Citrate Cross-Linked Gels with Strain Reversibility and Viscoelastic

Jun 27, 2014 - Biomaterials and Tissue Engineering Laboratory, School of Medical Science and ... found to be highest for the gels with the lowest cros...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/Langmuir

Citrate Cross-Linked Gels with Strain Reversibility and Viscoelastic Behavior Accelerate Healing of Osteochondral Defects in a Rabbit Model Paulomi Ghosh, Arun Prabhu Rameshbabu, and Santanu Dhara* Biomaterials and Tissue Engineering Laboratory, School of Medical Science and Technology, Indian Institute of Technology Kharagpur, Kharagpur 721302, India S Supporting Information *

ABSTRACT: Most living tissues are viscoelastic in nature. Self-repair due to the dissipation of energy by reversible bonds prevents the rupture of the molecular backbone in these tissues. Recent studies, therefore, have aimed to synthesize biomaterials that approximate the mechanical performance of biological materials with self-recovery properties. We report an environmentally friendly method for the development of ionotropically cross-linked viscoelastic chitosan gels with a modulus comparable to that of living tissues. The strain recovery property was found to be highest for the gels with the lowest cross-linking density. The force−displacement curve showed significant hysteresis due to the presence of reversible bonds in the cross-linked gels. Nanoindentation studies demonstrated the creep phenomenon for the cross-linked chitosan gels. Creep, hysteresis, and plasticity index confirmed the viscoelastic behavior of the cross-linked gels. The viscoelastic gels were implanted at osteochondral defect sites to assess the tissue regeneration ability. In vivo results demonstrated early cartilage formation and woven bone deposition for defects filled with the gels compared to nontreated defects. coupled with the development of fibrous tissue in the defect site.13 Moreover, existing biomaterials lack sufficient shear strength to withstand in vivo physiological load-bearing activities of native tissues without hardware supplementation.14 In this context, strain reversibility and viscoelasticity are desired mechanical properties of biomaterials employed for osteochondral defect management. Furthermore, scaffolds with a significant strain recovery property could endow cells with a better self-adapting capacity against stiff materials, which do not experience relevant strain.14 Zhang et al. chemically modified chitosan for the formation of dynamic hydrogels with a significant self-healing property.15 Owing to the structural resemblance to glycosaminoglycans of ECM, chitosan and its derivative are widely studied for bone, cartilage, and osteochondral tissue engineering applications by tailoring mechanical and biodegradation properties.16−18 Recently, our group demonstrated the ionotropic cross-linking of chitosan with multivalent ions for enhanced osteogenesis and biomineralization in vitro.19 Notably, citrate, a multivalent anion, accounts for 5.5 wt % of the organic fraction of bone and provides binding sites for calcium ions in apatite. Citrate-bound apatite influences nanocrystal thickening by stabilizing the size of nanocrystals.20 Furthermore, citric acid (CA) is reported to

1. INTRODUCTION Biological tissues such as skin, tendon, cartilage, and bone callus display viscoelastic behavior wherein energy dissipation by reversible bonds plays a pivotal role in autonomous damage healing.1−3 Reversible bonds dissociate upon loading by dissipating energy and then reform during the relaxation period.4 For instance, sacrificial reversible bonds found in collagen molecules of bone play an important role in the energy dissipation and fracture prevention mechanism.5 Similarly, dynamic association between water-extracellular matrix (ECM) molecules in articular cartilage is responsible for high-loadbearing activities.6 Most covalently cross-linked hydrogels lack the self-healing property; in contrast, reversible linkages such as electrostatic interactions, hydrophobic associations, and hydrogen bonds enables quick preparation and have the potential self-healing capability.7−10 Thus, the cross-linking of polyelectrolytic polymers using ionic solution leading to the formation of transient and multiple weak bonds might be a remarkable biomimetic approach for tissue regeneration. Osteochondral defects involve both bone and cartilage; therefore, scaffold-based tissue engineering should aim to mimic the natural environment of reparative osseous and chondrous tissues simultaneously.11 The nonlinear mechanical behavior of these tissues poses a serious challenge to regenerative medicine.12 In particular, the mechanical (creep, modulus, and yield stress) mismatch between the osteochondral injury area and scaffolds causes degenerative changes © 2014 American Chemical Society

Received: February 20, 2014 Revised: June 25, 2014 Published: June 27, 2014 8442

dx.doi.org/10.1021/la500698v | Langmuir 2014, 30, 8442−8451

Langmuir

Article

Figure 1. Rheological assessment of chitosan with different concentrations of citric acid at pH 7.4, 25 °C. (A) η* and (B) tan δ of chitosan as a function of time at constant strain and frequency of 0.01 and 1 Hz, respectively. (C) G′ of citrate−chitosan gels on strain sweep. (D) τy vs % crosslinker, data represented as mean ± SD. η*, complex viscosity; tan δ, phase angle; G′, shear modulus; τy, yield stress. according to our established method.18 The flow behavior of the solution was evaluated through viscosity measurements at different shear rates ranging from 0.1 to 100 s−1 at 25 °C using a Bohlin CVO rheometer (Malvern Instruments, Malvern, U.K.) with cone and plate geometry (CP 4°/40 mm diameter), maintaining a gap of 150 μm. The gelling behavior of chitosan solution was evaluated with different concentrations of CA: 1, 5, 10, 20, 40, and 60% w/v (Merck, India) at 25 °C. For the gelation kinetics study, the complex viscosity (η*) and phase angle (tan δ) were evaluated with time (t) sweep measurements in oscillatory mode at a constant strain and frequency of 0.01 and 1 Hz, respectively. Strain sweep was carried out on the formed chitosan gels at a constant frequency of 1 Hz to determine the linear viscoelastic region (LVR) by shearing them until structure breakdown occurred. The yield stress (τy) of the gels was calculated as the product of the yield strain (γy) and shear modulus (G′).

be used as a nontoxic cross-linker for many tissue engineering applications.21,22 In the present study, protonated glucosamine moieties of chitosan were ionotropically cross-linked with various concentrations of CA to investigate the viscoelastic and strain recovery properties of the gels. The elastic modulus, nanohardness, energy dissipation, creep, and resistance to plastic deformation were compared among gels of varied cross-linking densities. We hypothesized that transient ionic-bond-mediated strain-reversible hydrogels with an energy dissipation mechanism might be suitable as scaffolds in cyclic load-bearing applications. The hydrogels were implanted in the femoral condyle of rabbits and evaluated for their efficacy against nontreated defects toward osteochondral regeneration at different time periods.

