Tough , self-healable and tissue-adhesive hydrogel with tunable multifunctionality

An ideal hydrogel for biomedical engineering should mimic the intrinsic properties of natural tissue, especially high toughness and self-healing ability, in order to withstand cyclic loading and repair skin and muscle damage. In addition, excellent cell affinity and tissue adhesiveness enable integration with the surrounding tissue after implantation. Inspired by the natural mussel adhesive mechanism, we designed a polydopamine–polyacrylamide (PDA–PAM) single network hydrogel by preventing the overoxidation of dopamine to maintain enough free catechol groups in the hydrogel. Therefore, the hydrogel possesses super stretchability, high toughness, stimuli-free self-healing ability, cell affinity and tissue adhesiveness. More remarkably, the current hydrogel can repeatedly be adhered on/stripped from a variety of surfaces for many cycles without loss of adhesion strength. Furthermore, the hydrogel can serve as an excellent platform to host various nano-building blocks, in which multiple functionalities are integrated to achieve versatile potential applications, such as magnetic and electrical therapies. NPG Asia Materials (2017) 9, e372; doi:10.1038/am.2017.33; published online 14 April 2017


Preparation of hydrogels.
The PDA-PAM hydrogels were synthesized by the following procedures. (1) dopamine (DA) molecules went through an alkali-induced pre-polymerization process to form PDA chains by dissolving DA powder in a beaker containing NaOH aqueous solution (pH=11), and then allowing the DA to self-polymerize for 20 min in air atmosphere under stirring. (2) Acrylamide (AM), ammonium persulphate (APS), and N, N-methylenebisacrylamide (BIS) were mixed with the solution of PDA chains in an ice bath under stirring. After 10 min of mixing, the ice bath and stirrer were removed, and the AM was polymerized to form PDA-PAM hydrogels. Hydrogels with different weight ratios of DA/AM were synthesized to investigate the effects of DA on the properties of the hydrogels. The contents of those hydrogels are listed in Table S1. Note that more APS was needed when DA existed, because DA also consumed APS. In addition, PDA-PAM hydrogel cannot form when the weight ratio of DA/AM was higher than 8 wt. ‰ ( Figure S1), because the reductive DA molecules affected the activity of the initiator (APS) and thus retard the polymerization of AM monomers. Figure S1. Suitable DA contents (DA/AM ≤ 8 wt.%) resulted in solid hydrogels. Higher DA contents (DA/AM=10 wt. ‰, and 12 wt.%) resulted in viscous solution. 3 The self-healable multi-functional hydrogels were also synthesized by the same polymerization process as that of PDA-PAM hydrogels. Before polymerization, the functional nanoparticles (Fe 3 O 4 NPs and carbon black NPs) were added into DA solution to allow DA polymerizing on NP surfaces. In line with the positive spectra, the negative spectra of PAM, PDA and PDA-PAM hydrogel ( Figure S2B) also confirmed the presence of both PAM and PDA in the hydrogel. In fact, the characteristic fragments of PAM and PDA were found in the PDA-PAM hydrogel. For example, the fragments that are characteristic of the cyclic nitrogen in PDA (C 3 N -, and C 5 N -) 6 were all present in the hydrogel. In addition, the relative intensity of CN -/CNOof the hydrogel was remarkably higher than that of PAM, which resulted from the addition of PDA.

