An endoscopically compatible fast-gelation powder forms Janus-adhesive hydrogel barrier to prevent postoperative adhesions

Significance

Postoperative adhesions are one of the sequelae of almost all surgical procedures. An ideal antiadhesion barrier should be unilaterally adhesive on tissue and compatible with the endoscopic surgical procedure. Here, we develop a FJG powder with rapid water absorption and fast gelation ability. The Janus hydrogel barriers can be formed by stepwise hydrating the FJG powder with different solutions. The borate cross-linked viscoelastic hydrogel with good tissue compliance can be used as antiadhesion barriers for organs with different motion modalities. Most importantly, FJG powder is easy to store and can be applied in endoscopic procedures with a low-cost delivery device. These results emphasize the great potential of FJG powder for clinical translation.

Abstract

Postoperative adhesions occur widely in various tissues, bringing the risk of secondary surgery and increased medical burden. Hydrogel barriers with Janus-adhesive ability can achieve physical isolation of adjacent tissues and are therefore considered an ideal solution. However, integrating endoscopic delivery convenience and viscoelastic Janus hydrogel formation remains a great challenge. Here, we present a report of the in situ formation of Janus-adhesive hydrogel barrier using a sprayable fast-Janus-gelation (FJG) powder. We first methacrylate the polysaccharide macromolecules to break the intermolecular hydrogen bonds and impart the ability of rapid hydration. FJG powder can rapidly absorb interfacial water and crosslink through borate ester bonds, forming a toughly adhesive viscoelastic hydrogel. The Janus barrier can be simply formed by further hydrating the upper powder with cationic solution. We construct rat models to demonstrate the antiadhesions efficiency of viscoelastic FJG hydrogels in organs with different motion modalities (e.g., intestine, heart, liver). We also developed a low-cost delivery device with a standardized surgical procedure and further validated the feasibility and effectiveness of FJG powder in minimally invasive surgery using a preclinical translational porcine model. Considering the advantages in terms of therapeutic efficacy, clinical convenience, and commercialization, our results reveal the great potential of Janus-gelation powder materials as a next-generation antiadhesions barrier.

Postoperative adhesions are the abnormal fibrotic connection between an organ and the surrounding tissues, usually caused by wound clots, mesothelial injury, or foreign substances (e.g., gauze). Postoperative adhesions can occur in almost any part of the body and may cause lifelong risks such as organ movement disorder [e.g., intestinal obstruction (1) and restricted ventricular systole (2), Fig. 1A], chronic pain (3), and related complications (4, 5). Surgical adhesiolysis, as a second operation, cannot avoid the recurrence of adhesions and also lead to additional medical risks such as iatrogenic hemorrhage (6). Therefore, there is an increasing need for antiadhesion barriers that can serve as physical barriers between injured tissues and adjacent organs in clinical practice. Current commercial antiadhesion materials are mostly applied in the form of solid films (Seprafilm, Sanofi/Genzyme, and Interceed, Ethicon) or powders (Sepraspray, Sanofi/Genzyme), which form a sol-like barrier on the tissue surface (7, 8). However, these barriers exhibit limited clinical efficacy, mainly because of the poor retention in vivo due to the low tissue attachment capacity and fast degradation rate (9, 10).

Fig. 1.

Schematic diagram of FJG powder design and endoscopic delivery. A, Adhesions-prone organ with different motion modalities. B, Schematic diagram of different types of antiadhesion barriers. C, Comparison of the comprehensive properties of different antiadhesion barriers. D, Schematic diagram of the application of FJG powder in endoscopic surgery. Catheters (first from the left) are used for spraying different fluids or drainage; powder delivery device (second from the left) is used for delivering FJG powder; endoscopic instruments (forceps, scissors, etc., third from the left) are used to perform the procedure; endoscopes (fourth from the left) and display devices are used for imaging during operation. E, Schematic diagram of the gelation process of FJG powder. (i) FJG powder on the tissue surface through the delivery device; (ii) hydration of FJG powder with surface moisture to form an adhesive hydrogel layer. (iii) Spraying COS-PBA solution on the upper surface of the unhydrated powder. (iv) Formation of Janus hydrogel barrier. F, Schematic diagram of the Janus hydrogel network.

To enhance the in vivo retention and attachment of barriers, hydrogel films with Janus adhesion and controlled degradation have been extensively studied. Although these hydrogel films are effective in preventing adhesion formation in regularly shaped tissues such as tendons, their clinical application in the abdominal and cardiothoracic surgery is limited (11, 12). This is mainly because the hydrogel films are difficult to deploy to fully cover irregular or folded tissue surfaces (e.g., heart and intestine) and may rupture or self-adhere during endoscopic delivery in minimally invasive surgery (13, 14) (Fig. 1B). Sprayable or in situ polymerized hydrogels formed by dynamic interactions [e.g., supramolecular interaction (15), dynamic covalent bonds (16)] can effectively circumvent the difficulties of delivering film barriers under endoscopy. Besides, these materials also exhibit tunable mechanical viscoelasticity, making them motion-adaptable with the physiological motion of tissues (e.g., cardiac contraction, intestinal peristalsis, etc.) (15). However, these injectable/sprayable dynamic hydrogels are usually double-sided adhesives due to the absence of structural heterogeneity. This may lead to undesired tissue attachment or firm bonding between adjacent tissues, thereby restricting tissue movement or causing additional adhesions formation.

To integrate delivery convenience and asymmetric tissue adhesion, in situ formed Janus hydrogels have emerged as attractive candidates. A few strategies have focused on the formation of hydrogel barriers by in situ covalent cross-linking of precursor solutions, such as Michael-addition reactions or photo-initiated radical polymerization (17–19). However, irreversible covalent cross-linking will make hydrogels brittle and unadaptable to dynamic movement of organs, thereby losing function by fracturing or dislodging (20, 21). In addition, the liquid precursors may flow away on the irregular tissue surface before cross-linking, which will lead to the mispositioning of the barrier (22). Therefore, precise delivery and in situ formation of viscoelastic Janus hydrogels remains a great challenge (Fig. 1C).

