Electrospun Fibres of Chitosan/PVP for the Effective Chemotherapeutic Drug Delivery of 5-Fluorouracil

: Electrospun nanoﬁbrous mats consisting of chitosan (CS) and polyvinylpyrrolidone (PVP) were constructed. Tuning of solution and process parameters was performed and resulted in an electrospun system containing a 6:4 ratio of PVP:CS. This is a signiﬁcant increase in the proportion of spun CS on the previously reported highest ratio PVP:CS blend. SEM analysis showed that the nanoﬁbrous mats with 4 wt% CS/6 wt% PVP (sample E) comprised homogenous, uniform ﬁbres with an average diameter of 0.569 µ m. XPS analysis showed that the surface of the samples consisted of PVP. Raman and FTIR analysis revealed intermolecular interactions (via H-bonding) between PVP and CS. In FTIR spectra, the contribution of chitosan to CS/PVP complexes was shown by the downshift of the C=O band and by the linear increase in intensity of C-O stretching in CS. XPS analysis showed a smaller shift at the binding energy 531 eV, which relates to the amide of the acetylated functional groups. The obtained results demonstrate a sensitivity of Raman and FTIR tests to the presence of chitosan in PVP:CS blend. The chemotherapy drug 5-Fu was incorporated into the constructs and cell viability studies were performed. WST-8 viability assay showed that exposure of A549 human alveolar basal epithelial cells to 10 mg/mL 5-Fu loaded ﬁbres was most effective at killing cells over 24 h. On the other hand, the constructs with loading of 1 mg/mL of drug were not efﬁcient at killing A549 human alveolar basal epithelial cells. This study showed that CS/PVP/5-Fu constructs have potential in chemotherapeutic drug delivery systems.


Introduction
With advances in nanotechnology, substantial research is focused on the properties of nanoscale materials. Electrospinning is a technique which is widely used to manufacture nanofibers. Nanofibrous scaffolds have several applications ranging from filtration to biomedical applications. Electrospun fibres and scaffolds are finding their place as emerging devices in the delivery of cancer therapeutics [1,2]. The efficiency of electrospun fibres can be attributed to their large surface area to volume ratio characteristics and their 3-dimensional open porous system, the properties of which allow for high levels of drug loading and enhanced drug diffusion over time [3][4][5].
Chitosan (CS) is a widely studied natural polysaccharide because of its inherent non-toxicity, biocompatibility and biodegradability, and thus has been used in numerous biomedical applications including drug delivery [6]. While electrospinning of CS is commonly reported, it is notoriously difficult to electrospin and process in its pristine form. A

Preliminary Analysis/Solubility Testing
Initial tests were performed to examine the solubility of chitosan. Electrospinning of chitosan was unachievable using pure TFA. Glacial acetic acid was then introduced to the system in a 9:1 ratio which saw the production of fibres (see Table S1). The parameters that were altered were chitosan concentration, solvent and solvent concentration. Low-, medium-and high-molecular weight chitosan was investigated for its processability. The use of hydrolysed chitosan was also investigated. Briefly, 1 g of low molecular weight chitosan was added to 12.5 g of sodium hydroxide and 25 mL of water. This was heated to 95 • C for 48 h. After this, the chitosan was filtered off. The collected chitosan was subsequently washed with water, neutralized with acetic acid, rinsed again with water and then finally dried for 16 h at 16 • C. A range of solutions of different hydrolysed chitosan concentrations was made up in 90% acetic acid.

Chitosan-PVP Solutions
To create chitosan-PVP solutions, 6 wt% PVP was dissolved in TFA:acetic acid in a 9:1 ratio. This was stirred at room temperature for 3 h. Several solutions were made using 6 wt% PVP and chitosan in 9:1 TFA:glacial acetic acid. The concentrations of chitosan in solution were as follows: 0%, 1%, 2%, 3%, 4%. These solutions were stirred for 8 h at room temperature to ensure full homogeneity. The samples were labelled from A through to E, with A containing 0% CS 6% PVP, B containing 1% CS 6% PVP, C containing 2% CS 6% PVP, D containing 3% CS 6% PVP and E containing 4% CS 6% PVP (Table S1).

