to 250 nm was observed across evaluated N/P ratios. The complexes were switched from negatively charged to positively charged with N/P ≥ 8:1. Fig. 3 E and S9 showed TEM and SEM images of MSNs and AMSNs which suggested the diameters of the particles were 100 nm and that they had a radially aligned mesoporous architecture. The elemental mapping analysis (Fig. S10) showed that oxygen (O) and silicon (Si) are the main compositions, while the surface nitrogen (N) signal is due to APTES functionalization. Agarose gel electrophoresis analysis revealed that AMSNs efficiently complexed siRNA at N/P ratios above 8:1 (Fig. S11). In a significant way, AMSNs offered robust protection against RNase-mediated degradation of complexed siRNA ( Fig. 3 F). The optimized biophysical properties, which include suitable size parameters, positive surface charge and resistance to nuclease, facilitate cellular delivery of AMSNs/siRNA complexes. In order to eliminate the influence of functional gene silencing of on-target gene delivery, a non-targeting negative control siRNA (siNC) was employed to provide the optimal N/P ratio for AMSNs/siRNA-mediated gene delivery. As the N/P ratio increased, transfection efficiency in HMrSV5 cells improved but was accompanied by an increase in cytotoxicity (Fig. S12 and S13). The N/P ratio of 8:1 was chosen for subsequent experiments for maximum transfection efficiency with minimal cellular damage. The positive control for comparison was Lipofectamine 3000 (Lipo3000), a commercial transfection reagent. The results of quantitative study by flow cytometry (Fig. S14) state that uptake of AMSNs/siRNA and Lipo3000/siRNA was significantly more than naked siRNA. Also, the combined release from AMSNs/siRNA complexes increased to around 80% at 48 h, indicating their stability, and sustained release (Fig. S15). The enhanced delivery performance of AMSNs/siRNA complexes is a result of their optimized physicochemical characteristics. The FC@MT delivery system was constructed by incorporating AMSNs/siRNA complexes into an FCGCM hydrogel to inhibit peritoneal adhesion and achieve on-demand drug release in the peritoneal cavity. Post-CRS peritoneal injury microenvironment has an increased expression of TGF-β1 that facilitates MMT [ 10 , 64 ]. As shown in Fig. 4 A, the FC@MT system can sustain the uptake of AMSNs/siRNA complexes, effectively suppressing TGF-β1 expression and inhibiting MMT. The MTT assays (Fig. S16) and live/dead cell viability staining (Fig. S17) confirmed that FC@MT has a high biocompatibility to HMrSV5 cells but selective cytotoxicity to CT26 cells. The results of hemocompatibility assessment and SEM analysis (Fig. S18) indicated minimal hemolytic activity against erythrocytes, confirming the potential applicability of FC@MT in post-CRS applications. Fig. 4 Cellular uptake and lysosomal escape of FC@MT. (A) Schematic representation of FC@MT cellular uptake and inhibition of MMT; (B) Immunofluorescence micrographs of HMrSV5 cells treated with naked siRNA or FC@MT. Nuclei: blue (DAPI), siRNA: green (FAM-labeled), cytoskeleton: red (phalloidin). Scale bar: 40 µm; (C) Lysosomal escape ability of FC@MT in HMrSV5 cells. siRNA: green (FAM-labeled), lysosome: red (lyso-tracker). Scale bar: 40 µm; (D, E) RT-qPCR analysis of Cdh1 (D) and Tgfb1 (E) mRNA expression levels, * P < 0.05, ** P < 0.01, *** P < 0.001, mean ± SD ( n = 3); (F) Western blot assessing protein expression of MMT-associated biomarkers; (G) In vivo bioluminescence imaging of CT26-luc tumor cells within the peritoneal cavity. Fig 4 dummy alt text Cellular uptake and lysosomal escape of FC@MT. (A) Schematic representation of FC@MT cellular uptake and inhibition of MMT; (B) Immunofluorescence micrographs of HMrSV5 cells treated with naked siRNA or FC@MT. Nuclei: blue (DAPI), siRNA: green (FAM-labeled), cytoskeleton: red (phalloidin). Scale bar: 40 µm; (C) Lysosomal escape ability of FC@MT in HMrSV5 cells. siRNA: green (FAM-labeled), lysosome: red (lyso-tracker). Scale bar: 40 µm; (D, E) RT-qPCR analysis of Cdh1 (D) and Tgfb1 (E) mRNA expression levels, * P < 0.05, ** P < 0.01, *** P < 0.001, mean ± SD ( n = 3); (F) Western blot assessing protein expression of MMT-associated biomarkers; (G) In vivo bioluminescence imaging of CT26-luc tumor cells within the peritoneal cavity. To investigate the cell uptake efficiency of the released AMSNs/siRNA from FC@MT delivery system, HMrSV5 cells which were incubated with FC@MT were used for cytoskeletal and nuclear staining with phalloidin and DAPI. Refer to Fig. 4 B, when it comes to siRNA delivery efficiency, the FC@MT system performs a lot better than naked ones. The quantitative analysis of fluorescence intensities ratios (siRNA/nucleus; green/blue) corroborated the effective cellular uptake of AMSNs/siRNA complexes released from FC@MT (Fig. S19B). In CT26 cells, FC@MT-released AMSNs/siRNA complexes still showed efficient cellular uptake (Fig. S19A and S19C). TEM micrographs of CT26 cells co-incubated with FC@MT showed successful internalization of AMSNs/siRNA (Fig. S20). AMSNs were located in the endosomes and cytoplasm of CT26 cells. This demonstrates that the complexes escape the endosomal pathway to reach the cytoplasm where siRNA can act [ 65 ]. Rapid lysosomal degradation of naked siRNA greatly inhibits their endosomal escape and downstream gene-silencing effect, and thus, the capacity for lysosomal escape is an important metric for evaluation of siRNA delivery systems. Fig. 4 C showed that FC@MT-released AMSNs/siRNA complexes exhibited markedly reduced co-localization with lysosomes compared to naked siRNA in HMrSV5 cells, as evidenced by a significant decrease in the Pearson’s correlation coefficient (Fig. S21B). We consistently observed this enhanced escape phenotype in CT26 cells (Fig. S21A and S21C). Due to the established function of TGF-β1 as master regulator of MMT [ 66 ], we developed FC@MT delivery system to co-deliver TGF-β1 siRNA in order to suppress TGF-β1 expression and MMT. In the beginning, three TGF-β1 siRNA sequences were designed and fabricated, and the candidate which had the highest gene silencing efficiency was selected through RT-qPCR. As illustrated in the Fig. S22, TGF-β1 siRNA-96 silenced Tgfb1 more effective