Influence of residual composition on the structure and properties of extracellular matrix derived hydrogels

Materials Science and Engineering C 79 (2017) 793–801

Contents lists available at ScienceDirect

Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec

Influence of residual composition on the structure and properties of extracellular matrix derived hydrogels Jesús A. Claudio-Rizo a,b,c, Magdalena Rangel-Argote a,b, Laura E. Castellano a, Jorge Delgado a, José L. Mata-Mata b,⁎, Birzabith Mendoza-Novelo a,⁎ a b c

Departamento de Ingenierías Química, Electrónica y Biomédica, DCI, Universidad de Guanajuato, Loma del Bosque 103, 37150 León, Gto., Mexico Departamento de Química, DCNE, Universidad de Guanajuato, Noria alta s/n, 36050 Guanajuato, Gto., Mexico Ingeniería en Biotecnología, Universidad Politécnica de Pénjamo, Carretera Irapuato-La Piedad Km 44, 36921, Pénjamo, Gto., Mexico

a r t i c l e

i n f o

Article history: Received 14 December 2016 Received in revised form 15 May 2017 Accepted 16 May 2017 Available online 18 May 2017 Keywords: Extracellular matrix Hydrogel Type I collagen Biochemical composition

a b s t r a c t In this work, hydrolysates of extracellular matrix (hECM) were obtained from rat tail tendon (TR), bovine Achilles tendon (TAB), porcine small intestinal submucosa (SIS) and bovine pericardium (PB), and they were polymerized to generate ECM hydrogels. The composition of hECM was evaluated by quantifying the content of sulphated glycosaminoglycans (sGAG), fibronectin and laminin. The polymerization process, structure, physicochemical properties, in vitro degradation and biocompatibility were studied and related to their composition. The results indicated that the hECM derived from SIS and PB were significantly richer in sGAG, fibronectin and laminin, than those derived from TAB and TR. These differences in hECM composition influenced the polymerization and the structural characteristics of the fibrillar gel network. Consequently, the swelling, mechanics and degradation of the hydrogels showed a direct relationship with the remaining composition. Moreover, the cytocompatibility and the secretion of transforming growth factor beta-1 (TGF-β1) by macrophages were enhanced in hydrogels with the highest residual content of ECM biomolecules. The results of this work evidenced the role of the ECM molecules remaining after both decellularization and hydrolysis steps to produce tissue derived hydrogels with structure and properties tailored to enhance their performance in tissue engineering and regenerative medicine applications. © 2017 Published by Elsevier B.V.

1. Introduction The extracellular matrix (ECM) is the non-cellular component present within all tissues and organs that provides not only essential physical scaffolding for the cellular constituents but also initiates crucial biochemical and biomechanical cues, which are required for tissue morphogenesis, differentiation and homeostasis [1–2]. This matrix is composed of a variety of proteins and polysaccharides that are locally secreted and assembled into an organized network in close association with the surface of the cell that produced them [3]. Several studies have reported the preparation of hECM based on porcine dermis [4], bovine pericardium [5], porcine urinary bladder [6] and rat tail tendon [7] to mimic the structure and function of the ECM ex vivo. The polymerization of the hECM has allowed to develop biomedical hydrogels, which have been commonly used to create three-dimensional (3D) culture systems, to investigate the cell-ECM interactions and recently to build scaffolds for tissue engineering (TE) [5,8–9]. In this respect, several ⁎ Corresponding authors at: Loma del Bosque 103, C.P. 37150 León, Gto., Mexico. E-mail addresses: [email protected] (J.L. Mata-Mata), [email protected] (B. Mendoza-Novelo).

http://dx.doi.org/10.1016/j.msec.2017.05.118 0928-4931/© 2017 Published by Elsevier B.V.

advances have been made in the hydrogel design to deliver cells and biomolecules. This has allowed to get a more effectively harness of the cell–material interactions in regulation of cell fate and functions, and to modulate the environment of both normal and injured/diseased tissues towards regeneration [10]. Ideally, the hECM and their derivatives would retain the bioactivity associated with the ECM, and thus, the appropriate preparation and characterization of hECM will avoid purification steps that could remove the biomolecules present in the native ECM. Moreover, the choice of raw material tissue and the standardization of the hECM preparation could expand the biomedical utility of the ECM hydrogels. It is widely accepted that biomimetic hydrogels should include cell induction ligands such as growth factors and other biomolecules that can be delivered from injectable hydrogel systems [10]. The in situ gelation of hECM could be explored as a method that deliver definite and precise signals in an appropriate spatial and temporal manner [10–11]. To study the influence of the ECM composition on the properties of biomedical hydrogels, we started from the hypothesis that the tissue source and hECM composition could determine the polymerization kinetics (hydrogel formation), the structure and properties of the hydrogels obtained using hECM. With this in mind, four animal tissues

794

J.A. Claudio-Rizo et al. / Materials Science and Engineering C 79 (2017) 793–801

with differences in native composition in their (decellularized) lyophilized state were chosen, i.e., rat tail tendon (TR), bovine Achilles tendon (TAB), porcine small intestine submucosa (SIS) and bovine pericardium (PB). PB and SIS can be considered as ECM biomolecules-rich tissues [12–14], such as sulphated glycosaminoglycans (sGAG), fibronectin and laminin. Albeit it is recognized that TAB and TR have as a main component the type I collagen [15]. The hECM were obtained from the decellularized tissues using acid hydrolysis assisted by pepsin. The hECM compositions, the polymerization kinetics, viscoelastic properties, degradation, swelling, microstructure and in vitro biocompatibility were also evaluated. Ultimately, these properties were related to the hECM source and their composition. This work represents an essential stage in the design and standardization of hECM formulations that could enhance the biomedical applications of ECM hydrogels. 2. Experimental section 2.1. Materials The PB, SIS and TAB tissues were acquired in local slaughterhouses. The TR were obtained from the University of Guanajuato. Pepsin, type I collagenase, ethylenediaminetetraacetic acid (EDTA), 3-(4,5-dimethyl thiazol yl)-2,5-diphenyltetrazolium bromide (MTT), 2,2-dihydroxy-1,3indanedione (ninhydrin), t-octylphenoxypolyethoxyethanol (Triton X100) and other salts were purchased from Sigma-Aldrich. The LIVE/ DEAD® kit to assess viability/cytotoxicity and tissue extraction reagent I were purchased from Invitrogen. Human laminin and fibronectin, and human/mouse TGF beta-1 ELISA kits were purchase from eBioscience. 2.2. Decellularization of the tissues Approximately 4 g of tendons were excised from two adult Wistar rat tails and from one bovine leg. The TR and TAB tissue specimens were washed with fresh water to remove blood and surrounding muscular tissue. Approximately 5 g of SIS or PB were obtained from six 3 × 15 cm specimens from one intestinal tissue segment or from one 4 × 7 cm specimen from one pericardial sac. The SIS and TB tissue specimens were washed with fresh water to remove blood and surrounding muscular/fat tissue. All four tissue specimens were treated as follows: firstly, they were washed (3 ×, PBS). Then, they were left for 2 h in 10 mL of absolute ethanol under orbital stirring (OS), followed by a washing with 50 mL of 1% Triton × 100 (m/v) and 0.5% (m/v) EDTA (1 h for TR, and 18 h for TAB, SIS or PB; room temperature/RT, OS). After, the tissue specimens were rinsed (3×, 30 min, 50 mL of PBS). Subsequently, 50 mL of 10 mM Tris-HCl containing 0.5% (m/v) EDTA, 2.5 mg mL−1 RNase and 0.05 mg mL−1 DNase was added to the tissues (1 h for TR, 6 h for TAB, and 24 h for SIS or PB; RT, OS). Finally, the tissues were rinsed (30 min, 3×, PBS, RT, OS).

