Gelation of highly entangled hydrophobic macromolecular fluid for ultrastrong underwater in situ fast tissue adhesion.

Although strong underwater bioadhesion is important for many biomedical applications, designing adhesives to perform in the presence of body fluids proves to be a challenge. To address this, we propose an underwater and in situ applicable hydrophobic adhesive (UIHA) composed of polydimethylsiloxane, entangled macromolecular silicone fluid, and a reactive silane. The hydrophobic fluid displaced the boundary water, formed an in situ gel, bonded to tissues, and achieved exceptional underwater adhesion strength. Its underwater lap shear adhesion on porcine skin was significantly higher than that of cyanoacrylate and fibrin glues, demonstrating excellent water resistance. The burst pressure of UIHA on porcine skin was 10 times higher than that of fibrin glue. The cytocompatible UIHA successfully sealed ruptured arteries, skin, and lungs in rats, pigs, rabbits, and dogs. Together, the gelation of highly entangled hydrophobic macromolecular fluid provided a means to prepare underwater bioadhesives with strong bonding to tissues and excellent water resistance.

INTRODUCTION

Wet adhesives are applicable in all spectrums of tissue trauma where instant wound closure and hemostasis are needed, especially when time is critical in life rescue (–). Instant and strong wet/underwater adhesion performance is also required for wound closure in the scenarios of blood or body fluid exposure. However, adhesion to wet substrates or underwater surfaces is still a challenge, as water molecules in the boundary layer of the interface impede the direct contact between adhesives and substrates (, –). Inspired by marine mussels, sandcastle worms, tree frogs, clingfish, and geckos (–), several approaches were proposed to overcome surface water, including coacervation of polyelectrolytes, catechol bonding, dry hydrogel tapes absorption, or repelling water through micropattern (, , –).

Several recently reported bioadhesives for wet/underwater applications are hydrophilic polymers (, , , ). Since water is absorbed to the interface because of the bioadhesive hydrophilicity, adhesion either fails altogether or, if it does not, the adhesive forces are low. Another challenge with hydrophilic bioadhesives is their washout during in vivo application before cross-linking occurs. Moreover, hydrophilic adhesives have poor water resistance to avoid water swelling or diffusion, and the adhesion performance in wet conditions or underwater decreases considerably after 24 hours or longer (table S1). One strategy to address these challenges is to design hydrophobic bioadhesives that can repel water from the interface and establish a strong wet/underwater adhesion. Although the contribution of hydrophobic components is recognized in mussel’s wet adhesion, hydrophobic adhesives without hydrophilic components have not been explored for bioadhesives (–) because of the following intrinsic drawbacks: (i) water-immiscible hydrophobic fluid that tends to form isolated droplets in water due to capillary breakup from interfacial tensions (), (ii) nonspecific hydrophobic interactions are not as strong as covalent bonds (), (iii) most highly mobile hydrophobic adhesive materials required for diffusion to build strong interactions () are small organic molecules with cytotoxicity, and (iv) rapid curing is difficult to achieve. Recently, some hydrophobic bioadhesives were reported for wet/underwater environments containing some hydrophilic components. Lang et al. () proposed a hydrophobic liquid polyester prepolymer adhesive, which could be applied on wet biological tissue surfaces (). However, its adhesion strength to tissues was generally low (ca. 20 kPa) and primarily depended on noncovalent hydrogen bonding interactions. In addition, the underwater adhesion and water-resistant performances were not investigated. Han et al. () reported an Fe3+-induced dynamic hydrophobic hydrogel with an underwater adhesion strength of ~20 kPa, which is mainly attributed to the nonspecific interactions from C18-long aliphatic chains with substrates (). In addition, the underwater adhesion performance of the above hydrophobic adhesive decreased with a longer soaking time, which may be attributed to the presence of the hydrophilic components influencing the overall water resistance. A hyperbranched liquid polymer–based underwater adhesive containing hydrophobic backbone and hydrophilic catechol side branches that was solidified in the presence of water through spontaneous coacervation driven by hydrophobic aggregation was also reported (). However, the initial adhesion was low, and a longer curing time of 2 hours was needed. Collectively, literature reports to date indicate that in situ adhesives with short curing time, strong instant underwater adhesion, and high water resistance are still elusive (table S1) (, , , , –, , , –). Therefore, designing wet/underwater adhesives driven by the hydrophobic strategy is still challenging, especially for achieving strong instant adhesion under wet/underwater environments.

To form strong bioadhesion in wet/underwater conditions, we designed a water-immiscible hydrophobic fluid adhesive with high mobility and low surface energy to repel surface water through hydrophobic exclusion. The underwater and in situ applicable hydrophobic adhesive (UIHA) reported here diffused on the irregular surface of biological substrates and formed instantly strong interactions through mechanical interlocking when cross-linked. Furthermore, UIHA kept excellent water resistance attributed to hydrophobicity, and underwater adhesion was further enhanced via covalently bonding with substrates. The designed highly entangled hydrophobic macromolecular fluids in UIHA comprised three components: (i) macromolecular silicone fluid to provide dynamic entanglements, thus preventing underwater capillary breakup; (ii) reactive polydimethylsiloxane (PDMS) precursor for gelation; and (iii) a small amount of silane to covalently bridge the hydrophobic-hydrophilic interfaces (). Silicone has excellent tissue and blood compatibility and low cytotoxicity and has been widely used as a medical filling/implant material, as a drug delivery system, and as a wound care product (–). Despite these favorable features, silicone fluids have not been investigated for their potential as underwater bioadhesives. Here, we designed hydrophobic underwater adhesive based on silicone fluid because of its high flexible backbone for high mobility, high hydrophobicity for good water resistance, and low surface energy for repelling water from the substrate. The designed molecular composition of UIHA is different from regular silicone sealants, containing silane coupling agents and large amounts of entangled silicone macromolecules. Silane coupling agents were previously used as adhesion boosters through surface pretreatment of substrates before applying the adhesives (), but never for tissue substrates. They were rarely combined and used with PDMS simultaneously (), which is not suitable for rapid curing. Although this affects the curing efficiency and the adhesive cohesion, it fulfills the specific requirements for underwater in situ applications, enhances underwater adhesion by building covalent bonding with substrates in one pot, and improves the overall underwater adhesion performance.

