Coaxial nanofibers outperform uniaxial nanofibers for the loading and release of pyrroloquinoline quinone (PQQ) for biomedical applications

Pyrroloquinoline quinone (PQQ), present in breast milk and various foods, is highly recommended as an antioxidant, anti-inflammatory agent, and a cofactor in redox reactions in several biomedical fields. Moreover, PQQ has neuroprotective effects on nervous system disorders and immunosuppressive effects on different diseases. Herein, we report on the optimum fabrication of electrospun CS/PVA coaxial, core/shell, and uniaxial nanofibers. The morphological, elemental, and chemical structure of the fabricated nanofibers were investigated and discussed. PQQ, as a drug, was loaded on the uniaxial nanofibers and in the core of the coaxial nanofibers and the sustained and controlled release of PQQ was compared and discussed. The results revealed the privilege of the coaxial over the uniaxial nanofibers in the sustained release and reduction of the initial burst of PQQ. Remarkably, the results revealed a higher degree of swelling for CS/PVA hollow nanofibers compared to that of the uniaxial and the coaxial nanofibers. The coaxial nanofibers showed a lower release rate than the uniaxial nanofibers. Moreover, the CS/PVA coaxial nanofibers loaded with PQQ were found to enhance cell viability and proliferation. Therefore, the CS/PVA coaxial nanofibers loaded with PQQ assembly is considered a superior drug delivery system for PQQ release.


Introduction
The process of forming nanobers using an applied external eld has been reported for the rst time by Formhals in 1934. 1 This process is known as electrospinning, a technique in which bers are formed via charging polymer solutions electrically using high voltage and obtaining them at a specic feed rate. The obtained nanobers, from the aforementioned technique, have diameters that range from a few nanometers to several micrometers. 2 The applied external voltage causes repulsive forces within the polymer's solution to exceed the surface tension of a solution droplet at the tip of a needle. Consequently, it creates deformation upon the polymer solution's spherical droplet to elongate forming a Taylor cone, which in turn creates the strands of nanobers. In fact, the created nanobers still have to travel a sufficient distance from the needle to the collector to allow them to dry out from any remaining solvent, producing dry, well-separated solid nano-bers. 3 There are numerous factors that can affect the bers' morphology and diameter including the polymer viscosity, humidity, ow rate, tip-to-collector distance (TCD), and applied voltage. [4][5][6][7] Moreover, the accumulation of high surface area and high porosity nanobers to form a membrane is a perfect candidate to be used in a wide range of medical applications, including tissue engineering and drug delivery. 8 Coaxial electrospinning is an improvement of the uniaxial electrospinning process in which two blunt-tip needles are connected to a high voltage with the same or different ow rates. Coaxial electrospinning provides nanobers with a different and separate polymer solution for the core and shell, a tool used to fabricate hollow nanobers or bers with a varying core-shell. [9][10][11] In comparison with uniaxial nano-bers, coaxial bers have many advantages where the coaxial technique enables the spinning of the material, which is not easily spun via uniaxial electrospinning. 12,13 Moreover, in medical applications, sensitive bioactive substances such as antibiotics, drugs, DNA, and proteins, which are spun as the core solution, can be safely isolated from the harsh environment by the shell layer of nanobers. Subsequently, controlling the drug release is possible via tuning the thickness of the core and shell. 13,14 In addition, similar work has been done using novel biomaterial coaxial nanobers for bone and cartilage regeneration. 15,16 In the present study, chitosan (CS) and polyvinyl alcohol (PVA) were selected as core and shell polymers, respectively. CS is a natural biopolymer with a repeated polysaccharide unit, which consists of N-acetyl glucosamine and glucosamine linked by glycosidic bonds. 17 Many previous studies reported that the electrospinning of pure CS is not easy due to high viscosity, limited solubility and the strong molecular interaction of CS. 18,19 Therefore, to facilitate the synthesis of CS nanobers, chitosan/other polymer blends such as PVA have been used. 19,20 CS/PVA nanobers have been used in many biomedical applications such as tissue engineering, drug delivery and wound healing due to their antibacterial, biocompatible and biodegradable properties. Our study selects coaxial CS/PVA nano-bers loaded with the natural component pyrroloquinoline quinone (PQQ) as a promising material for biomedical applications such as a drug delivery system for many medical applications.
PQQ is an anionic, water-soluble compound that has been discovered in several fruits, vegetables, breast milk, and tissues of mammals. 21,22 PQQ was initially identied as a redox enzyme cofactor and later proved to be an essential nutrient for animal growth. It has been reported that the removal of PQQ from the diet causes growth impairment in animals. 23,24 As an essential nutrient, PQQ has been classied as a new B vitamin with many other benecial effects, such as anti-inammatory, 25 hepatoprotective, 26 cardioprotective, 27 and antioxidative properties. 28,29 Furthermore, much evidence shows that PQQ is a key player in the central and peripheral nervous systems. For example, PQQ shields the brain against reversible middle cerebral artery occlusion, 28 and increases the stimulus-response and the nerve growth factor levels in astroglial cells. 30 In addition, it enhances the locomotor function aer spinal cord injury in rats. This is achieved by lessening the secondary neuronal damage following primary mechanical injury 31 and it further improves the regeneration of rat sciatic nerves. 32 Consequently, PQQ is considered a perfect candidate for many medical applications and a key solution for complications, especially in the nervous system.
Herein, we demonstrate the fabrication of CS/PVA uniaxial and coaxial nanobers and compare their efficiency for the sustained release of pyrroloquinoline quinone (PQQ). The fabricated materials were fully characterized and compared. Also, the cell viability and proliferation of the cells on both classes of nanobers were evaluated for use in biomedical applications.

