Manipulating the mechanical properties of biomimetic hydrogels with multivalent host–guest interactions

Boguang Yang a, Zi Wei b, Xiaoyu Chen a, Kongchang Wei *ac and Liming Bian *abde
aDepartment of Biomedical Engineering, The Chinese University of Hong Kong, Hong Kong, China. E-mail: lbian@cuhk.edu.hk
bChina Orthopedic Regenerative Medicine Group (CORMed), Hangzhou, Zhejiang, China
cEmpa, Laboratory for Biomimetic Membranes and Textiles, Lerchenfeldstrasse 5, CH-9014 St., Gallen, Switzerland. E-mail: kongchang.wei@empa.ch
dDepartment of Pharmacology, Guangdong Key Laboratory for Research and Development of Natural Drugs, Guangdong Medical University, Zhanjiang, Guangdong, China
eShenzhen Research Institute, The Chinese University of Hong Kong, China

Received 1st August 2018 , Accepted 28th September 2018

First published on 29th September 2018


Biomimetic hydrogels with hierarchical network structures are promising biomaterials for tissue engineering due to their unique mechanical properties. One successful biomimetic strategy for facile construction of high-performance hydrogels is to incorporate reversible crosslinks as sacrificial bonds into chemical polymer networks. By mimicking the unfolding–refolding functions of the skeletal muscle protein titin, the reversible crosslinks can reinforce the otherwise weak and brittle hydrogels. However, the contribution of multivalent reversible crosslinks to the overall hydrogel mechanical properties has rarely been investigated. Herein we present the biomimetic hydrogels with multivalent host–guest interactions as reversible crosslinks, which provide not only energy storage capacity, but also elevated energy dissipation capacity to the dually crosslinked networks, therefore leading to the improved hydrogel ductility and tensile strength. Our results also reveal the manner of multivalent host–guest crosslinks contributing to the hydrogel mechanical properties, including gelation rate, energy storage and dissipation, tensile hysteresis, and fast spontaneous recovery.


Introduction

Polymeric hydrogels are hydrated three-dimensional networks of hydrophilic polymers.1 They share many structural and mechanical similarities with biological soft tissues and therefore have been developed for many biomedical applications.2–7 However, compared to solvent-free polymer networks such as elastic rubbers, hydrogels containing a large amount of water and relatively low polymer contents are usually much weaker.8,9 Such mechanical weakness severely limits the application of hydrogels in many biomedical fields involving considerable mechanical loadings, such as artificial cartilage, biomedical actuators, and injectable objects. In contrast, some biological tissues, such as muscles, often show excellent mechanical strength due to the hierarchical protein structures.10,11

One of the most successful strategies for facile construction of high-performance hydrogels is to mimic the structural properties of the skeletal muscle protein titin by incorporating reversible crosslinks into polymer networks. By mimicking the unfolding–refolding functions of titin, the reversible crosslinks can reinforce the otherwise weak and brittle hydrogels. Ever since Peppas and his co-workers discovered that chemical hydrogels can be significantly reinforced by polymer crystallization in 1976,12 many physical interactions have been introduced to chemical hydrogels to improve their stiffness and toughness. For example, alginate–Ca2+ interactions have been used to reinforce the otherwise weak chemical polyacrylamide (PAAm) hydrogels, resulting in highly stretchable and tough hydrogels.13 However, most of the studies have been focusing on improving the hydrogel stiffness, strength, and toughness by mimicking the force-induced unfolding of titin.14–18 Less effort has been made to control the time-dependent mechanical properties of the hydrogels by mimicking the refolding of the ruptured titin.19 For example, although the energy dissipation properties are improved by incorporating physical crosslinks as sacrificial bonds, the fast and complete self-recovery of the ruptured networks after loading has rarely been achieved.13,20–22 Such fast and complete self-recovery, however, is one of the most important time-dependent properties for the application of hydrogels involving continuous mechanical loading. Only a limited number of hydrogels crosslinked by host–guest interactions have shown muscle-like fast self-recovery after large deformation.23–25 This was attributed to the fast kinetics of the host–guest association and dissociation. Recently, Grindy and his co-workers also reported that the control of the time-dependent hydrogel properties can be achieved by tuning the metal-coordination crosslinking dynamics.26