this, 6 mg of samples were suspended in TER (24 h, 48 °C) in the presence of protease inhibitors (cOmplete™ ULTRA Tablets, Protease Inhibitor Cocktail), then they were homogenized using a pestle (5 min), centrifuged, aliquoted, and frozen. The DNA content in both all four hECM and native tissue extracts was evaluated using the purification kit Promega Wizard® genomic DNA. DNA was isolated, purified and rehydrated. The samples were evaluated in triplicate. The DNA concentration was determined by spectrophotometric measurements at 260 nm (Thermo Scientific MultiSkan Go) and expressed in nanograms per milliliter of hECM or native tissue extracts. The sGAG content in all four hECM and native tissue extracts was evaluated using a dye test with 1,9-dimethyl-methylene blue [16]. The sGAG concentration was determined by spectrophotometric measurements at 540 nm using chondroitin sulfate as standard (5–50 μg mL−1) and expressed in nanograms per milliliter of hECM or native tissue extracts. The laminin and fibronectin content in all four hECM and native tissue extracts was evaluated by enzyme-linked immunosorbent assays (ELISA) kits (ab108847 and ab119599; Abcam), according to the seller's specifications. Briefly, 5 μL of standard (0.004–1.0 μg mL−1 for fibronectin, 0.156–5 μg mL−1 for laminin) or sample were plated in triplicate. The concentrations of fibronectin and laminin were derived from the standard curve and expressed in picograms per milliliter of hECM or native tissue extracts. The technique of polyacrylamide gel electrophoresis (SDS-PAGE) was used to ensure the purity and molecular composition of type I collagen presents in the hECM. Commercial collagen standards (Sigma-Aldrich) and 7% polyacrylamide gels were used for separating proteins. The analysis was performed using a voltage of 50 V and run for 5 h. Finally, the gels were stained with 0.25% Coomassie blue R-250 solutions and destained with methanol/acetic acid (90:10).

2.5. Polymerization of hECM The hECM derived from TR, TAB, SIS and PB were polymerized under identical reaction conditions to produce 3D hydrogels or matrices. For this, 1 mL of hECM was placed in the ice bath and neutralized with 150 μL of PBS 10× (0.20 M ionic strength and pH 7.4). The neutral solutions of hECM were molded inside of 24-well plates. They were sealed with parafilm and incubated at 37 °C for 24 h to obtain the respective ECM hydrogels. Fig. 1 illustrates the steps in the gel preparation.

2.3. Hydrolysis of the tissues The decellularized SIS, PB and TAB tissues were hydrolysed in 0.01 M HCl solution containing pepsin (1 mg mL−1) (48 h, RT, OS). The hydrolysis of TR was carried out without pepsin. After that, all four hECM were stirred during additional 48 h at 4 °C to reach a complete tissue hydrolysis. The total protein concentration was determined by the BCA assay (bicinchoninic acid, ThermoScientific). The hECM solutions with 6 mg mL−1 of total protein were stored at 4 °C. 2.4. Biochemical composition analysis of hECM The composition analysis was carried out in either hydrolysates or extracts from decellularized or native tissue samples, respectively. To evaluate the composition of the native tissue samples, components were extracted from trimmed samples with the Tissue Extraction Reagent I (TER; Invitrogen) according to the seller's specifications. For

Fig. 1. Outline of the study objectives, indicating the steps in the preparation of the hECM with different composition, the subsequent polymerization at physiological conditions, and its effect on the ECM hydrogel properties.

J.A. Claudio-Rizo et al. / Materials Science and Engineering C 79 (2017) 793–801

2.6. Turbidimetric analysis The gelation kinetic was evaluated by turbidimetry using a Thermo Scientific MultiSkan Go UV–Vis spectrophotometer. The experiments were performed in triplicate. For each experiment, 200 μL of a neutral solution of hECM was added in a 96-well plate pre-heated to 37 °C for 40 min in the spectrophotometer heater. The absorbance at 405, 458, 515, 543 and 633 nm was measured to determinate the collagen fiber diameters (d) and mass-length ratios (μ) in the different formulations by the dependence of the turbidity with the wavelength [17]. 2.7. Evaluation of the concentration of α-amino groups (–NH2) Change [NH2] over gelation time in all four hydrogels was assessed by a ninhydrin test. The ninhydrin solution (1 mL, 1 wt%, citrate buffer, pH 5.0) was added to each ECM hydrogel. Then, samples were heated to 90 °C for 30 min. The mixture was cooled down to room temperature and diluted with 3 mL of distilled water. The absorbance of the solution was determined at 567 nm. The [NH2] was calculated using the molar extinction coefficient (ε = 1640 mL mmol−1 cm−1). 2.8. Evaluation of the chemical structure and thermal stability The chemical structure of the ECM hydrogel was assessed by infrared spectroscopy (ATR-FTIR, Spectrum One, Perkin Elmer). The ATRFTIR spectra were recorded on air-dried membranes at 16 cm−1 of resolution in a range from 4000 to 650 cm−1, using an average of 32 scans. On the other hand, lyophilized ECM hydrogels were placed in aluminum pans, sealed hermetically and heated from 5 to 100 °C with a heating rate of 10 °C min−1 using a differential scanning calorimeter (DSC, Q 200, TA Instruments). The denaturation temperature (Td, peak temperature) and denaturation enthalpy (ΔH) of the endothermic peak were calculated from the first DSC traces. 2.9. Rheological measurements The storage (G′) and loss (G″) moduli of each ECM hydrogel were obtained in small-amplitude oscillatory shear flow experiments using a HR-3 Discovery Hybrid Rheometer (TA Instruments) at 37 °C in a cone-plate geometry with diameter of 40 mm and cone angle of 0.5 deg.; with a solvent trap to prevent water evaporation. All experiments were performed using 10% of strain to ensure the linearity of the dynamic response. Moduli were measured as a function of the frequency (range 0.1–100 Hz). 2.10. Morphology observation The morphology of lyophilized gels (sponges) was characterized by an environment scanning electron microscope (ESEM, Fei Quanta 200). Before observation, the samples were frozen in liquid nitrogen, broken to get internal fragments, and finally attached to carbon tape. The pore size of each specimen was measured from SEM images using the Image Pro plus 6.0 software. 2.11. Swelling and degradation analysis Air-dried membranes derived from ECM hydrogels were immersed in distilled water at 25 °C. Then, weight changes over incubation time were assessed. The water uptake capacity of the specimens was reported as the percentage of the weight difference between swollen materials and air-dried materials. The in vitro degradation of hydrogels was performed using enzymatic (type I collagenase) or oxidizing (H2O2/ Co(II)) conditions. The hydrogels were incubated in collagenase solution (3 mL, 25 °C, activity of 75 U per each gel) or in oxidizing solution (3 mL, 25 °C, 5% v/v H2O2, 10 mM CoCl2 ∙ 6H2O), dipped into filter paper to remove the liquid excess and weighed at different time lapse.