Preparation of nanober solutions
The prepared solutions of CS and PVA were used for uniaxial and coaxial electrospinning as previously reported by Ibrahim et al. 33 Briey: (1) 3 g of CS (3% w/v) were dissolved in 100 ml of AA/distilled water solution (2% v/v) and heated in a water bath at 90 C with vigorous stirring for 2 h until complete dissolution, and used as a stock.
(2) 20 g of PVA (20% w/v) were dissolved in 100 ml of distilled water and heated in a water bath at 90 C with vigorous for 3 h until complete dissolution, and used as a stock.
(3) 70 ml of PVA/distilled water solution were mixed with 30 ml of CS/AA/distilled water solution in volume ratios of 30/70 of CS/PVA, with vigorous stirring till reaching a homogenous solution. 34 (4) Aer complete homogenization, 40 ml of distilled water were added to 60 ml of CS/PVA homogenous mixture and were allowed to vigorously stir for 1 h at 60 C to maintain the viscosity of the solution. 33,34 The pH and viscosity of various prepared solutions were adjusted using a pH meter and a viscometer.
(5) For drug loading, PQQ (15% w/v) was added to the mixture, where PQQ solution was prepared by dissolving in distilled water. The CS/PVA solution loaded with PQQ was homogenized using a homogenizer for 1 h at 3000 rpm. Finally, the solutions (CS/PVA and CS/PVA loaded with PQQ) were le on the bench overnight to get rid of air bubbles before the electrospinning process.

Electrospinning process of solutions
An electrospinning apparatus (MECC Nanon-01A, Japan) was used to fabricate nanobers with hollow and core/shell structures. The blunt-tip coaxial needle was composed of a 23 G (i.d.

Nanober testing
The elemental and morphological analysis of the fabricated nanobers were performed using a eld emission scanning microscope (FSEM, Zeiss SEM ultra 60) and a transmission electron microscope (TEM, JEM-1230, Jeol, Japan). To conrm the chemical bonding, Fourier Transform Infrared spectroscopy (FT-IR, Thermo Scientic Nicolet 8700, USA) was used to characterize the chemical structure of the uniaxial CS/PVA/PQQ, CS/ PVA hollow and CS/PVA/PQQ coaxial nanobers.