Although cyclodextrin-based host–guest interactions have been recently used to crosslink mechanically reinforced hydrogels, their contribution to the mechanical properties of dually crosslinked hydrogels have not yet been fully understood, especially when these host–guest crosslinkers are presented in a multivalent manner.25,27–30 Herein, we use β-cyclodextrin (β-CD) derived multivalent host–guest interactions to mimic the multiple unfoldable domains of titin and manipulate the mechanical properties of dually crosslinked PAAm hydrogels. Our results revealed that the multivalent effect significantly accelerate the gelation process. The obtained dually crosslinked hydrogels are much more stretchable than their chemically crosslinked counterparts. The reversible and multivalent host–guest crosslinks improve both the energy storage and dissipation of the hydrogels, as well as their self-recovery after large deformations. Moreover, the individual contribution of the two crosslinking types, namely irreversible and reversible crosslinking, to the overall hydrogel mechanical properties were examined by dynamic oscillatory rheometric measurements.

Results and discussion

Crosslinker-valency-dependent gelation kinetics of PAAm hydrogels

In our recent study, we have shown that the pre-assembly of the adamantane-functionalized hyaluronic acid (ADxHA, guest polymer) and the mono-acrylated β-cyclodextrin (mono-Ac-βCD, host monomer) can efficiently form host–guest complexes (termed host–guest macromers, HGM).31 Robust and bioactive biopolymeric hydrogels can be prepared from direct crosslinking of such HGM supramolecular structures. In this current study, we have prepared PAAm hydrogels by photo-polymerization of acrylamide monomers (AAm, 2 M) in the presence of N,N-methylenebisacrylamide (MBA, with the concentration of CMBA = 2 mM) as the chemical crosslinker and the pre-assembled HGM (with the concentration of CHGM* = 10 mM) as the supramolecular physical crosslinker (Scheme 1a). The concentration of HGM is expressed as CHGM*, meaning that the content of crosslinkable double bonds originated from HGM is equivalent to that of CHGM* mM MBA (ESI). The resultant hydrogel matrix contains the ductile synthetic polymers (PAAm) crosslinked by two types of crosslinks, which are the irreversible (MBA) and reversible (HGM) crosslinks, respectively. Therefore, such hydrogel was termed MBA–HGM dually crosslinked hydrogel (Scheme 1b and c). Accordingly, PAAm hydrogels crosslinked solely by low or high concentration of MBA were termed MBA-L (CMBA = 2 mM, CHGM* = 0 mM) or MBA-H (CMBA = 12 mM, CHGM* = 0 mM) hydrogels, and those crosslinked by HGM (CMBA = 0 mM CHGM* = 10 mM) was termed HGM hydrogels (Table 1).
image file: c8tb02021c-s1.tif
Scheme 1 Schematic illustration of the preparation of the MBA–HGM dually crosslinked hydrogels. (a) The structure of the pre-assembled host–guest macromer (HGM), acrylamide (AAm), and N,N-methylenebisacrylamide (MBA). (b) The polyacrylamide (PAAm) network crosslinked by irreversible chemical crosslinks (indicated by black arrows) and reversible host–guest crosslinks (indicated by red arrow). (c) Schematic illustration of both the irreversible and reversible crosslinks.
Table 1 The composition of the precursor solutions and the mechanical properties of the resultant hydrogels
Concentrationa [mM] Gelation speedb Dynamic modulusc
C AAm C MBA C HGM* t g −1 [min−1] G′ [kPa] G′′ [kPa]
a The content of the host–guest macromer (HGM) is expressed as CHGM*, meaning that the content of crosslinkable double bonds originated from HGM is equivalent to that of CHGM* mM MBA. b The gelation time (tg) was determined by the cross-over point of the storage modulus (G′) and loss modulus (G′′) during the UV curing process, and the gelation speed was defined as the reciprocal gelation time (tg−1). c The dynamic modulus of the hydrogels were determined by oscillatory rheometric time-sweep measurements (n = 3) after the hydrogels were cured with UV light (20 mW cm−2, 30 minutes).
MBA-L 2 × 103 2.0 0 0.167 ± 0.001 3.71 ± 0.03 0.25 ± 0.04
MBA-H 2 × 103 12.0 0 0.187 ± 0.001 21.27 ± 0.04 0.27 ± 0.01
MBA–HGM 2 × 103 2.0 10.0 0.229 ± 0.002 9.76 ± 0.01 2.00 ± 0.01
HGM 2 × 103 0 10.0 0.277 ± 0.002 6.82 ± 0.01 ±0.01


