Electrospinning of Nanodiamond-Modified Polysaccharide Nanofibers with Physico-Mechanical Properties Close to Natural Skins

Electrospinning of biopolymers has gained significant interest for the fabrication of fibrous mats for potential applications in tissue engineering, particularly for wound dressing and skin regeneration. In this study, for the first time, we report successful electrospinning of chitosan-based biopolymers containing bacterial cellulous (33 wt %) and medical grade nanodiamonds (MND) (3 nm; up to 3 wt %). Morphological studies by scanning electron microscopy showed that long and uniform fibers with controllable diameters from 80 to 170 nm were prepared. Introducing diamond nanoparticles facilitated the electrospinning process with a decrease in the size of fibers. Fourier transform infrared spectroscopy determined hydrogen bonding between the polymeric matrix and functional groups of MND. It was also found that beyond 1 wt % MND, percolation networks of nanoparticles were formed which affected the properties of the nanofibrous mats. Uniaxial tensile testing of the woven mats determined significant enhancement of the strength (from 13 MPa to 25 MP) by dispersion of 1 wt % MND. The hydrophilicity of the mats was also remarkably improved, which was favorable for cell attachment. The water vapor permeability was tailorable in the range of 342 to 423 µg·Pa−1·s−1·m−1. The nanodiamond-modified mats are potentially suitable for wound healing applications.


Introduction
Structural disruption in basement membranes of skin due to different types of injuries have to be cared and repaired as the skin is the largest organ of the body with many vital functions [1]. Elastic, biocompatible, and non-allergenic wound dressing materials are commonly utilized with an aim to keep the injured environment moist while protecting it from new infections [2]. With the same morphology and structure as the natural extracellular matrix (ECM), polymeric fibers are the most suitable materials that can be utilized for skin regeneration [3]. The porosity, oxygen permeability, mechanical durability, and hydrophilicity of fibrous membranes can be adjusted to practical needs by chemical and structural modifications [4].
So far, chitosan (CS) and CS-based fibers have been the most studied natural polymeric materials for wound healing and skin tissue engineering applications [5]. Not surprisingly, this natural polysaccharide offers a number of advantages over other natural and synthetic polymers. As a derivate of chitin [6], CS has proved to be biocompatible with non-allergenic activity and bactericidal on materials processing and characterizations, but further studies are underway to investigate the drug-loading ability of the MND-modified nanofibers structure. Figure 1 shows representative SEM micrographs of electrospun fibers containing different amounts of MND. The size distribution of the fibers was determined by image analysis and is shown in Figure 1 and Table 1. Randomly oriented fibers with diameters ranging from 80 to 300 nm were obtained by electrospinning the polysaccharide suspension (Figure 1a,b). Some mini-jets were also visible, which could be due to the high viscosity of the suspension, i.e., the electrical field could not overcome surface tension forces [37]. Introduction of diamond nanoparticles reduced the fiber diameters and yielded more uniform fibers with narrower size distribution (Table 1). It is possible that the effect of MND on the fiber diameters is related to the enhanced viscosity and conductivity of the suspension from which the fibers are drawn. At a given electric field strength, the higher liquid viscosity typically results in a weaker stretching of liquid jet while higher conductivity changes the shape of meniscus [38]. The balance between the effect of nanoparticles on the viscosity and conductivity determines the size of the fibers that leads to the formation of coarser fibers [39]. As it will be shown in the next section, no strong chemical interactions between MND and polymer occurs, but agglomeration of the nanoparticles are favorable, particularly at high concentrations. For instance, Figure 2 shows the formation of MND clusters in electrospun fibers. Rakha et al. [40] related particles' agglomeration to their surface functional groups. As the effective size of ultrafine MND increased, their effect on the viscosity decreased, so that finer fibers were attained [41].  Figure 1 shows representative SEM micrographs of electrospun fibers containing different amounts of MND. The size distribution of the fibers was determined by image analysis and is shown in Figure 1 and Table 1. Randomly oriented fibers with diameters ranging from 80 to 300 nm were obtained by electrospinning the polysaccharide suspension (Figure 1a,b). Some mini-jets were also visible, which could be due to the high viscosity of the suspension, i.e., the electrical field could not overcome surface tension forces [37]. Introduction of diamond nanoparticles reduced the fiber diameters and yielded more uniform fibers with narrower size distribution (Table 1). It is possible that the effect of MND on the fiber diameters is related to the enhanced viscosity and conductivity of the suspension from which the fibers are drawn. At a given electric field strength, the higher liquid viscosity typically results in a weaker stretching of liquid jet while higher conductivity changes the shape of meniscus [38]. The balance between the effect of nanoparticles on the viscosity and conductivity determines the size of the fibers that leads to the formation of coarser fibers [39]. As it will be shown in the next section, no strong chemical interactions between MND and polymer occurs, but agglomeration of the nanoparticles are favorable, particularly at high concentrations. For instance, Figure 2 shows the formation of MND clusters in electrospun fibers. Rakha et al. [40] related particles' agglomeration to their surface functional groups. As the effective size of ultrafine MND increased, their effect on the viscosity decreased, so that finer fibers were attained [41].