Chitosan-PVP-5-Fu Solutions
New 4 wt% CS/6 wt% PVP solutions were developed using the solvent system as above. 5-Fu was added to make three separate solutions in total. These were 1 mg/mL, 5 mg/mL and 10 mg/mL of drug in each CS/PVP solution.

Electrospinning Procedure
Each sample was loaded into a 5 mL syringe. This was then attached to the SprayBase ® pump system. Luer Lock tubing was connected to the end of the syringe. The other end of the tubing was connected to an 18-gauge needle, which was held in place on the SprayBase ® electrospinning system. The voltage was set at 15 kV, while the tip-collector gap was set at 15 cm. A flow rate of 0.1 mL h −1 was used to collect the electrospun fibres. The fibres were collected for three hours on aluminium foil on a flat plate collector. The fibres were then dried at 123 • C for 18 h in an oven.

Characterization
Each CS/PVP sample was removed from their aluminium foil and stored in air-tight tubes. Zeiss Gemini Ultra was used to perform Scanning Electron Microscopy (SEM) on each sample. ImageJ software was used to perform fibre diameter measurements, where 30 random fibres of each sample were measured. XPS analyses were performed on a ThermoFisher Scientific Instruments (East Grinstead, UK) K-Alpha+ spectrometer. XPS spectra were acquired using a monochromated Al Kα X-ray source (hν = 1486.6 eV). An X-ray spot of~400 µm radius was employed. Survey spectra were acquired employing a Pass Energy of 200 eV. High resolution, core level spectra for all elements were acquired with a Pass Energy of 50 eV. All high-resolution spectra were charge referenced against the C1s peak at 285 eV to correct for charging effects during acquisition. Quantitative surface chemical analyses were calculated from the high resolution, core level spectra following the removal of a non-linear (Shirley) background. The manufacturers Avantage software (Thermo Fisher Scientific, Waltham, MA, USA) was used which incorporates the appropriate sensitivity factors and corrects for the electron energy analyser transmission function. Raman measurements were carried out in backscattering geometry using a Renishaw 1000 micro-Raman system with motorized positional stage and Leica microscope. A He-Ne laser at wavelengths of 633 nm at a power of <10 mW (at the laser output) was used as the excitation source. The power was kept low to prevent the sample overheating. To reduce power further, laser filters were used with 50% of transmittance from the original laser power. The laser spot was focused on the sample surface (deposited onto cleaned Silicon wafer) using 50× magnification objectives with a long focus working distance. Since the angle of the scattering light cone with this objective is smaller than one obtained with short focus working distance objectives, the registered spectra even at a long time of signal accumulation (400 s) were relatively weak. Additionally, to register any spectrum reliably, it was necessary to focus the laser light on a single fibre and this resulted in low spectral intensity. Finally, Shimadzu IRAfinnity-1S was used to perform Fourier Transform Infrared Spectroscopy (FTIR) analyses using MIRacle 10 Single Reflection Attenuated Total Reflection (ATR) Accessory. The instrument was set at 100 runs to collect spectra in the region of 900-3900 cm −1 and 900-1900 cm −1 .

Cell Culture
A549 cells were cultured with Ham's F-12 medium containing 10% fetal bovine serum and 1% penicillin: streptomycin at 37 • C with an atmosphere of 95% air and 5% CO 2 . Cells in the logarithmic stage of growth were used for analysis.

Cell Viability Assay
Cell viability was determined using a Water-Soluble Tetrazolium-8 (WST-8) assay. Cells were seeded in a 96-well flat-bottomed plate at a density of 8 × 10 4 cells/mL with 100 µL per well. They were then incubated for 24 h and following this, exposed to electrospun constructs for an additional 24 h. After treatment, all media was removed from the cells. 100 µL of fresh media was added to the wells along with 10 µL of the WST-8 reagent. The plate was then incubated for two hours at 37 • C in a 5% CO 2 incubator. The plate was read at 465 nm using a microtiter plate reader. At least three replicates were performed for each test.

Electrospinning of Chitosan/PVP
Both solution and process parameters play a vital role in achieving the electrospinning of any polymer. Blend electrospinning is a common way of combining two polymers with desirable characteristics and a facile way of incorporating therapeutics into the resultant systems.
Prior to the preparation of blends, an initial investigation of set system parameters was performed. Electrospinning of PVP alone at a concentration of 6% with process parameters of 15 kV voltage, 15 cm tip collector gap and a flow rate of 0.1 mL/h resulted in PVP fibres which were homogenous with good morphology. Chitosan was then systematically introduced into the blend and each Chitosan/PVP blend was electrospun keeping the initial process parameters consistent (Table S1).