All the hydrogels were prepared on a rheometer equipped with a UV source (20 mW cm−2), enabling us to examine the gelation process with the strain-controlled (γ = 1%) dynamic oscillatory measurements at a single frequency (f = 1 Hz). The gelation time (tg) was determined by the cross-over point of the storage modulus (G′) and loss modulus (G′′) during the UV curing process (Fig. 1a), and the gelation speed was defined as the reciprocal gelation time (tg−1).32 Interestingly, although increasing the concentration of the chemical crosslinkers (MBA) from 2.0 mM to 12.0 mM leads to faster gelation, the gelation speed is only elevated by 11% (Fig. 1b, MBA-H). However, with constant content of crosslinkable double bonds (equivalent to that of 12.0 mM MBA), the gelation of MBA–HGM hydrogel is much faster than that of the MBA-H hydrogel (Fig. 2a). The gelation speed of MBA–HGM hydrogel is 23% higher than that of the MBA-H hydrogel (Fig. 1b), indicating the important influence of the crosslinker multivalency on the gelation kinetics of PAAm hydrogels. Such multivalent effect is also confirmed by the even higher gelation speed of the HGM hydrogel. Although the HGM hydrogel precursor solution (CMBA = 0 mM, CHGM* = 10.0 mM) contains less content of crosslinkable double bonds than MBA–HGM hydrogel precursor solution (CMBA = 2.0 mM, CHGM* = 10.0 mM), the gelation speed is 21% higher. One plausible explanation is that the existence of divalent MBA crosslinker in MBA–HGM hydrogels will decrease the average crosslinker valency (CVav.) and slow down the gelation process (Scheme S1, ESI). Such retardation on the gelation due to lower average crosslinker valency is more dominant compared with the acceleration due to higher total concentration of crosslinkable double bonds, thereby leading to the overall slower gelation of MBA–HGM hydrogel (CVav. = 9.5) compared with HGM hydrogel (CVav.= 38.9). Such a crosslinker valency dependency of PAAm gelation has rarely been reported before.


image file: c8tb02021c-f1.tif
Fig. 1 UV-initiated gelation of PAAm under different crosslinking conditions. (a) Gelation processes characterized by the single frequency strain controlled oscillatory rheometric measurements. (b) A summary of the gelation speed (tg−1: reciprocal gelation time) under different crosslinking conditions (sample number n = 3). The percentage values indicate the increase of gelation speed, the average crosslinker valency (CVav.) is defined as the average number of double bonds on each crosslinker molecule (Scheme S1, ESI).

image file: c8tb02021c-f2.tif
Fig. 2 The contribution of irreversible crosslinks (MBA) and reversible crosslinks (HGM) to the hydrogel mechanical properties. (a) Single frequency strain-controlled oscillatory measurement after 1800 s of UV curing. (b) A summary of the storage modulus (G′, left y axis) and loss modulus (G′′, right y axis) of the PAAm hydrogels (sample number n = 3). (c) Single frequency amplitude-sweep oscillatory measurements of the PAAm hydrogels. (d) The schematic illustration of the reversible manipulation of the PAAm hydrogel networks.

Reversible manipulation of PAAm hydrogel properties with multivalent host–guest interactions

