**1. Introduction**

Hydrogels can be produced by crosslinking polymers to form interpenetration network (IPN) with varying mechanical properties [1,2] for tissue engineering such as cardiac tissue repairs [3], controlling the fates of stem cells [1], and drug delivery, etc. [4]. Since hydrogels are highly permeable to nutrients and water-soluble metabolites, they can support cell growth and proliferation which are suitable for tissue engineering (TE). Hydrogels for TE usually consist of synthetic polymers, such as polyurethane and polyvinyl alcohol (PVA) [5,6], or natural polymers, such as HA and collagen [7,8].

HA is a natural glycosaminoglycan with carboxylic groups; it is an important component of the extracellular matrices (ECM) in various tissues and play important roles in cell proliferation and

migration [9]. For instance, the interactions of HA in a cardiac patch and CD44 of BMSC enhanced cardiac di fferentiations of BMSC in both cardiac gene and proteins expressions [10]. In addition, various methods to prepare HA-based hydrogels including oxidized-HA or methacrylated-HA have been investigated for TE of ECM of NP [11–13], respectively.

Various of SF-based membranes, sca ffolds or hydrogels have been extensively studied for the applications of TE because of its favorable biological responses, such as weak antigenic e ffects and inflammatory responses in-vivo [9,14–16]. For example, SF/HA patches laden with hBMSC promoted cardiac repair in a rat myocardial infarction (MI) model [10,16]. Developing suitable mechanical properties for hydrogels is also important to enhance cell proliferations and hBMSC di fferentiations for using various TE [17,18]. In this regard, the crosslinking tyrosine in SF to produce di-tyrosine bonds by HRP/H2O2 enzymatic reactions produced silk elastomers with sti ffness that are varied from 0.2 to 10 kPa [17]. However, the sti ffness of the aforementioned SF hydrogels was not suitable for the needs of some tissues such as human NP. Although the influences of molecular weights of SF on mechanical properties of SF hydrogels have been reported [17], the influence of total tyrosine contents in SF on the mechanical properties of SF-based hydrogels has not been investigated.

The intervertebral disc (IVD) absorbs shocks by transferring and dissipating loads to the ECM within the superior and inferior discs. IVD degeneration generally causes lower back pain, which is a common health problem. [7,18]. Currently, clinical treatments, such as spinal fusion and partial or total disc replacement cannot fully restore or maintain IVD structures and functions [7,18,19]. Since disc degeneration originates in nucleus pulposus (NP) regions, tissue engineering NP, that may mimic the structure of native NP tissue and possibly fully restore the functionality of healthy IVD discs, is grea<sup>t</sup> needed [18,19].

Various biomaterials have been investigated for TE of NP. They are: A. carbohydrate polymers including HA, dextran, chitosan and carboxymethylcellulose (CMC) to produce various hydrogels such as oxidized-HA or methacrylated-HA [11–13], dextran-chitosan-teleostean and methacrylated CMC hydrogels, respectively [7,18,19]; B. proteins including Type II collagen hydrogels, and laminin-based hydrogels, respectively [20,21]; C. hybrid of carbohydrate polymers and proteins, such as crosslinked oxidized HA/gelatin [12] and Type II collagen with a low molecular weight of HA [8,22]. However, the mechanical properties of some of the aforementioned biomaterials were not suitable for TE of NP, which need to be further processed [11,20,21]. For instance, the elastic modulus of the oxidized-HA hydrogels, using adipic acid di-hydrazide (ADH), are much lower than those of the native ECM of the NP [11]. Although hydrogels produced by oxidized HA-gelatin using ADH could improve the elastic modulus of the hydrogels (~11 kPa), the controlling the chemical reactions of HA/gelatin by ADH might not be easily carried out [12]. Recently, dextran-chitosan-teleostean triple-interpenetrating network and methacrylatedC MC hydrogels has been examined in a goa<sup>t</sup> model to support the mechanical functions of degenerative NP [18,19]. However, chitosan and CMC had not ye<sup>t</sup> been approved by the FDA for use in internal organs.