Fourier transform infrared (FT-IR) analysis.
The FT-IR spectra of pure PAM, pure PDA and PDA-PAM hydrogel (5 wt.% DA/AM) were analyzed to search for possible crosslinks between PDA and PAM. The spectra were recorded between 4000 and 500 cm -1 using a FT-IR spectrometer (Nicolet 5700, Germany). Before measurement, the samples were washed three times in deionized water and ethanol to eliminate reactant residues from the samples, and dried in vacuum chamber at 40 °C for 2 days.
As shown in FT-IR spectra ( Figure S3), the PAM hydrogel exhibited bands between 3000 cm -1 and 3500 cm -1 , corresponding to a stretching vibration of N-H, and at 1650 cm -1 for C=O stretching. The bands at 1620 cm -1 (N-H deformation for primary amine), 1450 cm -1 (CH 2 in-plane scissoring), 1420 cm -1 (C-N stretching for primary amide), 1350 cm -1 (C-H deformation), and 1120 cm -1 (NH 2 in-plane rocking) were also detected. The PDA showed a band near 1500 cm -1 for aromatic rings. The band at 3400 cm -1 resulted from the overlapping of hydroxyls and water adsorbed in PDA polymer and amine groups of PDA.
In comparison with pure PAM and PDA, the spectrum of the PDA-PAM hydrogel showed a new peak at 1258 cm -1 , corresponding to the C-N stretching in phenyl amines. The presence of this band indicates the interaction between -NH 2 groups of PAM and catechol groups of PDA, as shown in the black square in Fig. 1b- (1). Furthermore, the intensity of C-N of aliphatic amine (1197 cm -1 ) decreased, which also suggested the reaction of C-NH 2 with catechol groups of PDA. However, we could not assign the peak at 7.99 ppm to specific proton because they might belong to either catechol group or amino groups in DA.
The spectrum of AM in DMSO ( Figure S4bii) showed the characteristic resonance signals of methylene protons at 5.71 and 6.23 ppm. By the method of heavy water exchange, we obtained active hydrogen at 7.1 and 7.54 ppm, which might belong to the proton chemical shift of NH 2 group in AM ( Figure S4b).
After AM was mixed with the DA to form AM-DA complexation ( Figure S4c), the broad band of catechol groups in DA (8.87 ppm) split into two sharp peaks (8.85 ppm and 8.88 ppm). In addition, the active hydrogen in DA shifted from 7.99 ppm to 7.8 ppm ( Figure S4d).
These results demonstrated that there were interactions between the oligomers of DA and AM ( Figure S4d).  PDA-PAM hydrogel.

SEM morphology of the hydrogels.
The structures of the bulk PAM hydrogel and PDA-PAM hydrogel (DA/AM 8 wt.‰) were examined using a scanning electron microscope (SEM, JSM 6390, JEOL, Japan). Before examination, the hydrogels were freeze-dried. Then the dried hydrogels were cut to expose their inner structure, and the cross-section was observed.
SEM images revealed porous structure of dried hydrogels ( Figure S6). The pure PAM hydrogel had smooth surfaces. The PDA-PAM hydrogel had microfibril structures. The microfibrils in the hydrogel may be caused by the complexion of PDA and PAM chains, which interweaved three-dimensional structures through π-π interaction and hydrogen bonds.
The microfibrils in the PDA-PAM hydrogel might play as a bridge between two broken pieces to facilitate the broken hydrogel to self-heal. The microfibrils assisted healing process is similar to self-recovery of human skin upon damage with the aid of blood vascularization [6] .

Tensile tests.
The tensile tests were performed on a universal test machine (Instron, 5567, USA). The loading rate was 100 mm/min. The specimens had the width of 25 mm and the thickness of 3 mm. The gauge length between the clamps was 5 mm. The nominal stress (σ) was calculated as σ= F/A where F is the tensile load and A is the cross-sectional area.
The extension ratio (λ) was defined as the deformed length (l) relative to the original length l 0 λ= l/ l 0 . Tensile-recover test.
The PDA-PAM hydrogel (8 wt.‰ DA/AM) recovered its original length after it was pulled to an extension ratio of 20 and stored for a period of time ( Figure S8).

Time-dependent recoverability: tensile-recover-tensile test.
A tensile-recover-tensile test was conducted to characterize the time-dependent recoverability of the hydrogel. First, the specimens were pulled to achieve an extension ratio of 6, and then unloaded. After stored at different time intervals (10 sec, 10 min, 20 min, 30 min and 1 day), the specimens were pulled to 6 times of its initial length again. The tensile stress during the tensile-recover-tensile cycles is shown in Figure S9.
The pure PAM hydrogel ruptured at the second tensile test, which suggested that it was intrinsically brittle and could not recover its damaged internal structure caused by tension ( Figure S9a).
The PDA-PAM hydrogel was slightly weaker if the second loading was applied immediately (10 sec). However, the hydrogel recovered somewhat if the second loading was applied after it was stored a period of time, and can fully recover after a sufficient storage duration (1 day).
These results indicate that the internal damage of the hydrogel can be healed after the addition of DA, although the recovering process took a period of time ( Figure S9b).