RESULTS AND DISCUSSION

Adhesion mechanism and gelation behavior of macromolecular hydrophobic silicone fluids

The adhesion process of UIHA involves three phases, as shown in Fig. 1. In phase I, the entangled macromolecular hydrophobic fluid (PDMS precursor and silicone oil) is injected into water, contacts the substrate, displaces water (the weak boundary layer), and spreads on the surface owing to the low surface energy of silicone (, ). To confine the hydrophobic liquid within the applied areas, suppression of complete wetting and spreading is needed through viscous dissipation, thus requiring an entangled macromolecular silicone component for high viscosity (, ). Following an initial adhesion of the hydrophobic liquid thread, entangled bulky fluid continues to adhere to the substrate because of hydrophobic and viscous interactions. High viscosity is also required to keep the fluid continuous and hinder the capillary thinning and breakup induced by surface tension (, ). The reason is that capillary thinning is more gradual in viscoelastic fluids because of retardation of high viscosity and stress from extensional deformation, thus hindering the capillary breakup of the hydrophobic fluid after injection. In phase II, well-diffused fluid on the substrate solidifies through cross-linking of the PDMS network, and the adhesion is mainly attributed to hydrophobic interaction and interfacial interlocking of adhesives with irregular substrate surface (). Last, the interfacial adhesion is further enhanced by covalently bonding from silane hydrolysis in phase III.

Fig. 1.

Underwater adhesion mechanism illustration of UIHA adhesive.

(A) The in situ adhesion process of UIHA underwater, (B) the chemical reaction scheme inside UIHA and at the interface between UIHA and substrate, and (C) the proposed biomedical applications of UIHA. Unlike small molecules, gelation of macromolecular fluids shows obvious chain length dependence of the silicone fluids. Three types of silicone fluids with different viscosities (molecular weights), namely, silicone 500,000cs (silicone 500k), silicone 10,000cs (silicone 10k), and silicone 200cs (silicone 200), were chosen. Silicone 500k has a weight average molecular weight of ~260 kDa, much higher than its critical entanglement molecular weight (29 kDa; Mc,silicone) (), while the molecular weights for silicone 200 and silicone 10k are ~9.5 and ~60 kDa, respectively (). As shown in Fig. 2 (A and B), the gelation was assessed by the crossover of the storage modulus (G′) and loss modulus (G″), which shows obvious molecular weight and concentration dependence. Among the three silicone fluids, the lowest critical gelation concentration was 12 wt % for silicone 200, but 25 wt % for silicone 500k, suggesting that long-chain architecture hindered the formation of a cross-linked network. For the reactive PDMS precursor, the gelation time was only ~2 min at room temperature, which was delayed to ~7 min with the addition of 0.1 wt % silane to the PDMS, as the hydrosilylation reaction of silane competes with cross-linking and slows the formation of PDMS network (Fig. 2C). With the addition of silicone 500k, both G′ and G″ increased, but the gel point time was further delayed due to the steric effect. The amount of silane also had an important influence on UIHA gelation, which retarded gelation time and decreased gel modulus (Fig. 2D), as some cross-link points in the PDMS network were substituted by conjugation with silane (fig. S2). When the silane concentration reached 0.5 wt % of the PDMS precursor, the gel modulus decreased considerably, and the cross-linking reaction between the silane groups inside the UIHA took a longer time. The reason for this is that, although the water molecules are displaced, cross-linking may still take place slowly inside the matrix because of the mobility of flexible chains. The storage modulus (G′) of UIHA gels increased by ~27% when soaked underwater for up to 2 days (fig. S3), suggesting that the elastic cross-linking network became well developed over time, which may be due to the continuous cross-linking of silane groups inside UIHA.

Fig. 2.

Gelation behavior of macromolecular hydrophobic silicone fluids.

(A) Critical concentrations of PDMS precursor in various silicone fluids. (B) Time sweep rheological profiles of PDMS precursor/silicone fluid mixtures at their respective critical gelation concentrations (silicone 200: 12 wt %, silicone 10k: 20 wt %, and silicone 500k: 25 wt %). (C) Time sweep rheological profiles of PDMS precursor, silicone fluid/PDMS precursor mixture (25:75, wt %), and silicone 500k/PDMS precursor mixture (25:75, wt %) containing 0.1 wt % VTMS. (D) Time sweep rheological profiles of PDMS precursors containing different concentrations of silane. (E) Time sweep rheological profiles of silicone fluid 500,000/PDMS precursor mixture (25:75, wt %) containing 0.1 wt % VTMS in dry condition or underwater. (F) Gelation time versus temperature profiles of PDMS precursors and silicone 500k/PDMS precursor mixture (25:75, wt %) containing 0.1 wt % VTMS, respectively. (G) The hydrolysis kinetic of VTMS silane in an immiscible binary phase system of hydrophobic CDCl3/water. (H) Dynamic strain amplitude test of silicone 500k/PDMS precursor (75:25, wt %) at the shearing rate of 10 rad/s with alternating strain amplitudes of 1 and 400%. (I) Dynamic strain amplitude test of silicone 500k/PDMS precursor (65:35, wt %) at 10 rad/s with alternating strain amplitudes of 1 and 400%. (J) Dynamic strain amplitude test of silicone 200/PDMS precursor (75:25, wt %) at 10 rad/s with alternating strain amplitudes of 1 and 400%. (K) Photographs showing self-healing behavior of silicone gel (PDMS precursor/silicone 500k = 25:75, wt %). (L1 and L2) The SEM images of PDMS-silicone gel (25:75, wt %) morphology after silicone fluid was extracted using hexane. Photo credit: Y. Liu, University of Manitoba.

A fast cross-linking (~5 min) platinum-catalyzed PDMS precursor (Ecoflex 00-35) was used to prepare an organogel network, and the infused silicone fluid was used to consume dissipative energy by internal friction from fluid motion in the gel matrix (, ). A small amount of vinyltrimethylsilane [VTMS; <0.2 weight % (wt %)] was added, as vinyl groups are capable of reacting with PDMS precursors through hydrosilylation, while the silane groups hydrolyzed with water or hydroxyl groups on the substrate (Fig. 1B and fig. S1) ().