To fulfill the above requirements, we designed a sprayable fast-Janus-gelation (FJG) powder that can achieve in situ formation of Janus-adhesive hydrogels on tissue surfaces after two steps of simple operations (Fig. 1 D and E). FJG powder is a blend of lyophilized phenylboronic acid (PBA)-modified methacrylated alginate (AlgMA-PBA) and dopamine (DA)-modified methacrylated hyaluronic acid (HAMA-DA). Unlike traditional powder materials that form sol-like barriers after absorbing water, FJG powder can quickly hydrate with moisture to form an adhesive hydrogel layer on the tissue surface. Inspired by the neutralized ability of positively charged polymer solution (23), the dried upper powder is further hydrated by PBA-modified chitosan oligosaccharide (COS-PBA) solution to shield both the catechol and carboxyl groups to form a nonadhesive hydrogel layer (Fig. 1F). The formed viscoelastic FJG hydrogels can adapt to different organ movements and prevent postoperative adhesions in rats. In addition, FJG powder is easy to store and can be integrated into simple endoscopic procedures with our designed delivery device.

Fabrication of FJG Powder and Characterization of FJG Hydrogels

FJG powder is prepared by mixing and pulverizing the AlgMA-PBA and HAMA-DA of equal mass (SI Appendix, Fig. S1). Methacrylation of polysaccharides aims to break the hydrogen bonds and accelerate the hydration rate of the powder (24) (Fig. 2A). To verify this, we prepared powder composed of PBA-modified alginate (Alg-PBA) and DA-modified hyaluronic acid without methacrylation. After adding dropwise Dulbecco’s modified Eagle’s medium (DMEM) medium on the powder, the FJG powder quickly absorbs water and forms hydrogel that could be picked up by tweezers. On the contrary, the liquid is unable to penetrate the unmethacrylated powder and forms droplets only on the surface (Fig. 2B and Movie S1). Rheological testing also showed that the FJG powder forms a solid hydrogel within 3 s after exposure to phosphate buffer saline (PBS) (Fig. 2C). The rapid gelation of the powder benefits from the fast dissociation–polymerization time (<1 s) of the dynamic boronate bonds, which also endows the hydrogel with good polymer dissolution to ensure the rapid homogenization of the hydrogel formed by the unevenly distributed powder (25, 26). To verify the effect of the methacrylation degree on water absorption rate, we synthesized methacrylated hyaluronic acid (HAMA) with different methacrylation degrees (10%, 30%, and 50%) and methacrylated alginate (AlgMA) with a methacrylation degree of ~34% to further prepare FJG powders (FJG-10, FJG-30, FJG-50). The results showed that the FJG-10 presented a significant lowest water absorption rate, while both FJG-30 and FJG-50 could quickly absorb water to form hydrogels and there is no difference in the water absorption rate between these two groups (SI Appendix, Fig. S2A). Besides, the storage modulus of the hydrogels formed by different FJG powders is slightly decreased with increasing methacrylation degree, which may be due to the weakening of the noncovalent interactions in the hydrogel networks as induced by the disruption of intermolecular hydrogen bonds. To demonstrate the synergistic effect of methacrylation and fast gelation, we constructed powder consisting of HAMA and AlgMA, i.e., without the presence of boronate bonds. Interestingly, the hydration rate of the HAMA-AlgMA powder is higher compared to the unmethacrylated FJG powder but lower compared to the methacrylated FJG powder (SI Appendix, Fig. S2C). These results suggest that the synergy of methacrylation and boronate bonds facilitated fast gelation of polymer can endow the FJG powder with the ability of rapid hydration.

Fig. 2.

Hydration and mechanical properties of FJG powder and FJG hydrogel. A, Schematic diagram of hydration of FJG powder. B, Water absorption process of unmethacrylated powder and methacrylated FJG powder. The time series are annotated in the upper right corner of the images. C, Gelation time of the FJG powder hydrated with PBS at 37 °C. D, The storage modulus of FJG hydrogels hydrated with PBS and COS-PBA of different concentrations, n = 5 independent samples for each group. E, Fracture strain measured by tensile test. In the PBS group and 0.5% COS-PBA group, hydrogels do not fracture at 300% strain, all values in the two groups are set to 300% for the convenience of data presentation, with n = 5 independent samples for each group. F, swelling ratio of FJG hydrogels in PBS at 37 °C, n = 3 independent samples for each group. Data are shown as mean ± SD and compared by one-way ANOVA followed by Bonferroni’s post hoc test. ** and **** indicate P < 0.01 and P < 0.0001, respectively.

We further tested the effect of powder size on hydration. We obtained powders of different particle sizes by controlling the pulverization time (SI Appendix, Fig. S3A). We found that the particle size and size dispersity of the powder decreased with increasing pulverization time, while unchanged over 20 s (SI Appendix, Fig. S3B). We further compared the hydration ability of powders with different particle sizes. When the pulverization time was less than 20 s, the formed hydrogel had obvious unhydrated solids, which indicated that the insufficiently pulverized powder could not form homogeneous hydrogels quickly.

Considering that it is challenging to precisely control the amount of powder used in clinical practice, excessive liquid flushing is more convenient for clinical endoscopic surgery. Therefore, we assessed the hydration degree of FJG powders of different sprayed densities after flushing with excess liquid. Interestingly, hydrogels with a concentration of ~6.5% w/w are obtained regardless of powder density (SI Appendix, Fig. S4 A and B). Spraying COS-PBA solution with different concentrations (0.5 to 5% w/v) also does not affect the hydration degree of FJG powder (SI Appendix, Fig. S4C), which may be related to the dissolving and reaction kinetics of polymers. This indicates that the hydration degree of FJG powder is independent of the powder spraying density under the conditions of excess liquid infiltration.

