The Impact of Polyethylene Glycol-Modified Chitosan Scaffolds on the Proliferation and Differentiation of Osteoblasts

The objective of this study was to investigate the influence of polyethylene glycol (PEG) incorporated chitosan scaffolds on osteoblasts proliferation and differentiation. The chitosan polymer was initially modified by the predetermined concentration of the photoreactive azido group for UV-crosslinking and with RGD peptides (N-acetyl-GRGDSPGYG-amide). The PEG was mixed at different ratios (0, 10, and 20 wt%) with modified chitosan in 96-well tissue culture polystyrene plates to prepare CHI-100, CHI-90, and CHI-80 scaffolds. PEG-containing scaffolds exhibited bigger pore size and higher water content compared to unmodified chitosan scaffolds. After 10 days of incubation, the cell number of CHI-90 (1.1 × 106 cells/scaffold) surpasses that of CHI-100 (9.2 × 105 cells/scaffold) and the cell number of CHI-80 (7.6 × 105 cells/scaffold) were significantly lower. The ALP activity of CHI-90 was the highest on the fifth day indicating the favored osteoblasts' early-stage differentiation. Moreover, after 14 days of osteogenic culture, calcium deposition in the CHI-90 scaffolds (2.7 μmol Ca/scaffold) was significantly higher than the control (2.2 μmol Ca/scaffold) whereas on CHI-80 was 1.9 μmol/scaffold. The results demonstrate that PEG-incorporated chitosan scaffolds favored osteoblasts proliferation and differentiation; however, mixing relatively excess PEG (≥20% wt.) had a negative impact on osteoblasts proliferation and differentiation.


Introduction
Chitosan, the deacetylated derivative of chitin, has been broadly utilized for the fabrication of tissue engineering scafolds due to its nontoxicity, biodegradability, good biocompatibility [1], and resemblance to glycosaminoglycans [2]. However, chitosan still possesses some shortcomings for such a purpose, for instance, the mechanical properties of chitosan scafolds may not be suitable to match some specifc tissue engineering applications. Previously it was utilized an azide-based UV-crosslinking mechanism for crosslinking chitosan scafolds in order to increase the mechanical properties of chitosan scafolds [3][4][5]. Upon UV irradiation, the azido groups are converted into highly reactive nitrene groups which undergo direct insertion into C-H, O-H, and N-H bonds of nearby substance molecules [6]. Another shortcoming is that chitosan lacks bioactive signals equivalent to those existing in the extracellular matrix (ECM) for cell attachment, growth, and diferentiation. Incorporation of bioactive signals such as ECM adhesion proteins and cell-binding peptides into chitosan substrates can enhance cell adhesion [7][8][9]. RGDincorporated and crosslinked chitosan scafolds can be employed for mesenchymal stem cell proliferation and osteogenic diferentiation [3,4].
Tere are many reports aimed at improving chitosan properties by blending with natural or synthetic molecules. Park et al. [10] developed composite chitosan scafolds containing anionic carbohydrates. Te incorporation of chondroitin 4-sulfate or alginate in chitosan scafolds increased the compressive modulus of the scafolds and enhanced apatite formation. Furthermore, apatite-coated scafolds enhanced the spreading, proliferation, and osteogenic diferentiation of bone marrow stromal cells seeded on the scafolds. A report by Li et al. [11] using a 3aminopropyltriethoxysilane treatment for modifcation and biocompatibility of lyophilized chitosan porous scaffolds showed the silanization treatment with low 3aminopropyltriethoxy silane concentration showed no signifcant infuence on the morphology of chitosan scafolds, the porosity, and surface amino densities were increased after silanization whereas the swelling ratio was reduced, and the degradation ratio in PBS and antiacid degradation properties of the silanized chitosan scafolds were signifcantly improved. Chitosan doped with multiwalled carbon nanotubes has been used to create highly porous conductive scafolds [12]. Chitosan can also be modifed by the addition of hydrophobic alkyl chains along the hydrophilic backbone of the chitosan polymer and Cooney et al. [13] reported the scafolds produced from the unmodifed chitosan were more stable and rigid and possessed average pore diameters that were generally smaller than those fabricated from the hydrophobically modifed chitosan. Te generally larger pores in the butyl-modifed chitosan scafolds might be explained by increased phase separation rates due to the introduced hydrophobicity of the chitosan polymer. Te combination of hydrophobic groups opposed along an otherwise hydrophilic backbone creates internal forces that tend to fold or buckle the polymer chain, creating regions that exhibit micellar behavior [14].
One major drawback of chitosan in drug delivery is its low solubility. Chitosan is not soluble in aqueous solutions at neutral or alkaline pH, only soluble in aqueous acid solutions and a few organic solvents [15]. Hence, various chitosan derivatives have been prepared for the purpose of drug delivery [15,16]. Similarly, in this work it was argued that chitosan is a relatively hydrophobic material; however, in the natural tissues, the extracellular matrix is highly hydrated. Tus, the hydrophobic environment of chitosan scafolds might not be suitable for tissue growth. A number of studies indicate that both morphology and hydrophilicity infuence the attachment of cells onto the surface of a scafold [17]. Terefore, an increase in the hydrophilicity of chitosan scafolds might improve tissue engineering outcomes. Polyethylene glycol (PEG) possesses biodegradability, biocompatibility, less toxicity, and hydrophilicity and has been widely used in biomedical applications, including surface modifcation, bioconjugation, drug delivery, and tissue engineering [18,19]. Although there are interesting contributions to the preparation, characterization, and aggregation behavior of amphiphilic chitosan derivatives having poly-L-lactic acid side chains [20], there is no report on describing the osteogenic proliferation or diferentiation of cells, from hydrophilically modifed chitosan scafold. Hence, this study aimed to investigate the efect of adding polyethylene glycol into the RGD-conjugatedcross-linked chitosan scafolds on the proliferation and diferentiation of osteoblasts.