τy = γyG′

2. EXPERIMENTAL SECTION

(1)

For reproducibility, at least five samples of each of the hydrogels formed with different concentrations of CA were measured to evaluate η*, tan δ, and G′. Chemical characterization of the differently cross-

2.1. Gelation Kinetics. A chitosan (6 wt %, MW 710 000; >90% deacetylated, Marine Chemicals, Cochin) stock solution was prepared 8443

dx.doi.org/10.1021/la500698v | Langmuir 2014, 30, 8442−8451

Langmuir

Article

linked chitosan gels was determined by FTIR and ninhydrin assay (n = 3) as provided in the Supporting Information. 2.2. Strain Recovery. For each of the formed hydrogels, selfrecovery measurements were carried out at a frequency of 1 Hz using this procedure (applied shear force, expressed in terms of strain, duration in parentheses): 1% (150 s)−100% (150 s). The sequence of the above segments was repeated twice to examine the strain reversibility of the network. For reproducibility, five samples were measured for each cross-linked gel. 2.3. Nanoindentation Studies. Nanoindentation was carried out on cross-linked gels using a nanoindenter (TI 950 TriboIndenter, Hysitron Inc., USA) to determine the load vs displacement curve. The instrument was air calibrated with a constant electrostatic force. A total number of 10 points at a distance of 4 μm was employed for each sample. A maximum load of 100 μN was applied with a loading and unloading time period of 10 s. Plastic work done by the indentor on the cross-linked gels was determined by calculating the area enclosed by their respective load−displacement curve. Creep data was obtained at the same loading and unloading rate, with a loading time of 10 s, a hold time of 60 s, and an unloading time of 10 s with a force of 100 μN. Modulus, hardness, and resistance to deformation obtained were used for the comparative assessment of gels of different cross-linking densities. The plasticity index was calculated according to a previous report.23 The formula used was

ψ = A1/(A1 + A 2 )

3. RESULTS 3.1. Gelation Kinetics. Chitosan (6 wt %) dissolved in aqueous acetic acid was very viscous and exhibited shearthinning behavior due to the interchain entanglement of polymer molecules. The viscosity of the polymer solution was 7.6 Pa·s at a shear rate of 100 s−1 (data not shown). In this study, the complex viscosity (η*) of chitosan was evaluated with time sweep measurements before and after the addition of coagulant in an oscillatory mode at constant amplitude and frequency (Figure 1A). The gelation of chitosan occurred instantly upon addition of CA at neutral pH. The gelation kinetics of CA- chitosan could be divided into four parts: (i) liquidlike flow behavior before the addition of coagulant, (ii) immediate gelation after the addition of CA at 130 s as evident by a rapid rise in η* values, (iii) an increase in η* until a certain time, and finally (iv) a plateau region suggesting the saturation limit of the gels. Mechanistically, chitosan molecules in aqueous acetic acid are highly positively charged at pH 2.5 (pKa = 6.3) due to the protonation of amine groups.25,26 With the addition of CA at pH 7.4, the gelation of chitosan solution occurred both by competitive pH-induced neutralization as well as kinetically driven ionotropic complexation between polycationic chitosan and negatively charged citrate ions (Supporting Information FTIR data). At this pH, CA molecules are highly negatively charged (pKa = 3.1, 4.8, and 6.4) and formed an ion pair complex (COO−···NH3+) with polycationic chitosan instantaneously. Furthermore, the η* of the gels increased and tan δ values (Figure 1B) decreased with the respective increase in cross-linker concentration (C0) owing to a higher degree of ionotropic complexation and the formation of an interpenetrating gel network. The tan δ values for gels crosslinked with the highest C0 of CA almost reached zero, indicative of their elastic nature among all ionotropically cross-linked chitosan gels. Notably, tan δ values were always higher than zero for all citrate−chitosan hydrogels, suggesting the viscoelastic property of these gels. Figure 1C shows an increase in shear modulus (G′) with an increase in C0 of CA at constant frequency. This is mainly because upon increasing the C0 of the coagulant, the effective cross-linking density between chitosan and CA also increased (Table 1). According to the Maxwell model, citrate−chitosan gels can be described as a combination of springs and dampers. An increase in C0 and thus an increase in teh cross-linking density could be related to the increase in the number of springs and dampers in the gels. Accordingly, G′ increased at a higher cross-linking density due to the greater number of springs at small deformations (Table 1). The stress range over