Degree of swelling and degradation experiments
CS/PVA uniaxial, hollow and core/shell membranes were cut into small pieces of 3 Â 3 cm and the initial weight of nano-bers (W i ) was recorded. For the swelling test, the pieces of nanobers were immersed in phosphate buffer solution (PBS; pH 7.4) at 37 C for 1, 3, and 7 days, and the samples were weighed again immediately aer removing them from PBS and recorded as (W). Subsequently, the excess amount of PBS absorbed with lter paper and the weight of the nanober sample (W x ) were recorded. The degree of swelling of the nanobers was calculated using eqn (1): where W is the weight of swollen nanober samples by immersion in PBS, pH ¼ 7.4 at 37 C for 1, 2 and 3 days. W d is the weight of the dried nanober sample.
Regarding the degradation test, the nanobers were immersed in phosphate buffer solution (PBS; pH 7.4) at 37 C for 1, 3, and 7 days, and the nanobers were dried in a vacuum oven for 24 h and the weight of the dried nanober sample (W d ) was recorded. The degradation index (%) was calculated using eqn (2): where W i is the initial weight of the nanobers.

PQQ release experiment (in vitro)
Using a Franz diffusion cell apparatus (Automated Microette PlusTM; Hanson Research, USA) CS/PVA/PQQ uniaxial nano-bers, and CS/PVA and CS/PVA/PQQ coaxial nanobers were prepared to test drug release. The area of Franz diffusion cells was 1.76 cm 2 with a receptor compartment volume of 6.8 ml.
The test was performed using a cellulose membrane with a molecular weight cut-off value of 1400 g mol À1 (Sigma Aldrich, USA). The nanobers were placed in the donor compartment, while PBS with pH 7.4 was placed in the receptor compartment with adjusting temperature at 37 C and the stirring rate at 200 rpm. Samples (2 ml) were taken at 24, 72, 120 and 168 hours and diluted to measure the PQQ concentration by using a UVvisible spectrophotometer at a l max of 280 nm.

Antibacterial activity test
The antimicrobial activity of the nanober samples was determined using a modied Kirby-Bauer disc diffusion method. 35 Briey, 100 ml of the test bacteria were grown in 10 ml of fresh media until they reached a count of approximately 108 cells per ml for bacteria. 36 100 ml of microbial suspension was spread onto agar plates corresponding to the broth in which they were maintained. Isolated colonies of each organism that might be playing a pathogenic role should be selected from primary agar plates and tested for susceptibility by the disc diffusion method. 37,38 Gram(+) bacteria such as Staphylococcus aureus and Bacillus subtilis and Gram(À) bacteria such as Escherichia coli and Pseudomonas aeruginosa were incubated at 35-37 C for 24-48. The diameters of the inhibition zones were measured in millimeters. 35 Standard discs of ampicillin (antibacterial agent) served as positive controls for antimicrobial activity but lter discs impregnated with 10 ml of solvent (distilled water, chloroform, DMSO) were used as a negative control. Blank paper disks (Schleicher & Schuell, Spain) with a diameter of 8.0 mm were impregnated with 10 ml of tested concentration of the stock solutions. When a lter paper disc saturated with a tested chemical is placed on agar the chemical will diffuse from the disc into the agar. This diffusion will place the chemical in the agar only around the disc. The size of the area of chemical inltration around the disc is determined via the solubility of the chemical and its molecular size. If an organism is placed on the agar it will not grow in the area around the disc if it is susceptible to the chemical. This area of no growth around the disc is known as a "Zone of inhibition" or "Clear zone". For the disc diffusion, the calibers of the National Committee for Clinical Laboratory Standards 38 were used to quantitatively measure the inhibition zone diameters. The agar-based methods including tests such as Etest and disk diffusion can be good substitutes for broth-based methods due to their simplicity, higher accuracy, and the less time taken for the test. 39,40 MTT cytotoxicity assessment Regarding the evaluation of human broblast cell proliferation (in vitro), in a Fibrolife® serum-free medium, human dermal broblasts, adult (HDFa) (American Type Culture Collection® PCS-201-012™, USA) cell cultures were cultivated at 37 C and 5% CO 2 in T75 culture asks (Corning®, USA) for 24-27 hours. In order to determine the effect of CS/PVA hollow and CS/PVA/ PQQ core/shell nanobers on the proliferation of HDFa via 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assay the percentage viable cell was used 41 as follows: (1) CS/PVA uniaxial, hollow and core/shell nanobers were suspended (three replicates for each sample) with HDFa in a Fibrolife® serum free medium at a concentration of 5 Â 10 4 cell per well in Corning® 96-well tissue culture plates.
(2) Then the plates were incubated for 72 hours.
(3) Vehicle control with DMSO was run for each 96 well plate (untreated cells).
(4) Aer incubating for 72 h, the number of viable cells was then determined by adding 10 ml of 12 mM MTT stock solution (5 mg of MTT in 1 ml of phosphate buffer saline pH 7.4) into each well, including also the untreated control. 42,43 (5) The 96-well plates were then incubated at 37 C and 5% CO 2 for 4 hours.
(6) The cells were periodically viewed under an inverted microscope (Olympus BX63 Life Science, Japan) to detect the presence of intracellular punctuate purple precipitation. The optical density was measured at 490 nm with a microplate reader (680 XR reader, BIORAD, Hercules, CA, USA) to determine the number of viable cells.