Conjugation of Photoreactive Azido Groups or Peptides to
Chitosan. Azido groups were conjugated onto chitosan (molecular weight 50-190 kDa, 75-85% deacetylation) via a reaction forming covalent amide bonds between the amino groups of chitosan and the carboxyl groups of an azidobenzoic acid ester. Briefy, 17.6 mg 5-azido-2-nitrobenzoic acid N-hydroxysuccinimide ester was dissolved in 200 μL dimethylsulfoxide and then mixed with chitosan solution (0.1 g in 4.8 mL of 1% acetic acid), followed by 3 h incubation at room temperature. Te unreacted azido ester was removed by dialysis against deionized water through a seamless cellulose tube (MWCO 12,400 Da) in the dark for two days with the changes of deionized water every 12 h. After freeze-drying, the azido-conjugated chitosan (CHI-g-AZ) was kept at 4°C until use.
RGD peptides were conjugated onto chitosan molecules via a carbodiimide reaction according to a previously developed procedure [5]. Te graft ratio of RGD to chitosan (CHI-g-RGD) was estimated as 2.75 mol% with respect to the total moles of the amino groups of chitosan molecules.

Preparation of Chitosan and polyethylene Glycol Mixed
Scafold. A mixture of chitosan (unmodifed, CHI-g-AZ, and CHI-g-RGD) and PEG with a total concentration of 10 mg/mL in 1% acetic acid was prepared at diferent weight percentages (the composition and abbreviations listed in Table 1). Chitosan-PEG mixed scafolds were prepared by adding 70 μL/well of the unmodifed chitosan, CHI-g-AZ, CHI-g-RGD, and PEG mixture in 96-well tissue culture polystyrene (TCPS) plates. Briefy, the mixture was poured into 96-well TCPS plates (70 μL/well for cell culture experiments), followed by freeze-drying in the dark to form scafolds. Subsequently, chitosan substrates were crosslinked by UV irradiation for 30 min (wavelength range 280-380 nm). Te UV crosslinking time and the PEG dose were selected based on the previous reports [5,21].