than TGF-β1 siRNA-296 and TGF-β1 siRNA-622. We subsequently analyzed the inhibitory influence of FC@MT on MMT progression. The RT-qPCR analysis confirmed that the addition of FC@MT observed transcriptional modulation of MMT markers of HMrSV5 cells. The results showed that mRNA expression of Cdh1 was significantly increased. Whereas, the mRNA levels of Tgfb1, Col1a1, Vim and Acta2 showed significantly decreased ( Figs. 4 D–4E and S23). Western blot assays indicated consistent modulation at the protein level, with FC@MT treatment upregulating E-cadherin and downregulating TGF-β1, α-SMA and fibronectin ( Figs. 4 F and S24). This confirms that FC@MT inhibits MMT at the transcriptional and translational level rather than achieving a transient transcriptional effect [ 67 ]. Together, these multimodal data impose evidence for the effectiveness of FC@MT to disrupt TGF-β1-induced MMT through cooperative silencing of genes. Non-specific protein adsorption and cellular adhesion are critical pathophysiological factors contributing to postoperative adhesion formation [ 68 ]. To assess the effectiveness of anti-protein adsorption, rhodamine B-conjugated bovine serum albumin (BSA) and fibrinogen were incubated with the hydrogel surfaces (Fig. S25A). Compared to the poly-chondroitin sulfate hydrogel (PCS), the FC@MT has a significantly shallower fluorescence layer thickness, indicating the superior resistance against protein adsorption and penetration. The outstanding anti-fouling behaviors of FC@MT have been reconfirmed by the quantitative statistical analyses of fluorescence intensity and adsorbed protein mass (Fig. S25B and S25C). Cell adhesion resistance was assessed using two complementary tests: in vitro attachment of cells and in vivo peritoneal colonization. In tissue culture polystyrene (TCPS) controls, almost complete cellular coverage was observed in vitro , while FC@MT showed close to negligible cellular attachment after 24 h (Fig. S26A). In vivo , CT26-luc cells were inoculated intraperitoneally in the presence of FC@MT, followed by bioluminescence imaging at 48 h showing widespread peritoneal dissemination in control groups. On the other hand, the bioluminescent signals from the FCGCM and FC@MT groups were significantly lower ( Fig. 4 G), indicating lower CT26 cell colonization in the peritoneal cavity. Blood attachment assessments (Fig. S26B) indicated that the FC@MT exhibited low blood attachment after 2 and 4 h immersion. Moreover, bacterial adhesion tests (Fig. S26C), using E. coli live/dead staining, showed that FC@MT had strong antibacterial adhesion properties, with almost no adherent bacteria, compared to the large number of bacteria observed on TCPS. The findings indicate that FC@MT can inhibit non-specific protein adsorption and non-specific cellular adhesion, which may help to reduce adhesion formation. For evaluating the dual anticancer and antiadhesion potential of FC@MT, the orthotopic model of colorectal cancer was established successfully by injecting CT26-luc cells intraperitoneally into BALB/c mice ( Fig. 5 A). Seven days after inoculation, mice underwent CRS, and therapeutic agents were introduced at the resection site. The experimental groups were as follows: Model group (saline), the CGM group (CGM hydrogel), the 5-FU group (5-FU solution delivery), and the FC@MT group (FC@MT delivery system). Bioluminescence imaging with an in vivo imaging system was used for dynamic monitoring of tumor progression ( Fig. 5 B). Imaging data underwent a quantitative analysis that revealed that all mice had an accelerated growth of the tumor before CRS, and it showed signal attenuation right after CRS (Fig. S27). The presence of notable bioluminescence among the CGM and Model group by 7 d of CRS suggested aggressive recurrence of peritoneal metastasis. In contrast, treatment with both 5-FU and the FC@MT group showed a significant decrease in the signal intensity, confirming that peritoneal metastasis was significantly inhibited. Analysis of body weight changes (Fig. S28A) and survival (Fig. S28B) further demonstrated the superior antitumor effect of FC@MT relative to the Model group. Fig. 5 FC@MT synchronously inhibited post-CRS postoperative adhesion and peritoneal metastasis. (A) Schematic diagram of the post-CRS postoperative adhesion and peritoneal metastasis model and the FC@MT administration protocol; (B) In vivo imaging before and after CRS; (C-E) Photographs (C), weight measurements (D) and nodule counts (E) of resected tumor nodules, * P < 0.05, *** P < 0.001, mean ± SD ( n = 4–5); (F) Adhesion scores of tissues following tumor resection surgery ( n = 5); (G) H&E and Masson’s trichrome staining images of adhesion tissues, scale bar: 500 µm. Fig 5 dummy alt text FC@MT synchronously inhibited post-CRS postoperative adhesion and peritoneal metastasis. (A) Schematic diagram of the post-CRS postoperative adhesion and peritoneal metastasis model and the FC@MT administration protocol; (B) In vivo imaging before and after CRS; (C-E) Photographs (C), weight measurements (D) and nodule counts (E) of resected tumor nodules, * P < 0.05, *** P < 0.001, mean ± SD ( n = 4–5); (F) Adhesion scores of tissues following tumor resection surgery ( n = 5); (G) H&E and Masson’s trichrome staining images of adhesion tissues, scale bar: 500 µm. At the termination of the experiment, the tumors left over were collected by sacrificing mice ( Fig. 5 C). Assessment of antitumor efficacy was done through evaluation of tumor weight and counting of metastatic nodules ( Fig. 5 D–5E). The administration of FC@MT greatly reduced the spread of tumors within the peritoneum, as evidenced by the limited formation of nodules throughout the peritoneal cavity. Conversely, the Model and CGM groups showed large macroscopic metastatic deposits throughout the peritoneum and viscera. The tumor weight and metastatic nodule count of the FC@MT group showed no significant differences compared to those of the 5-FU group, suggesting comparable antitumor efficacy. At the same time, FC@MT showed significant anti-adhesion ability. Obvious peritoneal adhesion was not noted between the colonic and peritoneal surfaces in the FC@MT group, but the 5-FU group had significant adhesion ( Fig. 5 