Interactions between Nanoparticles and Polymer
In order to figure out the possible interactions between CS, BC, and MND particles, FT-IR spectroscopy was employed. Figure 3 shows FT-IR spectrum of the polymer blend, pristine MND particles, and electrospun fibers containing 3 wt % MND. The -OH stretching vibration of CS occurs at 3000-3500 cm −1 [25,42]. The peak around 2965 cm −1 is assigned to aliphatic C-H stretching vibration [31]. The N-H peak is for CS which overlaps the wide absorption peak of -OH group in range of 3000-3500 cm −1 . The peaks at 1343 cm −1 and 1467 cm −1 are attributed to symmetric and bending vibration of C-H group, respectively. Another peak at 1050 cm −1 is due to the stretching vibration of C-O-C group. As to the PEO, triplet peaks of the C-O-C stretching vibrations appear at 1148, 1101, and 1062 cm −1 . An absorption band at 2885 cm −1 is attributed to CH2 stretching vibration in PEO [43,44]. The peak of the glucose carbonyl of cellulose is detected at 1643 cm −1 . The peak at around 1045 cm −1 shows the C-O-C stretching vibration. Overall, the results are in a good agreement with previous studies on CS and BC composite films and firmly verify the presence of many intermolecular hydrogen and ionic bonds, as well as a few covalent bonds [24,25,42]. The absorption peaks of MND are seen at 3426 cm −1 , 1626 cm −1 , and 1107 cm −1 which are attributed to stretching vibration of -OH, -C=O, and the overlap of C-O groups with nitrogen groups, respectively. The peak at 1735 cm −1 indicates stretching vibration of functional groups of C=O and -COOH and the peak at 2341 cm −1 is related to the absorption of CO2 [45][46][47]. The FT-IR results determined there was no formation of new peaks as a result of blending the polymers with MND; only a slight red shift in characteristics peaks were observed that could be due to hydrogen bonding [48].

Interactions between Nanoparticles and Polymer
In order to figure out the possible interactions between CS, BC, and MND particles, FT-IR spectroscopy was employed. Figure 3 shows FT-IR spectrum of the polymer blend, pristine MND particles, and electrospun fibers containing 3 wt % MND. The -OH stretching vibration of CS occurs at 3000-3500 cm´1 [25,42]. The peak around 2965 cm´1 is assigned to aliphatic C-H stretching vibration [31]. The N-H peak is for CS which overlaps the wide absorption peak of -OH group in range of 3000-3500 cm´1. The peaks at 1343 cm´1 and 1467 cm´1 are attributed to symmetric and bending vibration of C-H group, respectively. Another peak at 1050 cm´1 is due to the stretching vibration of C-O-C group. As to the PEO, triplet peaks of the C-O-C stretching vibrations appear at 1148, 1101, and 1062 cm´1. An absorption band at 2885 cm´1 is attributed to CH 2 stretching vibration in PEO [43,44]. The peak of the glucose carbonyl of cellulose is detected at 1643 cm´1. The peak at around 1045 cm´1 shows the C-O-C stretching vibration. Overall, the results are in a good agreement with previous studies on CS and BC composite films and firmly verify the presence of many intermolecular hydrogen and ionic bonds, as well as a few covalent bonds [24,25,42]. The absorption peaks of MND are seen at 3426 cm´1, 1626 cm´1, and 1107 cm´1 which are attributed to stretching vibration of -OH, -C=O, and the overlap of C-O groups with nitrogen groups, respectively. The peak at 1735 cm´1 indicates stretching vibration of functional groups of C=O and -COOH and the peak at 2341 cm´1 is related to the absorption of CO 2 [45][46][47]. The FT-IR results determined there was no formation of new peaks as a result of blending the polymers with MND; only a slight red shift in characteristics peaks were observed that could be due to hydrogen bonding [48].  Figure 4 shows changes in the water contact angle of electrospun fibers with the MND concentration. The hydrophilicity at first decreased with the introduction of 1 wt % MND particles, but increased at the higher concentrations. The MND particles with many surface functional groups are hydrophilic but, at the high concentrations, the addition of these nanoparticles increases the hydrophobicity of the mats. It is suggested that changes in the contact angles were more likely due to the morphology/size variations of the fibers. It is known that the surface energy and surface geometric features influence the contact angles of fibrous mats [49,50]. Smaller fibers with agglomerated nanoparticles altered the hydrophilicity due to the roughness issue and surface features [51].