795

The degradation was expressed as the percentage of the weight difference of the samples before and after collagenase/oxidizing incubation. 2.12. Biocompatibility studies The metabolic activity of mouse RAW-264.7 macrophages or rat dermal fibroblasts (derived from primary culture) on each ECM gel was assessed. The cells were seeded on plastic culture dishes and incubated in a humidified atmosphere of 95% air and 5% CO2 at 37 °C using culture medium (RPMI for macrophages and DMEM for fibroblasts) supplemented with fetal bovine serum (FBS, 10%). The cell suspension (30,000 cells/gel) was added to the wells containing hydrogels and control wells (free of materials), and cultured for 1, 3 and 7 days. In the first instance, the cell viability was determined by the ability of cells to reduce MTT salts in water-insoluble formazan crystals. The MTT solution was added to sample-containing wells and cells were maintained under culture conditions for 3 h. Then, the medium was decanted, the blue formazan crystals were dissolved in 2-propanol and the absorbance of the supernatants was measured at 540 nm. The absorbance of MTT reduced by cells cultured in wells free of materials represents the 100% of metabolic activity (controls). To confirm the viability results, the cells were incubated with calcein-AM and ethidium homodimer-1 containing fluorescence reagents, then samples were taken with sterile forceps, placed carefully on cover slips and observed under a fluorescence microscope (AxiosKop40, Carl Zeiss) to record the green live cells and the red dead cells. The production of TGF-β1 by macrophages was measured in culture medium supernatants (after 1 and 3 days) using ELISA kits (eBioscience) according to manufacturer's specifications. The production of cytokines was normalized with respect to the cell number. 2.13. Statistical analysis All experiments were independently carried out at least three times. Mean and standard deviation (SD) are presented for each data set. Data sets were compared using analysis of variance (ANOVA). The difference of the means was checked with a Sidak-Holm post-hoc test and was considered statistically significant at level p b 0.05. 3. Results 3.1. Biochemical composition of hECM Fig. 2 shows the residual content of DNA, sGAG, laminin and fibronectin in all four hECM. The results indicated a higher concentration of residual DNA in hECM derived from SIS and PB than those derived from TR or TAB (Fig. 2a). The DNA reduction efficiency was 99, 98, 91 and 90% for hECM derived from TR, TAB, SIS and PB, respectively. In the case of sulphated polysaccharides and matrix glycoproteins, the results showed that the hECM derived from SIS contained the highest content of sGAG and laminin (Fig. 2b and d), while the hECM obtained from PB contained the highest fibronectin content (Fig. 2c). The reduction of laminin in all four hECM was close to 93% respect to amount extracted from native tissues, while preservation of sGAG or fibronectin in hECM was 80, 93, 89 and 92%, or 50, 59, 72 and 88% for TAB, TR, SIS and PB, respectively. The electrophoretic analysis of hECM revealed the typical banding patterns of type I collagen. The polypeptide chains of different molecular weight corresponding to α1 (I) (128 kDa), α2 (I) (112 kDa) and β (250–230 kDa) were observed in each hECM (Fig. 2e). 3.2. Polymerization and 3D network parameters of the ECM hydrogel Fig. 3 shows the results of the hECM polymerization, which were used to generate 3D matrices in hydrogel, were evaluated by both turbidimetry and amino groups quantification. The turbidimetric analysis

796

J.A. Claudio-Rizo et al. / Materials Science and Engineering C 79 (2017) 793–801

Fig. 2. Quantification of DNA (a), sGAG (b), fibronectin (c) and laminin (d) in decellularized hECM and native tissue extracts. (e) SDS-PAGE analysis indicating the bands of type I collagen in all four hECM. TR, TAB, SIS, and PB indicate the ECM hydrogels obtained from hECM of rat tail tendon, bovine Achilles tendon, porcine small intestinal submucosa, and bovine pericardium, respectively. Where MW are molecular weight markers. Data are expressed as mean values ± SD, n = 3. The difference of the means is significant (p b 0.05) among all marked groups using lines in the upper part of each graph (gray and black lines correspond to white and gray bars, respectively).