Characterization of Chitosan Scafolds.
Te morphology of chitosan-PEG scafolds was observed by scanning electron microscope (SEM) images (JSM-5310, JEOL, Japan). Te scafolds were frst dehydrated in graded series of ethanol solutions 30%, 50%, 70%, 90%, 95%, and 100% for 10 min each step followed by CO 2 critical point-drying. Samples were cut with a scalpel, coated with a gold layer on the section, and then observed with SEM at an acceleration voltage of 20 kV.
Te pore sizes of the scafolds were analyzed using an NIH Image J. Pores in SEM images were traced manually, and the enclosed areas and perimeters of pores were determined by the NIH Image J software. Te hydraulic diameters of the pores were determined by the following equation: pore diameter (D p ) � 4 × area/perimeter [5] more than 100 pores were counted for each type of sample.
Te compressive stress-strain properties of the scafolds were determined using a compressive testing machine (FGS-50V-H, NIDECSIMPO Corporation, Japan) and a digital force gauge (FGP-0.5, NIDECSIMPO Corporation, Japan). Te scafolds were subjected to an unconfned uniaxial compression to 70% strain at a compression velocity of 3 mm/s. Te continuous stress and peak stress were recorded and analyzed.
Dried chitosan-PEG scafolds were soaked in deionized water for 24 h. Te surface water contents on the scafolds were absorbed by a flter paper. Wet scafolds were weighed (W s ), and then placed in a 70°C oven overnight and weighed again (W d ). Te equation implemented to calculate water content is shown as follows: (1)

Culture of Osteoblasts on Chitosan Scafolds.
Standard sterile cell culture techniques were used for all cell experiments. Te animal procedure was followed by the ethical guidelines of Care and Use of Laboratory Animals (National Taiwan University, National Institutes of Health Publication No. 85-23, revised 1985) and was approved by the Animal Center Committee of National Taiwan University. Primary osteoblasts were isolated from neonatal rat calvariae according to the previously published procedure [22]. Te number and viability of the isolated osteoblasts were determined using a hemocytometer with trypan blue exclusion. Te isolated cells were cultured in standard T75 fasks to the second passage for the cell experiments. Prior to cell seeding, the chitosan-PEG scafolds were soaked in 70% ethanol for 30 min, followed by rinses with sterilized PBS three times. For cell culture on the scafolds, 20 μL of osteoblast suspension (1.5 × 10 7 cells/mL) was seeded onto scafolds, making the seeding density 3 × 10 5 cells per scafold. After 1, 5, or 10 days of culture, the cell-inoculated samples were analyzed for cell morphology cell numbers and alkaline phosphatase (ALP) activities.
After the cell culture, the morphology of cells in the scafolds after 5 days of incubation was observed by SEM. Te adhered osteoblasts were lysed with 0.1% Triton X-100 for 30 min. Cell proliferation was determined by the lactate dehydrogenase (LDH) method according to a reported protocol [9]. Intracellular alkaline phosphatase activities were assayed by determining the release of p-nitrophenol from 4-nitrophenyl phosphate disodium salt at pH 10.2, as reported previously [23].
Te cell doubling time (T 2 ) was calculated using the following equation: No is the number of cells at the beginning of the observation, and △N is the increase in the number of cells during the period of time of the length △t. Each division increases the number of cells by adding 1, and △N is also the number of cell divisions during the same period [24].

Mineralization Culture of Osteoblast/Scafold Constructs.
Osteoblast/scafold constructs were cultured for 5 days in the osteoblast culture medium, followed by 10 days of mineralization culture in the osteoblast diferentiation medium with daily replenishment of L-ascorbate (50 μg/mL). Te total amount of calcium deposition was determined using a calcium assay kit (Diagnostic Chemicals Limited, USA) [25].

Statistical Analysis.
Each experiment has been repeated at least three times. Te data were presented as mean-± standard deviation (SD). Te statistical assessment of signifcant variations was performed by Microsoft Excel 2010. Signifcance was assessed by one-way analysis of variance (ANOVA) and two-tailed Student-Newman-Keuls multiple comparisons. Te probability of p ≤ 0.05 was considered as a signifcant diference, where the symbol of * and * * marker represent p < 0.05 and p < 0.01, which is of signifcant diference statistically in 95% and 99% confdence level, respectively.