To investigate the effects of COS-PBA on the mechanical properties of FJG hydrogels, we performed rheological tests on hydrogels fully hydrated with different concentration of COS-PBA solution. The introduction of COS-PBA will increase the cross-linking degree of the hydrogel, as verified by the increased FJG hydrogel modulus (Fig. 2D) and the decreased tensile fracture strain (Fig. 2E and SI Appendix, Fig. S5 A and B) with increasing concentration of the COS-PBA solution. Increased COS-PBA also results in weakened dynamic properties, manifested as a decrease of the gel-point frequency and loss coefficient (tanδ) in the frequency sweep test (SI Appendix, Fig. S6 A and B). Spraying COS-PBA significantly reduces the hydrogel swelling, which will reduce the compression of the barrier on the tissue in the limited space in vivo (1, 15) (Fig. 2F). To balance the comprehensive effects of COS-PBA on the modulus, ductility, viscoelasticity, and swelling of hydrogels, we chose 1% COS-PBA solution for the following investigation. At such concentration, the formed FJG hydrogels have appropriate storage modulus (G′ = 4.23 ± 0.10 kPa), tensile fracture strain (~200%), and swelling ratio (~45% volume increase after soaking in PBS for 7 d), as well as good shear-thinning and self-healing properties endowed from the dynamic boronate bonds (27) (SI Appendix, Fig. S6 C and D). We next evaluated the biocompatibility and biostability of FJG hydrogels. FJG hydrogels have good biocompatibility, as verified by the high viability (>98%) of cells cultured in media containing FJG hydrogels (SI Appendix, Fig. S7 A and B). We also examined sections of different tissues after FJG hydrogel implantation by hematoxylin and eosin (H&E) staining and observed no histopathology (SI Appendix, Fig. S7C). FJG hydrogels could be degraded in vitro by hyaluronidase, which indicated that the FJG hydrogel is biodegraded in vivo (SI Appendix, Fig. S7D). In vivo fluorescence imaging of Cy7-labeled hydrogel showed that hydrogel has a residual of about 42% after implantation for 2 wk and about 23% after implantation for 3 wk (SI Appendix, Fig. S7 E and F). This suggests that the in vivo residence time of the hydrogel barrier can cover the critical period of adhesion occurrence (7 to 14 d).

Further, we evaluated the stability of FJG powder after long-term storage, which is necessary for commercial application promotion. Compared with hydrogels formed by freshly prepared FJG powder, there is no significant change in the modulus of hydrogels formed by FJG powder stored at room temperature for more than 6 mo (SI Appendix, Fig. S8). In addition, the hydrogels derived from long-term stored FJG powder have no obvious cytotoxicity (SI Appendix, Fig. S7 A and B). These results indicate that FJG powder is stable under a long-time conventional storage condition.

Janus-Adhesive Properties of FJG Hydrogels

To visualize the Janus-adhesive ability of FJG hydrogels, we first glued two pieces of trimmed pork using FJG powder with different hydration. Unhydrated FJG powder could bind and lift lower pork pieces instantly. In contrast, once the upper surface of the powder was sprayed with PBS or 1% COS-PBA, the lower pork could not be immediately adhered and lifted (Fig. 3A and Movie S2). This indicates that after hydration, FJG powder loses the capacity of quickly absorbing the interface moisture and adhesion (28). Further, we stuck the two pieces of pork together but did not lift them immediately. After 15 min, the pork sprayed with PBS could be lifted, while the pork sprayed with COS-PBA was still detached even after 2-h stacking (Fig. 3A). This indicates that after a short period of contact, the catechol groups on the surface of the PBS hydrated hydrogel can still establish adhesion to the tissue. To verify if the excess COS-PBA could effectively bind catechol groups while cationizing the upper surface of the hydrogel, we analyzed the chemical composition of the hydrogel surface after the hydration of COS-PBA using X-ray photoelectron spectroscopy (XPS). The N_1s peak indicates NH₃⁺ enrichment (~401.4 eV) on the hydrogel surface after being hydrated with COS-PBA but not after being hydrated with PBS (23) (Fig. 3B). This suggests that COS-PBA can rapidly integrate with the hydrogel network and cationize the hydrogel surface through dynamic boronate interaction.

Fig. 3.

Tissue adhesion and hemostasis of FJG powder. A, Adhesion ability of FJG powder, PBS sprayed powder, and 1% COS-PBA sprayed powder on pork adhered on glass slides. B, XPS spectra in the N_1s regions of hydrogel hydrated with PBS and COS-PBA. C and D, Interfacial toughness and adhesion strength of different adhesives on wet pig skin calculated from peeling test, n = 4 independent samples for each group. E, The FJG powder adhered to different tissues can resist water flushing and different deformations. (i) Stomach, (ii) heart, (iii) intestines, (iv) stretch, and (v) compression. The upper surface of the powder is sprayed with 1% COS-PBA solution mixed with blue dye. F, SEM images of liver tissue adhered with different hydrogels. The red arrows point out gap between tissue and hydrogel. G, Hemostasis of different materials in a liver resection model. H, Blood loss after hemostasis using different materials. In C, D, and H, each point in the column represents the data of a sample. Data are shown as means ± SD and compared using one-way ANOVA followed by Bonferroni’s post hoc test. * and **** indicate P < 0.05 and P < 0.0001, respectively.

To quantify the adhesion strength of the hydrogels, we performed shear and peeling experiments. We stuck two pieces of pig skin with FJG powder or prehydrated with PBS, 0.5% COS-PBA, and 1% COS-PBA, respectively, before sticking them. The powder exhibits a high adhesion strength of 47.56 ± 1.73 kPa and interfacial toughness of 103.4 ± 15.76 J/m² (Fig. 3 C and D and SI Appendix, Fig. S9 A and B). This can be further demonstrated by lifting heavy objects (SI Appendix, Fig. S9C and Movie S3). In addition, the hydrogel exhibits typical filamentous failure in the peeling experiment, indicating that the yield strength of FJG hydrogels is weaker than the adhesion strength (SI Appendix, Fig. S9D). Such adhesion is also stable as indicated by a slight decrease in adhesion strength (44.63 ± 1.37 kPa) and interfacial toughness (74.80 ± 9.50 J/m²) after 24 h of PBS soaking. However, when the upper surface of the powder is prehydrated, the adhesive strength and interfacial strength reach a minimum value (3.92 ± 1.78 kPa and 3.75 ± 0.20 J/m², respectively) at a COS-PBA solution concentration of 1%. The increased concentration (over 1%) does not change the adhesion strength of the top surface (SI Appendix, Fig. S9 E and F), possibly because of surface tension of the liquid and the interfacial friction. In addition, we tried the COS solution without grafting PBA. Although the upper surface of the hydrogel becomes unadhesive immediately after spraying COS, the adhesion strength between tissues increased after soaking in PBS for only 1 d (SI Appendix, Fig. S9G), indicating that COS could not be stably maintained on the hydrogel surface without PBA grafting. Considering mechanical properties together with Janus adhesion, we believe that COS-PBA solution with a concentration of 1% is the reasonable choice for surface hydration of FJG powder.