showed the sigmoidal polymerization curves featuring the stages of nucleation (phase lag), growth and plateau in all four hECM (Fig. 3a). Table 1 shows the lag time (tlag), polymerization rate during the growth phase (S), and average polymerization time (t1/2) during the polymerization process, as well as the fiber diameter (d) and the mass-length ratio (μ) of the 3D fibrillar matrices formed with each type of hECM. Interestingly, each hECM had different polymerization kinetic, suggesting differences in the fiber formation and fibrillar interactions. The lowest t1/2 corresponded to TR (statistically significant with respect to SIS and PB, p b 0.05), whereas the highest t1/2 was found for PB. The tlag followed the same trend. The highest S corresponded to TR, and the slowest S to PB. The collagen fibers formed from hECM derived from PB showed the largest diameters (statistically significant compared to TAB, TR and SIS, p b 0.05). The smallest fiber diameter was found in TR. The mass-length ratio of the fibrillar matrices followed the same trend. Fig. 3b shows the appearance of the hydrogels obtained from all four hECM after polymerization and aged during 24 h. The TR, TAB and SIS hydrogels appeared to

be more transparent, while the PB hydrogels appeared to be more opaque. Fig. 3c shows the change of the concentration of primary amines over polymerization time of all four hECM. The kinetic curves showed that the concentration of primary amines decreased until a plateau stage was reached. This plateau stage was achieved after 50, 60, 78 and 80 min of polymerization for TR, TAB, SIS and PB respectively. Based on these curves, the diminution in the concentration of free αamino groups from t = 0 to the plateau time was 0.76, 0.79, 0.84 and 0.90 mmol mL−1 for TR, TAB, SIS and PB respectively. 3.3. Chemical structure and thermal stability of ECM hydrogels Fig. 4a shows ATR-FTIR spectra of ECM materials in the lyophilized state. The signals of amide I (1700–1600 cm− 1, associated with the stretching vibration of C_O), amide II (1500–1400 cm−1, associated with the bending vibration of N\\H and stretching vibration of C\\N)

Fig. 3. Polymerization of hECM evaluated by changes to (a) absorbance and (c) free primary amine over gelation time. (b) The appearance of all four hydrogels. TR, TAB, SIS, and PB indicate the ECM hydrogels obtained from hECM of rat tail tendon, bovine Achilles tendon, porcine small intestinal submucosa, and bovine pericardium, respectively. Data are expressed as mean values ± SD, n = 3.

J.A. Claudio-Rizo et al. / Materials Science and Engineering C 79 (2017) 793–801 Table 1 Results for the turbidimetric kinetic study of hECM. hECM Kinetic parameter tlag/min t1/2/min S/min−1 d/nm μX10+10/ Da cm−1

TR

TAB

SIS

PB

10 ± 1d 15 ± 2d 0.017 ± 0.003d 126 ± 12d 106 ± 9d

15 ± 2 21 ± 3 0.015 ± 0.002c,d 133 ± 14b,d 109 ± 7b,d

17 ± 2c 25 ± 1c 0.011 ± 0.003c 142 ± 18c,d 114 ± 12c,d

18 ± 3 26 ± 3 0.009 ± 0.003 153 ± 16 122 ± 14

n = 3; data expressed as mean ± SD, the significant difference among groups is expressed as a: vs. TAB, b: vs. SIS, c: vs. TR, d: vs. PB at p b 0.05.

and amide III (1350–1250 cm−1, associated with the secondary conformation of collagen) were observed. Other signals associated with the bonds C\\H (2850–3000 cm−1), N\\H (3200–3600 cm−1) and O\\H (3360–3320 cm−1) associated with the collagen aminoacid skeletons were also detected. Some differences in the intensities of the signals of N\\H and O\\H were noticed, where the PB materials showed a clearer decrease in these intensities compared to the other materials. The absorption band corresponding to the carboxylic acid functional group (–COOH 1820–1760 cm− 1) was also observed in all four materials. The vibrational bands of proline were observed in the region of 1000– 850 cm−1, which is the characteristic region of the collagen α-helix structure. These bands were more intense in the PB materials. Fig. 4b shows the endothermic peaks associated with the fusion of type I collagen, where the collagen molecules transited from a triple helical conformation to a single-strand chain structure (gelatin). The peak temperature, so called denaturation temperature, was 49, 49.5, 48 and 52 °C for TR, TAB, SIS and PB, respectively. The denaturation enthalpy was 39, 42, 44 and 62 J/g for materials based on TR, TAB, SIS and PB, respectively. 3.4. Viscoelasticity and microstructure of ECM matrices Fig. 5a shows the change storage modulus (G′) with respect to the oscillation frequency in all four ECM hydrogels. The gel behavior was observed for all four ECM hydrogels, being that G′ was always greater than the viscous modulus (G″) (Fig. 5b). The PB hydrogels showed the highest G′ values. The G′ for SIS and TAB did not show significant difference between them, however, TR recorded the lowest G′ values. The SEM analysis showed that matrices produced by polymerization of all four hECM had a microstructure featured by porous networks. This microstructure is expected due to the entanglement and crosslinking of collagen fibers, which produced lamellar structures

797

with interconnected pores (Fig. 5c–f). Both TR and TAB matrices showed predominantly fiber-fiber junctions, with poor fibrillar entanglement and larger pores (Fig. 5c and d). On the other hand, the matrices based on SIS and PB (Fig. 5e and f) showed lamellar structures with interconnected porosity resulted of a high fibrillary entanglement. The average pore size was 507 ± 55, 480 ± 36, 302 ± 72 and 267 ± 48 μm, for matrices derived from TR, TAB, SIS and PB, respectively. 3.5. Water uptake and in vitro degradation of ECM hydrogels Fig. 6a shows that the ECM materials could absorb different amounts of water with respect to incubation time in PBS. After 1–2 days, the matrices did not absorb significant amounts of water. However, the matrix based on TAB absorbed a maximum capacity of water after 3 days, while those based on PB, SIS and TR did it after 9 days. Furthermore, the highest water absorption was observed in materials derived from PB and TAB, while the lowest water uptake was recorded on TR. Fig. 6b and b show the change of material mass over incubation time in presence of oxidizing agents (H2O2/Co(II)) and type I collagenase, respectively. The degradation under oxidizing medium resulted to be quite aggressive for ECM hydrogels. All four ECM hydrogels completely lost their mass after 60 min under the action of the oxidizing agent (Fig. 6b). However, the degradation under proteolytic conditions indicated that the hydrogels based on TR and TAB lost 50% of their initial mass at 2 h, reaching a complete degradation at 8 h and 12 h, respectively. A higher degradation resistance to collagenase was observed in hydrogels derived from SIS and PB. Both materials lost 75% of their initial mass after 8 h, and they were completely degraded after 14 h (SIS) or 16 h (PB). 3.6. In vitro biocompatibility of ECM hydrogels Fig. 7 shows the results of the in vitro biocompatibility of the all four ECM hydrogels in terms of cell viability, cell metabolic activity, and cytokine secretion mediated by the cell–material interactions. Viable macrophages, able to transform the calcein AM into fluorescent calcein by the intracellular esterase activity, were observed on the surface of all four ECM hydrogels after 1-day culture (Fig. 7a). The highest number of viable cells was observed in hydrogels derived from SIS. Moreover, dead cells (stained in red) was not observed in any ECM hydrogels. The metabolic activity of fibroblasts (Fig. 7b) or macrophages (Fig. 7c), analyzed by the enzymatic conversion of tetrazolium salts, was higher in hydrogels based on SIS after 1-day culture, with statistical significance respect to PB. After 3 and 7 days of culture, the metabolic activity was gradually decreased respect to control (cells without hydrogels) in all four ECM hydrogels.