C). As evidenced by the above discussion, scoring of adhesion revealing quantitative data support the effectiveness of FC@MT to prevent postoperative adhesion ( Fig. 5 F). Significantly higher adhesion severity was noted in the Model and 5-FU groups as compared to FC@MT groups. Histological examination of the tissues from the Model group confirmed heavy metastasis involvement ( Fig. 5 G). To evaluate potential systemic toxicity, the heart, liver, spleen, lungs and the kidneys were subjected to histological assessment. Histopathological examination did not reveal significant differences in mice treated with FC@MT (Fig. S29). The cardiac tissue showed well-organized myocytes without any abnormality. The hepatic sinusoid appears normal without any pathological abnormality. The splenic white and red pulps were within physiological limits. The perialveolar regions of the lungs did not show any infiltration of inflammatory cells or fibrosis. The glomeruli and renal tubules did not show degeneration or hemorrhage. As a whole, these outcomes indicate that the FC@MT delivery system has good in vivo biocompatibility. MMT is a peritoneal-specific kind of epithelial-mesenchymal transition (EMT) involved in the development of chronic peritoneal diseases, particularly peritoneal tumor metastasis and invasive endometriosis ( Fig. 6 A) [ 68 ]. The release of TGF-β1 siRNA downregulates TGF-β1 expression and halts MMT progression. Therefore, analysis of MMT-associated molecular markers in harvested tissues is necessary for evaluation of therapeutic efficacy. To better demonstrate how FC@MT inhibits peritoneal metastasis and adhesion formation, transcriptomic profiling was applied to identify differentially expressed genes (DEGs). Fig. 6 Underlying therapeutic mechanisms of FC@MT. (A) Schematic representation delineating the proposed mechanism of action of FC@MT in mitigating MMT; (B) Venn diagram illustrating the overlap of differentially expressed genes; (C) Principal component analysis scores plot depicting the clustering patterns; (D) Volcano plot presenting the differential gene expression profile between FC@MT group and Model group; (E-G) Relative gene expression of Acta2 (E), Tgfb1 (F), Fn1 (G), mean ± SD( n = 3); (H) Heatmap visualization of the expression patterns of the top-ranking differentially expressed genes; (I) KEGG pathway enrichment analysis comparing FC@MT group and Model group. Fig 6 dummy alt text Underlying therapeutic mechanisms of FC@MT. (A) Schematic representation delineating the proposed mechanism of action of FC@MT in mitigating MMT; (B) Venn diagram illustrating the overlap of differentially expressed genes; (C) Principal component analysis scores plot depicting the clustering patterns; (D) Volcano plot presenting the differential gene expression profile between FC@MT group and Model group; (E-G) Relative gene expression of Acta2 (E), Tgfb1 (F), Fn1 (G), mean ± SD( n = 3); (H) Heatmap visualization of the expression patterns of the top-ranking differentially expressed genes; (I) KEGG pathway enrichment analysis comparing FC@MT group and Model group. The transcriptomic analysis indicated that out of the 15,049 genes, the 9474 genes were filtered in all five experimental groups. 141 genes express uniquely in Model group ( Fig. 6 B). Principal component analysis showed clear separation of the gene expression profile for the FC@MT treated group from that of the Model group, indicating that FC@MT causes significant transcriptional reprogramming ( Fig. 6 C). According to comparative transcriptomic analysis between the Model and Normal groups, as depicted through volcano plots (Fig. S30), a significant number of genes were found to be upregulated in the Model group. Noteworthy among these are the classical markers of MMT, fibrosis and proliferation, such as Tgfb1, Acta2, Fn1 and Mki67 . FC@MT treatment significantly decreased the expression levels of these pathogenic genes as shown in Figs. 6 D–6G and S31. Substantial differences in gene expression were confirmed by hierarchical clustering analysis: FC@MT grouped with Normal, while Model has clear transcriptional differences ( Figs. 6 H and S32). Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) enrichment analyses revealed that multiple signaling pathways associated with tumor progression, ECM remodeling, and fibrosis were significantly upregulated following CRS ( Figs. 6 I, S33 and S34). The activation of pathways in the Model group of mice promoted peritoneal metastasis and adhesion formation which later corroborated the pathological phenotypes. Taken together, transcriptomic data show that the therapeutic efficacy of FC@MT relates to the inhibition of the TGF-β1-mediated MMT, ECM dysregulation and tumor proliferation, supporting the rational design of FC@MT delivery system. Western blot analysis showed a significant enhancement in the expression of mesenchymal markers (TGF-β1, α-SMA, vimentin and COL1A1) in the Model, CGM and 5-FU groups as compared to the Normal group ( Figs. 7 B–7D and S35A-S35C). In comparison, treatment with FC@MT considerably reduced the pro-MMT proteins’ expression. Crucially, FC@MT was confirmed as a reversal of MMT through E-cadherin, a canonical mesothelial cell marker ( Fig. 7 D). The corresponding genes, Acta2, Tgfb1 and Cdh1 were analyzed by RT-qPCR ( Fig. 7 E–7G and S35D-S35F) and showed similar transcriptional modulation. Immunofluorescence staining further verified our findings as tissues from the Model group showed prominent TGF-β1 and α-SMA fluorescence and low E-cadherin fluorescence, indicating robust MMT activation ( Fig. 7 H–7I). In contrast, the FC@MT-treated samples exhibited a reverse fluorescence intensity pattern with less mesenchymal marker expression and regained E-cadherin expression, indicating inhibition of MMT. ​Importantly, continuous E-cadherin-positive mesothelial cell layer restoration on the peritoneal surface was accomplished by FC@MT treatment, thus ensuring peritoneal mucosal repair. Fig. 7 Therapeutic efficacy of FC@MT in inhibiting MMT. (A) Schematic representation depicting the mechanism of FC@MT-mediated MMT inhibition; (B) Western blot analysis of α-SMA, TGF-β1, Vimentin, E-cadherin, and COL1A1 protein expression; (C, D) Quantitative