b Determination of the fracture energy (Gc)
Fracture energy was determined by the classical single edge notch test on the universal test machine (Instron 5567). The speed of the crosshead was 2 mm/sec. The specimen was fixed between clamps with the gauge length of l 0 = 5 mm ( Figure S10a). The thickness and the width of the specimen were 3 mm and 25 mm, respectively. During the testing, a pair of specimens was pulled: one specimen was unnotched and the other was notched ( Figure S10b).
The unnotched specimen was pulled to obtain the stress-strain curve, whereas the notched specimen was used to determine the critical extension ratio (λ c ), at which cracks expanded. As for the notched specimen, a notch with the length of 5 mm was made in the middle of the specimens ( Figure S10a). The fracture energy (Gc, J m -2 ) is calculated using the Equation (1) proposed by Greensmith for elastomers 7 .
where, a is the length of the crack; λc is the extension ratio at which cracks expand in the single edge notch tests. In the present study, the crack expands slowly and it is not easy to record the point at which notches turn into a running crack. Thus, the λc is determined when the force reaches maximum during the pulling of the notched specimen. W 0 is the strain energy density, which is calculated by integration of the stress versus engineering strain of un-notched samples, until λc (λc=ε c +1).
The method for Gc calculation is well established, and has been widely used to test fracture energy for elastomers 7 . According to the calculation, the fracture energy of pure PAM hydrogel is 579 J/m 2 , which is at the same level of that reported in previous study (250 J/m 2 ) 8 .
The consistence between our data and previous study proves that the calculation method and the data are reliable.

Rheological experiments.
The dynamic rheological tests of PDA-PAM hydrogels were characterized at room temperature using a Rheometric Scientific HAAKE (MARS, German) strain-controlled rheometer equipped with 20 mm parallel plates. The hydrogels were loaded into a 0.5 mm gap between the plates and allowed to relax until the normal force was zero. Water was applied to the outer edges of the gel to prevent drying. Strain amplitude sweeps (0.01-100%) were first performed to determine the linear viscoelasticity region. Dynamic frequency sweeps were performed at angular velocities ranging from 0.01 to 10 Hz at 1.0% strain amplitude (liner region). All rheological measurements were performed in triplicate.
The storage modulus (G') and the loss tangent (ratio of loss modulus G'' to storage modulus G') of the hydrogel were determined as the frequency changes. The loss tangent indicated relative degree of the viscosity to elasticity of the hydrogels ( Figure S11). The results showed that the G' of PDA-PAM hydrogel was lower than that of pure PAM hydrogel, revealing that the DA addition led to the decrease of elasticity of the hydrogel. The loss tangent of PDA-PAM hydrogel was higher than that of pure PAM hydrogel, which showed that DA addition resulted in the increase of the viscosity of the hydrogel. Although DA addition might limit the ability of these hydrogels to bear loads, the high viscosity also allowed the PDA-PAM hydrogels to self-healing during torn 9 .

Effects of crosslinker (BIS) contents on the properties of PDA-PAM hydrogels.
The maximum extension ratio (ER) of PDA-PAM hydrogels decreased with the BIS content, which indicated that the hydrogels become brittle when the BIS content increased (Table S2), while the tensile strength of both PAM and PDA-PAM hydrogels increased with the BIS content (Table S3). Self-healing ability tests.

Tensile-heal-tensile test.
The self-healing ability of hydrogel was quantitatively evaluated tensile-heal-tensile test (Fig.   4b, c). The hydrogel was stretched to break on the universal test machine (Instron 5567) and the crosshead was moved back to make the two broken pieces into contact; the two pieces of hydrogels were self-healed after 2 hours; the healed specimen was stretched to break again.