The underwater gelation behavior of UIHA is critical for biomedical applications. Figure 2E shows the gelation time, G′, and G″ are similar in a dry condition or underwater, suggesting excellent water tolerance of UIHA. As the hydrosilylation reaction rate is temperature dependent, the gelation of UIHA became faster when the temperature increased. As shown in Fig. 2F, the gelation of UIHA was over 30 min at 10°C, 7 min at 25°C, 2 min at 37°C, and 0.4 min at 50°C. Therefore, the components of UIHA were mixed and stored at low temperatures for an extended period but were cross-linked rapidly at a physiological temperature of 37°C. With increasing temperature, the cross-linking time could be shortened to less than 1 min. The temperature dependence and underwater gelation suggest the potential application of UIHA for emergency bleeding and trauma treatment. The hydrolysis kinetics of silane in a water/water-immiscible liquid binary phase system (water/chloroform-d) was examined by 1H–nuclear magnetic resonance (NMR) (Fig. 2G and fig. S4), and, as expected, the hydrolysis ratio increased with time.

In a highly entangled silicone fluid/PDMS gel system, the PDMS covalent cross-link formed the primary elastic network, which was strong and nonrecoverable. The chain entanglement of free silicone macromolecules formed the secondary dynamic network, which was weak but recoverable and self-healable, and therefore entangled silicone gels exhibited notable self-healing behavior. As shown in Fig. 2K (fig. S5 and movie S1), two pieces of silicone 500k/PDMS gels were easily merged by gentle contact, and the self-healing performance was determined by the fraction of the dynamic components. When the weight fraction of PDMS precursor was between 25 and 35 wt %, entangled silicone gels showed typical self-healing behavior, in which G″ was higher than G′ for the high-strain amplitude region (Fig. 2, H and I). With increasing PDMS weight fraction to 40 wt %, the entangled silicone gel became nonrecoverable, as increased cross-link density of the primary nonrecoverable network restrains mobility of the entangled silicone fluid. On the contrary, the gel with silicone 200 did not show self-healing activity because of reduced entanglement with shorter chains (Fig. 2J). The gelation of macromolecular silicone fluid also works for other PDMS precursors with various gelation times and modulus (fig. S6). The volume of the UIHA gel shrank ~30% after the silicone fluid was extracted by sonication in hexane but still kept the typical porous morphology of bulky gels, as shown from the scanning electron microscopy (SEM) images in Fig. 2L.

Underwater casting and writability of entangled UIHA

The highly entangled UIHA is suitable for underwater in situ bioadhesion, surgical sealing, and even electrical isolation. The water-immiscible entangled macromolecular fluid could form the continuous phase underwater with little influence from interfacial tension and corresponding capillary breakup because of the high viscosity. The critical entanglement concentration or weight fraction of high–molecular weight silicone fluid in reactive PDMS precursors was determined by rheological tests, as shown in Fig. 3A. The UIHA was also suitable for underwater writing (Fig. 3B). The viscosity of silicone 500k/PDMS mixtures under varying shear conditions showed shear-thinning viscoelastic behavior, as shown in Fig. 3C. As viscoelastic fluids, when UIHA adhesive was ejected and contacted the substrate, the momentum would not be dissipated instantly, and the fluid would apply a certain pressure on the substrates to repel the surface water layer and achieve close contact with the substrate. The underwater engagement pressure and the separation of normal stress between the moving plate and adhesive liquid were measured using an in-house built instrument, as shown in Fig. 3D. Figure 3 (E to I) shows the normal stress of liquid adhesive and control groups on stainless steel substrate, which was engaged/disengaged by another steel plate at a constant speed of 10 μm/s. The PDMS precursor showed a low-pressure and breakup adhesion on steel substrate because of its instant energy dissipation. With the addition of silicone 500k, the highly entangled fluid shows typical viscoelastic behavior with the normal stress of 1000 Pa (25 wt % silicone 500k) and 3000 Pa (50 wt % silicone 500k), respectively. The adhesion of adhesive liquid (pregel) during disengagement in water or dry condition is similar, indicating the full contact of adhesive and substrate where the boundary water barrier was completely displaced. Because of the excellent in situ underwater adhesion, UIHA could be applied for underwater in situ electrical isolation and water burst sealing, as shown in Fig. 3 (J and K) (fig. S7 and movies S2 and S3), which shows the potential future applications in bioelectronic implantation and sealing.

Fig. 3.

Underwater casting and writability of UIHA.

(A) The viscosity (as the shear rate approaches zero) versus PDMS precursor weight fraction (inset shows the photograph of PDMS precursor and UIHA written respectively on glass substrate underwater). (B) Digital photograph showing underwater writing of UIHA. (C) The curves of viscosity versus strain amplitude of various silicone fluid mixtures, including PDMS precursor, silicone 10k, silicone 500k/PDMS A [25:75, wt %; to prevent curing of liquid PDMS during the experiment, only E35A (PDMS) was used for the experiment, as E35A and E35B have similar viscosity profiles], silicone 500k/PDMS A (25:75 wt %), and silicone 500k. (D) Scheme showing measurement of normal stress of silicone mixture droplets engaging/disengaging a stainless steel plate at a constant speed of 10 μm/s. The normal stress of liquid droplets during plate engagement/disengagement at dry or underwater conditions. Liquid droplets include (E) PDMS precursor [for the engagement/disengagement experiments, only E35A (PDMS) was used to prevent curing], (F) silicone 500k/PDMS (25:75 wt %), (G) silicone 500k/PDMS (25:75 wt %), and (H) silicone 500k. (I) The normal adhesion stress of silicone mixture liquids on stainless steel plate underwater or in dry condition [**P = 0.0049; ***P = 0.0005; ****P < 0.0001; ns: P = 0.9703 (silicone 500k), ns: P = 0.2254 (silicone 500k: PDMS = 50:50), ns: P = 0.7925 (silicone 500k: PDMS = 25:75); n = 3]. (J) Under a water pressure of ~5 psi, a PP tube with punctured holes was in situ sealed with UIHA. (K) Underwater directly sealing/isolating leakage of a working circuit. Photo credit: Y. Liu, University of Manitoba.