Although the bioactivities of SF or HA/SF patches for cardiac repairs have been shown in-vitro, and in-vivo, respectively [10,16], the mechanical properties for HS-IPN hydrogels, which consisted of various amounts of tyrosine in SF, in terms of vary strains of SF in this study, and the weight ratios of SF to HA have not been investigated. Moreover, the potential of di fferentiations of induced hBMSC laden-HS-IPN hydrogels to NP cells has not been examined. Although using NHS/EDC to crosslink HA/SF to facilely prepare HA/SF hydrogels was recently reported [23], it was one step of our processes to prepare hydrogels. However, they did not examine the rheological properties of hydrogels. Using HRP/H2O2 to crosslink tyramine-substitute HA and 2% of SF to produce hydrogels with varying HA contents has also been reported by Raia et al. [24]. HA and SF-IPN hydrogels produced by them are configured by di-tyramine bonds in HA crosslinked network, di-tyrosine bonds in SF crosslinked network and tyramine-tyrosine bonds in HA/SF-IPN. The tyramine-substitute HA needed to be complexly and chemically synthesized for the research study, which was not a commercially available biomaterial.

To produce HS-IPN hydrogels, HA-crosslinked network hydrogels were first produced by using BDDE solutions to crosslink HA polymers (Scheme 1). The influences of varying strains of SF obtained from *B. mori* cocoons and, at a fixed quantities of HA, varying amounts of SF (expressed in wt. ratios of 5/1~5/7 for HA/SF, respectively) were crosslinked using HRP/H2O2 reactions to produce HS-IPN hydrogels. To further increase in-vitro stability of hydrogels, the carboxyl acids of HA in HS-IPN hydrogels were crosslinked with amine groups of added polyethyleneimine (PEI) and HS-IPN using *N*-(3-Dimethyl-amino-propyl)-*N*-ethylcarbodiimide hydrochloride)/*N*-Hydroxy-succinimide (EDC/NHS) reagents (Scheme 1). The differentiation of hBMSC laden on HS-IPN hydrogels, induced using TGF-β3 to NP cells were examined for TE of NP.

**Scheme 1.** The schematic procedures to develop HS-IPN hydrogel; HA polymer chains were first crosslinked by BDDE reagen<sup>t</sup> for several hours to produce HA crosslinked network hydrogel (blue dots); SF and a small quantity of PEI (polyethyleneimine) were added into the gel and well blended. SF polymer chains were then crosslinked by HRP/H2O2 enzymatic reactions in HA crosslinked network hydrogel to produce large quantities of di-tyrosine bonds, make SF crosslinked hydrogel, (brown dots or Tyr in the model hydrogels) and yield HS-IPN hydrogel. HS-IPN hydrogel and PEI were further crosslinked using EDC/NHS reagents to produce many amide bonds to stabilize the hydrogels.

#### **2. Materials and Methods**

#### *2.1. Fabrication of Interpenetrating Network HA-SF Hydrogels*

Various strains, strain A, B and C herein, of *B. mori* cocoons were the products of artificially cross breeding specific strains of *B. mori* to produce varying toughness of SF that might vary amino acid configurations of SF, including tyrosine contents. Each strain of *B. mori* was well-bred to control the quality of SF, including low variations in amino acid configurations of SF and its cocoon was gifted from MDARES (Miaoli Agricultural Research and Extension Station, Council of Agriculture, Executive Yuan, Miaoli, Taiwan). In addition, the stain A of SF was used in this study without being further specified.

Solution of SF (MW ~ 185 kDa) of strain A was prepared as described in early reports, such as degumming, de-solving in 9.3 M LiBr and removal of Li<sup>+</sup> using dialysis with DI water that were published by the authors' laboratory [10,16]. Briefly, silk cocoons were boiled in 0.02 M Na2CO for 0.5 h and then rinsed thoroughly in D.I. water to extract the glue-like sericin proteins from silk fibroin, degumming procedures. The extracted SFs were then dissolved in 9.3 M LiBr solution at 60 ◦C for 4 h to yield a 20% (w/v) solution, which was then dialyzed against D.I. water using a dialysis membrane

(MWCO 6000) (Spectra/Por 1, Repligen Co., Waltham, MA, USA) at room temperature to remove salt for 48 h [10,16].