Next, we investigated whether spraying COS-PBA solution affects the bottom surface adhesion of FJG hydrogels. FJG hydrogels hydrated with 1% COS-PBA solution could strongly adhere to different tissues and withstand water flushing and different deformations (Fig. 3E). Further, we freeze-dried the hydrogel-adhered liver tissues and analyzed the adhesion state of the hydrogel–tissue interface using scanning electron microscopy (SEM). FJG hydrogels have a dense interface with the tissue surface, regardless of whether the upper surface is sprayed with PBS or COS-PBA. However, when the prefabricated FJG hydrogels adhere to the tissue surface, there is a clear gap between the tissue and the hydrogel (Fig. 3F). These results indicate that FJG powder could tightly adhere to the tissue by absorbing interfacial moisture, while spraying COS-PBA on the top surface does not affect the adhesion of the hydrogel–tissue interface.

Considering the rapid gelation and high tissue adhesion of FJG powder, we believe that FJG hydrogel can quickly seal the wound and play a hemostasis role, which may seal the clots and reduce adhesions caused by the use of gauze (14, 29). To test this, we constructed a rat liver bleeding model to evaluate the hemostatic capacity of commercial fibrin glue and FJG powder. Rats treated with no hemostatic material lose 1.56 ± 0.16 g of blood before bleeding arrest (~3 min). The Fibrin glue stops bleeding within 10 to 15 s and results in a decreased blood loss of 0.50 ± 0.06 g. FJG powder stops bleeding immediately after application to the wound (<2 s) with minimal blood loss of 0.26 ± 0.06 g (Fig. 3 G and H). These results highlight the excellent hemostatic ability of FJG powder.

In Vivo Antiadhesion Efficiency of FJG Hydrogels in Rat Models

To comprehensively evaluate the antiadhesion efficacy of FJG hydrogels in different tissues, we selected three organs with different motion modalities to establish postoperative adhesion models (i.e., intestine, peristalsis; liver, quasi-static; heart, dynamic contraction). After surgery, the injured area was covered with commercial adhesion barriers Interceed or PBS-sprayed FJG powder (PBS/FJG, as bilateral adhesion hydrogels) or COS-PBA-sprayed FJG powder (COS-PBA/FJG, as Janus-adhesive hydrogels) or no treatment (as control). Rats were euthanized after 2 wk, and images were taken for double-blind evaluation of the degree of adhesions (Fig. 4A). Tissue samples were divided into several segments, and the average adhesion scores were given from 0 to 5 according to the classic clinical-scoring system with 0 representing no adhesion and 5 representing very dense and vascularized adhesions (30) (SI Appendix, Fig. S10).

Fig. 4.

Prevention of postoperative adhesions of different tissues in rat models. A, Gross observation of tissue adhesions in intestine, liver, and heart after 14 d. B–D, Average adhesion score of adhesions formation in different tissues. Each point in the column represents the data of a sample. n > 7 independent samples for each group (untreated, Interceed, PBS/FJG, and COS-PBA/FJG). Data are shown as mean ± SD and compared using one-way ANOVA followed by Bonferroni’s post hoc test. *, ***, and **** indicate P < 0.05, P < 0.001, and P < 0.0001, respectively. ns indicates no significant difference.

In the control group, we observed severe adhesions between organs and the abdominal wall or the thoracic cavity (average adhesion scores of 4.74 ± 0.36 for the intestine, 3.55 ± 0.64 for the liver, and 3.06 ± 0.78 for the heart, Fig. 4 B–D). The use of a commercial adhesion barrier only partially reduces the degree of intestinal adhesions (3.88 ± 0.43, P = 0.0162) but is ineffective for liver adhesions (3.52 ± 1.10) and epicardial adhesions (2.25 ± 1.10). Compared to the control group, the average adhesion score is significantly decreased in all hydrogel-treated groups. For the PBS/FJG group, part of the hydrogels has mild or moderate adhesions to the abdominal wall or thoracic cavity (average adhesion scores of 1.27 ± 0.73 for the intestine, 1.30 ± 0.61 for the liver, and 1.28 ± 0.71 for the heart). We also noticed that such adhesions can be easily detached by breaking the hydrogel (Fig. 4A). The adhesion score of the COS-PBA/FJG group is lower than that of the PBS/FJG group in the intestinal adhesions model (0.50 ± 0.55, P = 0.0223) and liver adhesions model (0.38 ± 0.50, P = 0.0490). For the epicardial adhesions model, there is no significant difference in adhesion score between the PBS/FJG group and the COS-PBA/FJG group (0.36 ± 0.36, P = 0.1226). One possible reason is that the dynamic contraction of the heart prevents the upper surface of the PBS/FJG hydrogel from adhering to the thoracic cavity. In addition, severe epicardial adhesions would restrict myocardial contraction, whereas it is not affected in the viscoelastic COS-PBA/FJG group (Movie S4).

