One-pot synthesis of highly mechanical and redox-degradable polyurethane hydrogels based on tetra-PEG and disulfide/thiol chemistry

Abstract

Highly mechanical hydrogels with stimuli degradability are promising scaffold materials for tissue engineering, due to their unique advantage of retaining mechanical strength in use, while being able to be readily removed by external stimuli after use. However, it has always been a big challenge to integrate both good mechanical properties and stimuli degradability into one single hydrogel. In this work, a series of tetra-PEG polyurethane hydrogels with tunable redox-degradability and a high compressive fracture strength has been synthesized by a one-pot method. The good mechanical properties are owed to an extremely uniform network of tetra-PEG, and the redox-degradability is realized using cystamine, which contains a highly DTT-sensitive disulfide bond. The mechanical strength of the as-prepared hydrogels reaches a megapascal range, and their complete degradation time can be flexibly adjusted from 4 to 22 days by controlling the proportion of cystamine. With the above properties, these hydrogels are believed to have potential bio-applications.

---

Introduction

Hydrogels are appealing for tissue engineering.¹ Stimuli-degradable hydrogels can be flexibly removed by an external stimulus after completing functions, and have a wide range of biological applications, including scaffolds for tissue engineering and regenerative medicine, barriers or adhesives between tissue and material surfaces, fillers for cavities caused by surgery, and reservoirs for drugs.^2–6 Broadly speaking, the most common stimuli-degradable hydrogels are those composed of hydrolyzable segments such as poly(lactic acid), poly(glycolic acid), and poly(ε-caprolactone) with water as the external stimulus. However, the slow but constant hydrolyzation process reduces the mechanical performance with time, affecting the use and effect of the materials. Therefore, many mechanically stable stimuli-degradable hydrogels have been designed using enzymatic hydrolysis,^7,11 photolytic cleavage,⁸ disulfide/thiol chemistry,⁹ and/or a combination of these methods.¹⁰ Among these, the disulfide/thiol chemistry method has the advantage that the disulfide bond is specifically cleavable under a mild external stimulus such as thermal, pH, light, dithiothreitol (DTT) and/or glutathione (GSH).⁹ Moriyama and coworkers developed a thiolated poly(ethylene glycol) (PEG) hydrogel that exhibits quick degradation after soaking in DTT solution.¹² Yang and coworkers prepared a GSH triggered degradable disulfide-containing PEG-based injectable hydrogel.¹³ Fairbanks and coworkers designed a hydrogel using thiol-functionalized 4-armed PEG macromolecules, which is degradable under the irradiation of low intensity UV light.¹⁴ Li and coworkers reported a polyurethane hydrogel crosslinked by azobenzene/α-cyclodextrin interaction and disulfide bonds, which underwent degradation under both light and reductive external stimuli.¹⁵ Though other similar hydrogel systems are developed,^16–19 the major drawback of these hydrogels is lack of mechanical strength, let alone the tedious and time-consuming multi-step polymerization process.

4-Armed PEG polyurethane (PU) hydrogel is one of the hydrogels with compressive strength reaching a megapascal range.^20–30 This is because the tetra-PEG hydrogels have an extremely uniform network structure that can behave cooperatively, thus the mechanical strength increases greatly.³¹ There are two advantages to introducing tetra-PEG into a PU system. First, the homogeneous network of the tetra-PEG hydrogel can be well preserved in the PU system, because the soft segments (tetra-PEG) and hard segments must connect alternately with each other, zero self-reaction helps to avoid micro-loops and micro-defects that weaken the gels. Second, tetra-PEG can react with hard segments without any further terminal modification, due to the high reactivity and efficiency between the isocyanate (–NCO) and hydroxyl (–OH) groups, greatly simplifying the difficult synthesis process.