Mechanical Properties of Electrospun Mats
Typical stress-strain curves of the electrospun mats are shown in Figure 5. The effect of MND on the mechanical properties are summarized in Table 2. It was found that the addition of 1 wt % MND particles significantly increased the elastic modulus and yield strength of the fibers along with  Figure 4 shows changes in the water contact angle of electrospun fibers with the MND concentration. The hydrophilicity at first decreased with the introduction of 1 wt % MND particles, but increased at the higher concentrations. The MND particles with many surface functional groups are hydrophilic but, at the high concentrations, the addition of these nanoparticles increases the hydrophobicity of the mats. It is suggested that changes in the contact angles were more likely due to the morphology/size variations of the fibers. It is known that the surface energy and surface geometric features influence the contact angles of fibrous mats [49,50]. Smaller fibers with agglomerated nanoparticles altered the hydrophilicity due to the roughness issue and surface features [51].  Figure 4 shows changes in the water contact angle of electrospun fibers with the MND concentration. The hydrophilicity at first decreased with the introduction of 1 wt % MND particles, but increased at the higher concentrations. The MND particles with many surface functional groups are hydrophilic but, at the high concentrations, the addition of these nanoparticles increases the hydrophobicity of the mats. It is suggested that changes in the contact angles were more likely due to the morphology/size variations of the fibers. It is known that the surface energy and surface geometric features influence the contact angles of fibrous mats [49,50]. Smaller fibers with agglomerated nanoparticles altered the hydrophilicity due to the roughness issue and surface features [51].

Mechanical Properties of Electrospun Mats
Typical stress-strain curves of the electrospun mats are shown in Figure 5. The effect of MND on the mechanical properties are summarized in Table 2. It was found that the addition of 1 wt % MND particles significantly increased the elastic modulus and yield strength of the fibers along with

Mechanical Properties of Electrospun Mats
Typical stress-strain curves of the electrospun mats are shown in Figure 5. The effect of MND on the mechanical properties are summarized in Table 2. It was found that the addition of 1 wt % MND particles significantly increased the elastic modulus and yield strength of the fibers along with ductility loss. Tjong et al. [52] related the mechanism of strength enhancement to the effect of nanoparticles on restricted movement of polymer chains. It was noted that the mechanical properties of the mats were close to that of natural skin, as reported in [27]. Meanwhile, the mechanical durability was degraded at high MND concentrations due to severe agglomeration of the nanoparticles. Herein, it should be noted that no chemical interactions occurred between the particles and the polymer matrix, so mechanical interlocking of the polymer chain by the nanoparticles could be the main mechanism of the enhanced durability.
Mar. Drugs 2016, 14, 128 6 of 12 ductility loss. Tjong et al. [52] related the mechanism of strength enhancement to the effect of nanoparticles on restricted movement of polymer chains. It was noted that the mechanical properties of the mats were close to that of natural skin, as reported in [27]. Meanwhile, the mechanical durability was degraded at high MND concentrations due to severe agglomeration of the nanoparticles. Herein, it should be noted that no chemical interactions occurred between the particles and the polymer matrix, so mechanical interlocking of the polymer chain by the nanoparticles could be the main mechanism of the enhanced durability.

Permeability
The water vapor permeability of films was determined by a gravimetric method at room temperature. It is noteworthy that permeability of wound dressings is an important factor in the skin repairing process in order to create the appropriate moisture in wound area and stop over-drying [17]. Figure 6 shows weight change of the mats over time per unit area. The slope of the lines yields the average value of permeability (Table 2). It is seen that the permeability of the electrospun mats decreases with increasing the ND concentration. It is known that the permeability is related to the size and structure of pores in the fibrous mats [53]. As shown in Figure 1, the membranes containing MND exhibited finer fibers with a more closed-packed structure; hence, the permeability was reduced.

Figure 5.
Stress-strain curves of electrospun mats containing different amounts of diamond nanoparticles.