We further assessed the tissue pathology by H&E staining and Masson’s trichrome (Masson) staining (Fig. 5A). In the control group and the commercial adhesion barrier group, tissues are tightly connected with the abdominal wall or thoracic cavity wall through fibrous connective tissue, and the arrangement of cells and extracellular matrix is disordered. Masson staining showed extensive collagen deposition in the adhesion sites. In the PBS/FJG group and the COS-PBA/FJG group, the tissue surface is uniformly covered with hydrogels. For PBS/FJG group, the hydrogel adheres tightly to the abdominal wall, and there is a small amount of collagen hyperplasia at the boundary. For COS-PBA/FJG group, the lower surface of the hydrogel adheres tightly to the tissue, while the upper surface is well defined with no obvious tissue attachment. These results indicate that the COS-PBA sprayed FJG hydrogels can stably adhere to different tissues and play a good antiadhesion effect.

Fig. 5.

FJG hydrogels reduced adhesion formation and acute inflammation in rat models. A, Representative H&E staining (Left) and Masson’s trichrome staining (Right) of tissue sections on day 14. The yellow dashed lines represent the boundaries between the proliferative adhesions and the organs or the implanted materials, and the blue dashed lines represent the boundaries between the implanted material and the organs. [Scale bar, 1 mm (the macro view) and 200 µm (the magnified view).] B–E, Serum concentration of PAI-1, t-PA, TNF-α, and TGF-β1 detected by ELISA. Each point in the column represents the data of a sample. n = 7 independent samples for each group. Data are shown as mean ± SD and compared using one-way ANOVA followed by Bonferroni’s post hoc test. *, **, and **** indicate P < 0.05, P < 0.01, and P < 0.0001, respectively. ns indicates no significant difference

We further investigated the mechanism by which FJG hydrogels prevent adhesions. Tissue plasminogen activator (t-PA) and plasminogen activation inhibitor (PAI-1) are essential for the balance of the fibrinolytic system, which determines most of the adhesion formation in vivo, especially for hemorrhagic injuries (31, 32). Besides, long-term inflammation after injury also interferes with fibrinolytic balance. Tumor necrosis factor-α (TNF-α) and transforming growth factor-β1 (TGF-β1) are two typical proinflammatory factors at the acute inflammatory stage, which have been proven to be associated with the occurrence of postoperative adhesions (33, 34). Given this, we examined blood samples of rats at the acute inflammatory stage (3 d after surgery) using enzyme-linked immunosorbent assay (ELISA) in the liver adhesions model. The results showed that the serum concentrations of TNF-α (Fig. 5B), TGF-β1 (Fig. 5C) and the antifibrinolytic PAI-1 (Fig. 5D) in the PBS/FJG group and COS-PBA/FJG group are lower than that in control and commercial adhesion barrier group while the pro-fibrinolytic t-PA has the opposite result (Fig. 5E).

Design and Test of the Gas-Assisted Endoscopic Delivery Device

To ensure the ease of use of FJG powder in clinical endoscopic surgery, we designed a gas-assisted endoscopic delivery device that can be used with a laparoscopic puncture (SI Appendix, Fig. S11 A and B). As the device is designed as a trigger opening, the operator can easily control the gas-assisted powder delivery by a spring trigger (SI Appendix, Fig. S11C and Movie S5). We first assessed the stability of device delivery. The rate of powder delivery is determined by the gas flow rate and the opening degree of the spring hole. Since the spring trigger can be flexibly adjusted in use, we tested the effect of the gas flow rate on the powder delivery rate at maximum pore opening (SI Appendix, Fig. S11D). The results showed that under the common carbon dioxide flow rate (5 L/min and 10 L/min), the powder can be continuously and stably ejected at an appropriate rate (66.56 mg/min and 96.58 mg/min, respectively). To assess the uniformity of powder delivery, we quantified the Fluorescein Isothiocyanate (FITC)-labeled FJG powder dispersion after spraying onto wet pig skin and used a commercial manual pressing device for comparison (SI Appendix, Fig. S11E). The total fluorescence intensity was used to approximate the amount of powder adhering to the tissue surface. Although the mean values of the two groups have no significant differences, the data for the commercial group have a large dispersion, indicating that the delivery volume of pulse compressions is not controllable (SI Appendix, Fig. S11F). There are fewer fluorescent spots on the tissue surface in the gas-assisted group, indicating the better concentration of the spray (SI Appendix, Fig. S11G). Data on the cover area and circularity of the central spots are also more discrete in the commercial group (SI Appendix, Fig. S9 H and I). In addition, we further validated the feasibility of delivering FJG powder to the liver surface using the gas-assisted device in vitro (SI Appendix, Fig. S11J and Movie S6). Taken together, these results indicate that our gas-assisted endoscopic delivery devices can stably and uniformly deliver FJG powder onto the target tissue surface.

Antiadhesion Efficiency of FJG Powder in a Porcine Model

We used a preclinical translational porcine model to verify the availability, convenience, and efficiency of FJG powder in clinical endoscopic procedures (SI Appendix, Fig. S12A). We designed a simple but standardizable endoscopic surgical procedure and performed surgery on pigs to verify the feasibility of endoscopic-assisted delivery of FJG powder and the effectiveness of FJG hydrogel in preventing postoperative adhesions in large animals (Fig. 6A). Videos and pictures showed that FJG powder can be uniformly delivered to the liver surface and exert a hemostatic effect (Fig. 6A and Movie S7). Then, the COS-PBA solution was sprayed on the powder through another catheter to form FJG hydrogel. Finally, the tissue surface was flushed and drained with saline. For the control group, the abdominal cavity was closed after hemostasis with gauze. For commercial adhesion barriers, the injured area was covered with an Interceed film after hemostasis. The pigs were euthanized, and the abdominal cavity was opened after 2 wk. Similar to the results in the rat experiment, severe adhesions are present in both the control (adhesion score of 4.00 ± 0.66) and commercial adhesion barrier groups (adhesion score of 3.58 ± 0.52). FJG hydrogel significantly reduces the degree of adhesions (adhesion score of 1.17 ± 0.80, Fig. 6 B and C). H&E and Masson staining also verified this (Fig. 6D). Further, we performed demonstration experiments of the endoscopic delivery of FJG powder on the intestinal surface (SI Appendix, Fig. S12B). These results demonstrate the availability of FJG powder in endoscopic surgery and its effectiveness in preventing liver adhesions in a porcine model.