Fig. 4. Chemical structure and thermal stability of ECM hydrogels assessed by (a) ATR-FTIR signals and (b) DSC thermograms, respectively. TR, TAB, SIS, and PB indicate the ECM hydrogels obtained from hECM of rat tail tendon, bovine Achilles tendon, porcine small intestinal submucosa, and bovine pericardium, respectively.

798

J.A. Claudio-Rizo et al. / Materials Science and Engineering C 79 (2017) 793–801

Fig. 5. Viscoelastic properties of ECM hydrogels assessed by oscillatory rheology, indicating storage (G′, a) and loss (G″, b) moduli as a function of frequency. TR, TAB, SIS, and PB indicate the ECM hydrogels obtained from hECM of rat tail tendon, bovine Achilles tendon, porcine small intestinal submucosa, and bovine pericardium, respectively. The microstructure of lyophilized materials assessed by SEM. Representative SEM micrographs of materials derived from TR (c), TAB (d), SIS (e) and PB (f).

Fig. 7d shows the secretion of TGF-β1, involved in the angiogenesis process, by RAW264.7 macrophages cultured in each ECM hydrogel. After 1-day culture, the secretion of TGF-β1 was higher in hydrogels derived from SIS and PB. The production of TGF-β1 increased in all four ECM hydrogels after 3 days of culture. The macrophages stimulated with hydrogels derived from SIS had the highest TGF-β1 secretion, being statistically significant compared to the control and TR. 4. Discussions The components remaining in hECM derived from TAB, SIS, TR and PB tissues can be determinants for engineering the structure, properties and biological behavior of the resulting gels. The amount of residual DNA in all four hECM was proportional to the DNA content in the native animal tissue (Fig. 2a). It has been reported that the presence of b10% of the original DNA could reduce or inhibit the xenogeneic immune responses in the hECM [18]. Also, the results indicated that laminins and sGAG, a proteoglycan component, are retained in similar proportions in all four hECM, i.e., SIS N PB N TAB N TR (Fig. 2b and d). These macromolecules remaining after ECM decellularization and hydrolysis are the two most abundant components of the basal lamina and are keys in the structure and properties of the ECM [16,19]. Instead, hECM derived from PB appears to retain the highest amount of fibronectin (Fig. 2c), a glycoprotein presents in both basal membranes and connective tissue with a key role in the cell adhesion and migration [20]. The tissue

decellularization using nonionic detergents as Triton × 100 does not produce structural changes in the ECM and/or significant loss of its native components [21]. Besides, other studies about the standardization for preparing hECM indicated that the hydrolysis step becomes decisive in the stability and permanence of adhesion proteins, as they can lose their structure and function [21]. The type I collagen composition in all four hECM was verified by the identification of the band α1 (I) (128 kDa) and the band α2 (I) (112 kDa). These two bands associated with polypeptides in triple helical configuration are associated with the formation of collagen fibers (polymerization) in vitro [22,23]. The identification of these components in the hydrolysates ensures the bioactivity associated with the native ECM biomolecules. Moreover, the cell responses to the tissue derived hydrogels could be fine-tuned by the differences in the 3D microenvironment generated by the residual composition and structure. All four hECM featured an in vitro polymerization, as studied by turbidimetry. These results showed that hECM with a higher residual content of ECM components, i.e., PB and SIS, exhibited higher time of nucleation and polymerization (Fig. 3a, Table 1). Previously, it was reported that the self-assembly of tropocollagen molecules is influenced by the presence of molecules such as proteoglycans, fibronectin and laminin by generating hydrogen bond interactions, which slows the nucleation stage [24]. This crosslinking mode between collagen and other ECM molecules during the fibrillar growth stage may be responsible for tailoring the 3D structure of the fibrillar gel network [25]. Moreover, the

Fig. 6. Profiles of swelling or degradation after incubation in PBS at 37 °C (a), or in oxidizing (5% v H2O2/CoCl2 10 mM) or enzymatic (type I collagenase, 70 U mg−1 per gel) solutions (c). TR, TAB, SIS, and PB indicate the ECM hydrogels obtained from hECM of rat tail tendon, bovine Achilles tendon, porcine small intestinal submucosa, and bovine pericardium, respectively. Data are expressed as mean values ± SD, n = 3.

J.A. Claudio-Rizo et al. / Materials Science and Engineering C 79 (2017) 793–801

799

Fig. 7. Influence of hECM composition on the biological response to materials. (a) Representative micrographs of macrophages on hydrogels (1 day) stained with green-fluorescent calcein indicating intracellular esterase activity and examined with a fluorescent microscope under a 10× objective. The metabolic activity of fibroblasts (b) and macrophages (c) on hydrogels after 1, 3 and 7 days of culture assessed by the MTT assay. (d) The secretion of TGF-β1 by macrophages after 1 and 3 days of culture. TR, TAB, SIS, and PB indicate the ECM hydrogels obtained from hECM of rat tail tendon, bovine Achilles tendon, porcine small intestinal submucosa, and bovine pericardium, respectively. Data are expressed as mean values ± SD, n = 3. The difference of the means is significant (p b 0.05) between all marked groups using lines in the upper part of each graph. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

self-assembly of other membrane proteins, such as fibronectin [26], can be related to the increase in fiber sizes (d) and the mass-length ratio (μ), for SIS and PB hydrogels (Fig. 2, Table 1). The fibronectin has been associated with collagen fibers generating interpenetrated networks modifying the structure and size of the fibrillar collagen matrix. The sGAG content can be also associated with the structural properties of the polymerized matrix. Previously, it was reported that a high concentration of proteoglycans, PEG and hyaluronic acid during the polymerization of type I inhibited the fibrillar collagen growth and decreased the fiber diameter [27–29]. Herein, the residual content of ECM biomolecules in all four hECM did not present inhibition of collagen polymerization. The decrease in the concentration of free amines can be associated with the processes of fibrillar growth and chemical crosslinking of the matrix components, which are carried out during the collagen polymerization, and they are the basis for generating the 3D fibrillar matrix (Fig. 3c). The N-terminal tropocollagen molecules reacted with the Cterminal tropocollagens by amide bond formation catalyzed by acid (Fig. 8) [30]. According to the polymerization kinetics of all four hECM (Fig. 3c), these processes were performed in b80 min, where the concentration of free amines was notoriously decreased. Later, the concentration of free amines was almost constant; this suggests that the events of collagen fibril growth and/or crosslinking were almost finished. During the maturation of hydrogels, at a longer time than 80 min, occur the physical entanglement processes of the fibers [31]. A lower content of