The MW of SF of strain A, B and C were determined by a SDS-PAGE method (Vertical Electrophoresis System, XCell SureLockTM Mini-Cell, Invitrogen, Waltham, MA, USA) with 150 V for 60 min. The bands of samples and molecular weight ladder (Himark Pre-Stained, Invitrogen, Waltham, MA, USA) were stained by a Coomassie Brilliant Blue R-250 (Sigma Corp., St. Louis, MO, USA). The images of the bands of the samples and standard molecular weight ladder were analyzed by an Image J (Windows version, Java 1.6.0-45, 32 bit mode, NIH, Bethesda, MD, USA). For amino acids analysis, 0.4 mg of dried SF samples (e.g., strand A) were hydrolyzed by 6 N HCL at 115 CC for 24 h. The analysis of amino acids of various samples were conducted by an amino acid analyzer (Hitachi L-8900, Tokyo, Japan).

To prepare HA hydrogel (H gel), HA (MW: 200 kDa, Lifecore Biomedical Inc., Chaska, MN, USA) was dissolved in NaOH at a concentration of 10%. 1,4-Butanediol diglycidyl ether (BDDE, Sigma Corp., St. Louis, MO, USA) was used for cross-linking reactions that was conducted at 37 ◦C for 6–8 h. The pH value of the solution was then adjusted to about 7.0 by adding HCl solution to terminate the crosslinking reaction [25,26]. After the HA-BDDE crosslinking reaction had been terminated, 0.15 mM PEI (polyethyleneimine, Mw. 25 KDa, Sigma, St. Louis, MO, USA), an appropriate amount of H2O2 (about 0.15 mM) and horseradish peroxidase (50 U/mL) (HRP, Sigma, St. Louis, MO, USA) and 5% SF solution were added to the H gels and well blended. SF in the aforementioned solutions were crosslinked and gradually turned to form the second interpenetration network, SF IPN-gels, because large quantities of di-tyrosine bonds of SF were produced by the reactions of HRP/H2O2, the enzyme reactions, in H gels for about 1 h at 37 ◦C [27].

To prepare SF hydrogel (S gel), an appropriate amount of H2O2 (about 0.15 mM) and horseradish peroxidase (50 U/mL) (HRP, Sigma, St. Louis, MO, USA) were added to 5% SF solution SF of strain A to induce enzymatic reaction for 1 h at 37 ◦C [27]. HS-IPN hydrogels were produced by blending the aforementioned H gels and S gels (Scheme 1).

To stabilize the HS-IPN hydrogels, 0.15 mM PEI in the hydrogels was further crosslinked by EDC/NHS (1:1) ( *N*-(3-Dimethyl-amino-propyl)- *N*-ethylcarbodiimide hydrochloride, C8H17 N3 HCL, EDC), Sigma Corp., St. Louis, MO, USA)/(*N*-Hydroxy-succinimide, C4H5NO3, NHS), Fluka Chemical Corp., Rochester, NY, USA) for around 30 min. at temperature lower than 20 ◦C to produce amide bonds of carboxyl groups of HA with amine groups in PEI and in SF, respectively (Scheme 1). In addition, to examine the influences of varying ratios of SF on viscoelastic properties of HS-IPN hydrogels, the weight ratios of HA (4% at final) and SF in the hydrogels were changed to 5:1, 5:3, and 5:7 (e.g., HS1-IPN, HS3-IPN, HS7-IPN (namely, the model hydrogels)), respectively.