Paper
Nanoscale Advances (7) The percentage of viability was calculated using the following eqn (3): where OD t is the mean optical density of the wells treated with the tested samples (drug or the selected formulation), while OD c is the mean optical density of the untreated cells.

Results & discussion
Morphological, structural, and elemental analysis The coaxial electrospinning technique was used to fabricate CS/ PVA hollow nanobers using acetic acid/distilled water solution as a solvent. On one hand, CS/PVA nanobers, as a shell, were synthesized under constant processing conditions where the concentration of CS/PVA mixtures was 50%, with a voltage of 27 kV, TCD of 15 cm, humidity 30-40%, and ow rate of 0.4 ml h À1 . On the other hand, the core was made hollow by lling the syringe with air using the same ow rate of the shell solution (0.4 ml h À1 ). Fig. 1 shows the as-synthesized hollow nanobers prepared under the aforementioned conditions, revealing the good uniformity with a ne structure that is free from beads. Note the transparency of the nanobers, which indicates the hollow core of the as-synthesized nanobers. Fig. 2a,b show the FESEM images of the coaxial CS/PVA//CS/ PVA/PQQ (shell//core) nanobers prepared under constant processing conditions (27 kV, TCD 15 cm, humidity 30-40% and ow rate of 0.3 ml h À1 for core solution and 0.4 ml h À1 for shell solution). Note that the coaxial nanobers exhibited uniform bers with a shell/core structure. The presence of core bers with a smaller diameter within the shell ber is evident. To further investigate the presence of the core/shell structure within the nanobers, the core/shell as well as the hollow nanobers were scanned by TEM. The TEM images shown in Fig. 2c and d depict the hollow and core/shell structure, respectively.
The morphology of the uniaxial CS/PVA/PQQ electrospun nanobers was observed by FESEM. Fig. 3a reveals that the fabricated bers have a smooth surface, uniform and completely bead-free. The elemental analysis of CS/PVA/PQQ is shown in Fig. 3b, indicating the existing elements as tabulated in the inset of Fig. 3b. The differences in the elemental peak intensities correspond to the initial elemental concentration used for the synthesis. It is quite lucid that carbon exhibits the most intense peak due to the high concentration of polymer used. The carbon constitutes the main backbone of all the polymer chains in the CS/PVA polymer. In addition, the remaining elements' (oxygen and nitrogen) peak intensities are directly proportionally to their atomic weight in the compound.
The FTIR spectra were collected to elucidate the chemical bonding for CS/PVA, PQQ, and CS/PVA/PQQ as shown in Fig. 4. The spectra of CS/PVA show a series of characteristic peaks for chitosan and PVA. The peak at 844 cm À1 conrms the presence of a saccharide structure as it is a major characteristic peak of the C-H bending in chitosan. Moreover, the 1096 cm À1 peak can be attributed to the C-O stretching vibrational mode in chitosan. Also, the peaks at 1375 and 1415 cm À1 are characteristic of the CH 3 symmetrical deformation mode. The peak at 1640 cm À1 can be ascribed to the amide bond in chitosan. The vibrational band between 2790 and 2980 cm À1 can be ascribed to the C-H stretching of the alkyl groups. Finally, the O-H stretching shown in the broad band from 2980-3660 cm À1 is   carboxyl groups. As is lucid from the CS/PVA/PQQ spectra, the reaction between PQQ and CS/PVA caused the minimization/ loss of most of the peaks except for the minor two peaks located at 1632 and 2363 cm À1 and the minor broad band at 3430 cm À1 . They were indicated on the gure due to their low transmission.