analysis of relative protein expression levels for TGF-β1 (C) and E-cadherin (D); (E-G) Rlative mRNA expression levels quantified by RT-qPCR for Acta2 (E), Tgfb1 (F) and Cdh1 (G), * P < 0.05, ** P < 0.01, *** P < 0.001, mean ± SD ( n = 3); (H) Immunofluorescence micrographs illustrating TGF-β1 (red) localization among experimental cohorts; scale bar: 400 µm; (I) Immunofluorescence micrographs depicting co-localization of α-SMA (red) and E-cadherin (green) across experimental cohorts; scale bar: 400 µm. Fig 7 dummy alt text Therapeutic efficacy of FC@MT in inhibiting MMT. (A) Schematic representation depicting the mechanism of FC@MT-mediated MMT inhibition; (B) Western blot analysis of α-SMA, TGF-β1, Vimentin, E-cadherin, and COL1A1 protein expression; (C, D) Quantitative analysis of relative protein expression levels for TGF-β1 (C) and E-cadherin (D); (E-G) Rlative mRNA expression levels quantified by RT-qPCR for Acta2 (E), Tgfb1 (F) and Cdh1 (G), * P < 0.05, ** P < 0.01, *** P < 0.001, mean ± SD ( n = 3); (H) Immunofluorescence micrographs illustrating TGF-β1 (red) localization among experimental cohorts; scale bar: 400 µm; (I) Immunofluorescence micrographs depicting co-localization of α-SMA (red) and E-cadherin (green) across experimental cohorts; scale bar: 400 µm. The subsequent development of peritoneal tumorigenesis and adhesive pathology is a widely neovascularized process that occurs during which dense connective tissue and abnormal blood vessels are found within adhesive foci and neoplastic tissue. Neovascularization is beneficial for adhesion formation as they provide nutrients to fibroblasts and tumor metastasis by facilitating the movement of tumor cells [ 69 ]. As illustrated in Fig. S36, the expression of vascular endothelial growth factor (VEGF) was significantly lower in the FC@MT group when compared to the control groups (Model, CGM, 5-FU). To further characterize the antineoplastic efficacy of FC@MT, we evaluated the tumor proliferation marker Ki67 and the epithelial tumor marker EpCAM. Western blotting, RT-qPCR, and immunofluorescence analyses (Fig. S37) consistently demonstrated a marked downregulation of Ki67 and EpCAM expression in the FC@MT group, in stark contrast to their significant elevation in the Model group. This concurrent suppression of tumor cell proliferation and epithelial tumor marker expression underscores FC@MT’s robust inhibitory effect on peritoneal metastasis.

Chondroitin sulfate (CS, Mn ≈ 70 kDa), citrulline (Cit), N-hydroxy-succinimide (NHS), N-(3-dimethylaminopropyl)-N'-ethylcarbodiimide hydrochloride (EDC·HCl), 5-FU, MPC, (3-aminopropyl)triethoxysilane (APTES) and Cy7 were purchased from Macklin Biochemical Technology Co., Ltd. (Shanghai, China). Glycidyl methacrylate (GMA), 2-hydroxy-4′-(2-hydroxyethoxy)-2-methylpropiophenone (I2959), tetrabutylammonium bromide (TBAB), triethylamine (TEA) and 4-dimethylaminopyridine (DMAP) were supplied by Heowns Biochem LLC (Tianjin, China). Tetraethoxysilane (TEOS), ethanol, hydrochloric acid (HCl, 38%), and hexadecyl trimethyl ammonium chloride (CTAC) were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). FAM-labeled TGF-β1 siRNA was synthesized by Sangon Biotech Co., Ltd. (Shanghai, China). Dulbecco’s Modified Eagle’s Medium (DMEM), Roswell Park Memorial Institute medium 1640 (RPMI-1640), and fetal bovine serum (FBS) were obtained from Sangon Biotech Co., Ltd. (Shanghai, China). Trypsin and antibiotics (100 units/ml penicillin and 100 units/ml streptomycin) were purchased from Servicebio Technology Co., Ltd. (Wuhan, China). All animal experiments complied with the ARRIVE guidelines and were conducted in accordance with the National Research Council’s Guide for the Care and Use of Laboratory Animals. Female BALB/c mice (6–8 weeks old, 18–22 g) were procured from Shaanxi Shaan Yao Medical Science & Biotechnology Co., Ltd. All animal experiments adhered to institutional ethical standards and were approved by the Xi’an Jiaotong University Ethical Committee (Protocol No XJTU 2019-003). CS-GMA was synthesized according to a previously reported method [ 27 ]. Briefly, 2 g CS was dissolved in deionized water, with TBAB and TEA added sequentially. After magnetic stirring for 1 h, GMA was gradually added, and stirring continued at room temperature for 24 h to complete the reaction. The resulting reaction solution was dialyzed against ultrapure water using a regenerated cellulose membrane (molecular weight cutoff: 3500 Da) for 3 days. Finally, the product was lyophilized to obtain CS-GMA. Subsequently, Cit was introduced into the CS-GMA solution along with DMAP and EDC·HCl. After a 24-h reaction period, the product was purified via dialysis against ultrapure water for 3 d and lyophilized to yield CS-GMA-Cit (CGC). The chemical structure of CGC was characterized using Fourier-transform infrared spectroscopy (FTIR) and proton nuclear magnetic resonance spectroscopy (¹H NMR, Bruker 400 MHz, Germany). The FCGCM hydrogel was fabricated via free radical copolymerization of CGC, MPC and 5-FU under 365 nm UV irradiation. A control CGM hydrogel (composed of CS-GMA and MPC without 5-FU loading) was prepared for comparative analysis. The microstructure of FCGCM was characterized using scanning electron microscopy (SEM; MAIA3 LMH, TESCAN). The equilibrium water content (EWC) of the FCGCM was quantitatively determined through gravimetric analysis. Rheological properties were evaluated using a rotational rheometer (MCR302, Anton Paar) with a 25-mm cone-plate geometry and 1 mm gap. Time sweep tests were conducted at 37 °C and 45 °C under oscillatory shear (1% strain, 1 Hz). Strain amplitude sweep tests (0.1%−1000% strain) were performed to determine the critical gel point. Shear-thinning behavior was assessed via steady shear viscosity measurements, and structural recovery was evaluated through three consecutive strain cycles (1%−500% strain, 1 Hz). Time sweep experiments were also conducted on FCGCM treated with urea (hydrogen bond disruptor) or NaCl (electrostatic screening agent) to clarify the contributions of hydrogen bonds and electrostatic interactions to hydrogel stability. For in vitro degradation, FCGCM hydrogel was incubated in a 75 U/ml chondroitinase solution at 37 °C with constant shaking (120 rpm). In the peritoneal microenvironment of CRC, inflammatory responses and activated macrophages can lead to the increased content of chondroitinase, which is a key endogenous enzyme responsible for degradation of CS-based biomaterials. The hydrogel weights before (W₀) and after (W) degradation were measured, and the weight change percentage was calculated as: Weight change ( % ) = W t / W 0 × 100 % For in vivo degradation, Cy7-labeled FCGCM was synthesized by covalent conjugation of Cy7 to CGC with EDC·HCl and DMAP as coupling agents. 