Tuning of Chitosan/PVP Blend Solution Parameters
A series of blends of PVP and chitosan were electrospun in a solvent system of TFA and Acetic Acid in the ratio of 9:1. The polymer concentrations were as follows: (A) 0% chitosan 6% PVP, (B) 1% chitosan 6% PVP, (C) 2% chitosan 6% PVP, (D) 3% chitosan 6% PVP and (E) 4% chitosan 6% PVP. While fibre formation could be achieved based on these PVP and chitosan concentrations, instrument parameters were next investigated and refined accordingly to produce a stable ejection of fibres from the electrospinner (Table S2).

Investigation of the Electrospinning Process Parameters on Fibres
Good-quality fibres were produced by setting the voltage to 15 kV, tip-collector gap to 15 cm and flow rate to 0.1 mL/h −1 , which resulted in a steady/stable jet ejection being emitted from the Taylor Cone located at the needle tip. PVP concentration could not be raised above 6%, while the chitosan concentration could not be raised above 4% since fibres were not formed when concentrations were raised above these values. It was necessary to use low molecular weight chitosan (mol. wt. 50,000-190,000) to eliminate the possibility of the solution becoming gel-like. The solvent system best suited was deemed to be TFA and acetic acid in a 9:1 ratio. It was necessary to set voltage at 15 kV; any higher distorted the Taylor Cone and any lower removed the Tylor Cone completely. The tip-collector gap was set at 15 cm while the flow rate was set at 0.1 mL/h −1 . This is a very slow flow rate which is important for the formation of chitosan/PVP fibres. A faster flow rate, as seen above, produces droplets from the Taylor Cone. It is also worth noting that the effect of temperature and humidity was initially studied and there were no significant changes to the spinnability of the polymers in our room-temperature setup.

Characterization of Electrospun Blends
Characterization of the CS/PVP fibres resulting from the electrospun blends was carried out by SEM, XPS, Raman and FTIR spectroscopy.

SEM Analysis
Each fibrous scaffold (A, B, C, D and E) was sputter-coated and analysed using scanning electron microscopy. Images were obtained for magnifications at 5 K (see Figure 1).

Investigation of the Electrospinning Process Parameters on Fibres
Good-quality fibres were produced by setting the voltage to 15 kV, tip-collector gap to 15 cm and flow rate to 0.1 mL/h −1 , which resulted in a steady/stable jet ejection being emitted from the Taylor Cone located at the needle tip. PVP concentration could not be raised above 6%, while the chitosan concentration could not be raised above 4% since fibres were not formed when concentrations were raised above these values. It was necessary to use low molecular weight chitosan (mol. wt. 50,000-190,000) to eliminate the possibility of the solution becoming gel-like. The solvent system best suited was deemed to be TFA and acetic acid in a 9:1 ratio. It was necessary to set voltage at 15 kV; any higher distorted the Taylor Cone and any lower removed the Tylor Cone completely. The tipcollector gap was set at 15 cm while the flow rate was set at 0.1 mL/h −1 . This is a very slow flow rate which is important for the formation of chitosan/PVP fibres. A faster flow rate, as seen above, produces droplets from the Taylor Cone. It is also worth noting that the effect of temperature and humidity weas initially studied and there were no significant changes to the spinnability of the polymers in our room-temperature setup

Characterization of Electrospun Blends
Characterization of the CS/PVP fibres resulting from the electrospun blends was carried out by SEM, XPS, Raman and FTIR spectroscopy.