Albeit with varied gelation speeds, the gelation of all hydrogels completed after 30 minutes of UV curing, as evidenced by the plateauing moduli (G′ and G′′) shown in Fig. 2a. The contribution of the divalent MBA crosslinker and the multivalent HGM crosslinker to the hydrogel modulus was then examined. The dynamic oscillatory measurements (Fig. 2a) were summarized in Fig. 2b. In contrast to the high dependency of gelation speed on the crosslinker valency, the dynamic modulus of hydrogels depends on both the crosslinker concentration and the crosslinking dynamics. The measurements on the hydrogels crosslinked solely by chemical MBA crosslinkers (MBA-L and MBA-H) showed that the chemical crosslinks (MBA) contributed to G′ in a concentration-dependent manner but barely contributed to G′′. This indicates that increasing the chemical crosslinking density will improve the energy storage property of the hydrogels but has minimal effects on the energy dissipation property. In contrast, compared to MBA-L hydrogels, the MBA–HGM hydrogels showed both higher G′ and G′′, indicating that the reversible HGM crosslinks contribute to both the energy storage and dissipation capability. Therefore, although the increase of the irreversible MBA crosslinker contents will increase the stiffness of the hydrogels, it will also significantly compromise the ductility of the hydrogels, making them much more brittle. This is evidenced by the dynamic oscillatory amplitude sweep showing a much narrower linear viscoelastic region (LVR) of MBA-H than that of MBA-L hydrogel (Fig. 2c). However, by replacing the surplus MBA in MBA-H hydrogels with reversible HGM crosslinks containing equivalent number of reactive groups, the MBA–HGM hydrogels showed improved stiffness and uncompromised ductility compared to the MBA-L hydrogels (Fig. 2c). Such simultaneous tuning of the energy storage and dissipation properties is reversible. In the presence of the competitive guest small molecules (amino-adamantane hydrochloride, ADA), the HGM crosslinkers in MBA–HGM hydrogels will be dissociated, yielding a similar network structure as that of MBA-L hydrogels (Fig. 2d). This is evidenced by the similar modulus (G′ and G′′) between “MBA–HGM +ADA” hydrogels and MBA-L hydrogels (Fig. 2a and b). Such a reversible tuning of the hydrogel mechanical properties also indicates that the simultaneous improvement of the energy storage and dissipation properties is mainly due to the introduction of HGM crosslinks rather than the introduction of additional biopolymers.

Reinforcement of PAAm hydrogels with multivalent host–guest interactions

The abovementioned findings under dynamic oscillatory loadings were further confirmed by the uniaxial tensile measurements (Fig. 3a and b). Although the MBA-H hydrogel was much stiffer (with elastic modulus of 34.8 kPa), the stretchability was much poorer than that of the MBA-L hydrogel (with elastic modulus of 10.1 kPa). With lower crosslinking density, albeit soft, the MBA-L hydrogel can be stretched to more than 8 times of its original length. In contrast, the brittle MBA-H hydrogel ruptured at stretching ratio less than 3 times of its original length with failure strength similar to that of the MBA-L hydrogel (40–45 kPa). However, The MBA–HGM hydrogel (with elastic modulus of 13.8 kPa) is much more stretchable than MBA-H hydrogel. It can be stretched up to 7.5 times of its original length. Meanwhile, the failure strength is around 90 kPa, 100% higher than that of the MBA-H hydrogel. This result is in accordance with the dynamic oscillatory amplitude sweep measurements, confirming that reversible HGM crosslinks can simultaneously improve the strength and ductility of the hydrogels. Such improvement is due to the effective energy dissipation enabled by the sacrificial host–guest crosslinks.
image file: c8tb02021c-f3.tif
Fig. 3 Tensile properties of PAAm hydrogels prepared with different crosslinking conditions. (a) Uniaxial elongation of PAAm hydrogels until failure. (b) The initial parts (λ < 0.5) of the tensile stress–strain curves from panel (a) are fitted with linear model for the calculation of the elastic modulus (E). (c) Cyclic tensile tests of PAAm hydrogels crosslinked by irreversible MBA crosslinkers (MBA-L) or by both MBA and reversible supramolecular HGM crosslinkers (MBA–HGM). (d) Self-recovery of the PAAm hydrogel crosslinked by both MBA and HGM crosslinkers (MBA–HGM) characterized by consecutive tensile cycles at a stretching ratio of 5 with different waiting times.

Cyclic tensile tests revealed that the MBA–HGM hydrogels can dissipate a large amount of the loading energy (22.8%, Fig. S1, ESI) under large deformation (stretching ratio λ = 5) as evidenced by the substantial hysteresis between the loading and unloading curves. In contrast, MBA-L hydrogels crosslinked solely by low concentration of chemical MBA crosslinks (CMBA = 2.0 mM, CHGM* = 0 mM) exhibited much lower stiffness under the same level of deformation, and the overlapping of the loading and unloading curves indicates that negligible energy was dissipated during the elastic deformation of the MBA-L hydrogel (Fig. 3c). Moreover, the fast self-recovery of the hysteresis in MBA–HGM hydrogels was confirmed by consecutive tensile cycles (Fig. 3d). Partial recovery was achieved if two consecutive tensile cycles were carried out without waiting interval, but nearly complete recovery was found if the waiting time was longer than 20 seconds. The minor compromise of the hydrogel mechanical strength after the original cycle may be due to some irreversible fractures of the chemical crosslinks. It is noteworthy that such fast self-recovery property of the MBA–HGM hydrogel was found unscathed by swelling, and the elastic modulus (12.1 kPa) of the swollen MBA–HGM (sMBA–HGM) is slightly lower than that of the original hydrogel (13.8 kPa) (Fig. 4a and b) but comparable to that of some human soft tissues (e.g. 8–17 kPa for muscles).33 This indicates that such hydrogels are promising prosthetic implant materials in locations where repeated biomechanical loading is considerable and in vivo swelling is inevitable. In addition, such excellent ductility and fast self-recovery of the MBA–HGM hydrogel was also demonstrated by substantial stretching and twisting without compromising the integrity (Movie S1, ESI).