free amines observed in PB and SIS hydrogels can be related to their richer composition of ECM biomolecules. Therefore, it can be deduced that the PB and SIS fibrillar matrices were more crosslinked and/or entangled, compared to TAB and TR matrices (Fig. 3a and c). Since the PB hydrolysate contained the highest fibronectin concentration, the typical IR bands for the fibrillar structure were increased, which suggests that fibronectin promotes the formation of type I collagen fibers (Fig. 4a). It has been proposed that the conformational changes of fibronectin promote the collagen fibril formation and its conversion into insoluble fibers [32]. A higher collagen denaturation temperature was associated with a higher crosslinking (Fig. 4b). It has been reported that crosslinking and associations of components in interpenetrated networks, i.e., proteoglycans, hyaluronic acid and collagen, are responsible for the increment in the collagen denaturation temperature [33]. Thus, the associations of the ECM components with the collagen affected the polymerization process and enhanced the crosslinking and entanglement of fibers in PB gels (Figs. 4 and 5). These results indicate that the enhanced fiber diameter, mass-length ratio and denaturation temperature were translated in to a significant increase in the storage modulus for the PB gels (Table 1, Fig. 5a). Each type of ECM appeared to produce different fiber-fiber interactions, such as fibrillar binding or fibrillar agglomerates, generating laminar structures with interconnect pores (Fig. 5c–f). For hydrogels derived from TAB and TR, low formation of fibrillar agglomerates was

Fig. 8. Scheme indicating the addition reactions of the collagen molecules to generate a fibrillar matrix.

800

J.A. Claudio-Rizo et al. / Materials Science and Engineering C 79 (2017) 793–801

observed. On the other hand, the materials obtained from SIS and PB evidenced laminar microstructures with interconnected porosity, indicating high formation of fibrillar agglomeration and entanglement, presumably because of their rich composition in ECM elements. The pore size can also be related to crosslinking properties and therefore with its mechanical properties [14,22]. A smaller pore size herein corresponded to PB matrices, which produced the highest fibrillar crosslinking and storage modulus (Figs. 3c and 5a). The diffusion of water into collagen networks was controlled by both electrostatic and hydrophobic interactions that can occur among the matrix components (Fig. 6a). A high content of polar molecules, such as sGAG, increased the water uptake of hydrogels based on collagen type I [34], however, the absorptive capacity can be limited by electrostatic interactions with other proteins [14]. The porosity of the materials did not show a direct relationship with the swelling capacity of the hydrogels. The collagen-based matrices are depolymerized by the action of radicals OH; oxidizing the collagen aminoacids as proline, hydroxyproline, glutamic and aspartic acid [35]. A relationship of the oxidative degradation rate with the structural characteristics and composition of the hydrogels was not detected. This indicates that the oxidative degradation mechanism was not controlled by the composition and microstructure of the materials (Fig. 6b). The type I collagenase is a metalloproteinase that degrades the α-helical regions of the collagen into small fragments; by the decleaving of the specific sites with sequence Pro-X-Gly-Pro. The aggregates and/or associations of matrix components can be related with the entanglements of the collagen fibers, retarding the penetration of the enzyme into the fibrillar matrix [36]. Hydrogels with a higher storage modulus and denaturation temperature had the smaller pore size and a lower proteolytic degradation rate, as observed in PB materials (Figs. 5 and 6c). The direct relationship among the structure parameters set forth herein can be harnessed for predict gel properties by measuring one of these factors. Fibroblasts and macrophages grew in the all four hydrogels under study (Fig. 7a), which is reasonable due to the natural origin and high biocompatibility associated with the ECM [37–40]. Nonetheless, differences in the metabolic activity evaluated by the MTT assay appeared to be associated with the composition of the materials. The hydrogel based on SIS, rich in sGAG, fibronectin and laminin, exhibited the highest metabolic activity of macrophages and fibroblasts after 1 and 3 days of culture (Fig. 7b–c). It has been proposed that the presence of ECM molecules in tissue biomaterials have a positive effect on cell growth and material biocompatibility [41]. The decrease of the metabolic activity of macrophages and fibroblasts observed after 7 days of culture can be associated with degradation events and cell-induced contraction in the materials, which could alter the cellular metabolic activity, showing viable cells but not necessarily showing proliferation [42]. The signaling molecules secreted by macrophages, such as TGFβ1 (Fig. 7d), are commonly monitored to study the cell response related to the cell–material interactions [43]. A sustained production of this cytokine indicates that macrophages can regulate both angiogenesis and vascularization stages of scaffolding [44]. The direct effects of ECM bioscaffolds upon macrophage activation could have implications for the use of site-specific ECM in therapeutic applications [45]. The content of bioactive molecules in materials derived from diverse source tissues can regulate the macrophage phenotypes. The enhanced secretion of TGF-β1 by macrophages cultured in hydrogels derived from SIS, exhibiting a higher content of sGAG and laminin, suggests that solubilized basement membrane components, which promotes cell recognition sites governing the cellular metabolism [19–20,43] are involved in the stimulation of macrophages. Together, these findings suggest that the ECM source and residual composition are determinants to establish the structure/property relationship in ECM hydrogels. There remain several challenges in controlling the structure and properties of these ECM hydrogels to fulfil requirements for the sustained drug delivery, cell delivery and tissue

engineering [46–49]. The selection of the ECM source along with the reassembly of the natural ECM components in the presence of synthetic polymers or inorganic nanoparticles can be explored to enhance the stability, mechanics and biological response of the hydrogels [50–56].

5. Conclusions The results of present study indicate that decellularized tissue hydrolysates keep the ECM soluble molecules sGAG, fibronectin and laminin along with collagen type I, in proportions that depend on the tissue origin. The presence of these biomolecules governs the collagen polymerization, thereby determining the structure and properties of the ECM gel network. In this regard, the swelling, mechanics and degradation are adjusted by the residual composition in ECM hydrogels. Ultimately, the cell cytocompatibility and the secretion of TGF-β1 by macrophages are enhanced in hydrogels with the highest residual content of ECM biomolecules.