Fig. 6.

Prevention of liver adhesions in porcine models. A, Schematic diagram and in vivo demonstration of in vivo endoscopic surgery procedure using FJG powders in pigs. B, Gross observation of tissue adhesions in the liver after 14 d and representative H&E staining and Masson’s staining of tissue sections on day 14. C, Average adhesion score of adhesions formation. n = 3 independent samples for each group (untreated, Interceed and COS-PBA/FJG). Data are shown as mean ± SD and compared using one-way ANOVA followed by Bonferroni’s post hoc test. * and ** indicate P < 0.05 and P <0.01, respectively.

Discussion

In this work, we achieved the in situ formation of a Janus-adhesive hydrogel barrier using a sprayable powder. The approach that asymmetrical gelation powder can resolve the long-standing contradiction between endoscopic delivery, Janus adhesion, and mechanical properties that limit the clinical application of antiadhesion barriers. To achieve rapid powder hydration and tissue adhesion, we partially disrupt the intermolecular hydrogen bonds of polysaccharides by methacrylation and introduce catechol groups into the network as cross-linking sites for borate bonds and tissue adhesion groups (35). Inspired by the electro-neutralization and antiadhesion effects of cations, we used COS-PBA solution to shield the catechol and carboxyl groups on the surface of FJG hydrogels to obtain Janus-adhesive hydrogels (Fig. 3). In particular, FJG powder is easy to manufacture and store on a large scale and can be applied in a standardized endoscopic surgical procedure, endowing its comprehensive advantage in clinical translation (Fig. 1C).

Janus adhesion of the barrier is critical for postoperative adhesion management. From the mechanical aspect, the dynamic tissues need to be dissociative from the adjacent tissues to perform motion functions (e.g., heart beating, tendon sliding, and gastrointestinal motility) (16, 36). For the therapeutic aspect, sticking adjacent tissues by barrier material may lead to extra adhesions in the boundary region of barriers (as verified by our results of the PBS/FJG group, Fig. 5), and the adhesion of materials to surrounding tissues may lead to barriers breakdown in the target injured area. Janus hydrogels can be easily fabricated in vitro by layered deposition, of which the manufacturing process is often complex or biologically harmful. Step-by-step hydration powders can solve this problem in a biofriendly approach. More importantly, although not shown in this work, the COS-PBA solution can be replaced with a liquid loaded with drugs or bioactive factors (e.g., exosomes or cytokines). This reveals a possible modular clinical treatment modality.

The mechanical property of the adhesion barriers is an overlooked factor that has recently received increasing attention. Low modulus, proper viscoelasticity (shear-thinning and self-healing), and minimized swelling are essential for most adhesion-prone tissues (15, 37). In this work, the dynamic borate bonds endow the FJG hydrogels with good viscoelasticity and excellent self-healing properties. Meanwhile, the introduction of 1% COS-PBA not only achieves Janus adhesion but also reduces the swelling ratio of the hydrogel (Fig. 2F). These together make FJG hydrogel competent as an antiadhesion barrier for various soft tissues, whether the dynamic heart and intestine or the quasistatic liver.

Endoscopic procedures have been widely used in surgical operations to reduce surgical trauma and postoperative adhesions (38). Therefore, barrier delivery that can be compatible with endoscopic procedures is critical for clinical ease of use. Although injectable hydrogels are considered to have good minimally invasive convenience, most injectable hydrogels need preoperative preparation of the precursor solution, which requires a sterile environment and additional surgical preparation. These will increase the cost of the procedure while failing to meet emergency needs. On the other hand, if prefabricated solutions are used, most natural macromolecules will spontaneously degrade in aqueous solutions, which will reduce their shelf life while increase the storage and transportation costs (39). In contrast, ready-to-use FJG powder can be stored at room temperature for a long time and work with our designed delivery device, which can deliver FJG powder uniformly and stably and prevent water vapor from entering the catheter. Although the device was fabricated by three dimensional (3D) printing in this study (costing ~$3 per set), it can be mass-produced by industrial machining for extremely low manufacturing costs (perhaps <$1 per set). Further, the sterilized device can be filled with powder and plastic-sealed for use as a disposable consumable. We also designed a standardized surgical procedure for FJG powder, integrating conventional surgical procedures such as wound hemostasis, barrier delivery, and peritoneal irrigation. Such an integrated clinical solution is undoubtedly beneficial to optimizing surgical programs and reducing medical costs.

In addition to clinical and commercial convenience, another appealing translational feature of FJG powder is that all the components are approved for clinical use by the US Food and Drug Administration. Sodium alginate, hyaluronic acid, and chitosan oligosaccharide have been widely used in wound dressings and implant devices (40–42). This will facilitate preclinical approval of FJG powder and may eventually classify it as a class III device. Although not covered in this work, FJG powder may also have extraordinary potential in other clinical areas such as wound closure, drug delivery, etc. Taken together, our results demonstrate the effectiveness and convenience of FJG powder in preventing postoperative adhesions and highlight its great potential for clinical and commercial translation.

Methods

Synthesis of FJG Powder and Preparation of FJG Hydrogels.