SEM Analysis
Each fibrous scaffold (A, B, C, D and E) was sputter-coated and analysed using scanning electron microscopy. Images were obtained for magnifications at 5 K (see Figure 1). The mean fibre diameter for sample A (6% PVP, 0% chitosan) was found to be 0.426 µm (SD of 0.115) after taking measurements of 30 random fibres. The mean fibre diameter is low since there is only PVP present. In theory, it is expected that the lowest concentration of polymer would produce the smallest diameter fibres, as is the case here. As shown by Yuya et al. [20], electrospinning of PVP with various concentrations leads to different fibre diameters. The article describes electrospinning PVP at 8%, 10% and 12% PVP, resulting in an average fibre diameter of 0.7, 0.9 and 1.2 µm, respectively. The mean diameter here of 0.426 µm, for sample A (6% PVP), may correlate with their analysis, which may mean that PVP alone is present in this sample. The highest concentration of polymer often suggests the highest average fibre diameter [21], however this is not the case here. The previous literature has suggested a non-linear relationship between solution concentration and fibre diameter [22]. This may be due to a lack of beading and heterogeneities present in samples D and E, thus presenting a narrower and smoother fibre.

XPS Analysis
XPS analysis gave a thorough indication of the surface composition of the fibrous composites along with the atomic percentage composition and electron configuration of the specified electrons within the atom. Each scaffold presented mainly C 1s, N 1s and O 1s characteristics (see Table 1). With regard to C 1s, N 1s and O 1s, there were no major chemical shifts that indicate the presence of CS on the surface of the materials. There are a couple of reasons why it was difficult to establish the presence of CS. CS is used in a smaller minority in the blend compared to PVP, with the highest being 4% CS, compared to that of 6% PVP. Additionally, in terms of the molecular weights used, PVP had a molecular weight of 1,300,000. This is a factor of 6.84 to 26 times higher in comparison to that of chitosan (mol wt 50,000-190,000). It may be possible that the surface of the fibrous composites is deficient in CS. Additionally, in terms of the XPS analysis, the ketone functional group (C=O) in PVP may be masking the polysaccharide backbone oxygen in chitosan. In addition, N-C=O groups in PVP may be masking the non-protonated amine or amide contained within chitosan, which one would expect to see at approximately 399.3 eV [23]. Samples B and C contained 2.22% and 2.11%, respectively, of F 1s (see Figure 2). The reason for this was due to residual of TFA molecules on the fibres. As the concentration of chitosan increases from 1% through to 4%, there were no significant chemical shifts regarding the C-C or C=O bonds (see Figure 2). This means it is difficult to highlight the presence of chitosan within the sample. which relates to the amide of the acetylated functional groups. While XPS proved beneficial in analysing the surface composition of each fibrous scaffold, it was difficult to pinpoint the exact location of chitosan.

Raman Analysis
Raman spectra registered in the region 300-1750 cm −1 of samples A-E with different CS content are shown in Figure 3a. The spectra were baselined, normalised to the intensity of the band at ~1670 cm −1 and then shifted along the Y-axis for convenience of presentation. From first glance, all spectra shown in Figure 3a demonstrate vibrational bands typical for PVP fibres (compare with the spectrum labelled as 0%) with the following major bands assignment: ring C-C vibration at 753 cm −1 , C-C vibrations at 850 cm −1 , C-C ring breathing at 934 cm −1 , C-C-backbone at 1023 cm −1 , CH deformation at 1372 cm −1 , CH2 scissoring at 1495 cm −1 and amide carbonyl (C=O) stretching vibrations at ~1670 cm −1 [24][25][26][27]. Similar band positions were observed earlier in different papers devoted to Raman study of complex PVP fibres [28,29]. At approximately 531 eV, it is possible to see the presence of N-C=O groups within only sample A and E, which are 0% CS and 4% CS, respectively. These relate to the amide carbonyl group found within PVP. There were no significant chemical shifts regarding the C=O and N-C=O groups as the concentration of chitosan increased from 0% to 4%. Previous literature stated that the oxygen of the polysaccharide backbone in chitosan should be located at approximately 532.7 eV [23]. Additionally, a smaller shift was seen at 531 eV, which relates to the amide of the acetylated functional groups. While XPS proved beneficial in analysing the surface composition of each fibrous scaffold, it was difficult to pinpoint the exact location of chitosan.