(3) Dynamic rheological tests
The self-healing ability was also confirmed by dynamic rheological tests using the two-step procedure of Krogsgaard et al 10 . Firstly, the hydrogels were broken by applying a 100% strain through a strain amplitude sweep, which caused their G'-values to diminish drastically ( Figure S12). The healing of the hydrogels was then monitored over time by continuing to oscillate the hydrogels at a low strain (0.01%) amplitude and low frequency (f =1 Hz) that allowed the hydrogels to recover. The PDA-PAM hydrogel nearly recovered its initial storage modulus G'-value (2707 Pa) within 30 sec after stopping large strain ( Figure S12a), while the pure PAM hydrogel did not recover its initial G'-value ( Figure S12b). The recovered storage modulus of PDA-PAM hydrogel was close to the initial value (2869 Pa), thereby confirming that the reversible nature of the non-covalent bonds between PDA chains could re-form after breakage and therefore impart the self-healing properties to the PDA-PAM hydrogel. The high reactive catechol groups in PDA also contributed to the self-healing property.  The results indicated that the increase of BIS content severely affected the adhesion strength of the PDA-PAM hydrogels ( Figure S13), which is because the increase of crosslinking density of the hydrogel restricts the mobility of polymer chains [8] . Consequently, the viscosity of the hydrogel decreased and the polymer chain was not easy to diffuse into the tissue surface and form intimate contact with tissue surfaces. Thus, increasing BIS content resulted in the decrease of adhesiveness of the hydrogel.

Gas chromatography for quantification the residual amount of acrylamide monomers
Before cell culture and in vivo implantation, the PDA-PAM hydrogel was purified and the residual amount of acrylamide monomers in the hydrogel was quantified by gas chromatography (GC) 11 .

Experiments:
GC condition: An A90 gas chromatograph (Echrom Technologies Co., shanghai, China) Hydrogen and air were used as detector gases at 30 and 300 ml/min, respectively.
Analytical procedure: Before measurement, the PDA-PAM hydrogels were firstly purified by soaking the hydrogel in the deionized (DI) water for one day and the DI water was refreshed every day. The purification process was repeated three times. Then the hydrogel was freeze-dried and ground to powder. Finally, 0.1 g of accurately weighted powder of PDA-PAM hydrogels was added into a mixture solution (10 ml, DI water: methanol = 2: 8) under vigorously vortex for 5 hours to extract residual AM monomers.
Working solution of AM was prepared by diluting the stock solution with the concentration of 10 mg/ml to a series of graded concentrations, including 1 mg/ml, 0.4 mg/ml, 0.2 mg/ml, 0.1 mg/ml, and 0.04 mg/ml, which was used to determine the standard curves of AM.
Results: Figure S14 shows the standard curve of AM, which has good linearity over the range of 0.04 -1 mg L −1 , the calibration equation is y = 287.64 x -5.479 with R 2 of 0.9987.
The method detection limit (MDL) was 0.0023 mg/ml. Figure S15 shows the FID chromatogram of the AM extracted from purified PDA-PAM hydrogel. A quite weak peak was observed at 2.733 minutes. The concentration of AM was below than MDL and could not be calculated, which demonstrated nearly no residual AM monomers in PDA-PAM hydrogel after three times of purification.  The optical density (OD) of DMSO extracts was read at 570 nm using an ELISA reader (MQX200, BioTEK, USA). The behavior of cells cultured on epidermal growth factor (EGF)-loaded PDA-PAM hydrogels were also investigated to confirm the in vitro bioactivity of the released EGF from the hydrogels. For obtaining the EGF-loaded hydrogel, the hydrogel (30 mg) was first sterilized in 75% alcohol, and then EGF (Shanghai Primegene Bio-Tech Co., Ltd., China, 10 μg/sample) were adsorbed into the hydrogel. Characterizing the affinity of extracellular matrix proteins to hydrogels.
The affinity of extracellular matrix proteins to hydrogels was experimentally determined by evaluating the adsorption and release of fibronectin (Fn). Fn was selected as the representative protein because it widely exists in ECM and promotes cell adhesion 12,13 . To quantify the adsorption of ECM proteins, PDA-PAM and PAM hydrogels were soaked in 1 ml of Fn in PBS (10 µg/ml) 14   The results demonstrated that Fn was fully adsorbed by the PDA-PAM hydrogel. The release profiles also demonstrated that no burst release was detected in both two cases, and the release of Fn from PDA-PAM hydrogel was slower than that from pure PAM hydrogel and, 3 2 as shown in Figure S17. Both the adsorption and release results demonstrated that the PDA-PAM hydrogel has higher protein affinity than PAM hydrogel, which is because PDA endows the hydrogels with abundant binding sites for immobilization of ECM proteins.