Acknowledgement We thank Maria C. Lona for assistance with cell viability assays. This work was supported by a doctoral scholarship from Mexico's National Council of Science and Technology (CONACYT) to J.A.C.R and the grants PDCAPN2015/1310 (CONACYT) and 1146/2016-2017 (University of Guanajuato). References [1] S.F. Badylak, The extracellular matrix as a biologic scaffold material, Biomaterials 28 (2007) 3587–3593. [2] A.D. Theocharis, S.S. Skandalis, C. Gialeli, N.K. Karamanos, Extracellular matrix structure, Adv. Drug Deliv. Rev. 97 (2016) 4–27. [3] A.H. Morris, T.R. Kyriakides, Matricellular proteins and biomaterials, Matrix Biol. 37 (2014) 183–191. [4] M.T. Wolf, K.A. Daly, E.P. Brennan-Pierce, S.A. Johnson, C.A. Carruthers, A. D'Amore, S.P. Nagarkar, S.S. Velankar, S.F. Badylak, A hydrogel derived from decellularized dermal extracellular matrix, Biomaterials 33 (2012) 7028–7038. [5] M. Climov, T. Leavitt, J. Molnar, D. Orgill, Natural biomaterials for skin tissue engineering, in: M. Albanna, J.H. Holmes (Eds.),Skin Tissue Engineering and Regenerative Medicine 2016, pp. 145–161. [6] D.O. Freytes, J. Martin, S.S. Velankar, A.S. Lee, S.F. Badylak, Preparation and rheological characterization of a gel form of the porcine urinary bladder matrix, Biomaterials 29 (2008) 1630–1637. [7] M.B. Bornstein, Reconstituted rat tail collagen used as substrate for tissue cultures on coverslips in maximow slides and roller tubes, Lab. Investig. 7 (1958) 134–137. [8] R.A. Callcut, M.J. Schurr, M. Sloan, L.D. Faucher, Clinical experience with Alloderm: a one-staged composite dermal/epidermal replacement utilizing processed cadaver dermis and thin autografts, Burns 32 (2006) 583–588. [9] S.M. Lien, L.Y. Ko, T.J. Huang, Effect of pore size on ECM secretion and cell growth in gelatin scaffold for articular cartilage tissue engineering, Acta Biomater. 5 (2009) 670–679. [10] W.S. Toh, X.J. Loh, Advances in hydrogel delivery systems for tissue regeneration, Mater. Sci. Eng. C 45 (2014) 690–697. [11] S.B. Seif-Naraghi, M.A. Salvatore, P.J. Schup-Magoffin, D.P. Hu, K.L. Christman, Design and characterization of an injectable pericardial matrix gel: a potentially autologous scaffold for cardiac tissue engineering, Tissue Eng. Part A 16 (2010) 2017–2027. [12] S.T. Kreger, B.J. Bell, J. Bailey, E. Stites, J. Kuske, B. Waisner, S.L. Voytik-Harbin, Polymerization and matrix physical properties as important design considerations for soluble collagen formulations, Biopolymers 93 (2010) 690–707. [13] B. Mendoza-Novelo, D.I. Alvarado-Castro, J.L. Mata-Mata, J.V. Cauich-Rodríguez, A. Vega-González, E. Jorge-Herrero, F.J. Rojo, G.V. Guinea, Stability and mechanical evaluation of bovine pericardium cross-linked with polyurethane prepolymer in aqueous medium, Mater. Sci. Eng. C 4 (2013) 2392–2398. [14] A. Lungu, M.G. Albu, N.M. Florea, I.C. Stancu, E. Vasile, H. Iovu, The influence of glycosaminoglycan type on the collagen-glycosaminoglycan porous scaffolds, Dig. J. Nanomater. Biostruct. 6 (2011) 1867–1875. [15] L.P. Gartner, Textbook of Histology, 4, Elsevier, 2015 134–185. [16] J.P. Hodde, S.F. Badylak, A.O. Brightman, S.L. Voytik-Harbin, Glycosaminoglycan content of small intestinal submucosa: a bioscaffold for tissue replacement, Tissue Eng. 2 (1996) 209–217. [17] M.E. Carr, J. Hermans, Size and density of fibrin fibers from turbidity, Macromolecules 11 (1978) 46–50. [18] M.L. Wong, L.G. Griffiths, Immunogenicity in xenogeneic scaffold generation: antigen removal versus decellularization, Acta Biomater. 10 (2014) 1806–1816. [19] J.P. Hodde, R. Record, R. Tullius, S.F. Badylak, Fibronectin peptides mediate HMEC adhesion to porcine-derived extracellular matrix, Biomaterials 23 (2002) 1841–1848.

J.A. Claudio-Rizo et al. / Materials Science and Engineering C 79 (2017) 793–801 [20] B. Brown, K. Lindberg, J. Reing, D.B. Stolz, S.F. Badylak, The basement membrane component of biologic scaffolds derived from extracellular matrix, Tissue Eng. 12 (2006) 519–526. [21] T.W. Gilbert, T.L. Sellaro, S.F. Badylak, Decellularization of tissues and organs, Biomaterials 27 (2006) 3675–3683. [22] S.T. Kreger, B.J. Bell, J. Bailey, E. Stites, J. Kuske, B. Waisner, S.L. Voytik-Harbin, Polymerization and matrix physical properties as important design considerations for soluble collagen formulations, Biopolymers 93 (2010) 690–707. [23] J.A. Ramshaw, J.A. Werkmeister, H.A. Bremner, Characterization of type I collagen from the skin of blue grenadier, Arch. Biochem. Biophys. 267 (1988) 497–502. [24] K.E. Kadler, A. Hill, E.G. Canty-Laird, Collagen fibrillogenesis: fibronectin, integrins, and minor collagens as organizers and nucleators, Curr. Opin. Cell Biol. 20 (2008) 495–501. [25] T. Velling, J. Risteli, K. Wennerberg, Polymerization of type I and III collagens is dependent on fibronectin and enhanced by integrins α11β1 and α2β1, J. Biol. Chem. 277 (2002) 37377–37381. [26] A.M. Milan, R.V. Sugars, G. Embery, R.J. Waddington, Modulation of collagen fibrillogenesis by dentinal proteoglycans, Calcif. Tissue Int. 76 (2005) 127–135. [27] R.K. Singh, D. Seliktar, A.J. Putnam, Capillary morphogenesis in PEG-collagen hydrogels, Biomaterials 34 (2013) 9331–9340. [28] T.D. Sargeant, A.P. Desai, S. Banerjee, A. Agawu, J.B. Stopek, An in situ forming collagen–PEG hydrogel for tissue regeneration, Acta Biomater. 8 (2012) 124–132. [29] Y.L. Yang, L.J. Kaufman, Rheology and confocal reflectance microscopy as probes of mechanical properties and structure during collagen and collagen/hyaluronan selfassembly, Biophys. J. 96 (2009) 1566–1585. [30] D.E. Birk, M.V. Nurminsaya, E.I. Zycband, Collagen fibrillogenesis in situ: fibril segments undergo post-depositional modifications resulting in linear and lateral growth during matrix development, Dev. Dyn. 202 (1995) 229–243. [31] J. Kopp, M. Bonnet, J.P. Renou, Effect of collagen crosslinking on collagen-water interactions (a DSC investigation), Matrix 9 (1989) 443–450. [32] P. Singh, C. Carraher, J.E. Schwarzbauer, Assembly of fibronectin extracellular matrix, Annu. Rev. Cell Dev. Biol. 26 (2010) 397–419. [33] Q. Lv, K. Hu, Q. Feng, F. Cui, Fibroin/collagen hybrid hydrogels with crosslinking method: preparation, properties, and cytocompatibility, J. Biomed. Mater. Res. A 84 (2008) 198–207. [34] N. Annabi, J.W. Nichol, X. Zhong, C. Ji, S. Koshy, A. Khademhosseini, F. Dehghani, Controlling the porosity and microarchitecture of hydrogels for tissue engineering, Tissue Eng. B Rev. 16 (2010) 371–383. [35] H.A. Gruber, E.F. Mellon, Oxidation products of amino acids and collagen, Anal. Biochem. 66 (1975) 78–86. [36] Y. Wang, W. Zhang, J. Yuan, J. Shen, Differences in cytocompatibility between collagen, gelatin and keratin, Mater. Sci. Eng. C 59 (2016) 30–37. [37] P. Lu, K. Takai, V.M. Weaver, Z. Werb, Extracellular matrix degradation and remodeling in development and disease, Perspect. Biol. 3 (12) (2011) 50–58. [38] Y. Yao, S. Yang, C. Zhang, X. Yu, Biocompatibility of compounds of extracellular matrix and thermally reversible hydrogel, Mater. Sci. Ed. 22 (3) (2007) 439–442. [39] C.M. Murphy, F.J. O'Brien, Understanding the effect of mean pore size on cell activity in collagen-glycosaminoglycan scaffolds, Cell Adhes. Migr. 4 (2010) 377–381.