Raman Analysis
Raman spectra registered in the region 300-1750 cm −1 of samples A-E with different CS content are shown in Figure 3a. The spectra were baselined, normalised to the intensity of the band at~1670 cm −1 and then shifted along the Y-axis for convenience of presentation. From first glance, all spectra shown in Figure 3a demonstrate vibrational bands typical for PVP fibres (compare with the spectrum labelled as 0%) with the following major bands assignment: ring C-C vibration at 753 cm −1 , C-C vibrations at 850 cm −1 , C-C ring breathing at 934 cm −1 , C-C-backbone at 1023 cm −1 , CH deformation at 1372 cm −1 , CH 2 scissoring at 1495 cm −1 and amide carbonyl (C=O) stretching vibrations at~1670 cm −1 [24][25][26][27]. Similar band positions were observed earlier in different papers devoted to Raman study of complex PVP fibres [28,29]. Chemosensors 2021, 9, x FOR PEER REVIEW 8 of 19 However, more detailed analysis revealed that the position of the C=O band is downshifted gradually with the increase in CS content and approaching the shift of 13 cm −1 for sample E (40% CS and 60% PVP) (see Figure 3b). This shift is due to the hydrogen bonding (H-bonding) between PVP and CS molecules. Similar shifts in Raman spectrum of different polymer blends and complexes were demonstrated in the literature [29][30][31][32]. In particular, the significant downshift of carbonyl band was attributed by Hao et al. [31] to Hbonding in complex Gantrez/PVP. The authors provided a detailed explanation why these polymers can form the H-bonding in complex while it does not exist in their pure states. In addition to that, Suknuntha et al. [33] performed a molecular modelling simulation to identify the interacting molecular groups in CS/PVP blends and concluded that the most favourable site for H-bonding formation is with the oxygen atom of C=O group of PVP. The H-bonding with C=O groups in PVP can be formed by the -OH and -NH2 groups from chitosan. In addition to H-bonding, CS and PVP may also interact with each other through electrostatic interactions between the negatively charged PVP oxygen atom (N + = C-O-) and the ionizable amino group of CS [13,34]. Further support for the effect of CS presence on CS/PVP fibre composite was obtained from considering the range of spectra of 2730-3100 cm −1 representing the CH and CH2 stretching vibrations bands [26,28,35], see  However, more detailed analysis revealed that the position of the C=O band is downshifted gradually with the increase in CS content and approaching the shift of 13 cm −1 for sample E (40% CS and 60% PVP) (see Figure 3b). This shift is due to the hydrogen bonding (H-bonding) between PVP and CS molecules. Similar shifts in Raman spectrum of different polymer blends and complexes were demonstrated in the literature [29][30][31][32]. In particular, the significant downshift of carbonyl band was attributed by Hao et al. [31] to H-bonding in complex Gantrez/PVP. The authors provided a detailed explanation why these polymers can form the H-bonding in complex while it does not exist in their pure states. In addition to that, Suknuntha et al. [33] performed a molecular modelling simulation to identify the interacting molecular groups in CS/PVP blends and concluded that the most favourable site for H-bonding formation is with the oxygen atom of C=O group of PVP. The H-bonding with C=O groups in PVP can be formed by the -OH and -NH 2 groups from chitosan. In addition to H-bonding, CS and PVP may also interact with each other through electrostatic interactions between the negatively charged PVP oxygen atom (N + = C-O-) and the ionizable amino group of CS [13,34]. Further support for the effect of CS presence on CS/PVP fibre composite was obtained from considering the range of spectra of 2730-3100 cm −1 representing the CH and CH 2 stretching vibrations bands [26,28,35], see Figure 4a,b. This spectral range fitting with three vibrational bands is shown in Figure 4a for sample A (0% CS), while it is shown in Figure 4b for sample E (4% CS in CS/PVP fibre composite). This spectral range has been investigated in various works [32,[35][36][37][38][39][40][41][42][43] for extracting different characteristics of complex blends and macromolecules, including the order/disorder of acyl chain, the degree of acetylation (DA) and degree of deacetylation (DD). for sample A (0% CS), while it is shown in Figure 4b for sample E (4% CS in CS/PVP fibre composite). This spectral range has been investigated in various works [32,[35][36][37][38][39][40][41][42][43] for extracting different characteristics of complex blends and macromolecules, including