Epidermal growth factor (EGF) release experiment.
EGF  EGF was sustained released during whole period, and no burst release was observed in the initial stage ( Figure S18). The release of Fn from PDA-PAM hydrogel was slower than that from pure PAM hydrogel. These results indicated that EGF was stably immobilized in the hydrogel. The stable immobilization of growth factor was attributed to PDA in the hydrogel.
PDA enabled the chemical conjugation of biomolecules containing primary amine or thiol groups via imine formation or Michael addition reaction as demonstrated by previous studies, and physically immobilized growth factors through non-covalent interaction [12][13][14] . Yoshihiro et al. [15] reported that PDA treated titanium and stainless steel surfaces effectively immobilized the epidermal growth factor and promoted cell growth efficiently. Shin et al. [16] also demonstrated that secondary ligation of VEGF via deposited polydopamine layer on the  Intuitive observation indicated that the PDA-PAM hydrogel accelerated the wound healing.
After 15-day treatment, the regenerated skin tissue covered by the PDA-PAM hydrogel was comparable to that covered by the EGF-loaded hydrogel (Fig. 6e). The new skin tissue on both the PDA-PAM hydrogel and the EGF-loaded hydrogel covered the whole wound.
The size of the wound on each photograph was measured. The wound closure was quantitatively evaluated by the ratio of the area of the 15-day and 5-day to that of 0-day, which is shown in (Fig. 6d). The results demonstrated that the size of the wound tended to decline during the 15-day implantation. Both PDA-PAM hydrogels with and without EGF had a wound healing ratio of ~80%, higher than that of blank group (60 %).

Histological examinations
Histological examination further indicated that the PDA-PAM hydrogels accelerated skin tissue regeneration compared with that of blank groups after 15 day-treatment, as shown by the H&E and Masson's trichrome staining images in Fig. 6f. The wound treated with PDA-PAM hydrogels were covered by a complete and thick epidermis (Fig. 6f-1). No inflammatory reaction appeared at the interface between new skin tissue and hydrogels), while the wound area generates a few microvessels (Fig. 6f-2). Masson's trichrome staining images showed that the PDA-PAM hydrogels treated wound areas showed adequate collagen deposition ( Fig. 6f-3). More organized aligned collagen fibers, and more mature fibers were found in the PDA-PAM hydrogel than those in the blank groups ( Fig. 6f-9).
When the hydrogel loaded with EGF, more mature tissue grew into the wound area as indicated by the mature glands grown from the glandular cavity ( Fig. 6f-5). Masson's trichrome staining showed that the collagen fibers grown in EGF-hydrogel treated wound were compact and aligned in order and parallel to the epidermis (Fig. 6f-6).
For blank group, a small amount of inflammatory reaction appeared, and the collagen fibers in the wound area were randomly arranged ( Fig. 6f-7 ,8,9).
In summary, the EGF-free PDA-PAM hydrogel facilitated a quicker wound healing process than that of the blank groups. The good wound healing ability of the PDA-PAM hydrogel was attributed to four reasons. Firstly, the high tissue adhesiveness of PDA-PAM hydrogel guaranteed the hydrogel intact fixation with surrounding tissue and protects the wound sites from infection during the entire healing period. Secondly, the excellent cell affinity of the hydrogel facilitied cells adhesion, attachment and migration. Thirdly, the hydrogel had good affinity to ECM proteins ( Figure S14), and the protein adsorption could further improve cell attachment, which established positive feedback loop and finally created suitable ECM microenvironments for cell adhesion 14 . Finally, the hydrogel stably immobilizes epidermal growth factor (EGF) that enhanced the migration of fibroblasts to the wound area and promoted complete skin regeneration for superior treatment of dermal wounds ( Figure S15).