0.5 mL Cy7-FCGCM was intraperitoneally injected into BALB/c mice ( n = 3 per time point), and longitudinal monitoring was performed on Day 0, 3, 7 and 10 via near-infrared fluorescence imaging (MARS, Shanghai Artemis Intelligent Imaging). Complementary 0.5 T MRI (MiNiMR60) with T1-weighted sequences was used to quantify volumetric changes of subcutaneously implanted FCGCM at the same time points. The 5-FU release profile was determined in PBS and a simulated peritoneal fluid (SPF) model (pH 6.8, high MMP activity) mimicking the postoperative peritoneal microenvironment [ 28 ]. 1 ml FCGCM hydrogel was placed in 50 ml centrifuge tubes with 10 ml release medium. At predetermined time intervals, 200 µl supernatant was collected and replaced with fresh medium. The concentration of released 5-FU was quantified using a UV–Vis spectrophotometer (PerkinElmer Lambda 950) following the Lambert-Beer law. Molecular structures of the target compounds were retrieved from the PubChem database, and topology files were generated using Multiwfn [ 29 , 30 ], Sobtop [ 31 ] and ORCA 6.0.1 [ [32] , [33] , [34] , [35] , [36] , [37] , [38] , [39] , [40] , [41] , [42] , [43] ]. Molecular Dynamic simulations (MDS) were performed with GROMACS in a 10 nm × 10 nm × 10 nm cubic periodic boundary cell containing 50 molecules of 5-FU, 50 molecules of citrulline, and 200,000 molecules of water, using GAFF and OPLS-AA force fields for a 200 ns trajectory to quantify intermolecular hydrogen bonds [ 44 ]. Density functional theory (DFT) calculations were conducted with ORCA 6.0.1: geometry optimization used the B3LYP/DFT-D3(BJ)/def2-TZVP method, and single-point energy calculations adopted revDSD-PBEP86-D4/2021/def2-TZVPP. The Solvation Model based on Density implicit solvation model was applied throughout, and intermolecular binding energy was calculated as: E binding = E complex − ( E partA + E partB ) AMSNs were synthesized following reported protocols [ 45 , 46 ]. Briefly, CTAC and TEA were dissolved in deionized water and stirred at 95 °C for 1 h, then TEOS was added dropwise. The resulting MSNs were purified, and 50 mg MSNs were dispersed in ethanol with 25 µl APTES added for 24 h of amine functionalization to obtain AMSNs. AMSNs were characterized by SEM/TEM (morphology), XPS/EDS (elemental composition), and BET (specific surface area and pore size distribution). TGA was performed using a simultaneous thermal analyzer. FTIR spectra were recorded on a Vertex 70 spectrometer (Bruker, Germany) with samples pressed into KBr pellets and scanned 32 times over the range of 4000 to 400 cm −1 . AMSNs and siRNA (25 pmol) were mixed at varying nitrogen-to-phosphate (N/P) ratios in diethylpyrocarbonate (DEPC)-treated water and incubated at room temperature for 30 min to form AMSNs/siRNA complexes. The N/P ratio was defined as the molar ratio of cationic amino groups (N) on AMSNs to anionic phosphate groups (P) in siRNA. The hydrodynamic particle size and zeta potential of the AMSNs/siRNA complexes were characterized by dynamic light scattering (DLS) using a Malvern Zetasizer Nano ZSE (Malvern, UK) at 25 °C. Additionally, the morphological features of the complexes were visualized using TEM. AMSNs/siRNA complexes were prepared at N/P ratios of 0–16. Their siRNA binding capacity was evaluated via 3% agarose gel electrophoresis, and RNase resistance was assessed by incubating complexes (N/P = 2:1) with 5/10/20 µg/ml RNase for 10 min. The siRNA cumulative release rate from complexes was measured in PBS at 4–96 h. The supernatant was collected and measured using a microplate reader to calculate the siRNA content. The cytotoxicity of AMSNs/siRNA complexes against HMrSV5 cells detected by the MTT method. Briefly, HMrSV5 cells were seeded into a 96-well plate at a density of 1 × 10 4 cells per well and incubated for 24 h. The medium was then replaced with fresh medium containing complexes at various N/P ratios. After incubation, the medium was removed, and 100 µl MTT solution was added to each well. Following incubation, the MTT solution replaced with dimethyl sulfoxide and the optical density at 492 nm of each well was measured using a microplate reader (Bio-Rad, Berkeley, CA). HMrSV5 cells were seeded in 35-mm coverslip chambers at a density of 1 × 10 5 cells per well in complete culture medium. After incubation, the medium was replaced with serum-free medium containing naked TGF-β1 siRNA, Lipo3000/siRNA complexes, or AMSNs/siRNA complexes. Transfection efficiency was quantitatively assessed by flow cytometry. AMSNs/siRNA complexes were mixed with FCGCM solutions and crosslinked through 30 min of UV irradiation to form FC@MT. Non-specific protein adsorption was evaluated using bovine serum albumin (BSA) and fibrinogen (2 mg/ml) as representative proteins [ 47 ]: FC@MT disk samples were incubated with protein solution for 2 h, rinsed with PBS, and bound proteins were detached with 1% sodium dodecyl sulfate (SDS) for quantification. Protein concentration was then quantified using a Micro BCA protein assay kit (Applygen). For fluorescence-based analysis, BSA and fibrinogen were conjugated with Rhodamine B for fluorescent labeling. After 3 h of co-incubation, the thickness of the adsorbed protein layer was assessed via fluorescence microscopy. The anti-adhesive properties of FC@MT were evaluated using HMrSV5, L929, RAW 264.7, and CT26 cells. FC@MT disks were placed in a 24-well plate and seeded with cells at a density of 5 × 10 4 cells per well, followed by incubation for 24 h. After removing the culture medium, non-adherent cells were eliminated by gentle washing with PBS. Adherent cells were then visualized using an inverted microscope (ICX41, Shunyu Instruments). Tissue culture polystyrene (TCPS, Corning) served as the control substrate for comparative analysis. The anti-blood adhesion activity was evaluated to further assess the antifouling capability of FC@MT. FC@MT was fabricated into cylindrical constructs and immersed in a 2% blood dilution for 2 and 4 h at 37 °C. Subsequently, the FC@MT were gently rinsed three times with PBS to remove non-adherent blood before observation and imaging. To evaluate the antibacterial adhesion capacity of FC@MT using Escherichia coli live/dead staining, FC@MT samples were incubated with an E. coli suspension (1 × 10 7 CFU/ml) at 37 °C for 4 h. The samples were then gently rinsed with PBS to remove non-adherent bacteria and stained with a live/dead bacterial staining kit for 15 min in the dark. Fluorescent imaging was performed using a laser confocal microscope to observe and quantify adherent live and dead bacteria. CT26-luciferase (CT26-luc) cells were enzymatically detached and resuspended in RPMI-1640 culture medium at a density of 5 × 10 6 cells/mL. Female BALB/c mice (6–8 weeks old) underwent a 12-h fasting protocol prior to intervention. The mice then received intraperitoneal administration of FCGCM hydrogel or FC@MT nanocomposites, followed by orthotopic implantation of CT26-luc cells. In vivo bioluminescence imaging was performed 48 h post-inoculation using an IVIS® Imaging System. The cytotoxicity of FC@MT against L929, HMrSV5 and CT26 cells was detected by the MTT method. The cells were seeded into 96-well plates and incubated for 24 h. Subsequently, the medium was replaced with fresh medium containing FC@MT powder (50–1600 µg/ml). After incubation, the medium was removed, and 100 µl MTT solution was added to each well. Following incubation, the MTT solution replaced with dimethyl sulfoxide and the optical density at 492 nm of each well was measured. For live/dead cell viability assessment, FC@MT-treated cells were incubated with the Calcein-AM/PI live/dead cell fluorescent staining kit (Servicebio) for 30 min. Cellular fluorescence was then visualized using an inverted fluorescence scanning microscope (Olympus FV3000). Hemocompatibility evaluation was performed to assess the biocompatibility profile of FC@MT. FC@MT specimens were equilibrated in saline solution for 3 d to obtain extraction media. The extracts were then incubated with freshly prepared erythrocyte suspensions at 37 °C. Negative and positive controls were established using physiological saline and 1% Triton X-100 solution, respectively. After incubation, samples were centrifuged at 3000 rpm for 10 min to pellet intact erythrocytes. Hemolytic activity was quantitatively determined by spectrophotometric measurement of hemoglobin release at an absorbance of 540 nm. HMrSV5 cells were plated in 6-well culture plates at a density of 1 × 10 6 cells per well and allowed to adhere overnight. After cell attachment, the complete growth medium was replaced with serum-free medium containing AMSNs/siRNA complexes. Cellular internalization was allowed to proceed for 24 h under standard culture conditions (37 °C, 5% CO 2 ). Following the incubation period, cells were washed three times with PBS to remove uninternalized complexes, then fixed and sectioned. The sections were mounted on copper grids for TEM analysis. HMrSV5 and CT26 cells were cultured in 6-well plates until they reached approximately 70% confluency. Lyophilized FC@MT was reconstituted in medium and added to the cell cultures for co-cultivation. After a 6-h incubation, the cells were fixed and stained for the cytoskeleton and nuclei using phalloidin and 4′,6-diamidino-2-phenylindole (DAPI), respectively. Cellular uptake of FAM-labeled siRNA was then visualized using confocal fluorescence microscopy. HMrSV5 and CT26 cells were seeded in coverslip chambers and allowed to adhere overnight. Subsequently, the cells were incubated with FC@MT for 6 h. As controls, parallel groups were treated with naked siRNA for the same duration. After washing with PBS, the cells were stained with LysoTracker Red (L8010, Solarbio) for 120 min and rinsed again with PBS. Finally, the cells were observed under a fluorescence microscope. Quantitative real-time PCR (RT-qPCR) was used to evaluate the gene silencing efficiency of the three synthesized TGF-β1 siRNAs. CT26 cells were seeded into each well of 6-well plates and incubated at 37 °C with 5% CO 2 for 24 h in a cell culture incubator. TGF-β1 siRNAs (TGF-β1-96, −296 and −622) at a concentration of 50 nM were complexed with AMSNs; the mixture was diluted in 250 µl serum-free medium and incubated at room temperature. The medium was replaced with the prepared AMSNs/siRNA complexes, followed by incubation for 6 h. Total RNA was extracted from CT26 cells using TRIzol reagent. First-strand cDNA synthesis was performed using a commercial reverse transcription kit (Takara). RT-qPCR was conducted on a CFX96 Real-Time PCR Detection System (Bio-Rad, USA). Relative Tgfb1 mRNA expression levels were normalized to GAPDH and calculated using the comparative 2 −ΔΔCT method. The sequences of the designed and synthesized TGF-β1 siRNAs are listed in Table 1 . Table 1 Sequences of designed and synthesized TGF-β1 siRNAs. Table 1 dummy alt text Gene Sequences (5′ to 3′) Sense (5′−3′) Antisense (5′−3′) Tgfb1 -mus-296 GAAGCGGACUACUAUGCUATT UAGCAUAGUAGUCCGCUUCTT Tgfb1 -mus-96 GCAAGACCAUCGACAUGGATT UCCAUGUCGAUGGUCUUGCTT Tgfb1 -mus-622 GAACCAAGGAGACGGAAUATT UAUUCCGUCUCCUUGGUUCTT NC siRNA UUCUCCGAACGUGUCACGUTT ACGUGACACGUUCGGAGAATT Sequences of designed and synthesized TGF-β1 siRNAs. For in vitro MMT inhibition assays, HMrSV5 cells were incubated with CT26 cell supernatant for 24 h to induce MMT. After a 6-h incubation, the culture medium was replaced with 500 µl fresh DMEM supplemented with 10% FBS, and the cells were cultured for an additional 48 h. Total RNA was isolated from HMrSV5 cells using TRIzol reagent. Tgfb1 mRNA expression levels were calculated using the comparative 2 −ΔΔCT method. Primer sequences are provided in Table 2 . Table 2 Primer sequences utilized for the RT-qPCR analysis of target genes in HMrSV5 cells. Table 2 dummy alt text Primer Sequences (5′ to 3′) Gapdh -Forward GTGGACCTGACCTGCCGTCTAG