Hence, in this work, we proposed an one-pot method to synthesize robust and redox-degradable PU hydrogels. Tetra-PEG, which endows the hydrogel with good biocompatibility, was used as a soft segment; and hexamethylene diisocyanate (HDI) end-capped cystamine (Cys), which provides redox-degradability due to its disulfide bond, was used as a hard segment. The facile preparation process of the PU hydrogels is briefly described in Scheme 1. First of all, Cys, the chain extender, was end-capped by HDI to form the hard segment HDI–Cys–HDI (yellow segment in Scheme 1). Then tetra-PEG was added to the mixture as the soft segment and crosslinker to achieve the final PU hydrogels, termed PEG–(S–S)_X–HDI, where X = 0, 1, 2, and 3, referring to the number of S–S bonds between every two PEG segments in the network, so the terms of the series of PU hydrogels are further abbreviated to X0, X1, X2, and X3, respectively.

Scheme 1 Preparation process of PEG–(S–S)_X–HDI hydrogels, and the corresponding hydrogel photograph with X from 1 to 3, abbreviated to X1 (A), X2 (B), and X3 (C).

Results and discussion

Fourier Transform Infrared (FTIR) and Raman analysis demonstrate the successful preparation of the PEG–(S–S)_X–HDI hydrogels. From the FTIR spectra of the X1 hydrogel (Fig. 1A), the disappearance of the –NCO peak (2276 cm⁻¹), and the emergence of –C=O peaks (1715 and 1639 cm⁻¹) and the O–C–O peak (1096 cm⁻¹) indicate that the HDI has completely reacted with Cys and tetra-PEG in the PU hydrogel. Since a disulfide bond (S–S) is a symmetrical structure, corresponding IR absorption is hardly observed. The existence of the disulfide bond was further proved by the Raman spectra at 537 and 580 cm⁻¹ (Fig. 1B), meaning Cys (the chain extender) has been successfully incorporated into the PU hydrogels. Moreover, by normalizing the absorption intensity of Raman spectra at 844 cm⁻¹ (C–O stretch of tetra-PEG) (Fig. 1C), the content ratio of S–S bonds in X1, X2, and X3 was confirmed as approximately 1 : 2 : 3. Afterwards, the exact content of S in each hydrogel was detected by an Inductive Coupled Plasma Emission Spectrometer (ICP), see Table 1. The ideal content of S refers to the calculated value of corresponding hydrogel with a perfect structure. Conversion amounts were obtained by dividing the content of S with the ideal content. It’s found that, firstly, the corresponding S content appears to be 0.59%, 1.13%, and 1.76% in X1, X2, and X3, indicating the amount of S–S bonds in the hydrogels can be quantitatively controlled. Secondly, the S conversion of X1, X2, and X3 reaches 92.2%, 88.2%, and 91.7%, respectively, meaning that Cys is used in a high efficiency.

Fig. 1 (A) FTIR spectra of HDI, PEG, and freeze-dried X0 and X1 hydrogels. (B) Representative Raman spectrum of the X1 hydrogel, since the spectra of X2 and X3 are similar to X1. (C) Local enlarged Raman spectra of the X1, X2, X3 hydrogels from 450 cm⁻¹ to 950 cm⁻¹. The two bands in the inset image in the left top refer to S–S stretching.

Table 1 Content of S in each PEG–(S–S)_X–HDI hydrogel, X from 0 to 3, tested by ICP. Conversion amounts were calculated by dividing the content of S with the ideal content

	X = 0	X = 1	X = 2	X = 3
Content of S/%	0	0.59	1.13	1.76
Ideal content of S/%	0	0.64	1.28	1.92
Conversion amounts/%	—	92.2	88.2	91.7

---

Random crosslinking points are one of the main hindrances to the strength of hydrogels.³² However, PEG–(S–S)_X–HDI hydrogels, whose frameworks consist of homogeneous 4-arm PEG, have subtly avoided this problem. Before mechanical property examination, all the samples were immersed in deionized water to reach a fully swollen state. As X increases from 1 to 3, the equilibrium water content (EWC) increases from 93.93%, to 94.85%, to 95.01%. Fig. 2A and B show the compressive and tensile stress–strain curves of the PEG–(S–S)_X–HDI hydrogels, and one can see that the hydrogels present a high compression fracture strength of 0.6–5.9 MPa, and 74–88% deformation at break and high tensile fracture strength of 35–53 kPa, and 431–511% deformation at break. Rheology studies show that the PEG–(S–S)_X–HDI hydrogels are of high elasticity, as their storage modulus is two orders higher than their loss modulus (Fig. 2C). Fatigue tests reveal that the PU hydrogels have good reliability. Take the X3 hydrogel as an example (Fig. 2D); it can perfectly recover its mechanical properties after 50-cycle testing under the conditions of 0–83% deformation, and a 5 mm min⁻¹ compression rate. We note that the content of Cys in the hard segments plays a significant role in the mechanical strength of the PEG–(S–S)_X–HDI hydrogels, which is increased with higher Cys concentration. An explanation is that a higher Cys component increases the association energy of the hydrophobic interaction, because water is a good solvent for the PEG segments but poor for the hard segments, and the large amount of extra strong hydrophobic interactions contribute a lot to the total mechanical strength. We suppose that it is the aggregation of the hard segments through hydrophobic interactions that makes the hydrogel opaque (Scheme 1A–C).