Type of scafolds Chitosan CHI-g-AZ CHI-g-RGD PEG
International Journal of Biomaterials 3 PEG was examined with SEM. All the scafolds exhibited an open pore microstructure with interconnectivity. Te pore structure of the scafolds at diferent PEG concentrations is similar to each other ( Figure 1). Te average pore sizes of CHI-100, CHI-90, and CHI-80 were 33.3 ± 7.4, 42.2 ± 8.2, and 46.9 ± 8.6 μm, respectively, indicating the pore sizes of the scafolds were signifcantly increased (p < 0.05) with the increasing PEG contents in chitosan scafolds (Figure 1). It was suspected that since PEG is more hydrophilic than chitosan, more water molecules surround PEG and form larger ice crystals during the freezing step than pure chitosan scafolds. As a result, after lyophilization chitosan/PEG scafolds contain larger pores compared with pure chitosan scafolds. Te compressive properties of the chitosan-PEG scaffolds were next evaluated (Figure 2(a)). Te compressive stresses of all scafolds increased with increasing strain until a maximum at the end of the compression (70% strain). Te maximum compression stress of CHI-100, CHI-90, and CHI-80 scafolds was 56.1 ± 2.0, 46.9 ± 1.6, and 41.3 ± 7.0 kPa, respectively. Te incorporation of PEG signifcantly decreased the stifness of chitosan scafolds (p < 0.05). Tis situation is most visible in the CHI-80. It is not surprising because PEG is less stifness material compared to chitosan [27]. A similar argument by Cheng et al. [28] explains the blend of PNIPAM with PEG hydrogels exhibits a lower mechanical strength than pure PNIPAM. Tanuma et al. [29] reported that the PEG-cross-linked chitosan hydrogel flm swelling ratio increases with the decrease of molecular weight of PEG with the same content sample, and the degradation rate of chitosan component was found to be infuenced by the content and molecular weight of PEG. An increase in the total PEG content resulted in a considerable increase in the degradation rate.
Te water contents in CHI-100, CHI-90, and CHI-80 scafolds were next determined. Te water uptake of the chitosan scafolds was signifcantly (p < 0.05) increased with increasing PEG contents from 4476 to 6025% (Figure 2(b)) dry weight basis. Besides the hydrophilicity of the added PEG, chitosan-PEG scafolds have higher pore size and more water storage space as a result the ratio of water absorption had a signifcant diference (p < 0.05) with unmodifed chitosan. Te previous study on incorporating PEG into Alginate/Elastin composite matrix indicates water content increased with an increase in PEG content [30]. Similarly, Wan et al. [31] reported that the introduction of PEG segments enhanced the surface hydrophilicity of the poly-llactide-polyethylene glycol copolymers. Likewise, several modifcations (chemical, mechanical, and structural) of hyaluronic acid hydrogels have been conducted in the fabrication of artifcial extracellular matrix [32]. Since hyaluronic acid has negative charges, it can absorb large amounts of water and swell up to 1000 times in volume [33], However, chitosan is claimed for inadequate moisture availability, thus this study is the frst on improving the hydrophilicity of chitosan scafold via hydrophilic polymer along with the increase of pore size and water content.
Overall, the incorporation of PEG increases the pore sizes and the water-uptake ability of chitosan scafolds but sacrifces the scafold's stifness. Te Chitosan-PEG scafolds with appropriate hydrophilicity were expected in favor of mass transportation, and then cell proliferation and differentiation. It was expected that cell proliferation would be much improved by increasing the hydrophilicity of the three-dimensional scafolds, which even outweighed the disadvantages of the weaker mechanical property. Next, it was examined the efect of PEG incorporation in the culture of osteoblasts.