Gapdh -Reverse GAGTGGGTGTCGCTGTTGAAGTC Cdh1 -Forward GCCATCGCTTACACCATCCTCAG Cdh1 -Reverse CTCTCTCGGTCCAGCCCAGTG Tgfb1 -Forward AGCAACAATTCCTGGCGATACCTC Tgfb1 -Reverse TCAACCACTGCCGCACAACTC Acta2 -Forward TCGTGCTGGACTCTGGAGATGG Acta2 -Reverse CCACGCTCAGTCAGGATCTTCATG Col1a1 -Forward TGGCAAAGAAGGCGGCAAAGG Col1a1 -Reverse AGGAGCACCAGCAGGACCATC Vim -Forward TGAATGACCGCTTCGCCAACTAC Vim -Reverse CTCCCGCATCTCCTCCTCGTAG Primer sequences utilized for the RT-qPCR analysis of target genes in HMrSV5 cells. Protein concentrations were determined using a BCA assay and normalized prior to loading. Proteins were separated by SDS-PAGE and transferred onto PVDF membranes. The membranes were incubated with primary antibodies overnight at 4 °C, followed by incubation with species-matched HRP-conjugated secondary antibodies for 2 h at room temperature. ECL luminescent reagent was used for protein band visualization. Female BALB/c mice (6–8 weeks old) were intraperitoneally inoculated with 1 × 10 6 CT26-luc cells (5 × 10 6 cells/ml). Seven days post-inoculation, CRS was performed to resect all macroscopic tumor nodules under isoflurane anesthesia, with the procedure limited to 2 min and strict hemostasis. Mice were randomly divided into four groups (Model: sterile saline; CGM: CGM hydrogel; 5-FU: 5-FU solution; FC@MT: FC@MT system), and the corresponding agents were applied to the surgical site. After an additional 7-d recovery period, the mice were humanely euthanized for terminal tumor excision and subsequent pathological analysis. Tumor progression was dynamically monitored via in vivo bioluminescence imaging. Adhesion formation was assessed using a standardized scoring system ( Table 3 ). Table 3 Clinical adhesion scoring criteria. Table 3 dummy alt text Criteria Adhesion extent 0 No adhesion 1 One thin filmy adhesion 2 More than one thin adhesion 3 Thick adhesion with focal point 4 Thick adhesion with plantar attachment or more than one thick adhesion with focal point 5 Very thick vascularized adhesion or more than one plantar adhesion Clinical adhesion scoring criteria. Resected adhesion and tumor tissues were fixed in 4% paraformaldehyde, paraffin-embedded, and sectioned. Tissue sections were subjected to comprehensive staining, including Masson’s trichrome, hematoxylin and eosin (H&E), and immunofluorescence. Major organs (heart, liver, spleen, lung, kidney) were systematically harvested and processed for Masson staining. Transcriptomic profiling was performed to assess the inhibitory effect of FC@MT on MMT progression. RNA integrity was examined with an Agilent 2100 Bioanalyzer utilizing the RNA Nano 6000 Assay Kit. Total RNA served as the initial material for library construction, with polyadenylated mRNA first purified from the total RNA pool and then fragmented via divalent cation-induced cleavage under elevated temperature conditions. First-strand cDNA was synthesized using random hexamer primers and an RNase H-dependent reverse transcription approach, after which RNase H and DNA polymerase I were used for second-strand cDNA synthesis. cDNA fragments of 370–420 bp in length were isolated by size selection with the AMPure XP system. The obtained library fragments were subjected to PCR amplification and subsequent purification, and final quality control was conducted on the Agilent 2100 Bioanalyzer to validate appropriate fragment size distribution and concentration before high-throughput sequencing. Total RNA and protein were isolated from adhesion tissue samples in each experimental group. The mRNA and protein expression levels were determined via RT-qPCR and western blot assays, respectively. The primer sequences used for amplification are detailed in Table 4 . Table 4 Primer sequences for RT-qPCR analysis of target genes in tissues from BALB/c mice. Table 4 dummy alt text Primer Sequences (5′ to 3′) Gapdh -Forward GGCAAATTCAACGGCACAGTCAAG Gapdh -Reverse TCGCTCCTGGAAGATGGTGATGG Cdh1 -Forward GGCACTCTTCTCCTGGTCCTG Cdh1 -Reverse GATGGTGATGATATGAGGCTGTGG Tgfb1 -Forward ACCGCAACAACGCCATCTATGAG Tgfb1 -Reverse GGCACTGCTTCCCGAATGTCTG Acta2 -Forward GCGTGGCTATTCCTTCGTGACTAC Acta2 -Reverse CGTCAGGCAGTTCGTAGCTCTTC Col1a1 -Forward GACAGGCGAACAAGGTGACAGAG Col1a1 -Reverse CAGGAGAACCAGGAGAACCAGGAG Vim -Forward GCGTGCGGCTGCTTCAAG Vim -Reverse TCTCGTTGGTGCGGGTGTTC Vegfa -Forward GTGACAAGCCAAGGCGGTGAG Vegfa -Reverse CGATGATGGCGTGGTGGTGAC Mki67 -Forward GCCTGCCCGACCCTACAAAATG Mki67 -Reverse CACTCATCTGCTGCTGCTTCTCC Epcam -Forward TGGACCTGAGAGTGAACGGAGAG Epcam -Reverse ACACCACCACAATGACAGCGATG Primer sequences for RT-qPCR analysis of target genes in tissues from BALB/c mice. Statistical analyses were performed using GraphPad Prism 8.0.1. Quantitative data are expressed as mean ± standard deviation (SD). Comparisons among multiple groups were analyzed by one-way analysis of variance followed by appropriate post hoc tests. A two-tailed P value < 0.05 was considered statistically significant.

To meet the significant clinical challenges of postoperative adhesion and peritoneal metastasis after CRS, we cleverly designed an integrated therapeutic platform (FC@MT) consisting of zwitterionic antifouling FCGCM hydrogel and AMSNs/TGF-β1siRNA complexes. FC@MT employs a dual-loading strategy that has been rationally designed. It immobilizes the chemotherapeutic agent 5-FU within the FCGCM hydrogel through hydrogen bonds. Also, it encapsulates TGF-β1 siRNA in AMSNs to ensure efficient cellular uptake and lysosomal escape. Sustained release of 5-FU via hydrogen bonding removes residual tumor cells and prevents metastatic dissemination in CRS mice. At the same time, TGF-β1 siRNA in combination with chemotherapy acts adoptively to downregulate TGF-β1 expression, inhibiting MMT responsible for peritoneal adhesion and metastasis. In comparison to the improved antitumor and anti-adhesion efficacies of individual components found in current existing formulations, the entity possessing such unique therapeutic functions will manifest negative systemic toxicity. Overall, these characteristics suggest that FC@MT has great translation potential as a promising clinical strategy for the combined management of two major complications of CRC, namely postoperative adhesion and peritoneal metastasis.