Permeability
The water vapor permeability of films was determined by a gravimetric method at room temperature. It is noteworthy that permeability of wound dressings is an important factor in the skin repairing process in order to create the appropriate moisture in wound area and stop over-drying [17]. Figure 6 shows weight change of the mats over time per unit area. The slope of the lines yields the average value of permeability (Table 2). It is seen that the permeability of the electrospun mats decreases with increasing the ND concentration. It is known that the permeability is related to the size and structure of pores in the fibrous mats [53]. As shown in Figure 1, the membranes containing MND exhibited finer fibers with a more closed-packed structure; hence, the permeability was reduced.

Cell Viability Assessment
Many studies have shown that diamond particles are cytocompatiable [54][55][56]. To be certain that embedding of these particles in the fibrous structure of CS/BC prepared by electrospinning does not degrade biocompatibility, MTT assay was performed. Figure 7 shows the cell viability of electrospun mats containing different amounts of MND after 24 and 72 h incubation. The viability varied in the range of 75%-90% of the control sample. As compared with previous studies [8], the prepared mats exhibited better cytocompatibility to L929 cells. The cell viability decreased when increasing the amount of MND beyond 2 wt %. This observation might be related to the agglomeration of the nanoparticles as well as refined fibrous structure with different pore sizes.

Cell Viability Assessment
Many studies have shown that diamond particles are cytocompatiable [54][55][56]. To be certain that embedding of these particles in the fibrous structure of CS/BC prepared by electrospinning does not degrade biocompatibility, MTT assay was performed. Figure 7 shows the cell viability of electrospun mats containing different amounts of MND after 24 and 72 h incubation. The viability varied in the range of 75%-90% of the control sample. As compared with previous studies [8], the prepared mats exhibited better cytocompatibility to L929 cells. The cell viability decreased when increasing the amount of MND beyond 2 wt %. This observation might be related to the agglomeration of the nanoparticles as well as refined fibrous structure with different pore sizes.

Cell Viability Assessment
Many studies have shown that diamond particles are cytocompatiable [54][55][56]. To be certain that embedding of these particles in the fibrous structure of CS/BC prepared by electrospinning does not degrade biocompatibility, MTT assay was performed. Figure 7 shows the cell viability of electrospun mats containing different amounts of MND after 24 and 72 h incubation. The viability varied in the range of 75%-90% of the control sample. As compared with previous studies [8], the prepared mats exhibited better cytocompatibility to L929 cells. The cell viability decreased when increasing the amount of MND beyond 2 wt %. This observation might be related to the agglomeration of the nanoparticles as well as refined fibrous structure with different pore sizes.

Materials
Chitosan, with an average molecular weight of 90-150 kDa and degree of deacetylation of 75%-85%, was supplied by Sigma-Aldrich Co. (St. Louis, MO, USA). Bacterial cellulose nanofibers were purchased from Nano Novin Polymer Co., Sari, Iran. Medical grade diamond nanoparticles with an average particle size of 2-6 nm (Grade PL-D-G, purity > 87%) were supplied by PlasmaChem GmbH (Berlin, Germany). Acetic acid (100%) was obtained from Merck Co., Darmstadt, Germany. Polyethylene oxide (PEO), with a molecular weight of 900 kDa, was provided by Sigma-Aldrich Co.
Polyethylene oxide (PEO), with a molecular weight of 900 kDa, was provided by Sigma-Aldrich Co. (St. Louis, MO, USA).

Preparation of Fibrous Mats
To prepare polymeric solutions for electrospinning, an aqueous solution of chitosan in acetic acid (3 wt %) was prepared. The solution was blended with a bacterial cellulose gel (5 wt %) dissolved in acetic acid (90%). A polyethylene oxide solution in deionized water (5 wt %) was separately prepared and added to the CS/BC suspension. The final composition of the suspension was 45:45:10 CS:BC:PEO. The aim of PEO addition was to reduce the viscosity of the gel and facilitate electrospinning [57].
MND-modified fibrous mats were prepared by electrospinning the polymeric suspension mixed with MND in a concentration of 1, 2, and 3 wt % of the polymer. The diamond nanoparticles were dispersed in deionized water by employing an ultrasonic bath (Wise Clean, Model: WUC-D10h using power of 690 W) for one hour. The dispersion was added to the polymeric suspension and sonicated for one hour.
A single jet electrospinning apparatus (Model HVPS-35/500, ANSTCO, Tehran, Iran) was employed to prepare fibrous mats. The suspension was fed into a horizontally aligned syringe with a needle orifice of 0.55 mm in inner diameter. The distance between syringe and collector was fixed at 10 cm and the applied voltage was 20-22 kV. The electrospinning feeding rate was 0.3 mL/h and the suspensions were spun toward a rotating drum (1000 rpm). The randomly oriented nanofibers were collected on an aluminum foil.