image file: c8tb02021c-f4.tif
Fig. 4 (a) The tensile stress–strain curve of swollen MBA–HGM (sMBA–HGM) hydrogel being stretched to break. The initial part is fitted with linear model for the calculation of the elastic modulus (EsMBA–HGM = 12.1 kPa). (b) The tensile stress–strain curves of repeated stretching cycles of the sMBA–HGM hydrogel without resting interval in between cycles. (c) The contribution of the irreversible and reversible crosslinks to the spatial and temporal control over the hydrogel mechanical properties.

Therefore, by the dynamic oscillatory measurements and uniaxial tensile tests, we have proven that the irreversible chemical crosslinks mainly contribute to the spatial control over the hydrogel network structures and the crosslinking density-dependent hydrogel mechanical properties such as the energy storage property. However, the reversible host–guest crosslinks contribute to both the spatial and temporal control over the hydrogel mechanical properties, thereby regulating the time-dependent properties of the hydrogels (Fig. 4c).

Individual contribution of the chemical and physical crosslinks to PAAm hydrogel properties

To better quantify the individual contribution of the two crosslinking types to the hydrogel mechanical properties, we carefully examined the frequency dependent dynamic modulus of the hydrogels. Inspired by the study of polymer networks crosslinked by two types of metal-coordination interactions,19 we used the Maxwell–Weichert model consisting of two Maxwell elements (a spring in series with a dashpot) in parallel for describing the dually crosslinked hydrogels (Fig. 5a), where image file: c8tb02021c-t1.tif represents the contribution of the chemical crosslinks in MBA–HGM hydrogels to the overall dynamic modulus of the hydrogel, and likewise, the image file: c8tb02021c-t2.tif represents the contribution of the reversible host–guest crosslinks. Therefore, the overall dynamic modulus of the MBA–HGM dually crosslinked hydrogels can be described as the following equations:
 
image file: c8tb02021c-t3.tif(1)
 
image file: c8tb02021c-t4.tif(2)

image file: c8tb02021c-f5.tif
Fig. 5 The contribution of the irreversible and reversible crosslinks to the hydrogel mechanical properties, including the energy–storage (G′) and energy–dissipation properties (G′′) of the polyacrylamide hydrogels. (a) The Maxwell–Weichert model consisting of two Maxwell models in parallel for describing the dually crosslinked hydrogels. (b–d) Oscillatory frequency-sweep rheometric measurement of PAAm hydrogels crosslinked under different conditions.

We compared the measured dynamic modulus of the MBA–HGM dually crosslinked hydrogels, G′(MBA–HGM) and G′′(MBA–HGM), to those of the chemical or physical hydrogels. Herein, G′(MBA-L) and G′′(MBA-L) are the measured dynamic modulus of the MBA-L hydrogels, G′(HGM) and G′′(HGM) are the measured dynamic modulus of the HGM hydrogels. Interestingly, we found that G′(MBA–HGM) matched quite well with the simple linear combination of G′(MBA-L) and G′(HGM) within the whole frequency range (Fig. 5b, green and black data points):

G′(MBA–HGM) = G′(MBA-L) + G′(HGM)

This result indicates that the two types of crosslinkers contribute independently to the energy storage property of the dually crosslinked hydrogel. The individual contribution of one type of crosslinker is not affected by the presence of the other type. Both the chemical (MBA) and the physical (HGM) crosslinks in the MBA–HGM hydrogels contribute to the energy storage property as they do to the individually crosslinked networks:

image file: c8tb02021c-t5.tif

image file: c8tb02021c-t6.tif

More interestingly, G′′(MBA–HGM) was found to be very close to G′′(HGM), indicating the substantial contribution of the reversible host–guest crosslinks to the energy dissipation property of the MBA–HGM hydrogel (Fig. 5c). However, a close examination revealed that the contribution of the chemical crosslinks was frequency dependent (Fig. 5d):