Self-healable, conductive hydrogels.
The PDA-PAM hydrogels could be tuned to conductive hydrogels by incorporating carbon black (CB) nanoparticles in the hydrogels ( Figure S20a).
The conductive hydrogel was synthesized by the same polymerization process as that of PDA-PAM hydrogels. CB NPs were added into DA solution to allow DA to polymerize on nanoparticles surfaces. Thus, the PDA functionalized CB NPs can well disperse in the hydrogels (Fig. 7b-i). The CB NPs incorporated hydrogels with various CB NPs content (5 wt. % -50 wt.% to AM) were finally obtained by the same polymerization process as that of PDA-PAM hydrogels (Table S4). CB NPs-PAM hydrogels without PDA grafting were also prepared as controls. The conductivity of hydrogels sample was measured by a two-probe method using a potential state (CHI 660, USA). The effect of water content on the conductivity was also studied. Three kinds of CB NPs-PDA-PAM hydrogels containing 70 wt.%, 75 wt.%, and 80 wt.% of water were prepared, as listed in Table S5. The CB NPs content in the hydrogels were kept at 20 wt.%.

Results:
The hydrogel incorporated with CB NPs displayed homogeneous morphology at the microscopic level using SEM (JSM 6390), which suggested that the CB NPs were well dispersed into the polymeric matrix and covered with PDA microfibrils ( Figure S20b). The conductivity of PDA-PAM hydrogel was 10 -7 S/cm, which was nearly insulated. After CB NPs were incorporated, the conductivity of the hydrogels significantly increased ( Figure   S20c). The inset of Figure S20c showed that increasing CB NPs contents led to the increase of the conductivity of CB NPs-PDA-PAM hydrogels. The conductivity of the hydrogels increased with water content ( Figure S20d).
To investigate self-healing and stretchable ability, a battery-powered circuit with an LED was used ( Fig. 7b-ii). (1) The CB NPs-PDA-PAM hydrogel with 20 wt.% of CB NPs was worked as a conductor to connect circuit, and the LED was bright; (2-3) the hydrogel was stretched, and the light of LED gradually became weak during the process; (4) the hydrogel was stretched until to break and the LED was off; (5) the fractured two parts of hydrogel were put into contact to heal for 2 hours and the LED was bright again; (6) the healed hydrogel was stretched again. These results showed that the conductive hydrogels still had self-healing and stretchability. The resistance of CB NPs-PDA-PAM hydrogel with 20 wt.% of CB NPs during stretching was measured. The hydrogel was embedded into two parallel titanium 4 1 electrodes and was connected into electrical loop. The electrical resistance was recorded when 2, 4, and 7 of extension ratio was applied. The relative electrical resistance change (ΔR/R 0 ) was calculated by following Equation S2 .
Where R 0 and R denote the resistance without and with applied strain, respectively.
As shown in Figure S20e, The resistance of the hydrogel increased during tension. When the extension ratio was 2, the electrical resistance of the hydrogel increased 18%. When the extension ratio was 7, the electrical resistance of the hydrogel increased 86%.

Self-healable, magnetic hydrogels.
The PDA-PAM hydrogels can be transformed to the magnetic hydrogels by incorporating NPs incorporated hydrogel can be attracted by a magnet, and tightly adhere on the magnet, and then can be stretched 7 times of its initial length ( Figure S21d). These results demonstrate that the hydrogel with Fe 3 O 4 NPs incorporation still has adhesiveness, magnetism and stretchability. In order to prove the self-healing ability of the magnetic hydrogels, the hydrogel blocks with (black) and without (brown) Fe 3 O 4 NPs were put into contact and self-healed to form a hydrogel rope. Then the hydrogel rope adhered on the nails and was bended by two magnets. These results showed that the Fe 3 O 4 NPs incorporated hydrogels had self-healing and magnetic properties.  The NPs distributed heterogeneously in the PAM hydrogel without PDA.