the order/disorder of acyl chain, the degree of acetylation (DA) and degree of deacetylation (DD). In our case, we suggest using the ratio of high-frequency bands to demonstrate the contribution of the presence of CS to the vibrational band at ~2930 cm −1 with respect to the band at 2981 cm −1 . It was noted that CS/PVP composites show a complex region of CH and CH2 stretching vibrations with contribution from both type of molecules [36]. Therefore, the ratio of intensities I2930/I2981 was chosen to characterise the influence of the addition of CS on Raman spectra of CS/PVP composite. This is in accordance with the literature; the CH stretching bands of stronger intensity appears in CS at ~2885 and at ~2930 cm −1 [35,37,38], while in PVP fibres, the most intense bands, depending on molecular weight and state of PVP, are located in the range 2822-2830, 2875-2890, and 2940-2980 cm −1 [24,29]. Pure PVP of high molecular weight (K30) shows the most intense CH stretching bands in the Raman spectrum at 2880, 2930 and 2980 cm −1 [25]. We believe that the ratio of peak intensities (I2930/I2981) or integrated intensities (A2930/A2980) can be considered as a measure of the CS content in the CS/PVP fibre composite, based on the fact that the contribution to the intensity of band at ~2830 cm −1 will be provided by both types of molecules, while the band at 2981 cm −1 is mainly due to the contribution of PVP molecules. Figure  5a,b show the dependence of intensity ratios (A2930/A2981) and (I2930/I2981) vs. CS content. It can be seen that these ratios increase in linear range with the increase in CS content from above 1% CS. Figure 5c shows the dependence of the position of the peaks, obtained during the fitting and support that the fitting of these complex bands was performed reliably.
In addition to the changes observed in Raman spectra of PVP/CS fibre composites at the increase in CS content, relatively small shifts in the peak position of some other bands were also observed. In particular, the band of pyrrolidone group shifted from 1424 cm −1 in sample A (at 0% CS) to 1427 cm −1 in sample E (at 4% CS); the CH bending band seen at 1371 cm −1 in sample A (0%CS/6%PVP) shifted to 1374 cm −1 in sample E (4%CS/6% PVP).
Therefore, we can conclude that despite initial analysis showing that a small percentage presence of CS in PVP does not affect the Raman spectra significantly, a more detailed analysis of band position and deconvolution of some complex bands can result in reliable information on CS content in CS/PVP composite fibres. One of the differences in larger In our case, we suggest using the ratio of high-frequency bands to demonstrate the contribution of the presence of CS to the vibrational band at~2930 cm −1 with respect to the band at 2981 cm −1 . It was noted that CS/PVP composites show a complex region of CH and CH 2 stretching vibrations with contribution from both type of molecules [36]. Therefore, the ratio of intensities I 2930 /I 2981 was chosen to characterise the influence of the addition of CS on Raman spectra of CS/PVP composite. This is in accordance with the literature; the CH stretching bands of stronger intensity appears in CS at~2885 and at~2930 cm −1 [35,37,38], while in PVP fibres, the most intense bands, depending on molecular weight and state of PVP, are located in the range 2822-2830, 2875-2890, and 2940-2980 cm −1 [24,29]. Pure PVP of high molecular weight (K30) shows the most intense CH stretching bands in the Raman spectrum at 2880, 2930 and 2980 cm −1 [25]. We believe that the ratio of peak intensities (I 2930 /I 2981 ) or integrated intensities (A 2930 /A 2980 ) can be considered as a measure of the CS content in the CS/PVP fibre composite, based on the fact that the contribution to the intensity of band at~2830 cm −1 will be provided by both types of molecules, while the band at 2981 cm −1 is mainly due to the contribution of PVP molecules. Figure 5a,b show the dependence of intensity ratios (A 2930 /A 2981 ) and (I 2930 /I 2981 ) vs. CS content. It can be seen that these ratios increase in linear range with the increase in CS content from above 1% CS. Figure 5c shows the dependence of the position of the peaks, obtained during the fitting and support that the fitting of these complex bands was performed reliably. contribution from CS molecules to the higher-frequency range of Raman spectra (2730-3100 cm −1 ) could be because the intensities of Raman bands for CS in this region are much stronger in comparison with the low-frequency region. This follows from our measurements and correlates with previous studies [37,38].