#### *2.2. Characterizations of HS-IPN Hydrogels*

#### 2.2.1. Attenuated Total Reflectance-Fourier Transform Infrared (ATR-FTIR) Spectra of the Hydrogels

After various HS-IPN hydrogels were freeze-dried at −50 ◦C, their transmission spectra of the samples were examined using an ATR-FTIR instrument (IRA ffinity-1, Shimadzu Co, Kyoto, Japan) with a resolution of 4 cm<sup>−</sup><sup>1</sup> at wave numbers 400–4000 cm<sup>−</sup><sup>1</sup> [10]. Detailed procedures for the measurements can be referred elsewhere. The spectra were analyzed using the built-in standard software package (IRA ffinity-1, Shimadzu Co, Kyoto, Japan) [10].

#### 2.2.2. Swelling Ratios of H, S, Crosslinked H/PEI and the Model Hydrogels

To study the swelling ratios of various gels, the samples were placed in phosphate saline (PBS) allowed for unrestricted deformation swelling test. H/PEI crosslinking hydrogels were fabricated after H gels were produced, followed by crosslinked 0.015 mM PEI in the gels by EDC/NHS crosslinking reagen<sup>t</sup> for 30 min. The purpose of producing H/PEI crosslinking hydrogels was to examine swelling property of the gels after amide bonds formation between the carboxyl groups in HA gels and amine

groups of PEI. The weights of hydrogels were measured at time of 0, 0.5, 1, 4, 24 and 48 h, Wt, after they were immersed into PBS for the aforementioned time. Their surfaces were gently wiped before they were weighted. The swelling ratio of a hydrogel was calculated by the following equation while Wnet was the weight of the hydrogel before immersed into PBS:

$$\text{Swelling ratio } (\%) = \frac{\text{W}\_{\text{t}} - \text{W}\_{\text{net}}}{\text{W}\_{\text{net}}} \times 100\% \tag{1}$$

#### 2.2.3. Morphology of HS-IPN Hydrogels

To observe the morphology and the pore structure of HS-IPN hydrogels, all hydrogels were frozen at −50 ◦C and then at liquid nitrogen, further dried in vacuum for several days, so-called freeze-dry technique, to produce sca ffolds. They were cut by a surgical knife to obtain the cross-section for observing the pore size distributions and characterizing the network in the sca ffolds. All samples were coated with Pt and imaged with scanning electron microscopy (SEM, JSM-7600F, JEOL, Tokyo, Japan) at an accelerating voltage of 3–10 kV. The pore sizes of gel were calculated by ImagePro software (IPP7.0, Media Cybernetics, Rockville, MD, USA).

#### *2.3. Rheological Studies of Vary Compositions of HS-IPN Hydrogels*

The viscoelastic properties of various hydrogels were determined using a rheometer with an oscillatory mode (Discovery-HR1, TA Instruments, New Castle, DE, USA) at 37 ◦C. Oscillation frequency sweep tests using a parallel-plate sensor with a gap of 1 mm and loaded by 0.1 mL of HS-IPN hydrogels were carried out by the rheometer. The elastic modulus (G) and viscous modulus (G"), phase shift angle (δ) and complex modulus (|G\*|) of vary compositions for HS-IPN hydrogels were determined at a fixed strain of 0.01 rad with varying angular frequencies of 1–100 rad/s, as reported in other studies [11,12,28].

#### Compressive Modules of the Model Hydrogels

An MTS Model 858 Bionix Test System (MTS Corp., Eden Prairie, MN, USA) with a 5 kg load cell was used to perform confined compression tests on the porcine NP and l HS7-IPN hydrogels (or the model hydrogels). An acrylic cylinder with a diameter of 10 mm and a height of 30 mm was made and used as the indenter. A hollow, cylindrical acrylic container with an outer diameter of 30 mm, an inner diameter of 10 mm, and a height of 20 mm, was use in the confined compression test. A cavity with a diameter of 25.5 mm and a depth of 2.5 mm was cut into the bottom of the container and a porous plate was placed at the bottom of it. Porcine NP and hydrogel specimens were prepared and used to fill the acrylic container to a height of 5 mm. The confined compression tests were then carried out with five specimens for each material. A preload of 5 N (63.7 kPa) was applied to the indenter for 10 min to obtain a pressure balance on each specimen. In each confined compression test, following the preload phase, the indenter was moved at 0.1 mm/min until a final strain of 5% for the model hydrogels and the porcine NP while that of 15% (large strain compression) performed for the model hydrogels was mainly used for the comparisons with other studies. The load-displacement curve that was thus obtained for each specimen was then converted to a stress–strain (<sup>σ</sup>–<sup>ε</sup>) curve. The elastic modulus E was calculated as E = σ/ε for the linear portion of the stress-strain curve.