(2)

where A1 is the area encompassed between loading and unloading curves and A2 is the area encompassed by the unloading curve (viscoelastic recovery). It follows that ψ = 1 for a fully plastic deformation, ψ = 0 for a fully elastic deformation, and 0 < ψ < 1 for viscoelastic−plastic materials. 2.4. In Vivo Studies. 2.4.1. Implantation Surgery in the Rabbit Model. The cytocompatibility of citrate cross-linked gels was assessed in vitro as provided in the Supporting Information. In vivo studies of CA cross-linked gels were performed under the compliance of Institutional Animal Ethical Committee guidelines of the Indian Institute of Technology, Kharagpur, India. Briefly, New Zealand white male rabbits weighing approximately 2 kg were maintained in a wellaerated atmosphere with adequate diet supplements. Prior to the start of the experiment, rabbits were anesthetized with an injection of ketamine hydrochloride. For the operation, bilateral knees of the animals were washed and disinfected with povidone−iodine (Apollo Pharmacy). Defects of 4 mm diameter and 5 mm depth were created in the weight-bearing areas of femoral condyles using a sterile drill bit. The cavities were washed with saline during and after drilling, followed by the insertion of sterilized hydrogels of appropriate size into the defect site. Nontreated defects were used as a control. The defect site was closed using a 4-0 vicryl suture. Four animals per treatment groups were used. After surgery, animals moved freely in the cage and were encouraged to move out of the cage every day. 2.4.2. Gross, Histological, and Histomorphological Evaluation. Rabbits were sacrificed at appropriate time intervals to retrieve specimens with surrounding tissues for gross evaluation. After gross examination, specimens were fixed in formalin, decalcified with 14% ethylene diamine tetraacetic acid, dehydrated in an ethanol series, passed through xylene, and embedded in paraffin. Sections of 3 μm thicknesses were sliced, followed by toluidine blue staining for the detection of cartilage deposition and H & E and von Kossa staining for cellularity and subchondral bone formation, respectively. Semiquantitative histomorphological analysis was done according to the ICRS Visual Histological Assessment grading scale24 as provided in the Supporting Information. A representative score for each parameter was determined by averaging the scores of two blinded observers. 2.5. Statistical Analysis. All data in this study were expressed as the mean ± standard deviation. Statistical analysis was performed using one-way ANOVA followed by tukey post hoc analysis in OriginPro 8 software for mechanical studies. The student’s t test was used to compare histological scores between the defects filled with citrate gels and nontreated control groups. Significance was determined at p < 0.05.

Table 1. Effect of Cross-Linker Concentration on the Degree of Cross-Linking, Shear Modulus, and Yield Stress of the Formed Gels CA %a 1 5 10 20 40 60

% CDb 24.6 36.7 45.0 56.3 78.3 83.0

± ± ± ± ± ±

3.8 4.2 3.1 6.1 8.2 8.5

G′ (kPa)c 2.8 4.3 9.2 20.4 46.0 78.2

± ± ± ± ± ±

0.1 0.4 0.7 4.9 9.2 6.6

τy (kPa)d 0.3 0.6 1.6 4.9 16.7 24.3

± ± ± ± ± ±

0.05 0.08 0.4 0.9 2.4 3.5

a CA, citric acid. bCD, cross-linking density obtained from ninhydrin assay. cG′, shear modulus at 1% strain and 1 Hz frequency. dτy , yield stress; values are represented as mean ± SD.

8444

dx.doi.org/10.1021/la500698v | Langmuir 2014, 30, 8442−8451

Langmuir

Article

which G′ is independent of the applied shear strain is the linear viscoelastic region (LVR).27 In the present study, LVR of citrate−chitosan gels progressively extended with an increase in the cross-linking density of the gels. Beyond LVR, G′ abruptly decreased, indicating the breakdown of the network structure. Interestingly, the yield strain (γy) could be obtained by extrapolating the linear G′ region beyond the point where G′ starts decreasing.28 Yield stress (τy) values (Table 1) were calculated from γy and G′ following eq 1. Similar to G′, γy and τy of the gels also increased with increasing cross-linking density (Figure 1D). 3.2. Strain Recovery. In Figure 1C, the G′ response characterized by the linear stress−strain curve was obtained until 1% applied strain, beyond which the network started to break down. Accordingly, for strain reversibility measurements of cross-linked networks, parameters for alternative relaxation and strain were chosen as 1 and 100%, respectively (Figure 2A). Two distinct zones at lower and upper ranges of the complex modulus due to sequential strain and relaxation,

respectively, were evident for all ionically cross-linked gels. Complex modulii increased with the increase in C0 of CA used for gelation. This is mainly due to the greater gel strength associated with the respective increase in the degree of crosslinking. At a large-amplitude oscillatory force (100%), the network structure of ionotropically cross-linked gel formed by 5% CA was disrupted and the complex modulus decreased from 4 kPa to 280 Pa. When applied strain was reduced to 1%, gels formed by 5% CA regained the cross-linked structure, leading to the recovery of the complex modulus to its initial values, indicating that the strain-induced damage is self-mendable in nature. Notably, the strain recovery was also remarkable for gels with the highest cross-linking density (gels formed using 60% CA), though not complete. Mechanistically, during the application of a large-amplitude oscillatory force, the network structure of ionotropically cross-linked gel breakdown via disruption of ion pairs/reversible bonds was regained when the strain was released. However, the extent of structure breakdown was significantly higher for 60% cross-linked gels in the impeding kinetically driven recovery of the complex modulus. Moreover, upon removal of strain, random ion pair formation in abundant citrate ions occurred that led to an unorganized structure. Thus, the degree of ionic cross-linking has a direct influence on the extent of strain reversibility and could be correlated to the selfrepairing property of various biological tissues. Strain recovery is an important biomaterial property as most tissues in the human body are continuously undergoing cyclic mechanical strain.29−31 Cyclic loading and strain are reported to stimulate the anabolic metabolism of bone as well.32−34 Ionotropically CA cross-linked gels (5%) were further treated with NHS-EDC to obtain covalently CA cross-linked gels and were compared to native 5% ionic gels for strain recovery (Figure 2B). The formation of covalently cross-linked gel was confirmed by the formation of amide bonds as evidenced by FTIR spectroscopy (Supporting Information). Though the modulus of covalently cross-linked gels was higher than that of ionotropically cross-linked gels with a similar cross-linking density, the network structure for covalently cross-linked gels failed to regain the structure to the same extent after the strain was released. This is mainly due to the irreversible nature of cross-linking in gels formed via covalent bonding. Thus, it was confirmed that the existence of ionic interactions in the gels is primarily responsible for the strain recovery property. Initial attempts to form ionic gels with monovalent ions failed, indicating the need for multivalent ionic agents for ionotropic gelation (data not shown). CA is a six-carbon compound wherein the two COO− groups associate electrostatically with NH3+ groups of chitosan and pendant carboxyl/hydroxyl groups of CA may form multiple H bonds among themselves and OH moieties of chitosan, all of which could significantly contribute to the strain recovery property of the gels. 3.3. Nanoindentation Study. The modulus and viscoelasticity of the gels were also derived using nanoindentation at random points. Figure 3A shows the force−displacement curve of chitosan gels with varying degrees of cross-linking. Gels cross-linked with 5% CA demonstrated a greater displacement than 20 and 60% cross-linking, indicating more deformation when an equivalent load of 100 μN was applied. Interestingly, 5% CA cross-linked gels had a very low elastic limit after which the gels experienced a maximum displacement of up to ∼540 nm with a permanent deformation of ∼450 nm. In contrast, gels formed with 60% CA displayed a relatively higher elastic