Dry and underwater adhesion performance of UIHA on different surfaces

The UIHA shows impressive instant underwater adhesion performance as shown in Fig. 4A and movie S4. The shear adhesion on glass (25 mm × 15 mm) of underwater in situ–coated UIHA can support a weight of 5 kg under a flowing water condition, as shown in Fig. 4B (movies S5 and S6). The shear adhesion of UIHA on glass, PDMS, and porcine skin substrates without silane demonstrated that silicone 500k/PDMS (75:25, weight ratio) has an optimal shear adhesion (Fig. 4, C and E). With the introduction of 0.1 wt % silane to the PDMS, the lap shear adhesion on glass doubled to 164.8 ± 20.4 kPa in 15 min because of the covalent bonds between silane and abundant hydroxyl groups on the glass, and the average values continued to increase but were not significant with extended time as shown in Fig. 4 (F and G). Considering both shear adhesion and gelation time, PDMS precursors/silicone 500k (75:25, weight ratio) with 0.1 wt % silane could form strong adhesion in a short time, and thus, this ratio was used in UIHA. Figure 4 (H to K) presents the underwater adhesion performance of UIHA on the glass and porcine skin substrates. The shear adhesion strength of UIHA on glass soaked in water for 2 days increased to 216 ± 33.8 kPa, exhibiting excellent adhesion and stability. Unlike the glass substrate, the instant adhesion strength of UIHA on porcine skin increased from 25.4 ± 2.7 to 34.0 ± 2.3 kPa (~33% improvement) with silane coupling agents. After 2 days underwater, the adhesion strength increased nearly threefold to 89.1 ± 7.4 kPa, attributed to the covalent bonding between the silane and the hydroxyl groups on porcine skin (fig. S8). The adhesion strength increase on porcine skin is lower than that on the glass, which may be due to the slower formation of covalently bridging bonds on porcine skin, as the glass surface has abundant hydroxyl groups (Fig. 4K). Therefore, a longer time is needed for silane on the interface to react with the porcine skin substrate to build stronger adhesion. Overall, the adhesion strength of UIHA with silane increased to ~3-fold on both glass and porcine skin substrate after 2 days of underwater exposure, compared with UIHA without silane. To gain a comparative insight into the adhesive strength of UIHA, we evaluated three types of currently available medical glues [cyanoacrylate, fibrin glue, and GelMA (methacrylated gelatin)]. As shown in Fig. 4L, UIHA showed significantly higher shear adhesion strength in a wet environment and higher water stability on porcine skin substrates than the comparators. Although the instant adhesion (i.e., as prepared) of the cyanoacrylate glue is higher than that of UIHA, its water resistance is poor, and its underwater adhesion after 1 and 2 days of soaking is significantly lower than that of UIHA. Furthermore, the application of cyanoacrylate led to tissue surface hardening, which may not be suitable for soft tissue applications. The adhesion strengths of cyanoacrylate, fibrin glue, and GelMA significantly decreased when soaked underwater, while the lap shear adhesion strength of UIHA increased after underwater exposure. The adhesion fracture energy of UIHA on glass and porcine skin tissue substrates was also evaluated, as presented in Fig. 4 (M and N), and UIHA exhibited excellent interfacial toughness on glass, reaching 848 ± 122 J/m2 after underwater exposure for 2 days. On porcine skin, the interfacial toughness of UIHA was 46.6 ± 14.4 J/m2, and further increased to 115.2 ± 13.8 J/m2 after underwater exposure for 2 days, demonstrating its applicability as tissue adhesives (, ). The lap shear adhesion strengths of previously reported tissue adhesives on porcine skin substrates under dry conditions are presented in fig. S9 (, , , –).

Fig. 4.

The dry and underwater adhesion performance of UIHA.

(A) The photographs show a weight of 50 g adhered to a glass substrate with UIHA in 5 min underwater. (B) The photographs show the underwater in situ adhesion of UIHA on glass, affording a weight of 5 kg. (C and D) The lap shear adhesion of PDMS/silicone 500k mixtures with different weight fractions (25:75, 50:50, and 75:25) on (C) glass substrate (*P = 0.0136; ***P = 0.0084; n = 3), (D) PDMS substrate (****P < 0.0001; n = 3), and (E) porcine skin substrate (*P = 0.0312; ***P = 0.0001; n = 3). (F) The lap shear adhesion stress-strain curves of UIHA containing different amounts of silane on a glass substrate. (G) The lap shear adhesion strength of UIHA containing different amounts of silane on glass substrate (*P = 0.0154; **P = 0.0024; ns: P = 0.3722; n = 3). (H and I) The lap shear adhesion stress-strain curves of UIHA with/without silane (0.1 wt %) stored underwater for a different period on (H) glass or (I) porcine skin substrates. (J and K) The lap shear adhesion of UIHA with/without silane (0.1 wt %) soaked underwater for a different period on (J) glass substrate (n = 3; P values are presented in table S2) or (K) porcine skin substrate (n = 3, P values are presented in table S3). (L) The lap shear adhesion of cyanoacrylate, fibrin glue, GelMA, and UIHA on porcine skin substrate soaked underwater for a different period (n = 3; P values are presented in table S4). (M and N) The interfacial toughness of UIHA soaked underwater for a different period on (M) glass substrate (n = 3; ns: P = 0.07741; ***P = 0.00092) or (N) porcine skin substrate (n = 3; *P = 0.01561; **P = 0.00398). Photo credit: Y. Liu, University of Manitoba.

In vitro burst pressure and in vivo adhesion of UIHA to arteries in multiple animal models