Osteoblast Culture on the Chitosan Scafolds.
Te nontoxicity of the Chitosan scafold has been afrmed [34]. In this study, the chitosan-PEG scafold showed good cell adhesion on all the used scafold formulations (Figure 3). Te cells on the pure chitosan scafold (Figure 3(a)) are few and separately adhered on the surface, while the cells on chitosan-PEG (Figures 3(b) and 3(c)) are more aggregated which indicated the favored environment for cells proliferation. Cell proliferation is the process of multiplying the number of cells, and in this process, mitochondria gained a central role in the regulation of cell proliferation [35]. It was found that the addition of PEG decreased one-day cell adhesion to the chitosan-based scafolds (Figure 4(a)). It is not surprising because PEG is a well-known nonadhesive material [36,37]. After fve days of incubation (Figure 4(a)), it was observed that the trend of cell number was still the same on the frst day; however, the cell number in all scafolds signifcantly (p < 0.05) improved and the doubling time of cells were 41.5, 22.6, and 23.3 h on CHI-100, CHI-90, and CHI-80 scafolds, respectively. After 10 days of incubation, the cell numbers of CHI-90 (1.1 × 10 6 cells/scaffold) surpass that of CHI-100 (9.2 × 10 5 cells/scafold), while the cell number of CHI-80 (7.6 × 10 5 cells/scafold) was signifcantly lower than the cell numbers of CHI-100 and CHI-90 (p < 0.001). During this period, the doubling time of cells of CHI-100, CHI-90, and CHI-80 scafolds was 66.4, 21.5, and 926 h, respectively, indicating that the rate of cell proliferation of CHI-90 remained fast. However, the cell proliferation rate of CHI-100 and CHI-80 decreased, especially CHI-80. PEG in chitosan scafolds provides well hydration environment. As a result, it may enhance the difusion of nutrients, bio-factors, and wastes. Hence, it might be the main reason CHI-90 scafolds could maintain low doubling time. On the other hand, during incubation, it was observed that the CHI-80 scafold was too soft that it might afect the cell proliferation of osteoblasts. It was reported before by Tanuma et al. [29] that the degradation rate of the chitosan component was found to be infuenced by the content and molecular weight of PEG. An increase in total PEG content resulted in a considerable increase in the degradation rate.
Te osteogenic diferentiation of osteoblasts on the chitosan-PEG scafolds was investigated by early and late osteogenic markers. Alkaline phosphatase (ALP), an essential enzyme for ossifcation, is an early bone marker protein, and one of the most frequently used markers to demonstrate osteoblast diferentiation [38]. Te fnal stage of osteoblast diferentiation is mineralization, at which a mineral matrix containing mainly calcium phosphate is secreted and deposited by mature osteoblasts.
In this study, after osteogenic culture for one day, the cellular ALP activity of CHI-80 was the highest, followed by CHI-90 and CHI-100 (Figure 4(b)). However, after fve days of incubation, the ALP activity of CHI-100 and CHI-90 increased signifcantly (p < 0.001) compared to their frst day, respectively, and exceeded the values of CHI-80. After ten days of incubation, the ALP activity was decreased in all the samples. Te ALP activity of CHI-90 was the highest on the ffth day, indicating the favored osteoblasts' early-stage diferentiation.
After the osteoblasts were cultured in the osteogenic medium for 2 weeks, the total amounts of calcium in CHI-90 and CHI-100 were quantifed as 2.7 and 2.2 μmol/scafold; whereas, the amount in CHI-80 was 1.9 μmol/scafold International Journal of Biomaterials ( Figure 5). Te results indicate that calcium deposition between the CHI-100 and CHI-90 had a signifcant diference (p < 0.05), suggesting that osteoblast diferentiation is enhanced with the optimal amount of PEG (10%). However, excess PEG (20%) signifcantly decreased osteoblasts mineralization. For future work, it is suggested to investigate optimizing hydrophilic polymer doping onto a chitosan scafold.

Conclusion
Te impact of PEG-incorporated chitosan scafolds on osteoblasts diferentiation and proliferation has been demonstrated in this study. Te characteristic analysis of PEGcontaining scafolds exhibited bigger pore size, weaker mechanical properties, and higher water content compared to unmodifed chitosan substrates. Te cultured osteoblasts on the PEG-chitosan scafold showed better cell proliferation and diferentiation than that of the chitosan scafold. However, adding more PEG (≥20% wt.) into the scafolds has no beneft on the proliferation and diferentiation of osteoblast. Taken together, these results indicate that adding hydrophilic molecules such as polyethylene glycol at an optimum amount (10% wt) into chitosan changed the characteristic of the scafolds and improved the proliferation and diferentiation of osteoblast. Te biocompatibility, safety, and biodegradability of the chitosan make it an excellent scafold candidate, and in the near future will witness its crucial role in biomaterials and tissue engineering.

Data Availability
All data used to support the fndings of this study are included within the article.

Conflicts of Interest
Te authors declare that there are no conficts of interest.