Figure 2. Complex modulus of step−strain measurements for (A) chitosan gels cross-linked ionically with different concentrations of citric acid and (B) ionically and covalently cross-linked gels with similar cross-linking densities. 8445

dx.doi.org/10.1021/la500698v | Langmuir 2014, 30, 8442−8451

Langmuir

Article

Figure 3. (A) Representative force−displacement curve of chitosan gels formed with 5, 20, and 60% citric acid. The inset shows the round-shaped nose for the loading−unloading peak of citrate−chitosan gel due to the delayed viscoelastic response. (B) Plastic work done by the indentor on the gels, n = 10 (p < 0.05). (C) The introduction of a hold segment in the force−displacement curve reduces the creeping effect of citrate cross-linked gels. (D) Time-dependent creep curves showing three distinct zones.

limit with a maximum displacement of up to ∼160 nm and a permanent deformation of 150 nm. Thus, chitosan gels formed with a lower degree of cross-linking experienced a higher permanent deformation with a larger hysteresis. However, the extent of reversible displacement/recovery was better for gels with a lower degree of cross-linking. For instance, the extents of recovery were 90 and 10 nm for 5 and 60% cross-linked gels, respectively. Dissipated energy, Uhys, in terms of plastic work done by the indenter on the gels is depicted in Figure 3B. From the force− displacement curve, it is evident that Uhys decreased with an increase in the cross-linking density, which is mainly due to the pronounced elastic response and associated stiffness. In other words, gels having a low degree of cross-linking displayed stronger viscoelastic behavior. Plastic work done by the indentor on 5, 20, and 60% CA gels are 115 471 ± 4719, 51 165 ± 4561, and 36 118 ± 2934 μN·nm, respectively. There are reports for varied hysteresis-associated self-recovery properties of gels owing to their differential cross-linking density and the

nature of cross-linking. A previous study has reported chemically cross-linked poly(acrylic acid) hydrogels with mechanical hysteresis for a range of deformation due to energy dissipation associated with ionic interactions.35 The inset of Figure 3A illustrates the round-shaped nose of a 5% CA cross-linked gel at the corresponding loading− unloading peak. The characteristic nose was also observed in the case of gels with higher cross-linking densities, though to a lesser extent, that is, less convex than 5% CA cross-linked gels. This is because immediately after the unloading segment starts, the penetration depth of the indenter on the polymer slightly increases even though the imposed load decreases at a constant rate. In other words, a delayed viscoelastic response of the cross-linked gels was evident during the initial phase of the unloading segment. This is in good agreement with a previous report.36 However, the nose, also known as a creeping effect, leads to error while calculating other derived quantities such as the average hardness, elastic modulus, plasticity index, and 8446

dx.doi.org/10.1021/la500698v | Langmuir 2014, 30, 8442−8451

Langmuir

Article

Table 2. Nanoindentation-Derived Values for Differently Cross-Linked Gelsa CA%b

ψc

H (MPa)d

Er (MPa)e

H/Er

H3/Er2 (MPa)f

5 20 60

0.83 ± 0.06 0.84 ± 0.06 0.98 ± 0.07

4.9 ± 0.25 11.1 ± 0.30 58.6 ± 0.51

85.9 ± 0.74 174.5 ± 3.4 645.9 ± 12.2

0.056957 ± 0.0012 0.063457 ± 0.0012 0.090681 ± 0.002

15.9 ± 0.38 44.6 ± 0.47 481.7 ± 10.1

a Values are represented as mean ± SD; nanoindentation values of 20 and 60% CA cross-linked gels were significantly different from those of 5% CA cross-linked gels (p < 0.05). bCA, citric acid. cψ, plasticity index. dH, average hardness. eEr , average elastic modulus. fH3/Er2, resistance to plastic deformation.