Repair of arterial rupture and lung leakage is still a challenge for bioadhesives (, , –). To evaluate the adhesive strength of UIHA, we first performed in vitro burst pressure test in blood vessel and lung models. A punctured hole 2 mm in diameter on porcine skin was then placed on a pressure chamber and sealed by either UIHA or fibrin glue. The sealed skin samples were then kept underwater for 24 hours or at ambient conditions before the burst pressure was measured. As shown in Fig. 5A, the burst peak pressure for UIHA was 120.60 ± 9.36 kPa; its underwater burst pressure increased to 136.80 ± 4.08 kPa after 24 hours (P = 0.0285). In contrast, the fibrin glue had burst pressures of 11.27 ± 3.79 kPa (P < 0.0001) and 10.57 ± 1.59 kPa (P < 0.0001) for samples kept at ambient condition and underwater for 24 hours, respectively. It is evident that the UIHA showed superior sealing strength with nearly 12-fold enhancement. A very tight bonding occurred in the interaction interfaces between UIHA and porcine carotid artery ex vivo through SEM (Fig. 5B). Meanwhile, UIHA displayed superior antiswelling ability in the aqueous phase (fig. S10). For in vivo tests, UIHA was applied to the incision of the carotid artery in the rat model. The puncture created by a 25-gauge needle (around 0.5 mm in diameter) in the vessel (around 1 mm in diameter) led to arterial blood spurting, after which 10 μl of UIHA was applied to the puncture for 3 min (Fig. 5C, i and ii, and movie S7). After the vessel clamps were released, no visible leakage was detected, and no artery hematoma was observed for 24 hours after surgery. After 3 days, micro–computed tomography angiography (CTA) three-dimensional reconstruction of the carotid artery demonstrated 100% patency for all UIHA-sealed arteries, similar to the normal artery (Fig. 5C, iv). On day 3, ultrasound (Fig. 5C, v, and movie S8) revealed that UIHA (around 0.386 mm in the thickest part) adhered to the wall of the vessel, and no vascular stenosis or thrombus was formed. From hematoxylin and eosin (H&E) staining (Fig. 5C, vi), a thin fibrous capsule of about 50-μm thickness was observed around the UIHA-sealed vessels after 8 weeks. The immunohistological analysis showed neither CD68+ macrophages (Fig. 5D, i) nor CD3+ lymphocytes (Fig. 5D, ii) on day 56, suggesting the absence of inflammatory reaction after 8 weeks.

Fig. 5.

In vitro and in vivo UIHA adhesion to arteries.

(A) Burst pressure of UIHA, UIHA underwater for 24 hours, fibrin glue, and fibrin glue underwater for 24 hours (n = 3). (B) SEM images of the interface between the porcine artery and UIHA (ex vivo). (C) Rat carotid artery incisions sealed by UIHA in vivo (n = 8). Incisions before (i) and after sealing on days 0 (ii) and 56 (iii). (iv) Micro-CTA at day 3, showing the blood flow in the UIHA sealed site (left) and the normal control without operation (right). (v) Ultrasound image demonstrating UIHA-adhered carotid arterial wall and UIHA after 3 days. (vi) H&E staining of the adhesive interface between the rat artery and UIHA on day 56. (D) Fluorescence immunohistochemical analysis of the UIHA-sealed arteries after 56 days showing no noticeable local macrophage (CD68) (i) and lymphocyte (CD3) infiltration (ii). Femoral artery incision sealed by UIHA on (E) rabbit and (F) beagle dog (n = 3). Femoral artery incisions before (i) and after sealing (ii). (G) Porcine femoral artery incisions sealed by UIHA (n = 5). Porcine artery incisions before (i) and after sealing on days 0 (ii) and 56 (iii). Doppler images in cross section (iv) and vertical section (v and vi) were used to demonstrate blood flow after sealing. Photo credit: G. Guan and Y. Li, Army Medical University.

Encouraged by the above results, we made a longitudinal incision by a scalpel in the femoral artery of a rabbit (around 2 mm) (Fig. 5E, i and ii) and a beagle dog (about 2 to 3 mm) (Fig. 5F, i and ii). As can be observed, UIHA blocked the blood leakage efficiently. To further evaluate UIHA’s strong adhesion, we applied UIHA to seal the porcine femoral artery incisions (Fig. 5G, i and ii, and movie S9). When a 2- to 3-mm longitudinal incision was created by a scalpel, blood spurted from the injury, and then, UIHA was applied to seal the arterial injury (about 5 to 10 min). Hemorrhage was not observed after removing the hemostatic clips, and hematoma was not detected within 24 hours after surgery. Doppler imaging revealed the blood flow in the cross section (Fig. 5G, iv) and vertical section (Fig. 5G, v), and thrombus formation was not detected on day 3. The average blood flow velocity of UIHA-sealed arteries (62.72 ± 6.99 cm/s) was similar to that of the sham group (69.82 ± 3.78 cm/s, P = 0.5606). Thus, the UIHA was able to seal the artery effectively by maintaining adhesive strength and biocompatibility.

In vivo performance of UIHA to seal lung, skin, and bone injuries in rats and pig models

Because of the strong performance of UIHA in repairing ruptured arteries of four different animal models, we examined its potential use to repair other tissues such as the lung, skin, and skull bone. When UIHA was applied to injuries of lung and skin tissues, tightly bonded interfaces were found between UIHA and the lung (Fig. 6A, i and ii) as well as between UIHA and the skin (Fig. 6B, i and ii). UIHA can be applied on a patch to a complex physiological environment, such as in the lung, where high pressure and hemorrhage following an incision are fatal. The cross-linked UIHA patch (around 5 mm in diameter) coated with a thin layer of pregelled UIHA sealed the blood-leaking lungs of rats effectively and stopped the bleeding (Fig. 6A, iv, and movie S10). After 7 days, UIHA remained adhered to the wound site (Fig. 6A, v), and the in situ water immersion test showed the absence of air bubbles. On the basis of UIHA-sealed lung leakage on rats, we made a 1-cm incision on the porcine lung using a scalpel (Fig. 6B, i). UIHA patch (approximately 1.2 cm in diameter) and UIHA liquid were combined (Fig. 6B, ii and iii, and movie S11) to seal the leaking porcine lung where air leakage and bleeding were stopped. To demonstrate the potential application on skin, we made a cut of 2.5-cm transverse incision on rat abdominal skin, where a large tension environment could be built (Fig. 6B, iii). UIHA can completely close the wound gap (Fig. 6B, iv, and movie S12). Wounds with UIHA (Fig. 6B, v) presented satisfying healing compared with suturing groups (Fig. 6B, vi). UIHA can also seal hard tissues, such as rat skull (Fig. 6C, i and ii, and movie S13). Micro-CT revealed that the skull cracks diminished during bone regeneration and the growing integration with the host over 30 days (Fig. 6D, iii). Moreover, UIHA was found to be noncytotoxic (fig. S11) and elicited minimal host inflammatory response at the interface of UIHA and the tissue (fig. S12). The UIHA in this study did not degrade during the 12 weeks in vivo follow-up, since the implants had no significant weight loss (fig. S12). Notwithstanding this, it was safe in the body, which is an important application consideration. Furthermore, there is no evidence that the UIHA impedes adjacent tissue functionality (figs. S11 and S12). Hydrophobic and biodegradable underwater in situ adhesives could also be another alternative future direction to the present study.