Fig. 2 (A) Compression stress–strain curves of the X1 (black), X2 (red), and X3 (blue) hydrogels. (B) Tensile stress–strain curves of the X1, X2, and X3 hydrogels. (C) Rheology analysis of the X1, X2, and X3 hydrogels. (D) Fatigue tests of the X3 hydrogel as a representative.

To investigate the degradability of the PEG–(S–S)_X–HDI hydrogels, we treated them with 10 mM DTT. First of all, quantitative degradation analysis of the X1, X2, and X3 hydrogels with (Fig. 3A) and without (Fig. 3B, controlled group) DTT treatment was carried out by measuring the weight loss of the hydrogels. It was found that the weights of hydrogels treated with DTT decreased obviously over time, owing to the cleavage of S–S bonds in the hard segments, but those of the controlled samples remain unchanged, showing that the hydrogels have a rapid redox-degradation property. Meanwhile increasing the Cys component in the hard segments can effectively accelerate the degradation rate, for instance X1 (22 d) < X2 (9 d) < X3 (4 d). Therefore, it is possible to tune the degradation rate by controlling the proportion of Cys in the hydrogel. Once redox-degradation starts, the hydrogels became soft (inset images in Fig. 3C–H) and lost their mechanical strength. But the parallel samples without DTT treatment retained a high mechanical strength. This property ensures that the hydrogels can retain a constant mechanical strength when in use, and be easily removed after use.

Fig. 3 Degradation process of the PEG–(S–S)_X–HDI hydrogels. Weight loss curves of the X1, X2, and X3 hydrogels (A) with and (B) without 10 mM DTT treatment. The network morphology evolution by SEM of the redox-degraded X1 hydrogel at (C) 0 d, (D) 1 d, (E) 6 d, (F) 10 d, (G) 14 d, (H) 19 d. The insets are photographs of the corresponding hydrogel.

According to the morphology evolution of the redox-degradation process of the X1 hydrogel (Fig. 3C–H), the degradation pattern of the PU hydrogels is that of bulk erosion.³³ This is because the high hydrophilicity of PEG allows the rapid permeability of DTT solution into the hydrogel network, making the network structure gradually become larger and sparser, and eventually collapse. Finally, we evaluated the anti-adhesion property of the hydrogels, as this is important to prevent post-surgical adhesion.³⁴ L929 cells were cultured on X1, X2, and X3 and tissue culture polystyrene (TCPS) for 24 h, and the surfaces of the substrates were observed by a microscope (Fig. 4A–D). Compared to TCPS, the cells adhered on the substrates of X1, X2, and X3 were quite few, which illustrates that the series of PU hydrogels exhibits outstanding anti-cell adhesive properties. It’s well known that the super-high hydrophilicity of tetra-PEG segments helps to form a hydration layer on the surface of the PU hydrogels, which thereby restrains cells from attaching to the surface.³⁵ However, there is a trend that with the increase of Cys content, the number of cells adhered on the surface augments, which is mainly contributed to by the increase of hydrophobic segments.

Fig. 4 (A–D) Morphology of L929 cells adhered on the TCPS, X1, X2, and X3 hydrogels after culturing for 24 h. Red arrows point out L929 cells adhered on the substrate. (E) Cell viability after being incubated in X1, X2, and X3 hydrogel solutions for 24 h.