Degree of swelling and degradation
To assess the physical properties of the fabricated nanobers for biomedical applications, the degree of swelling and degradation are considered critical parameters that need to be quantitatively evaluated. The results of the degree of swelling are represented in Fig. 5a    This journal is © The Royal Society of Chemistry 2020 Nanoscale Adv., 2020, 2, 3341-3349 | 3345 Paper Nanoscale Advances compared to that of the uniaxial and the coaxial nanobers. For the coaxial nanobers, the presence of CS/PVA/PQQ in the core decreased the swelling ratio of the coaxial nanobers on the seventh day of the test from 440.77% for CS/PVA hollow and 400.10% for CS/PVA/PQQ uniaxial nanobers to 300.09%. Furthermore, this reduction in the degree of swelling of the nanobers can be ascribed to the presence of a core structure within the shell, resulting in increasing the rigidity of the nanobers. Fig. 5b shows the degree of degradation of the as-synthesized nanobers. The slower the rate of degradation, the slower the release of PQQ, which is the most favorable characteristic in medical applications and sustained drug release systems. The results showed that the degradation index of CS/PVA hollow nanobers is around 6.74% for day one, 11.52% for day three, 14.13% for day ve, and 16.77% for day seven. For uniaxial CS/ PVA/PQQ nanobers, the degradation index recorded was 5.01, 8.99, 11.99, and 14.88% for the 1st, 3rd, 5th, and 7th day, respectively. However, the degradation rate of PVA/CS//CS/PVA/ PQQ nanobers is around 3.907% for day one, 7.90% for day three, 10.12% for day ve, and 13.22% for day seven. Interestingly, it was clearly observed that the degradation rate for CS/ PVA//CS/PVA/PQQ coaxial scaffolds was much slower than that of uniaxial and hollow nanobers. The presence of CS/PVA in the core of coaxial nanobers was helpful to reduce the degradation rate of bers, because it takes more time to degrade this core part in comparison with the time consumed to degrade the uniaxial and hollow nanobers.