FTIR Spectroscopy
Due to the presence of environmental moisture when using our ATR-FTIR setup, the measured IR spectra in the regions from 1300 cm −1 to 2000 cm −1 and around 3000-3900 cm −1 are modulated by the narrow rotational bands of water (see examples shown in Figure 6a). Therefore, we mainly focus our discussion on the region 900-1300 cm −1 , where some vibrational bands typical for CS may present [12,[39][40][41][42][43]. Before moving to this discussion, it is worth noting that the carbonyl band (C=O), which can be seen clearly in Figure 6a, is gradually downshifted from its original position at 1661 cm −1 in sample A to 1557 cm −1 for sample E. This confirms the influence of intermolecular interactions (namely H-bonding) on the position of C=O band and was previously described in a number of In addition to the changes observed in Raman spectra of PVP/CS fibre composites at the increase in CS content, relatively small shifts in the peak position of some other bands were also observed. In particular, the band of pyrrolidone group shifted from 1424 cm −1 in sample A (at 0% CS) to 1427 cm −1 in sample E (at 4% CS); the CH bending band seen at 1371 cm −1 in sample A (0%CS/6%PVP) shifted to 1374 cm −1 in sample E (4%CS/6% PVP).
Therefore, we can conclude that despite initial analysis showing that a small percentage presence of CS in PVP does not affect the Raman spectra significantly, a more detailed analysis of band position and deconvolution of some complex bands can result in reliable information on CS content in CS/PVP composite fibres. One of the differences in larger contribution from CS molecules to the higher-frequency range of Raman spectra (2730-3100 cm −1 ) could be because the intensities of Raman bands for CS in this region are much stronger in comparison with the low-frequency region. This follows from our measurements and correlates with previous studies [37,38].