Figure 4. Gross appearance of (A) the drilled site; defect sites (B) filled with citrate cross-linked chitosan gels and (C) left untreated after 6 weeks. The defect filled with gels showed partial regeneration with cartilaginous tissue.

were highest for chitosan gels cross-linked with 60% CA followed by gels cross-linked with 20 and 5% CA, respectively. H and Er increased 11- and 7-fold, respectively, when the concentration of cross-linker increased from 5 to 60%. Thus, a correlation was obtained between the cross-linking density and elastic modulus/hardness of the polymer. This may be because a high degree of cross-linking forms a stiffer network that restricts stress deformation and eventually increases modulus and hardness values. This is in agreement with a previous report.35 The partitioning between energy dissipation by elastic and plastic deformation (H/Er) was found to be proportional to the cross-linking densities of chitosan gels. A relatively low value of H/Er for 5% CA cross-linked gels indicated that a greater fraction of work was consumed in plastic deformation. A similar finding was reported by Choudhary et al.39 H/Er could be correlated to Figure 3B, where plastic work done by the indentor was highest for chitosan gels cross-linked with 5% CA. Alternatively, H/Er was highest for gels cross-linked with 60% CA (p < 0.05 with respect to 5% CA cross-linked gels). Moreover, the resistance to plastic deformation, H3/Er2 was also highest for 60% CA cross-linked gels. 3.4. In Vivo Response. Though gels cross-linked with different concentrations of CA were cytocompatible (Supporting Information), gels formed with 5% CA were chosen for animal studies due to high strain reversibility and a low plasticity index (i.e., viscoelastic behavior). Furthermore, employing a scaffold with a low modulus might mimic the conditions experienced during osteochondral reparative stages.14,40−43 Figure 4A shows a representative image of the defects that were created in the femoral condyle of rabbits. Optical images of the specimens from experimental (defects filled with citrate gels) and control (nontreated defects) groups after 6 weeks are shown in Figure 4B,C respectively. No major sign of inflammation/infection was observed. Boundaries between the defects and the surrounding tissue were distinct in both groups. In the experimental group, defects were filled with shiny white repair tissue, whereas the nontreated defect

resistance to plastic deformation. The creeping effect could be eliminated by allowing a holding time for the indenter at a static load for attaining equilibrium during deformation before the unloading segment begins.23 In the present experiment, when the cross-linked gels were exposed to a load of 100 μN for 60 s, the differently crosslinked gels exhibited different degrees of creep response as shown in Figure 3C. An increasing degree of cross-linking caused a shallower depth of penetration of nanoindenter into the gels during the holding period, mainly due to an increase in the stiffness of the samples. Interestingly, after a hold time of 60 s, the unloading curve for 5% cross-linked gels displayed a maximum displacement of up to 1069 nm with permanent deformation at 749 nm. In contrast, gels with higher degree of cross-linking depicted a steeper unloading curve. For instance, 60% CA cross-linked gels showed maximum displacement and permanent deformation at 235 and 200 nm, respectively. Figure 3D depicts the creep displacement of the gels with three distinct zones. In zone 1, gels crept in a steady manner, after which in zone 2 the creep rate increased until displacements of ∼600 and 99.5 nm for 5 and 60% CA cross-linked gels were reached, respectively. This zone represents the onset of plastic deformation and strain softening. Beyond zone 2, the creep displacement continued to accumulate. Zone 3 represents different extents of strain hardening of the gels at different cross-linking densities. This is in agreement with a previous report.37,38 Hysteresis and creep phenomena are typical viscoelastic plastic responses of the gels. Furthermore, the plasticity index (ψ) of the differently crosslinked gels was determined using the force−displacement curve. The ψ values of the cross-linked gels ranged from 0 to 1, indicating that the gels were viscoelastic in nature (Table 2). Evidently, gels formed using 5% CA were the most viscoelastic in nature (p < 0.05). This is consistent with Figure 1B, where phase angle values were larger than zero, indicating the viscoelastic property of the gels. Hardness and elastic modulus values of the gels were inferred from the force−displacement curve and are presented in Table 2. The average hardness (H) and elastic modulus (Er) value 8447

dx.doi.org/10.1021/la500698v | Langmuir 2014, 30, 8442−8451

Langmuir

Article

Figure 5. Histological sections of the defects filled with CA cross-linked gels (A−C, G−I) and the nontreated empty defect sites as controls (D−F, J−L) 2 weeks (A−F) and 6 weeks (G−L) after surgery. The defects were stained with toluidine blue (dotted lines, defect margins; stars, gels; triangles, chondrocytes; arrows, osteoblast cells; BM, bone marrow; CA, citrate cross-linked chitosan gels; FC, fibrocartilage; NC, native cartilage; RC, repaired cartilage; TB, trabecular bone). Original magnification: (A, D, G, J) 5×, scale bar 50 μm; (B, C, E, F, H, I, K) 20×, scale bar 40 μm.

in defects filled with citrate gels as compared to empty defects indicating active bone formation. Calcium deposition was absent from the experimental and control groups after 2 weeks. Faint positive staining for von Kossa was obtained in nontreated defects after 6 weeks; in contrast, defects filled with viscoelastic citrate gels displayed prominent mineralization. Moreover, partially degraded gels were still visible in subchondral regions after 6 weeks (Figures 5 G,I and 6). Histological grading scores were used to assess regenerated tissues semiquantitatively (Table 3). Histomorphometric analysis indicated that defects filled with citrate gels underwent early integration with native cartilage compared to nontreated empty defects. On the whole, the experimental group scored significantly higher than the control group (p < 0.05) after 6 weeks of implantation.

sites showed reddish irregular tissue formation with surface depression. Decalcified tissue sections were investigated histologically at regular intervals (Figure 5). At 2 weeks, no obvious cartilaginous tissue was formed in the citrate-gel-filled/nontreated defects (Figure 5A,B,D,E). Also, trabecular bone was poorly formed in both experimental and control groups (Figure 5C,F). After 6 weeks, the proteoglycan expression for the cartilaginous matrix was observed in the defects filled with citrate cross-linked gels (Figure 5G). Toluidine blue staining showed metachromasia at the margins of the defect filled with gel. However, the cellular arrangement was irregular and the typical zonal architecture of cartilage was not observed (Figure 5H). However, the border with normal cartilage was distinct, with the regenerated cartilage integrated mostly with the underlying subchondral bone. In terms of subchondral bone repair, a continuous layer of trabecular bone was formed below the cartilage layer (Figure 5H,I). The healing of the subchondral area is critical to supporting the overlying neocartilage tissue.11,46 On the contrary, defect sites in the control group were mainly filled with immature repair tissue (Figure 5J,K). Fibroblast-like cells were identified in the empty defect with minimal proteoglycan deposition. Moreover, subchondral ossification was insufficient in the nontreated defects (Figure 5L). This was further validated by H & E and von Kossa staining of subchondral bone regions (Figure 6). A greater number of cuboidal osteoblast-like cells lining the trabecular bone was evidenced by H & E staining after 6 weeks