G′′(MBA–HGM) = G′′(HGM) (f < 2.0 Hz)

G′′(MBA–HGM) = G′′(MBA) + G′′(HGM) (f > 2.0 Hz)

This result indicates that at low frequencies, the MBA–HGM hydrogels dissipates the loading energy mainly by temporary dissociation of the host–guest complexes; the deformation of the static polymer network stabilized by chemical crosslinkers is elastic and does not dissipate much loading energy. However, at high frequencies, the recovery of the deformed static network may not be able to keep up with the oscillating strains; therefore, the static polymer network crosslinked by chemical crosslinks also contributes to the energy dissipation:

image file: c8tb02021c-t7.tif

image file: c8tb02021c-t8.tif

Conclusions

In summary, we used multivalent host–guest crosslinkers to reinforce the chemically crosslinked hydrogels, and we showed that the reversible host–guest crosslinks can simultaneously tune the energy storage and dissipation properties of the hydrogels. The dually crosslinked hydrogels showed significantly improved mechanical strength, ductility, and self-recovery. Moreover, we carefully examined the individual contribution of the two crosslinking types to the overall mechanical properties of the hydrogels. Our finding revealed that the chemical crosslinks contributed mainly to the energy storage property, while the host–guest crosslinks significantly contributed to both the energy storage and dissipation properties. Insights from this study may inspire us to develop more advanced hydrogel materials with well-tailored overall mechanical properties for a broad range of applications.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

Project 31570979 was supported by the National Natural Science Foundation of China. The work described in this paper was supported by a General Research Fund grant from the Research Grants Council of Hong Kong (Project No. 14205817). This research was also supported by the project BME-p3-15 of the Shun Hing Institute of Advanced Engineering, The Chinese University of Hong Kong. This work was supported by the Health and Medical Research Fund, the Food and Health Bureau, the Government of the Hong Kong Special Administrative Region (Reference No. 03140056 & 04152836). This research was supported by the Chow Yuk Ho Technology Centre for Innovative Medicine (The Chinese University of Hong Kong). This research is supported by an Innovation Technology Fund (TCFS, GHP/011/17SZ), Hong Kong.