In vitro PQQ release
In order to verify the previous results and better compare the uniaxial and coaxial nanobers, a release study of the PQQ was carried out by loading PQQ in the core of the CS/PVA//CS/ PVA coaxial nanobers as well as by loading the surface of CS/ PVA uniaxial nanobers, see Fig. 6. The concentration of the loaded PQQ in uniaxial and coaxial nanobers for the drug release study was kept constant equivalent to 3.5 mg ml À1 . 44 During the rst hour of the test, 20% of PQQ from the uniaxial nanobers was released. However, only 10% of PQQ was released from the coaxial nanobers without initial burst release. Moreover, the PQQ release from the core of the coaxial nanobers reached 57.2% aer 168 hours of the experiment. In contrast, in the same time interval of the test, a cumulative release of 74% of PQQ occurred from the uniaxial nanobers. Note that the coaxial nanobers have a lower release rate than the uniaxial nanobers. Since PQQ exists in the core of coaxial nanobers, the CS/PVA shell acted as a barrier to slow down the rate of release of PQQ. In comparison, the uniaxial nanobers, in which PQQ was loaded on their surface, have shown a rapid release. The difference in the release prole shown in Fig. 6 is due to the presence of the CS/PVA shell layer. Normally, this shell layer acts as a barrier to mass transfer and thus reduces the release rate of PQQ from the coaxial nanobers in comparison with the uniaxial nanobers. 44 Antibacterial test Table 1 and Fig. 7 show the inhibitory activity of CS/PVA hollow, CS/PVA/PQQ uniaxial and CS/PVA//CS/PVA/PQQ coaxial nanobers in comparison with the antibacterial agent ampicillin. Generally, all the fabricated nanobers inhibited the growth of the tested bacteria as indicated by the   Nanoscale Advances Paper inhibition zone. The data in Table 1 show that CS/PVA hollow nanobers showed the least inhibition of the growth of all tested bacteria. However, it still indicates a higher inhibition effect aer loading PQQ on coaxial and uniaxial nanobers. The results suggest that PQQ loaded on CS/PVA nanobers as uniaxial or coaxial nanobers can be used as a potential antibacterial scaffold for medical applications. Two main modes of action contribute to the resultant antibacterial effect indicated by the inhibition zones tabulated in Table 1. Firstly, chitosan can bind to the bacterial cell wall due to its negative charge. Subsequently, this binding can cause cell disruption and attachment to the DNA, which prevents DNA replication and cell death. 45 Secondly, PQQ is naturally used in bacteria to change glucose into gluconic acid, which has a pH of 5-5.5. 46 Since PQQ is added in excess to the bacteria (loaded on the nanobers), more glucose is changed into gluconic acid, which consumes high energy. The bacteria used in the test work were at neutral to slightly alkaline pH. Thus, the accumulation of gluconic acid causes slight impairment of the bacterial behavior, motility, and growth. 47 The aforementioned concluded modes of action can be conrmed by the less inhibition zone diameter of the hollow CS/PVA nanober due to the absence of PQQ. There is not much difference between the inhibition of Gram positive and Gram negative bacteria, which could be ascribed to the broad spectrum of the investigated coaxial/uniaxial nanobers. In addition, it could also be attributed to the antibacterial mechanism that does not interact with the cell membrane.

MTT cytotoxicity assessment
In order to evaluate whether the loading of PQQ on CS/PVA nanobers will be effective for many medical applications, the effect of CS/PVA hollow and PQQ loaded on CS/PVA nanobers on proliferation of HDFa cells was examined using MTT assay, Fig. 8. The CS/PVA/PQQ nanobers showed a signicant increase in the proliferation of HDFa cells aer 72 h of incubation, followed by CS/PVA hollow nanobers (p < 0.05) (Fig. 8).
The difference in the cell growth rate and cell numbers between CS/PVA and CS/PVA/PQQ nanobers proves the role of PQQ loaded on CS/PVA nanobers in increasing the proliferation of HDFa cells. Additionally, the comparison between uniaxial and coaxial nanobers on the proliferation of HDFa cells reveals that coaxial nanobers showed an increase in cell growth more than the uniaxial nanobers, which could be attributed to low cytotoxicity. In addition, other work including Lu et al. has reported that the structure of the coaxial bers resembles the extracellular matrix, which enhances cell proliferation and growth. 48 Furthermore, the MTT assay results showed that the nanober scaffold had no cytotoxicity to HDFa and did not cause inhibition of proliferation or differentiation of HDFa.  This journal is © The Royal Society of Chemistry 2020 Nanoscale Adv., 2020, 2, 3341-3349 | 3347 Paper Nanoscale Advances

Conclusion
In conclusion, electrospun CS/PVA coaxial, core/shell, and uniaxial nanobers were successfully fabricated. FESEM imaging revealed the good uniformity of the fabricated nano-bers with a ne structure that is free from beads. Moreover, their TEM images proved the hollow and core/shell structures. The elemental composition and the chemical bonding for CS/ PVA, PQQ, and CS/PVA/PQQ were elucidated using EDX and FTIR analyses. The results proved the privilege of CS/PVA coaxial nanobers in comparison with uniaxial nanobers when loaded with PQQ for biomedical applications. The release study of PQQ showed that CS/PVA coaxial nanobers have a lower release rate than uniaxial nanobers with less initial release burst in comparison with uniaxial nanobers. In addition, coaxial nanobers have lower degradation and swelling rates than uniaxial nanobers. Furthermore, it enhances the proliferation of HDFa cells. The obtained results strongly suggest the fabricated CS/PVA coaxial nanobers loaded with PQQ in the core for use in many applications as an alternative to many complicated and expensive drug systems.

Conflicts of interest
There are no conicts to declare.