4. DISCUSSION It has long been known that vigorous cyclic loading results in cartilage fibrillation accompanied by degenerative changes in osseous tissue in joint regions.45 Serink et al. reported daily impact loading to cause osteoarthritis within 3−6 weeks for rabbit knees.46 In this context, ionotropically cross-linked chitosan gels that withstand alternate strain and relaxation could be advantageous for application under conditions experiencing high strain and cyclic load. In the present study, citrate−chitosan gels were implanted in weight-bearing osteochondral defects of the femoral condyle in the rabbit model and were compared with empty nontreated defects. An 8448

dx.doi.org/10.1021/la500698v | Langmuir 2014, 30, 8442−8451

Langmuir

Article

Figure 6. Representative hematoxylin−eosin and von Kossa−nuclear fast red stained sections depicting the comparison of in vivo subchondral bone formation in citrate-gel-filled defects and nontreated defects at 2 and 6 weeks, respectively (stars, gels; arrowheads, osteocyte in lacuna; arrows, osteoblast lining cells; BM, bone marrow; CD, calcium deposits; TB, trabecular bone). Scale bar 50 μm.

curve by nanoindentation. Each of these measured parameters reflects different properties of the gels with varied cross-linking densities.23,47 The increase in cross-linking density of the gels led to the decrease in Uhys as well as viscoelastic behavior. This was further validated by the creep data and plasticity index of the gels. Alternately, resistance to plastic deformation and yield stress increased with increasing cross-linking densities of the gels. Thus, citrate cross-linked gels with tailorable mechanical properties along with prominent viscoelastic behavior could be employed for tissue engineering needs. However, longer studies should be performed to assess tissue healing under full weight bearing. Though the present results cannot be extrapolated to other tissue types and larger animals, it provides an indication of healing via the biomimetic approach. In summary, we demonstrated the ability of strainreversible gels to mediate early osteochondral repair in rabbits.

Table 3. Histological Scores for Specimens Retrieved after 6 Weeksa score parameter

experimental groupb

control groupc

surface matrix cell distribution cell population viability subchondral bone cartilage mineralization total

2.56 ± 0.15 1.96 ± 0.15 1.83 ± 0.76 2.70 ± 0.26 2.16 ± 0.35 1.86 ± 0.30 13.07 ± 3.2

0.35 ± 0.15 0.32 ± 0.16 0.33 ± 0.15 1.60 ± 0.53 1.93 ± 0.12 0.75 ± 0.09 5.28 ± 1.86d

For each parameter, values are expressed as mean ± SD. bDefects with citrate cross-linked gels. cEmpty defects. dp < 0.05 versus experimental group. a

abundant deposition of proteoglycan in cartilage regions was observed with trabecular bone formation in the subchondral area of the defects filled with citrate gels. The early repair of osteochondral defects could be due to the resemblance of the gels with the ECM components and strain reversibility under repetitive shear force. The strain reversibility in turn is attributed to the transient ionic bonds that break and reform upon application and withdrawal of shear force, respectively. Besides strain reversibility, there are several other biomechanical considerations39 such as modulus, hardness, and dissipated energy during loading−unloading for osteochondral tissue regeneration. In the present study, dissipated energy (Uhys) and resistance to plastic deformation along with hardness and moduli were derived from the load−displacement

5. CONCLUSIONS Chitosan gels of various cross-linking densities were formed with different concentrations of citric acid at neutral pH via electrostatic interactions. The gels exhibited increases in the modulus and yield stress and a decrease in the phase angle with an increase in the cross-linking densities as evidenced by rheological studies. Strain recovery was prominent in ionically cross-linked gels with low cross-linking densities. On the contrary, covalently cross-linked citrate−chitosan gels with similar cross-link densities did not exhibit a significant strain recovery property owing to irreversible linkages present in the network. Similar to rheological assessments, nanoindentation 8449