#### *2.4. Cytotoxicity of the Model Hydrogels*

To evaluate the biocompatibility of the model hydrogels, MTT (3-(4,5-dimethylthiazol-2-yl)-2,5- diphenyl-tetrazolium bromide), Sigma–Aldrich, St. Louis, MO, USA) assays were performed to test the cytotoxicity of vary ratios of extraction mediums, taken from supernatants, which was obtained by incubation the hydrogels with L929 fibroblasts for 24 h according to the guidelines of international standard organization (ISO) 10993-5 [29]. L929 fibroblasts were purchased from the Bio-resource Collection and Research Center (BCRC, Hsin-Chu, Taiwan) and cultured in Dulbecco's modified

Eagle's medium that contained 10% horse serum at 37 ◦C in a 5% CO2 incubator. After L929 cells were cultured with vary volume ratios of extraction mediums to cultural medium (e.g., 1/8, 1/4, 1/2 and 1.0 (or extraction medium only), namely dilution ratios) for 24 h, MTT assay to the cells were carried out to examine the cell viabilities as early report of this lab [29].

#### *2.5. Di*ff*erentiations of Induced hBMSC to NP in the Model Hydrogels*

To evaluate the potential of the model hydrogels for NP regeneration, human bone marrow-derived mesenchymal stem cell (hBMSC) from passage 6~8 were used according to our previous study [10]. The density of hBMSC of 1 × 10<sup>7</sup> cell/mL was seeded on the surface layer of a 1.0 cm<sup>2</sup> the model hydrogels to produce the cell-laden hydrogels. The viability of hBMSC in the model hydrogels after 3 d of cultivation were stained with Live/Dead or Viability/Cytotoxicity kit (Invitrogen, Waltham, MA, USA) according to the manual instructions. Live/Dead stain of the cells was observed by a laser confocal scanning microscopy (LCSM) (Olympus FV1000, Olympus Corp., Tokyo Japan) [10].

To induce the chondrogenesis of hBMSC in the model hydrogels, after the cell proliferations in the hydrogels for 3 d, the culture medium was changed to chondrogenic di fferentiation medium containing TGF-β3 which was purchased from Lonza Corp. (Lonza, Gaithersburg, MD, USA). The di fferentiation medium was changed every two days, and followed by culture for 7 d and 14 d, respectively, for chondrogenesis study of hBMSC.

2.5.1. Immuno-Histochemical (IHC) Analysis of Specific Protein Expressions of the Di fferentiations of Induced hBMSC to NP

IHC analysis of specific protein expressions the di fferentiations of hBMSC to NP after induced by TGF-β3 was carried out at 7 d and 14 d. All samples were washed using PBS and fixed by 4% paraformaldehyde. After the fixation, they were moved from the cultural wells and then embedded in para ffin and sectioned into sections with a thickness of 5 μm. The sections were then stained with Alcian blue for examining the depositions of glycosaminoglycan (GAG) and collagen type II, respectively. Cell-free hydrogel was stained as a negative control.

2.5.2. Real-Time PCR for Specific Gene Expressions of the Di fferentiations of Induced hBMSC to NP

To determine the specific gene expressions of the di fferentiations of induced hBMSC to NP in the model hydrogels, the ECM-related gene expression, including Aggrecan (AGN), Collagen type I (Col I) and Collagen type II (Col II) were assessed by real-time PCR (StepOnePlus ™ Real-Time PCR System, Applied Biosystems, Foster, CA, USA) at 7 d and 14 d, according to other studies [11,12,19]. GAPDH was used as the internal control. The relative gene expression was calculated as 2−ΔΔCt .