Cytotoxicity analysis was also carried out, and the cell viabilities of X1, X2, and X3 all maintained a high level (>80%) after incubation together with the corresponding hydrogels for 48 hours (Fig. 4E). This experiment proves that the PU hydrogels have negligible cytotoxicity, because the hydrogel networks are composed of biocompatible tetra-PEG. This property may further make it possible to use the hydrogels as tissue filler or structural biomaterials.^36,37

Conclusions

We used a one-pot method to successfully synthesize a series of tetra-PEG polyurethane hydrogels, which present tunable redox-biodegradability from 4 to 22 days in 10 mM DTT solution by controlling the Cys content in the hard segments, high compressive strength reaching 5.9 MPa, anti-cell-adhesion properties and low cytotoxicity. Because the PU hydrogels can preserve high mechanical strength in use, while being able to be completely removed after use by an external reducing agent such as DTT, they are promising for various bio-applications, for instance, as structural biomaterial, tissue filler, barriers between material and tissue, etc.

Experimental

Materials

4-Arm polyethylene glycol (4-arm PEG, M_n = 20 000, hydrogel value = 11.2 mg KOH per g) was obtained from Xiamen Sinopeg Biotech Co. Ltd, Xiamen, China, and was dried at 85 °C under vacuum for 2 hours. Hexamethylene diisocyanate (HDI) and cystamine dihydrochloride (Cys·2HCl) were purchased from J&K. Catalyst, stannous octanoate and dichloromethane (CH₂Cl₂) were obtained from Adamas Reagent, Ltd. The CH₂Cl₂ was further dried by CaH₂. Triethylamine (TEA) and ethyl alcohol absolute were purchased from Shanghai Chemical Reagent Corporation. For the biological experiments, all glassware was dried overnight under vacuum at 110 °C before use. Bovine serum albumin (66 kDa, >98% purity), fetal bovine serum (FBS), and Dulbecco’s modified Eagle medium (DMEM) were purchased from Gibco BRL. HeLa cells were obtained from the Cell Resource Center of Shanghai Biological Sciences Institutes. Water used in the experiments was purified to a resistivity higher than 18.2 MΩ cm using a Hitech system.

Preparation of the PEG–(S–S)_X–HDI hydrogels

A solution of HDI in CH₂Cl₂ (0.1 g mL⁻¹), TEA and Cys·2HCl was added into CH₂Cl₂ (2 mL). The amount of Cys·2HCl was determined by the X value. When X = 1, HDI was twice more than Cys·2HCl in mol, when X = 2, 1.5 times, when X = 3, 1.33 times. Moreover, the mol of Cys·2HCl was 2X times more than that of 4-arm PEG. The mixture was then stirred in an ultrasonic bath (50 W, 40 kHz) at room temperature for 15 min. Then 4-arm PEG (0.2 g, 0.00001 mol) was added into the mixture and after PEG was completely dissolved, a solution of stannous octanoate in CH₂Cl₂ (0.1 g mL⁻¹, 10 L) was added. The resulting mixture was stirred in the ultrasonic bath again for 15 min, and then kept at 50 °C for 24 hours for gelation. The resultant gels were successively dipped in ethyl alcohol absolute and deionized water, which were refreshed every day, for 3 days to remove the residual reagent completely. Finally, the PEG–(S–S)_X–HDI hydrogels had been successfully prepared.

Gel characterization

Raman spectra of the freeze-dried PEG–(S–S)_X–HDI gels were recorded on a Paragon 1000 (Perkin Elmer) spectrometer, with the excited wavelength at 780 nm. FTIR spectra of the freeze-dried PEG–(S–S)_X–HDI gels, X from 1 to 3, were recorded on a Paragon 1000 (Perkin Elmer) spectrometer. The content of S in each hydrogel was examined by ICP, iCAP 6000 Radial, Thermo. Scanning electron microscopy (SEM) images of the PEG–(S–S)_X–HDI hydrogels were obtained by a Philips Sirion 200 instrument, the hydrogel samples were dried under liquid nitrogen, and the fracture surface was chosen for observation by SEM. The EWC of the X1, X2, and X3 hydrogels was calculated through the following equation:

EWC = (W_wet − W_dry)/W_wet × 100%

here W_wet refers to the weight of the fully swelled sample, W_dry refers to the weight of the fully dried sample.