dx.doi.org/10.1021/la500698v | Langmuir 2014, 30, 8442−8451

Langmuir

Article

(10) Wang, M.; Janout, V.; Regen, S. L. Hyper-thin organic membranes with exceptional H2/CO2 permeation selectivity: importance of ionic crosslinking and self-healing. Chem. Commun. 2011, 47, 2387−2389. (11) Shao, X. X.; Hutmacher, D. W.; Ho, S. T.; Goh, J. C.; Lee, E. H. Evaluation of a hybrid scaffold/cell construct in repair of high-loadbearing osteochondral defects in rabbits. Biomaterials 2006, 27, 1071− 1080. (12) Edelsten, L.; Jeffrey, J. E.; Burginx, L. V.; Aspden, R. M. Viscoelastic deformation of articular cartilage during impact loading. Soft Matter 2010, 6, 5206−5212. (13) Simon, T. M.; Jackson, D. W. Articular cartilage: injury pathways and treatment options. Sports Med. Arthrosc. 2006, 14, 146−154. (14) Willie, B. M.; Petersen, A.; Schmidt-Bleek, K.; Cipitria, A.; Mehta, M.; Strube, P.; Lienau, J.; Wildemann, B.; Fratzl, P.; Duda, G. Designing biomimetic scaffolds for bone regeneration: why aim for a copy of mature tissue properties if nature uses a different approach? Soft Matter 2010, 6, 4976−4987. (15) Zhang, Y.; Tao, L.; Li, S.; Wei, Y. Synthesis of Multiresponsive and Dynamic Chitosan-Based Hydrogels Synthesis of Multiresponsive and Dynamic Chitosan-Based Hydrogels. Biomacromolecules 2011, 12, 2894−2901. (16) Hutmacher, D. W. Scaffolds in tissue engineering bone and cartilage. Biomaterials 2000, 21, 2529−2543. (17) Datta, P.; Ghosh, P.; Ghosh, K.; Maity, P.; Samanta, S. K.; Ghosh, S. K.; Mohapatra, P. K.; Chatterjee, J.; Dhara, S. In Vitro ALP and Osteocalcin Gene Expression Analysis and In Vivo Biocompatibility of N-Methylene Phosphonic Chitosan Nanofibers for Bone Regeneration. J. Biomed. Nanotechnol. 2012, 9, 1−10. (18) Ghosh, P.; Rameshbabu, A. P.; Dhara, S. 2,5-Dimethoxy 2,5dihydrofuran crosslinked chitosan fibers enhances bone regeneration in rabbit femur defects. RSC Adv. 2014, 4, 19516−19524. (19) Pati, F.; Kalita, H.; Adhikari, B.; Dhara, S. Osteoblastic cellular responses on ionically crosslinked chitosan-tripolyphosphate fibrous 3D mesh scaffolds. J. Biomed. Mater. Res., Part A 2013, 101, 2526−2537. (20) Hu, Y. Y.; Rawal, A.; Schmidt-Rohr, K. Strongly bound citrate stabilizes the apatite nanocrystals in bone. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 22425−22429. (21) Jiang, W.; Pan, H.; Cai, Y.; Tao, J.; Liu, P.; Xu, X.; Tang, R. Atomic force microscopy reveals hydroxyapatite-citrate interfacial structure at the atomic level. Langmuir 2008, 24, 12446−12451. (22) Taguchi, T.; Saito, H.; Uchida, Y.; Sakane, M.; Kobayashi, H.; Kataoka, K.; Tanaka, J. Bonding of soft tissues using a novel tissue adhesive consisting of a citric acid derivative and collagen. Mater. Sci. Eng: C 2004, 24, 775−780. (23) Briscoey, B. J.; Fiori, L.; Pelillo, E. J. Nano-indentation of polymeric surfaces. Phys. D: Appl. Phys. 1998, 31, 2395−2405. (24) Cui, W.; Wang, Q.; Chen, G.; Zhou, S.; Chang, Q.; Zuo, Q.; Ren, K.; Fan, W. Repair of articular cartilage defects with tissueengineered osteochondral composites in pigs. J. Biosci. Bioeng. 2011, 111, 493−500. (25) Albanna, M. Z.; Bou-Akl, T. H.; Blowytsky, O.; Walters, H. L.; Matthew, H. W. Chitosan fibers with improved biological and mechanical properties for tissue engineering applications. J. Mech. Behav. Biomed. Mater. 2013, 20, 217−226. (26) Pati, F.; Adhikari, B.; Dhara, S. Development of chitosantripolyphosphate fibers through pH dependent ionotropic gelation. Carbohydr. Res. 2011, 346, 2582−2588. (27) José-Moura, M.; Figueiredo, M. M.; Gil, M. H. Rheological Study of Genipin Cross-Linked Chitosan Hydrogels. Biomacromolecules 2007, 8, 3823−3829. (28) Yanez, J. A.; Shikata, T.; Lange, F. F.; Pearson, D. S. Shear Modulus and Yield Stress Measurements of Attractive Alumina Particle Networks in Aqueous Slurries. J. Am. Ceram. Soc. 1996, 79, 2917−2921. (29) Den-Buijs, J. O.; Ritman, E. L.; Dragomir-Daescu, D. Validation of a Fluid-Structure Interaction Model of Solute Transport in Pores of Cyclically Deformed Tissue Scaffolds. Tissue Eng., Part C 2010, 16, 1145−1156.

measurements revealed a higher modulus of the gels with increasing cross-link density. Additionally, the hardness, resistance to plastic deformation, and plasticity index also increased with increasing cross-linking densities of the gels. From the plasticity index values, it is inferred that highly crosslinked gels exhibit higher plasticity whereas gels with a lower degree of cross-linking tend to be more viscoelastic. Furthermore, in vivo studies indicated that gels with the strain-reversibility property supported early osteochondral regeneration.



ASSOCIATED CONTENT

S Supporting Information *

Chemical characterization of ionically and covalently crosslinked citrate−chitosan gels, a table showing the histomorphological scoring criteria, and in vitro cytocompatibility tests of citrate−chitosan gels with different cross-linking densities. This material is available free of charge via the Internet at http:// pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS A fellowship from the Council of Scientific and Industrial Research (CSIR), Government of India, New Delhi, is acknowledged for P.G. and A.P.R. Financial aid from the Defence Research & Development Organisation (DRDO) and CSIR, Government of India, New Delhi, is acknowledged.