Fig. 6.

Ex vivo and in vivo adhesive properties of UIHA on the lung, skin, and skull.

(A) Ex vivo and in vivo tests on a rat lung leakage model sealed by UIHA-coated patches (n = 5). (i) H&E and (ii) SEM images of the interface between lung and adhesives. Cut sites (iii) before and (iv) after sealing. On day 7, images of the lung leakage sites (v) with UIHA. (B) In vivo test on a porcine lung leakage model sealed by a UIHA-coated patch (n = 3). UIHA patch (ii) and cut sites (i) before and (iii) after sealing. (C) Ex vivo and in vivo tests on a rat skin incision sealed by UIHA (n = 5). (i) H&E and (ii) SEM images of the adhesive interface on rat skin. Cut sites before (iii) and after sealing (iv). Images of the skin leakage sites (v) with UIHA and (vi) without UIHA sealing on day 5. (D) In vivo test on a skull injury sealed by UIHA (n = 5). Skull injury (i) before and (ii) after sealing. Micro-CT images of bone integration and reconstructive (iii) with UIHA sealing after 1 month. Photo credit: G. Guan and Y. Li, Army Medical University.

Together, the UIHA exhibited exceptional adhesion for in situ hemostasis and tissue repair of arteries, lungs, bone, and skin. In UIHA, hydrophobic elastomer chains interwoven with macromolecular organic viscous fluid created an in situ underwater tissue/organ sealing and wound closure capability (, ). The underlying adhesion mechanism shed light on the design strategies for tissue sealants, surgical glues, and even implantation of bioelectronics under extreme environments.

MATERIALS AND METHODS

Materials

Ecoflex 00-35 (E35A/E35B, AB components curable PDMS, platinum catalyzed) and Ecoflex 00-50 (E50A/E50B, AB components curable PDMS, platinum catalyzed) were obtained from Smooth-On Inc. Slygard 184 PDMS was from Dow Corning. Silicone fluids were ordered from Beijing Haibeisi Tech. VTMS was obtained from Sigma-Aldrich. Fresh porcine skin tissue was purchased from a local meat market and was stored in a −20°C freezer before use. α-Cyanoacrylate (Guangzhou Baiyun Medical Adhesive Company), fibrin glue (Guangzhou Beixiu Biotechnology), and GelMA [synthesized according to the previously published paper in our group ()] were also used.

Preparation of UIHA

All procedures were performed in an ice bath, and all materials were precooled on ice. In a typical preparation, 50 mg of VTMS and 950 mg of Ecoflex 00-35 B (E35B) were mixed in a 2-ml polypropylene (PP) centrifuge microtube to obtain a mixture containing 5 wt % silane, and the microtube cap was tightly sealed until used. A total of 750 mg of E35A, 500 mg of silicone 500k, and 30 mg of the above mixture containing 5 wt % silane and 720 mg of E35B were weighed into another 2-ml PP microtube mixed in sequence. The mixture was well mixed with a thin rod in an ice bath and then centrifuged for 15 s at 5000 revolutions per minute to remove the bubbles and obtain the UIHA containing 0.1 wt % silane (0.1 wt % is the weight ratio of silane to the sum of E35A and E35B). A freshly prepared adhesive mixture was used immediately for all experiments.

Rheological characterization

All rheological measurements were conducted on a TA Instruments rheometer (Discovery Hybrid HR-1) equipped with a stainless steel cone plate at 2° angle and a 20-mm diameter geometry or a flat plate of 8-mm diameter geometry. For the time sweeping tests, samples endured a constant shearing rate of 10 rad/s with a strain of 0.5% under various temperatures from 10° to 50°C. To measure the underwater gelation behavior, the plate was immersed in the water reservoir of ~5-mm depth, and the measurements were performed when the water temperature reached equilibrium with the rheometer set temperature. In the oscillation frequency sweeping tests, the frequency swept from 0.001 to 1000 rad/s at 25°C. In the viscosity measurements of the mixtures of reactive PDMS precursor/silicone fluids, only E35A was mixed with silicone fluids to prevent viscosity increase due to cross-linking reactions during measurements rather than E35A/B components, as E35A and E35B have similar viscosity and rheological profiles.

For the strain alternating experiments, all gel samples were performed under a constant shear rate of 10 rad/s, with the strain alternating between 0.1 and 400%. The period of every step is 200 s, and there were eight steps or four cycles in total. The strain sweeping experiments were also performed under similar conditions, and the strain ramped from 1 to 4000% under a constant shear rate of 10 rad/s. The underwater engagement/disengagement experiments were implemented on a rheometer with either a flat bottom plate or the upper plate is the steel stainless cone plate (2°) angle with a 20-mm diameter geometry (the plate was considered as a flat plate for calculation). The initial gap between the two plates was 3.2 mm, and 150 μl of PDMS/silicone liquid mixture was added onto the center of the bottom flat plate to ensure that the liquid could fill the entire gap when the gap distance is minimum. During experiments, the upper plate approached the bottom plate at a speed of 10 μm/s until the minimum gap reached 200 μm, and then the upper plate started to disengage.

The storage modulus (G′) of UIHA soaked underwater over time was measured through oscillation time sweep rheological tests. The UIHA gels were cured at 37°C for 30 min for complete cross-linking of E35A and E35B before the test. All samples (20-mm diameter and 300-μm gap distance) were tested at 25°C, with a constant strain of 0.5% and a shear rate of 10 rad/s. Then, all samples were soaked underwater and measured again under the same conditions after 1 and 2 days. The excess surface water was wiped off by a paper towel, and samples were further dried under vacuum for 1 hour before experiments to remove surface water completely.