Colorectal cancer (CRC) is one of the most common malignant tumors globally and the second leading cause of cancer-related death [ 1 ]. Cytoreductive surgery (CRS) has been demonstrated to play a first-line role in treating early-stage CRC, aggregating excellent results in selected patients [ 2 ]. Postoperative peritoneal adhesions are not only a common postoperative complication but also associated with tumor recurrence and peritoneal metastasis [ 3 ]. In addition, these adhesions limit the possibilities of further reoperations and the efficacy of CRS [ 4 ]. The existing clinical therapeutic strategies offer solutions for peritoneal adhesion and metastasis as separate problems when they often occur at the same time and share the same pathogenic mechanisms. For this reason, integrated therapeutic approaches that alleviate these complications in parallel are an urgent clinical need. Fibroblast proliferation at injured peritoneal sites is proven core contributors to formation of peritoneal adhesion and metastasis [ 5 , 6 ]. Peritoneal mesothelial cells undergo myofibroblast differentiation and mesothelial-mesenchymal transition (MMT) due to surgical injuries [ 7 ]. Earlier studies have shown that MMT is instrumental in creating these two complications: it incites peritoneal adhesion by ramping up extracellular matrix (ECM) synthesis and angiogenic factor [ [8] , [9] , [10] , [11] , [12] ], and stimulates the generation of cancer-associated fibroblasts, which is a cell population that is a key driver of peritoneal metastatic progression [ 13 ]. Since MMT is a common mediator of peritoneal adhesion and metastasis, designing and developing targeted delivery systems that can interfere with MMT will be of great clinical significance to improve the outcome of CRS in patients. Transforming growth factor-β1 (TGF-β1) acts as the main cytokine controlling the occurrence of MMT among several factors which cause MMT. On one hand, TGF-β1 promotes MMT by stimulating aberrant ECM synthesis and deposition, and down regulating mesothelial markers (E-cadherin) and up regulating myofibroblast markers (α-SMA) in peritoneal mesothelial cells, promoting adhesion formation [ 14 , 15 ]. Conversely, aberrant TGF-β1 overexpression in malignant tumors induces tumor immune evasion and drives tumor microenvironment remodeling to facilitate metastasis [ 16 , 17 ]. Therefore, targeted suppression of TGF-β1 expression is a promising strategy for MMT inhibition. Compared with small-molecule inhibitors and antibody-based therapeutics, RNA interference (RNAi) technology features key advantages of cost-effectiveness, high target specificity, and low drug resistance risk [ 18 ]. Small interfering RNA (siRNA) exerts therapeutic effects via sequence-specific post-transcriptional gene silencing, and its ultrahigh specificity makes it an ideal tool for precise inhibition of TGF-β1-mediated pathological signaling [ 19 , 20 ]. Hydrogels, as three-dimensional hydrophilic polymer networks with high water content and excellent biocompatibility, are promising biomaterials for localized drug delivery in the peritoneal cavity due to their structural similarity to the native ECM and tunable physicochemical properties. Among them, zwitterionic hydrogels are classic antifouling materials that form a dense hydration layer via electrostatic interactions with water molecules, effectively repelling non-specific protein adsorption and cell adhesion—key pathophysiological processes in peritoneal adhesion and metastasis [ [21] , [22] , [23] , [24] , [25] , [26] ]. However, most traditional zwitterionic hydrogels exhibit inherent biological inertness and lack specific bioactivities, and their single physical antifouling function cannot meet the therapeutic demand in the complex postoperative CRC microenvironment. The rational integration of zwitterionic hydrogels with bioactive therapeutic components to construct multifunctional composite systems, combining physical antifouling barriers with active biological therapies, is expected to achieve synergistic anti-adhesion and anti-metastasis effects, representing a novel therapeutic paradigm for CRC postoperative complications. To address this abovementioned clinical need, we designed and engineered the FC@MT multifunctional hydrogel-nanoparticle co-delivery system. Firstly, 2-methacryloyloxyethyl phosphorylcholine (MPC, a zwitterionic monomer), citrulline-conjugated chondroitin sulfate, and 5-fluorouracil (5-FU) were used to fabricate the FCGCM hydrogel. Concurrently, APTES-modified mesoporous silica nanoparticles (AMSNs) were prepared to encapsulate TGF-β1 siRNA for efficient cellular uptake and lysosomal escape. The AMSNs/TGF-β1 siRNA complexes were further encapsulated into the FCGCM hydrogel to form the FC@MT co-delivery system via UV crosslinking ( Scheme 1 A). Then a clinically relevant CRS mouse model of CRC was established to evaluate the in vivo therapeutic efficacy of FC@MT in inhibiting postoperative peritoneal adhesion and tumor metastasis ( Scheme 1 B), and systematically elucidated the underlying molecular mechanism by which FC@MT exerts synergistic therapeutic effects through silencing TGF-β1 to block MMT progression ( Scheme 1 C). This study aims to provide a novel and effective integrated therapeutic strategy for the clinical management of postoperative CRC complications and to offer a feasible design concept for the development of multifunctional zwitterionic hydrogel-based drug delivery systems. Scheme 1 Overview of FC@MT delivery system. (A) Preparation of FC@MT delivery system, in which AMSNs/TGF-β1 siRNA complexes are encapsulated; (B) Application of FC@MT in murine models was marred by the occurrence of abdominal adhesion and peritoneal metastasis; (C) Cellular uptake of AMSNs/TGF-β1 siRNA complexes and the silencing effects of TGF-β1 siRNA and downstream genes. Scheme 1 dummy alt text Overview of FC@MT delivery system. (A) Preparation of FC@MT delivery system, in which AMSNs/TGF-β1 siRNA complexes are encapsulated; (B) Application of FC@MT in murine models was marred by the occurrence of abdominal adhesion and peritoneal metastasis; (C) Cellular uptake of AMSNs/TGF-β1 siRNA complexes and the silencing effects of TGF-β1 siRNA and downstream genes.

The authors declare that there is no conflicts of interest.