REFERENCES

(1) Amendola, V.; Meneghetti, M. Self-healing at the nanoscale. Nanoscale 2009, 1, 74−88. (2) Fantner, G. E.; Oroudjev, E.; Schitter, G.; Golde, L. S.; Thurner, P.; Finch, M. M.; Turner, P.; Gutsmann, T.; Morse, D. E.; Hansma, H.; Hansma, P. K. Sacrificial Bonds and Hidden Length: Unraveling Molecular Mesostructures in Tough Materials. Biophys. J. 2006, 90, 1411−1418. (3) Levental, I.; Georgesa, P. C.; Janmey, P. A. Soft biological materials and their impact on cell function. Soft Matter 2006, 3, 299− 306. (4) Gulyuza, U.; Okay, O. Self-healing polyacrylic acid hydrogels. Soft Matter 2013, 9, 10287−10293. (5) Thompson, J. B.; Kindt, J. H.; Drake, B.; Hansma, H. G.; Morse, D. E.; Hansma, P. K. Bone indentation recovery time correlates with bond reforming time. Nature 2001, 414, 773−776. (6) Seror, J.; Merkher, Y.; Kampf, N.; Collinson, L.; Day, A. J.; Maroudas, A.; Klein, J. Articular Cartilage Proteoglycans As Boundary Lubricants: Structure and Frictional Interaction of Surface-Attached Hyaluronan and Hyaluronan-Aggrecan Complexes. Biomacromolecules 2011, 12, 3432−3443. (7) Peppas, N. A.; Huang, Y.; Torres-Lugo, M.; Ward, J. H.; Zhang, J. Physicochemical foundations and structural design of hydrogels in medicine and biology. Annu. Rev. Biomed. Eng. 2000, 2, 9−29. (8) Wang, Q.; Mynar, J. L.; Yoshida, M.; Lee, E.; Lee, M.; Okuro, K.; Kinbara, K.; Aida, T. High-water-content mouldable hydrogels by mixing clay and a dendritic molecular binder. Nature 2010, 463, 339− 343. (9) Gasnier, A.; Royal, G.; Terech, P. Metallo-Supramolecular Gels Based on a Multitopic Cyclam Bis-Terpyridine Platform. Phys. Chem. Chem. Phys. 2013, 15, 7338−7344. 8450

dx.doi.org/10.1021/la500698v | Langmuir 2014, 30, 8442−8451

Langmuir

Article

(30) McDonald, S. J.; Dooley, P. C.; McDonald, A. C.; Schuijers, J. A.; Ward, A. R.; Grills, B. L. Transient expression of myofibroblast-like cells in rat rib fracture callus. Acta Orthop. 2012, 83, 93−98. (31) Kinner, B.; Gerstenfeld, L. C.; Einhorn, T. A.; Spector, M. Expression of smooth muscle actin in connective tissue cells participating in fracture healing in a murine model. Bone 2002, 30, 738−745. (32) Duncan, R. L.; Turner, C. H. Mechanotransduction and the functional response of bone to mechanical strain. Calcif. Tissue Int. 1995, 57, 344−358. (33) Simmons, C. A.; Matlis, S.; Thornton, A. J.; Chen, S.; Wang, C. Y.; Mooney, D. J. Cyclic strain enhances matrix mineralization by adult human mesenchymal stem cells via the extracellular signal-regulated kinase (ERK1/2) signaling pathway. J. Biomech. 2003, 36, 1087−1096. (34) Arnsdorf, E. J.; Tummala, P.; Kwon, R. Y.; Jacobs, C. R. Mechanically induced osteogenic differentiation–the role of RhoA, ROCKII and cytoskeletal dynamics. J. Cell Sci. 2009, 122, 546−553. (35) Miquelard-Garnier, G.; Creton, C.; Hourdet, D. Strain induced clustering in polyelectrolyte hydrogels. Soft Matter 2008, 4, 1011− 1023. (36) Menèík, J. Nanoindentation in Materials Science; InTech, 2012; doi: 10.5772/2829, published online. (37) Dreher, M. L.; Nagaraja, S.; Bui, H.; Hong, D. Characterization of load dependent creep behavior in medically relevant absorbable polymers. J. Mech. Behav. Biomed. Mater. 2014, 29, 470−479. (38) Wornyo, E.; Gall, K.; Yang, F.; King, W. Nanoindentation of shape memory polymer network. Polymer 2007, 48, 3213−3225. (39) Choudhary, N.; Kaur, D. Effect of Ti addition on the structural, mechanical and damping properties of magnetron sputtered Ni-Mn-Sn ferromagnetic shape memory alloy thin films. J. Phys. D: Appl. Phys. 2012, 45, 495304. (40) Giannoudisa, P. V.; Einhorn, T. A.; Marsh, D. Fracture healing: the diamond concept. Injury 2007, 38, S3−S6. (41) Marsell, R.; Einhorn, T. A. The biology of fracture healing. Injury 2011, 42, 551−555. (42) Bax, B. E.; Wozney, J. M.; Ashhurst, D. E. Bone Morphogenetic Protein-2 Increases the Rate of Callus Formation after Fracture of the Rabbit Tibia. Calcif. Tissue Int. 1999, 65, 83−89. (43) Schindeler, A.; McDonald, M. M.; Bokko, P.; Little, D.G. Bone remodeling during fracture repair: The cellular picture. Semin. Cell Dev. Biol. 2008, 19, 459−466. (44) Aviv-Gavriel, M.; Garti, N.; Füredi-Milhofer, H. Preparation of a Partially Calcified Gelatin Membrane as a Model for a Soft-to-Hard Tissue Interface. Langmuir 2013, 29, 683−689. (45) Kerin, A. J.; Coleman, A.; Wisnom, M. R.; Adams, M. A. Propagation of surface fissures in articular cartilage in response to cyclic loading in vitro. Clin. Biomech. 2003, 18, 960−968. (46) Serink, M. T.; Nachemson, A.; Hansson, G. The effect of impact loading on rabbit knee joints. Acta Orthop. Scand. 1977, 48, 250−262. (47) Leong, P. L.; Morgan, E. F. Measurement of fracture callus material properties via nanoindentation. Acta Biomater. 2008, 4, 1569− 1575.

8451

dx.doi.org/10.1021/la500698v | Langmuir 2014, 30, 8442−8451