1H-NMR characterization

All 1H-NMR characterizations were carried out on a Bruker Avance 300 spectrometer. Samples were dissolved in deuterated chloroform at a concentration of ~10 mg/ml. To evaluate the hydrolysis behavior of VTMS in a hydrophobic/hydrophilic immiscible binary phase system, 100 mg of VTMS was dissolved in 2 ml of CDCl3 in a 50-ml centrifuge tube, and then 40 ml of deionized (DI) water was added into the tube. The mixture was then left standing for layer separation. At different time points, the aliquots were collected from the CDCl3 layer, diluted to 10 mg/ml, and dried with anhydrous sodium sulfate powder before NMR characterization.

FTIR-ATR characterization

Fourier transform infrared (FTIR) characterizations were performed on a Thermo Scientific Nicolet iS-10 FTIR spectrometer equipped with an attenuated total reflection (ATR) accessory. The resolution was 4 cm−1, and the number of scans was four. The porcine skin tissue sample was cut into thin strips, and fats were removed by razor blades as much as possible. After repeatedly washing with water, the porcine skin strip was lyophilized to remove water completely.

SEM for gel morphology characterization

The morphology UIHA gels were visualized by SEM using an FEI Nova NanoSEM 450 with an operating voltage of 15 V. To remove the silicone fluid from the tested gels, samples were washed by sonicating in hexane in a bath sonicator for 3 hours per day for 5 days, with solvent exchanging twice a day. A notch was cut on the sample edge, and then the sample was torn apart. Samples were sputter coated by a thin layer of gold for SEM imaging.

Lap shear adhesion test

All the lap shear adhesion tests were performed using an Instron universal tester (Instron 5965) equipped with a loading cell of 1 kN. For all sample substrates, the surface was cleaned with ethanol and DI water before coating. For the test on glass substrates, thin cover glass slides (22 mm × 22 mm) were used as substrates, and ~25 μl of sample was coated on the area of 22 mm × 4 to 6 mm and cured for 15 min at 25°C. The ends of glass slides held by clamps were taped with paper to prevent slipping during the test, and two clamps should be aligned to avoid internal stress. For tests on PDMS substrates, Sylgard 184 was used to prepare the PDMS substrate in accordance with the supplier’s instructions. The Sylgard 184 PDMS were cut to strips of 50 mm × 10 mm, and ~25 μl of mixture was applied onto a region of 10 mm × 10 mm for each sample. For tests on porcine skin tissue, porcine skin was first thawed and then cut to stripes of 50 mm × 10 mm. The fat tissue and hair on porcine skin substrates were removed with a razor blade. The cleaned porcine skin tissue stripes were soaked in DI water and stored in a fridge at 4°C before use. To prepare samples, ~100 μl of mixture was coated on the area of 10 mm × 10 mm, and an external pressure of ~600 Pa was applied on each sample to prevent the porcine skin stripe bending. Samples were fully cured at 25°C for 15 min and then tested directly or soaked in water for certain periods before the test. For each sample group (n = 3), lap shear adhesion tests of cyanoacrylate, fibrin glue, and GelMA adhesives were conducted on porcine skin substrates under the same conditions. A total of 50 μl of each kind of adhesives was applied on 10 mm × 10 mm precleaned porcine skin tissue surface and cured under room temperature with a gentle pressure of ~600 Pa. For GelMA adhesive, the applied concentration of the aqueous solution is 25 wt %, which was cured with 2.5 μl of ammonium persulfide (500 mg/ml) aqueous solution and 2.5 μl of tetramethylethylenediamine (20 volume %) aqueous solution.

Measurements of interfacial fracture toughness (adhesion energy)

The adhesion energy of UIHA was measured by a 180° peeling test using an Instron universal tester equipped with a load cell of 1000 N. For the test on porcine skin tissue, two pieces of porcine skin strips (10 mm × 80 mm) were adhered together with UIHA and measured after preparation or soaked underwater for 1 or 2 days. For the test on a glass substrate, UIHA patches (30 mm × 120 mm) were first cured, and then a region of 25 mm × 50 mm was glued with UIHA adhesive, adhered to a glass slide (25 mm × 75 mm), and measured after curing or soaked underwater for 1 or 2 days. All tests were performed at a constant peeling rate of 50 mm/min. The plateau force is the average force in the plateau region of the force curve, and the adhesion energy is calculated as 2 × force (plateau)/width.

Underwater in situ sealing/isolation of electronics

The underwater in situ isolation model of electronics was built in-house. A circuit (3 V) with a broken isolation layer was soaked in salt water (1 M CaCl2 solution), which bridges the leaking area and another light-emitting diode bulb indicator to form another circuit. To isolate the leaking area underwater, UIHA was injected onto the broken isolation layer area to completely reseal the electronic circuit.

In situ sealing of water burst in balloons and tubes

A water balloon was prepared by filling nitrile latex with water. One hole was created on the balloon by puncture with a needle (20 gauge). A UIHA patch (15 mm × 15 mm) was prepared in advance and then coated with a thin layer of UIHA liquid. The UIHA patch was adhered to the hole and gently finger pressed for a few seconds to stop the water from leaking. The burst test system was built with an air compressor and air pressure controller connected with a PP tube (inner diameter: 3 mm; outer diameter: 4 mm). A ~5-psi pressure was applied to a red color water-filled PP tube, and the other end of the tube was sealed completely. A hole was punctured from one side of the PP tube by a 20-gauge needle for the water outlet. The system was set on a hotplate of 37°C to mimic the body temperature. A UIHA patch (~10 mm × ~5 mm) was prepared in advance and then coated with a thin layer of UIHA liquid. The patch was allowed to adhere across the hole perimeter of the PP tube and gently pressed manually for 2 min to seal the punctured tube.

Ex vivo burst pressure measurements

Ex vivo burst pressure of UIHA was obtained by following standard protocol for measuring surgical sealants. It was performed on a custom-made pressure chamber equipped with a digital manometer and a syringe pump. Porcine skin tissues were purchased from a local market. The adipose tissue was removed, and a 2-mm diameter punctured hole was created. Two-hundred microliters of UIHA solution was injected onto the defect through a syringe. Samples were fully cured at 37°C for 30 min and then tested directly or soaked in water for certain periods before testing. After gelation, the pressure was applied by pumping phosphate-buffered saline (PBS) via a syringe pump at a rate of 0.75 ml/min, and the pressure was recorded by the manometer.