Materials Characterization
The size and morphology of the fibers were studied by field-emission scanning electron microscopy (FE-SEM; Hitachi Ltd., Tokyo, Japan, F-System 4160) at an acceleration voltage of 20 kV. Gold coating was applied before imaging. Fourier transform infrared (FT-IR) analysis was performed by a Spectrum 400 Perkin Elmer (Waltham, MA, USA) in the range of 450-4000 cm´1 with a resolution of 1 cm´1. A universal tensile testing machine (Instron 5566, Norwood, MA, USA) was employed to evaluate the mechanical properties i.e. ultimate Tensile Strength (UTS) and elongation to break, based on ASTM Standard D638. The examined mats had a thickness of 30˘2 µm and a gauge length of 20 mm. The crosshead speed was 5.0 mm/min. Each test was repeated three times and the average values were recorded.
To determine the hydrophilicity of the membranes, water contact measurement was performed by employing an OCA 15 plus video-based optical contact angle meter (Data Physics Instruments GmbH, Filderstadt, Germany). Water vapor permeability (WVP) was gravimetrically determined according to ASTM E96 desiccant method at room temperature. Films with an area of 3.5 cm 2 were prepared and overlaid on cups filled with silica gel. The cups were placed in a desiccator containing saturated salt solution of ammonium nitrate with relative moisture 65%. At several time periods (1-24 h), the weight change was measured. Water vapor transmission (WVT) and permeability (WVP) were determined by [23]: where G is the weight gain of the cup at time t, A the permeation area, ∆P the vapor pressure difference across the mat-calculated based on the vapor saturation pressure (S) and the relative humidity inside (100%) and outside the cup (49%)-and T the average thickness of the mats.

Cell Viability
Cell viability was evaluated using the standard 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay protocol. The assay is based on the conversion of MTT into formazan crystals by living cells, which determines mitochondrial activity. Briefly, 5ˆ10 5 mouse skin fibroblast cells (L929) (National Cell Bank, Iran Pasture Institute) were seeded onto the specimens with a 6-well plate and incubated at 37˝C in 5% CO 2 for 24 and 48 h. After each interval, 200 mL of MTT solution (Sigma, St. Louis, MO, USA, 5 mg/mL) in 1ˆDulbecco's phosphate-buffered saline (PBS; Sigma, St. Louis, MO, USA) was added to each well and the cells were incubated for another 4 h. Upon removal of the MTT solution, the formed formazan crystals were solubilized with isopropanol for 15 min. Absorbance was read at the wavelength of 570 nm. The data were reported separately for each well by an ELISA reader (BioTek Microplate Reader, BioTek Company, Winooski, VT, USA). An average of triplicate wells was calculated, and the standard deviation for each sample was calculated based on Student's t-test (p < 0.05). To observe the morphology of the adherent cells, the films were washed by PBS three times and then immersed in 3% glutaraldehyde PBS solution for 30 min for cell fixing. The films were dehydrated in ascending series of ethanol aqueous solutions (50 to 100 percent) at room temperature. The specimens were kept overnight in a desiccator to remove any moisture. The growth of the cells was observed after 24 h incubation.

Conclusions
Chitosan/bacterial cellulose nanofibrous membranes modified with diamond nanoparticles were fabricated by electrospinning. It was shown that uniform mats with an average fiber diameter of about 130 nm could be prepared at the spinning distance of 120 mm, applied voltage of 20 kV, and CS/BC ratio of 1:1. The effect of MND particles on the electrospinning of CS/BC blends was also studied. It was shown that MND particles reduced the average fiber diameter. The mechanical properties under tensile loading, the hydrophilicity, and water vapor permeability of the nanocomposites were also examined. An enhancement in the mechanical strength with a decrease in strain to failure was measured due to the addition of MND to the CS/BC nanofibers. Introduction of MND particles was accompanied with a gradual decrease in the hydrophilicity of the nanofibrous membranes as well. The water vapor permeability of the CS/BC nanofibers was also decreased. The cell viability of fibrous membrane after a 1-day incubation was >90% as compared with the control. We concluded that the fibrous mat containing 1% MND could be a promising candidate for wound dress and tissue engineering applications. Meanwhile, in vivo studies and pre-clinical investigations are required to explore the suitability of developed mats for biomedical applications.