Rheological measurements

Rheological characterization of the PEG–(S–S)_X–HDI hydrogels was determined by rheological measurements. The storage and loss moduli of all PEG–(S–S)_X–HDI hydrogel samples were measured over the frequency range 0.1 < w < 100 rad s⁻¹ with a strain amplitude of 2% at 25 °C by a rotational rheometer (Bohlom Instruments, Advanced Rheometer) with plate/plate arrangement (25 mm in diameter and 1–1.5 mm in the gap width), within the linear viscoelastic range, in which G′ and G′′ are independent of strain.

Mechanical test of the hydrogels

Compression and tensile tests were operated on a universal test machine (Instron 4465) at room temperature. The compression test samples were cylindrical, with a 10 mm diameter and 5–8 mm height. The compression speed was 2 mm min⁻¹. For tensile tests, the corresponding samples were 7–8 mm in diameter and about 25 mm in effective length. The elongation speed was 100 mm min⁻¹. For recovery tests, the compression loading rate was 5 mm min⁻¹.

Degradation of the PEG–(S–S)_X–HDI hydrogels

PEG–(S–S)_X–HDI hydrogels samples weighing from 1.0–1.5 g were immersed in a solution of DTT (10 mM) in PBS (pH = 7.4, 0.01 M) at 37 °C. The weight of samples was measured periodically, and the solution of DTT was refreshed every three days.

Biocompatibility of the PEG–(S–S)_X–HDI hydrogels

Cell adhesion study. The three kinds of hydrogels, named PEG–(S–S)_X–HDI (X = 1, 2, and 3), were cut into 8 mm × 8 mm squares, and then rinsed in PBS solution, refreshed daily, for 3 days. The samples were sterilized before being added to a 24-well plate with cultured L929 cells in a density of 1 × 10⁵ cell per well. The culture medium of L929 cells consisted of DMEM with 1 wt% penicillin/streptomycin solution, 10 wt% FBS and 5% CO₂ at 37 °C. After the cells were cultured for 24 hours, all the hydrogel samples were rinsed twice with PBS to remove unattached cells and observed under a Nikon-C1 laser scanning confocal microscope.

Cytotoxicity. The toxicity analysis of the PEG–(S–S)_X–HDI (X = 1, 2, and 3) hydrogels was tested by MTT assay. 1 × 10⁵ 3T3/balb cells per well were seeded in a 96 well plate. Cells were incubated at 37 °C, 5% CO₂ atmosphere for 24 h. Afterwards, the PEG–(S–S)_X–HDI (X = 1, 2, and 3) hydrogels were respectively added in triplicate wells for each sample before being incubated at 37 °C, in 5% CO₂ atmosphere for 24 h. After the incubation, cells were washed and replenished with a fresh culture medium, and were further incubated for 2 h. Next, MTT was added to the cells at a final concentration of 0.5 mg mL⁻¹ and incubated for another 4 h. In order to dissolve the resulting formazan crystals, 100 μL DMF solubilization solution was added to each well. Absorbance was measured at a wavelength of 570 nm. The cell viability was assessed by dead/live double staining.

Acknowledgements

This project is supported by the National Science Fund of China (20974061) and the Shanghai Leading Academic Discipline Project (no. B202).