Notes and references

  1. O. Wichterle and D. Lím, Nature, 1960, 185, 117–118 CrossRef.
  2. D. Seliktar, Science, 2012, 336, 1124–1128 CrossRef CAS PubMed.
  3. A. S. Hoffman, Adv. Drug Delivery Rev., 2012, 64, 18–23 CrossRef.
  4. W. Shen, X. Chen, J. Luan, D. Wang, L. Yu and J. Ding, ACS Appl. Mater. Interfaces, 2017, 9, 40031–40046 CrossRef CAS.
  5. J. Sun, X. Liu, Y. Lei, M. Tang, Z. Dai, X. Yang, X. Yu, L. Yu, X. Sun and J. Ding, J. Mater. Chem. B, 2017, 5, 6400–6411 RSC.
  6. L. Li, C. Wang, Q. Huang, J. Xiao, Q. Zhang and Y. Cheng, J. Mater. Chem. B, 2018, 6, 2474–2480 RSC.
  7. H. Huang, J. Xu, K. Wei, Y. J. Xu, C. K. Choi, M. Zhu and L. Bian, Macromol. Biosci., 2016, 16, 1019–1026 CrossRef CAS.
  8. K. S. Anseth, C. N. Bowman and L. Brannon-Peppas, Biomaterials, 1996, 17, 1647–1657 CrossRef CAS.
  9. A. van Oosten, P. A. Galie and P. A. Janmey, GELS HANDBOOK: Fundamentals, Properties and Applications Volume 1: Fundamentals of Hydrogels, World Scientific, 2016, pp. 67–79 Search PubMed.
  10. S. Labeit and B. Kolmerer, Science, 1995, 270, 293–296 CrossRef CAS.
  11. M. S. Kellermayer, S. B. Smith, H. L. Granzier and C. Bustamante, Science, 1997, 276, 1112–1116 CrossRef CAS.
  12. N. A. Peppas and E. W. Merrill, J. Polym. Sci., Part A: Polym. Chem., 1976, 14, 441–457 CAS.
  13. J. Y. Sun, X. H. Zhao, W. R. K. Illeperuma, O. Chaudhuri, K. H. Oh, D. J. Mooney, J. J. Vlassak and Z. G. Suo, Nature, 2012, 489, 133–136 CrossRef CAS PubMed.
  14. J. P. Gong, Y. Katsuyama, T. Kurokawa and Y. Osada, Adv. Mater., 2003, 15, 1155–1158 CrossRef CAS.
  15. K. Haraguchi and T. Takehisa, Adv. Mater., 2002, 14, 1120–1124 CrossRef CAS.
  16. Y. Okumura and K. Ito, Adv. Mater., 2001, 13, 485–487 CrossRef CAS.
  17. T. Sakai, T. Matsunaga, Y. Yamamoto, C. Ito, R. Yoshida, S. Suzuki, N. Sasaki, M. Shibayama and U.-i. Chung, Macromolecules, 2008, 41, 5379–5384 CrossRef CAS.
  18. H. Kamata, Y. Akagi, Y. Kayasuga-Kariya, U. Chung and T. Sakai, Science, 2014, 343, 873–875 CrossRef CAS.
  19. M. Rief, M. Gautel, F. Oesterhelt, J. M. Fernandez and H. E. Gaub, Science, 1997, 276, 1109–1112 CrossRef CAS.
  20. T. L. Sun, T. Kurokawa, S. Kuroda, A. Bin Ihsan, T. Akasaki, K. Sato, M. A. Haque, T. Nakajima and J. P. Gong, Nat. Mater., 2013, 12, 932–937 CrossRef CAS PubMed.
  21. Q. Chen, L. Zhu, C. Zhao, Q. M. Wang and J. Zheng, Adv. Mater., 2013, 25, 4171–4176 CrossRef CAS PubMed.
  22. Y. Yang, X. Wang, F. Yang, H. Shen and D. Wu, Adv. Mater., 2016, 28, 7178–7184 CrossRef CAS PubMed.
  23. J. Liu, C. S. Tan, Z. Yu, N. Li, C. Abell and O. A. Scherman, Adv. Mater., 2017, 29, 1605325 CrossRef.
  24. J. Liu, C. S. Y. Tan, Z. Y. Yu, Y. Lan, C. Abell and O. A. Scherman, Adv. Mater., 2017, 29, 1604951 CrossRef.
  25. K. Wei, X. Chen, R. Li, Q. Feng and L. Bian, Chem. Mater., 2017, 29, 8604–8610 CrossRef CAS.
  26. S. C. Grindy, R. Learsch, D. Mozhdehi, J. Cheng, D. G. Barrett, Z. B. Guan, P. B. Messersmith and N. Holten-Andersen, Nat. Mater., 2015, 14, 1210–1216 CrossRef CAS PubMed.
  27. T. Kakuta, Y. Takashima and A. Harada, Macromolecules, 2013, 46, 4575–4579 CrossRef CAS.
  28. T. Kakuta, Y. Takashima, M. Nakahata, M. Otsubo, H. Yamaguchi and A. Harada, Adv. Mater., 2013, 25, 2849–2853 CrossRef CAS.
  29. M. Tan, Y. L. Cui, A. D. Zhu, H. Han, M. Y. Guo and M. Jiang, Polym. Chem., 2015, 6, 7543–7549 RSC.
  30. M. Nakahata, Y. Takashima and A. Harada, Macromol. Rapid Commun., 2016, 37, 86–92 CrossRef CAS.
  31. K. Wei, M. Zhu, Y. Sun, J. Xu, Q. Feng, S. Lin, T. Wu, J. Xu, F. Tian and J. Xia, Macromolecules, 2016, 49, 866–875 CrossRef CAS.
  32. S. Mal, P. Maiti and A. K. Nandi, Macromolecules, 1995, 28, 2371–2376 CrossRef CAS.
  33. H. Saraf, K. Ramesh, A. Lennon, A. Merkle and J. Roberts, J. Biomech., 2007, 40, 1960–1967 CrossRef CAS.

Footnotes

Electronic supplementary information (ESI) available: Experimental details [PDF] and Movie S1. See DOI: 10.1039/c8tb02021c
These authors contribute equally.

This journal is © The Royal Society of Chemistry 2019