SEM characterization of the tissue-adhesive interface

Samples with the surrounding tissue were fixed overnight with glutaraldehyde and lyophilized. The samples were then mounted onto an aluminum holder and sputter coated with gold. SEM images of the samples were obtained on an emission SEM (ZEISS crossbeam 340-47-76) at 10 to 20 kV.

Swelling ratios study

The swelling ratios of UIHA at different weight fractions were calculated by dividing the measured weights of the samples after incubation at 37°C in PBS by their corresponding dry weights at different times.

In vitro cytocompatibility of UIHA

The cytocompatibility of UIHA at different weight fractions was examined by using endothelial cells and a live/dead assay. In brief, endothelial cells were seeded and cultured on the surface of the UIHA for 24 hours at 37°C and 5% CO2. Cell viability test was performed with a live/dead viability/cytotoxicity kit for mammalian cells. An inverted fluorescence microscope (EVOS FL Auto, Life Technologies) was applied to image live (green stain) and dead (red stain) cells. ImageJ software was used to calculate the cell viability by dividing the number of live cells by the total number of cells. CCK-8 assay (Sangon Biotech) test was also carried out to quantify the cell viability in accordance with the instruction provided by the manufacturer.

Animal experiments

All animal experiments were carried out in accordance with the regulations of ethical approval for research involving animals and were approved by the Ethics Committee of the Third Military Medical University, China.

In vivo biocompatibility and biodegradation of UIHA

Subcutaneous implantation was carried out with male Sprague-Dawley rats (200 to 250 g). Rats were anesthetized with 1 to 1.5% isoflurane. Incisions of 1.5 cm were made, and separate subcutaneous pockets were created on the dorsum of the rat. The implanted materials were gelled, weighed, sterilized, and implanted into the dorsal subcutaneous pockets (n = 5, 30 to 60 mg). The skin incisions were closed by suturing. At days 14, 28, 56 and 84, the animals were euthanized by isoflurane (2.0 to 2.5%) inhalation, and the implants with the surrounding tissue were explanted for further histological analysis. For in vivo biodegradation evaluation, the surrounding tissue beside the implants was peeled off, and then the residues were weighed. The degradation rate was measured on the basis of the changes of weights before and after implantation, which was calculated with the following equation: (Wbefore − Wafter)/Wbefore in percentage.

Incision closure of rat carotid artery with UIHA

The carotid artery incision sealing capacity of UIHA was tested on rats. Rats (n = 8) were anesthetized as described before. Under sterile conditions, the neck was incised, and the carotid artery was exposed and blocked by two vascular clamps. An incision was made in the vessel by a 25-gauge needle. UIHA solution (10 μl) was applied to the incision. After UIHA gelation for 3 min, the vascular clamps were released, and the artery incision was closed by UIHA. After 3 days, micro-CTA and ultrasound with color Doppler (VisualSonics, Vevo 2100) were performed to evaluate blood flow.

Incision closure of femoral artery in large animal models

Closure of femoral artery incisions was tested on rabbits, canines, and pigs. Anesthesia of rabbits (n = 3), beagle canines (n = 3), and mini pigs (n = 5) were induced with pentobarbital sodium (1%) and then maintained by the inhalation of 2.0 to 2.5% isoflurane. Surgical preparations were performed as described previously. In brief, the skin was incised, and the femoral artery was exposed and controlled by two vascular clamps proximally and distally. A 2- to 3-mm incision was created with a blade scalpel. The UIHA solution (20 μl on rabbits and 200 to 300 μl on canines and pigs) was applied to the wound. After 5 to 10 min of UIHA curing, the vascular clamps were released, and no bleeding was detected. Eight weeks later, ultrasound with color Doppler (Esaote MyLab System, Esaote) was performed to evaluate blood flow.

Leakage sealing of rat and porcine lung with UIHA

The lung leakage sealing capacity of UIHA was tested on rats (n = 5) and mini pigs (n = 3). Anesthesia was performed as described above. Breathing was maintained by a ventilator. After a right lateral thoracotomy, an incision was generated on the lung with a 25-gauge needle. Air bubbles and blood flow were detected from the defect in an immersion test with warm PBS. A UIHA solution-coated UIHA patch was used to stop the bleeding and seal the pulmonary injury. After 5 to 7 min for UIHA curing, the sealing effect of UIHA on lung leakage was evaluated by submerging the defect in warm PBS.

Closure of rat skin incision with UIHA

Skin incision sealing capacity of UIHA was tested on rats (n = 5). After anesthesia, the abdomens of the rats were shaved and disinfected with ethanol. A 2.5-cm transverse incision on rat skins was generated, and 10 μl of UIHA solution was added to the edges of the incision. After 3 to 5 min for UIHA curing, the incisions were sealed effectively. In the control group, skin incisions underwent regular suturing closure (4-0 nonresorbable suture).

Rat skull injury sealing with UIHA

The skull injury sealing capacity of UIHA was tested on rats (n = 5). After anesthesia, the top of the skulls was exposed. Craniotomy was operated to generate a square incision of 5 mm × 5 mm, and 10 μl of UIHA solution was added to the defected area. In the control group, craniotomy was operated without UIHA treatment.

Micro–computed tomography angiography

For CTA analysis, 3 days after the operation on the carotid arteries, the rats were anesthetized as described before. A thoracotomy was performed to provide good exposure for intravascular contrast agent (iohexol injection, Yangzijiang Pharmaceutical Group) injection. The rats were euthanized by anesthetic dose, and a micro-CT scanner (Quantum FX, PerkinElmer) was used to evaluate the patency of the rat carotid arteries.

Histology and immunohistology

UIHA adhesive and surrounding tissue were used for histological analysis. The sections were fixed with 4% paraformaldehyde/PBS at 4°C overnight and then processed for H&E staining. Anti-CD68, anti-CD3 (Abcam), primary antibodies with Alexa Fluor 568 conjugated (Life Technologies), and secondary antibodies were applied to immunofluorescence staining. The sections were further stained by Hoechst 33342 (Invitrogen) for nuclei. The H&E-stained sections were imaged with a Leica microscope. The immunofluorescence-stained sections were imaged with a Zeiss confocal microscope.

Statistical analysis

For each experiment, at least three samples were tested, and data were presented as means ± SD (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001). One-way analysis of variance (ANOVA) test was performed, followed by Tukey’s test for statistical analysis (OriginPro 8.6).