References

1. O. H. Brekke and I. Sandle, Nat. Rev. Drug Discovery, 2003, 2, 52–62.
2. S. Y. Park, Y. Lee, K. H. Bae, C. H. Ahn and T. G. Park, Macromol. Rapid Commun., 2007, 28, 1172–1176.
3. J. P. Jung, J. Z. Gasiorowski and J. H. Collier, Biopolymers, 2010, 94, 49–59.
4. C. Majidi, Soft Robotics, 2014, 1(1), 5–11.
5. S. Ashley, Sci. Am., 2003, 289(4), 52–59.
6. N. Shpaisman, L. Sheihet, J. Bushman, J. Winters and J. Kohn, Biomacromolecules, 2012, 13, 2279–2286.
7. O. D. Krishna and K. L. Kiick, Biopolymers, 2010, 94, 32–48.
8. K. Haraguchi and T. Takehisa, Adv. Mater., 2002, 14(16), 1120.
9. T. Sakai, T. Matsunaga and Y. Yamamoto, et al., Macromolecules, 2008, 41(14), 5379–5384.
10. J. P. Gong, Y. Katsuyama, T. Kurokawa and Y. Osada, Adv. Mater., 2003, 15(14), 1155–1158.
11. T. A. Holland, E. W. H. Bodde and V. Cuijpers, et al., Osteoarthritis Cartilage, 2007, 15(2), 187–197.
12. S. C. Rizzi and J. A. Hubbell, Biomacromolecules, 2005, 6(3), 1226–1238.
13. S. C. Rizzi, M. Ehrbar and S. Halstenberg, et al., Biomacromolecules, 2006, 7(11), 3019–3029.
14. A. M. Kloxin, A. M. Kasko and C. N. Salinas, et al., Science, 2009, 324, 59–63.
15. P. J. Gong, Trends Biochem. Sci., 2003, 28(4), 210–214.
16. H. Jia, Z. Li and X. Wang, et al., J. Mater. Chem. A, 2015, 3(3), 1158–1163.
17. P. Y. Sun, L. Y. Tian, Z. Zheng and X. L. Wang, Acta Polym. Sin., 2009, 8, 803–808.
18. Y. C. Wang, F. Wang and T. M. Sun, et al., Bioconjugate Chem., 2011, 22(10), 1939–1945.
19. S. A. Guelcher, Tissue Eng., Part B, 2008, 14(1), 3–17.
20. F. Meng, W. E. Hennink and Z. Zhong, Biomaterials, 2009, 30(12), 2180–2198.
21. T. McPherson, A. Kidane, I. Szleifer and K. Park, Langmuir, 1998, 14, 176–186.
22. D. Eglin, D. Mortisen and M. Alini, Soft Matter, 2009, 5(5), 938–947.
23. E. Behravesh, M. D. Timmer and J. J. Lemoine, et al., Biomacromolecules, 2002, 3(6), 1263–1270.
24. H. Koo, G.-w. Jin, H. Kang, Y. Lee, H. Y. Nam, H. S. Jang and J. S. Park, Int. J. Pharm., 2009, 374, 58–65.
25. X. Z. Shu, Y. Liu and F. Palumbo, et al., Biomaterials, 2003, 24(21), 3825–3834.
26. A. Shimizu, T. Suhara and T. Ito, et al., Surg. Today, 2014, 44(2), 314–323.
27. P. Tos, A. Crosio and I. Pellegatta, et al., Muscle Nerve, 2015, 53, 304–309.
28. Z. Zhang, J. Ni and L. Chen, et al., J. Biomed. Mater. Res., Part B, 2012, 100(6), 1599–1609.
29. T. McPherson, A. Kidane, I. Szleifer and K. Park, Langmuir, 1998, 14, 176–186.
30. J. A. Ledon, J. A. Savas and S. Yang, et al., Am. J. Clin. Dermatol., 2013, 14(5), 401–411.
31. M. Mehdizadeh, H. Weng, D. Gyawali, L. Tang and J. Yang, Biomaterials, 2012, 33, 7972–7983.
32. T. McPherson, A. Kidane, I. Szleifer and K. Park, Langmuir, 1998, 14, 176–186.
33. H. Koo, G.-W. Jin, H. Kang, Y. Lee, H. Y. Nam, H.-S. Jang and J.-S. Park, Int. J. Pharm., 2009, 374, 58–65.
34. Z. Zhang, J. Ni, L. Chen, L. Yu, J. Xu and J. Ding, Biomaterials, 2011, 32, 4725–4736.
35. T. McPherson, A. Kidane, I. Szleifer and K. Park, Langmuir, 1998, 14, 176–186.
36. F. Luo, T. L. Sun, T. Nakajima, T. Kurokawa and J. P. Gong, et al., Adv. Mater., 2015, 27, 2722–2727.
37. T. L. Sun, T. Kurokawa, S. Kuroda, A. B. Ihsan, T. Akasaki and J. P. Gong, et al., Nat. Mater., 2013, 12, 932–937.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra04320h

---