801

[40] C. Helary, I. Bataille, A. Abed, C. Illoul, A. Anglo, L. Louedec, D. Letourneur, A. Meddahi-Pellé, M.M. Giraud-Guille, Concentrated collagen hydrogels as dermal substitutes, Biomaterials 31 (2010) 481–490. [41] S.F. Badylak, Decellularized allogeneic and xenogeneic tissue as a bioscaffold for regenerative medicine: factors that influence the host response, Ann. Biomed. Eng. 42 (2014) 1517–1527. [42] K.W. Ng, D.T. Leong, D.W. Hutmacher, The challenge to measure cell proliferation in two and three dimensions, Tissue Eng. 11 (2005) 182–191. [43] J.L. Stow, P.C. Low, C. Offenhäuser, D. Sangermani, Cytokine secretion in macrophages and other cells: pathways and mediators, Immunobiology 214 (2009) 601–612. [44] A.B. Roberts, M.B. Sporn, R.K. Assoian, J.M. Smith, N.S. Roche, L.M. Wakefield, U.I. Heine, L.A. Liotta, V. Falanga, J.H. Kehrl, Transforming growth factor type beta: rapid induction of fibrosis and angiogenesis in vivo and stimulation of collagen formation in vitro, Proc. Natl. Acad. Sci. U. S. A. 83 (1986) 4167–4171. [45] J.L. Dziki, D.S. Wang, C. Pineda, B.M. Sicari, T. Rausch, S.F. Badylak, Solubilized extracellular matrix bioscaffolds derived from diverse source tissues differentially influence macrophage phenotype, J. Biomed. Mater. Res. A 105 (2017) 138–147. [46] Y. Wu, X. Chen, W. Wang, X.J. Loh, Engineering bioresponsive hydrogels toward healthcare applications, Macromol. Chem. Phys. 217 (2) (2016) 175–188. [47] E. Ye, X.J. Loh, Polymeric hydrogels and nanoparticles: a merging and emerging field, Aust. J. Chem. 66 (9) (2013) 997–1007. [48] J.A. Claudio-Rizo, B. Mendoza-Novelo, J. Delgado, L.E. Castellano, J.L. Mata-Mata, A new method for the preparation of biomedical hydrogels comprised of extracellular matrix and oligourethanes, Biomed. Mater. 11 (3) (2016), 035016. . [49] X.J. Loh, J. Li, Biodegradable thermosensitive copolymer hydrogels for drug delivery, Expert Opin. Ther. Pat. 17 (8) (2007) 965–967. [50] Q.Q. Dou, S.S. Liow, E. Ye, R. Lakshminarayanan, X.J. Loh, Biodegradable thermogelling polymers: working towards clinical applications, Adv. Healthc. Mater. 3 (2014) 977–988. [51] P. Thoniyot, M.J. Tan, A.A. Karim, D.J. Young, X.J. Loh, Nanoparticle–hydrogel composites: concept, design, and applications of these promising, multi-functional materials, Adv. Sci. 2 (2015) 1400010. [52] E. Ye, P.L. Chee, A. Prasad, X. Fang, C. Owh, V.J.J. Yeo, X.J. Loh, Supramolecular soft biomaterials for biomedical applications, Mater. Today 17 (4) (2014) 194–202. [53] J.A. Claudio-Rizo, M. Rangel-Argote, P.U. Muñoz-González, L.E. Castellano, J. Delgado, G. Gonzalez-García, J.L. Mata-Mata, B. Mendoza-Novelo, Improved properties of composite collagen hydrogels: protected oligourethanes and silica particles as modulators, J. Mater. Chem. B 4 (40) (2016) 6497–6509. [54] B.Q.Y. Chan, Z.W.K. Low, S.J.W. Heng, S.Y. Chan, C. Owh, X.J. Loh, Recent advances in shape memory soft materials for biomedical applications, ACS Appl. Mater. Interfaces 8 (16) (2016) 10070–10087. [55] M. Rangel-Argote, J.A. Claudio-Rizo, L.E. Castellano, A. Vega-González, J.L. MataMata, B. Mendoza-Novelo, ECM–oligourethane–silica hydrogels as a local drug release system of dexamethasone for stimulating macrophages, RSC Adv. 7 (17) (2017) 10443–10453. [56] E.A. Appel, J. Barrio, X.J. Loh, O.A. Scherman, Supramolecular polymeric hydrogels, Chem. Soc. Rev. 41 (2012) 6195–6214.