Methacrylation of alginate (molecular weight: ~30,000, Macklin) and hyaluronic acid (molecular weight: ~90,000, Meilunbio) was first carried out using methacrylic anhydride according to previously reported (43, 44). Then, AlgMA-PBA, HAMA-DA, and COS-PBA precursors were synthesized through ethyl-dimethyl-aminopropyl carbodiimide (EDC, Aladdin) and N–hydroxy-succinimide (NHS, Aladdin) coupling chemistry, respectively (45, 46). For AlgMA-PBA, AlgMA (1 g) was dissolved in 100 mL 2-morpholinoethanesulfonic acid buffer at 25 °C. Then, 500 mg 3-aminophenylboronic acid (PBA, Aladdin) was dissolved in 10 mL dimethyl sulfoxide (DMSO) in a centrifuge tube and added to AlgMA solution with 700 mg EDC and 500 mg NHS. The solution pH was maintained from 4.5 to 5.5 for 24 h using 1 M NaOH. After the reaction, the product was precipitated with 10-fold absolute ethanol and redissolved in water. Then, the solution was purified by dialysis against deionized (DI) water for 7 d and then lyophilized. For HAMA-DA, HAMA (1 g) was dissolved in 100 mL PBS buffer at 25 °C with 500 mg dopamine hydrochloride ( Aladdin), 500 mg EDC, and 350 mg NHS. The solution pH was maintained from 4 to 5 for 24 h using 1 M NaOH and then purified by dialysis against DI water at pH = 3 for 3 d. After dialysis, the solution pH was adjusted to 7 using 1 M NaOH and then lyophilized. For COS-PBA, 5 g chitosan oligosaccharide (COS, molecular weight: ~3,000, Shanghai Yuanye Bio-Technology) was dissolved in 80 mL D-H solvent (DMSO/H₂O = 1:1). Then, 1.65 g 4-carboxyphenylboronic acid (PBA) was dissolved in 20 mL D-H solvent in a centrifuge tube with 1.90 g EDC and 1.37 g NHS. The mixture was allowed to react for 30 min at room temperature to activate the carboxyl groups and then added to the COS solution. The solution pH was maintained from 4.5 to 5.5 for 24 h using 1 M NaOH. After the reaction, the solution was purified by dialysis against 1% Na₂CO₃ solution for 1 d and DI water for 3 d and then lyophilized. To confirm successful grafting, the precursors were analyzed by proton NMR (1H NMR) using a 400 MHz JEOL NMR Spectrometer (SI Appendix, Fig. S1). FJG powder was prepared by mixing and pulverizing the lyophilized AlgMA-PBA and HAMA-DA of equal mass. To form FJG hydrogels, FJG powder was hydrated with PBS buffer or COS-PBA solution. For cell culture and animal experiments, all precursors were filtrated through a 0.22-μm membrane filter (Millex-GP; Millipore) before lyophilization.

Characterization of FJG Powder Hydration.

Pig skin was used as a carrier to place different masses of powder in a unit area (1 to 5 mg/cm²) and then sprayed excess PBS or COS-PBA solution (over 2 mL/cm²) on the upper surface. After that, the excess liquid was removed through filter paper This experiment was repeated on a nonadhering Teflon plate to scrape and weigh the obtained hydrogels to determine the hydration ratio of FJG powder. To test the water absorption rate of the powders, we added dropwise 200 μL PBS on 10 mg powder and used filter paper to absorb the unhydrated PBS after 2 s. The water absorption rate was calculated as

$$\frac{m_{\mathit{hyddrogel}}-m_{\mathit{powder}}}{m_{\mathit{powder}}}\times 100\%.$$

Mechanical Characterization.

For rheology tests, an MCR 302 rheometer (Anton Paar) with 15-mm stainless steel parallel plates was used. The samples were prepared by mixing FJG powder with a different solution. The parallel plate gap is set to 1 mm. Then, a frequency sweep test was performed on the hydrogels at a strain of 0.1% and frequency from 0.1 Hz to 10 Hz. Strain sweeps were performed to confirm the linear elastic regime and characterize the shear-thinning property and self-healing property of hydrogels. To characterize the gelation rate, the FJG powder was attached to the upper plate by double-sided tape and then contacted with the COS-PBA solution located on the lower plate. Once the designated location was reached, a time scan was started to determine the modulus change of the hydrogel. Gelation time was defined as the time point when the storage modulus of the FJG hydrogel exceeded the loss modulus. Tensile and fracture tests were carried out using a universal tensile testing machine (Bose) Hydrogels were prepared in a strip mold with a length of 10 mm, a width of 10 mm, and a thickness of 2 mm. The strips were fixed on a stretching clamp at a spacing of about 4 mm. The strips were stretched up to 12 mm (a strain of 300%) at a velocity of 1 mm/s.

Swelling and Degradation of FJG Hydrogels.

Swelling of the hydrogels was assessed by measuring wet weights of the hydrogels after preparation and after incubating in PBS at 37 °C for 7 d. To characterize enzyme degradation of the hydrogels, the hydrogels were soaked in hyaluronidase solution (Solarbio, 1,000 U hyaluronidase for 500 μL hydrogel) at 37 °C and the wet weights were recorded. In vivo fluorescence images of Cy7-labeled hydrogels were taken by an animal imaging system (AniView600, Guangzhou Biolight Biotechnology). The Cy7-labeled hydrogel was constructed by replacing 20% of AlgMA-PBA with Cy7-labeled AlgMA-PBA. Cy7-labeled AlgMA-PBA was synthesized by the reaction of Cy7-amine (Duofluor Inc.) with AlgMA-PBA. Briefly, 200 mg AlgMA-PBA was dissolved in 20 mL water and 10 mg Cy7-amine was added to AlgMA-PBA solution with 140 mg EDC and 100 mg NHS. The solution pH was maintained from 4.5 to 5.5 for 24 h using 1 M NaOH. After the reaction, the product was purified by dialysis against DI water for 3 d and then lyophilized.

Characterization of Janus-Adhesive Properties of the Hydrogels.

To characterize the Janus adhesion ability of FJG hydrogels, trimmed pork (2.5 cm by 2.5 cm) was glued to a glass slide and FJG powder was placed on the pork. The powder was prehydrated with PBS or cos-PBA. Then, the two pieces of pork were put together and lifted up. For shearing and pealing tests, trimmed pork pieces were glued with different adhesives with a bonding area of 2.5 cm × 2.5 cm. After stabilizing for 15 min, pork pieces were separated at a speed of 10 cm/min and the force–displacement curve was recorded. The adhesive strength was defined as the maximum stress reached before failure and adhesive toughness was calculated as integral of the stress-displacement curve. The chemical compositions of the hydrogel surface were measured using an X-ray photoelectron spectrometer (Thermo Fisher ESCALAB Xi+). A monochromatic Al Kα X-ray was used as an excitation source (hν = 1486.6 eV) running at 15 kV and 10 mA. Microscopic images of the hydrogel–tissue interface were obtained by SEM (SU3500, Hitachi). To characterize the hemostasis ability, one-third of the rat liver was excised and treated with different agents. The amount of bleeding was determined by weighing on absorbent paper.

Cell Culture and In Vitro Experiments.

Human umbilical vein endothelial cells (HUVECs) were cultured in DMEM added with 10% fetal bovine serum (Gibco/Thermo Fisher Scientific) and 1% Penicillin-Streptomycin (Gibco/Thermo Fisher Scientific) at 37 °C in 5% CO₂. To evaluate the cytocompatibility of the hydrogels, hydrogels were presoaked in DMEM for 24 h and the extract was used to culture cells for 48 h. Cell viability was evaluated using a Live/Dead Viability Kit (Invitrogen) according to the manufacturer's instructions.

In Vivo Postoperative Adhesions Models of Rat.

Male Sprague–Dawley rats (Laboratory Animal Center of Xi’an Jiaotong University) (weight 200 to 250 g) were used to establish postoperative adhesion models. Rats were first anesthetized with 1.5% sodium pentobarbital solution and mechanically ventilated through tracheal intubation. For the intestinal adhesion model, the caecum and adjacent abdominal wall were scraped with gauze and tweezers until the epidermis was damaged. Then, the cecum was sutured to the abdominal wall with 6-0 sutures to secure the position of the cecum. For the liver adhesion model, the liver surface was wounded with forceps. For the epicardial adhesion model, the epicardial surface was scraped with a 30G needle. After surgery, the injured area was covered with different barriers. To evaluate the adhesion scores, organs were divided into several segments (three segments for intestine, four segments for liver, and four segments for heart). The average score of each segment was used as the indicator of the adhesion degree. Adhesion scores were given from 0 to 5 according to the classic clinical-scoring system (0 for no adhesions; 1 for a few filmy adhesions; 2 for numerous filmy adhesions; 3 for moderate adhesions; 4 for dense adhesions; and 5 for very dense and vascularized adhesions). Serum samples of rats were collected on day 3 after surgery for ELISA analysis. All ELISA kits were purchased from Proteintech. After 14 d, rats were euthanized, and the harvested tissues were fixed in a 4% (w/v) formaldehyde solution, paraffin-embedded, and sectioned. H&E staining and Masson's trichrome staining of sections were conducted using standard methods.

Gas-Assisted Delivery Device Design and Testing.

The delivery device consists of a long delivery catheter (length of 30 cm and diameter of 1 cm) and a vial with a spring valve, which could be assembled by threading or nesting. The tail of the delivery catheter and the bottom of the vial are connected to the same CO₂ source. The delivery catheter is kept ventilating to prevent the water vapor in the abdominal cavity from entering the catheter. The device was designed in SolidWorks and manufactured by 3D printing (WeNext Technology Co., Ltd.). All tests are carried out at a gas pressure of 5 kPa. The delivery device was sterilized with EtO before surgery.

In Vivo Preclinical Porcine Model.

All animal experimental procedures were approved by the Xi’an Jiaotong University Biomedical Ethics Committee (No. 2022-1034). Pigs weighing 30 to 40 kg were injected intravenously with propofol to induce anesthesia. Mechanical ventilation was provided by endotracheal intubation, and anesthesia was maintained with isoflurane (0 to 5% in oxygen) during the surgery. Standard laparoscopic interventional procedures were carried out after skin disinfection. CO₂ gas was inflated into the abdominal cavity to bulge the abdominal wall and expose the surgical field. The liver surface was scratched 10 times with gauze, and a wound (1 to 2 cm) was cut out to induce adhesions. The FJG powder was delivered to the liver surface by the delivery device according to the following procedure: first, delivering the powder 2 to 5 cm from the tissue surface until the target area is uniformly covered with the powder; second, withdrawing the delivery device and spraying the COS-PBA solution onto the powder-covered area using another catheter; and third, flushing the hydrogel-covered area with saline. Finally, draining the excess fluid from the abdominal cavity under negative pressure. During the spraying process, the gas source pressure of the delivery device was kept at 5 kPa, while the intraabdominal air pressure was kept constant by dynamic gas pressure detection. After surgery, the wound was closed with 2-0 sutures and wrapped with gauze to reduce infection. On the 14th day, all pigs were euthanized, and the tissues were harvested.

Microscopy and Image Analysis.

Unless specifically mentioned, bright-field images were taken with an Olympus IX2-UCB microscope, and fluorescence images were taken with an Olympus FV3000 confocal laser-scanning microscope. Tissue sections are imaged by a slide scanner (Pannoramic DESK, 3DHISTECH). Images were analyzed using Image J (NIH).

Statistics and Software.

Statistical analyses were performed using GraphPad Prism 9.1 (GraphPad Software). Multiple comparisons among the three or more groups were performed using one-way ANOVA with Bonferroni post hoc testing, and differences between two groups were analyzed by a two-tailed Student’s t test. P < 0.05 was taken as the threshold for significant differences. All data are shown as mean ± SD. The delivery device is designed with SolidWorks 2016. All chemical structures are drawn by Integle Chemistry Draw (Integle), and all schematics are drawn and licensed by Biorender (XM24RDX6IF).

Supplementary Material

Appendix 01 (PDF)

Movie S1.

Demonstration of water absorption of different powder.

Movie S2.

Demonstration of FJG powder hydrated with different solutions for bonding tissues.

Movie S3.

Demonstration of lifting heavy objects with FJG powder bonded pig skin.

Movie S4.

Demonstration of the effect of epicardial adhesions and FJG hydrogel on myocardial contraction.

Movie S5.

Demonstration of the working mechanism of the delivery device.

Movie S6.

Demonstration of in vitro delivery of FJG powder using the delivery device.

Movie S7.

Demonstration of the procedure of in vivo delivery of FJG powder to liver and intestine via endoscopic surgery.

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.