FTIR Spectroscopy
Due to the presence of environmental moisture when using our ATR-FTIR setup, the measured IR spectra in the regions from 1300 cm −1 to 2000 cm −1 and around 3000-3900 cm −1 are modulated by the narrow rotational bands of water (see examples shown in Figure 6a). Therefore, we mainly focus our discussion on the region 900-1300 cm −1 , where some vibrational bands typical for CS may present [12,[39][40][41][42][43]. Before moving to this discussion, it is worth noting that the carbonyl band (C=O), which can be seen clearly in Figure 6a, is gradually downshifted from its original position at 1661 cm −1 in sample A to 1557 cm −1 for sample E. This confirms the influence of intermolecular interactions (namely H-bonding) on the position of C=O band and was previously described in a number of papers [13,29,43,44] devoted to FTIR investigation of polymer fibre composites and blends. As was reported in earlier studies of different research groups [29,45,46], the H-bonding in complex polymers is considered to be strong if the carbonyl band position is located below 1661 cm −1 . This is also in agreement with the Raman data described above; however, the downshift of C=O band in FTIR spectra of the investigated CS/PVP fibre composites is not as dramatic. One of the reasons for this difference may lie in the fact that during micro-Raman spectroscopy investigations, a laser beam (of 1-2 µm in diameter) was focused on one fibre, and, therefore, only a small area of the sample surface was tested, whereas in case of ATR-FTIR measurements, the signal from a much larger surface and depth of sample was collected and averaged.
In general, the bands at 1000-1200 cm −1 are attributed to the saccharide structure of chitosan and seen in particular at 1154, 1073 and 1032 cm −1 [40] in pure CS. FTIR spectra, truncated in the region of 900-1370 cm −1 , for all investigated CS/PVP composites, are shown in Figure 6b. From this figure it is clear that even at 1% CS, some additional bands to that, demonstrated in the IR spectrum of 0% CS, appear in the region 1000-1100 cm −1 , as well as some noticeable changes in peak positions and relative intensity of vibrational bands in the region 1100-1200 cm −1 . In particular, C-C stretching is located at 1131 cm −1 , while skeletal vibrations involving the C-O stretching typical for CS saccharide structure are observed at 1030 and 1076 cm −1 [39][40][41][42]. The behaviour of these bands, as well as the band at 1175 cm −1 was analysed versus CS content after fitting the spectra in the range 900-1300 cm −1 using Renishaw Wire 3.2 software (see examples in Figure 7a,b). The results extracted from the spectral fitting are summarized in Figure 8a-d. In particular, the dependence of the absorbance peak intensity vs. CS content is shown for C-O band at 1030 cm −1 in Figure 8a, while the integrated intensity of the band at 1076 cm −1 vs. CS content is shown in Figure 8b. Both graphs demonstrate a linear increase in band intensity when the CS content is increased. In addition to that, the band at around~1130 cm −1 shows the downshift from its original position at 1137 cm −1 for 0% of CS to 1130 cm −1 for sample E (4% CS/6%PVP). Only a slight increase in the intensity of this band is seen from Figure 8c.
below 1661 cm −1 . This is also in agreement with the Raman data described above; however, the downshift of C=O band in FTIR spectra of the investigated CS/PVP fibre composites is not as dramatic. One of the reasons for this difference may lie in the fact that during micro-Raman spectroscopy investigations, a laser beam (of 1-2 μm in diameter) was focused on one fibre, and, therefore, only a small area of the sample surface was tested, whereas in case of ATR-FTIR measurements, the signal from a much larger surface and depth of sample was collected and averaged. In general, the bands at 1000-1200 cm −1 are attributed to the saccharide structure of chitosan and seen in particular at 1154, 1073 and 1032 cm −1 [40] in pure CS. FTIR spectra, truncated in the region of 900-1370 cm −1 , for all investigated CS/PVP composites, are shown in Figure 6b. From this figure it is clear that even at 1% CS, some additional bands to that, demonstrated in the IR spectrum of 0% CS, appear in the region 1000-1100 cm −1 , as well as some noticeable changes in peak positions and relative intensity of vibrational  Figure 8b. Both graphs demonstrate a linear increase in band intensity when the CS content is increased. In addition to that, the band at around ~1130 cm −1 shows the downshift from its original position at 1137 cm −1 for 0% of CS to 1130 cm −1 for sample E (4% CS/6%PVP). Only a slight increase in the intensity of this band is seen from Figure 8c. In addition, the band at 1178 cm −1 demonstrates a slight downshift to 1175 cm −1 and a significant increase in integrated intensity vs. CS content (Figure 8c). In accordance with FTIR bands assignment, the vibrational bands in this region may belong to CH2 twist in PVP molecules [24] and to the antisymmetric stretching of C-O-C bridge [40,43], to C-C stretching vibration [41] and to some combinational bands [35] in chitosan. At the same time, it can be also seen from Figure 7a,b that the vibrational band at ~1200 cm −1 , which exists in the original spectrum for sample A (0%CS/6%PVP composite), does not demonstrate a significant change in the peak position or the peak intensity at higher CS content. This band can be used as a reference for estimating the ratio of peak intensities in order to prove the increase in intensity for band at 1175 cm −1 . The graph of dependence of the integrated intensities ratio of band A1175 to typical PVP band at 1200 cm −1 (i.e., A1175/A1200) vs. CS content was plotted in Figure 8d. As can be seen from Figure 8d, this ratio also demonstrates a reasonable linear dependence with increasing CS content and confirms that the band fitting procedure was reliable.   In addition, the band at 1178 cm −1 demonstrates a slight downshift to 1175 cm −1 and a significant increase in integrated intensity vs. CS content (Figure 8c). In accordance with FTIR bands assignment, the vibrational bands in this region may belong to CH 2 twist in PVP molecules [24] and to the antisymmetric stretching of C-O-C bridge [40,43], to C-C stretching vibration [41] and to some combinational bands [35] in chitosan. At the same time, it can be also seen from Figure 7a,b that the vibrational band at~1200 cm −1 , which exists in the original spectrum for sample A (0%CS/6%PVP composite), does not demonstrate a significant change in the peak position or the peak intensity at higher CS content. This band can be used as a reference for estimating the ratio of peak intensities in order to prove the increase in intensity for band at 1175 cm −1 . The graph of dependence of the integrated intensities ratio of band A 1175 to typical PVP band at 1200 cm −1 (i.e., A 1175 /A 1200 ) vs. CS content was plotted in Figure 8d. As can be seen from Figure 8d, this ratio also demonstrates a reasonable linear dependence with increasing CS content and confirms that the band fitting procedure was reliable.
Therefore, based on FTIR spectral analysis, we can not only get the confirmation on the presence of CS in CS/PVP fibre composites but also a quantitative estimation of the CS content using reference linear dependence of the intensity of C-O bands at 1030 and 1076 cm −1 . FTIR investigation also provides a solid support to the conclusion drawn in the previous Section based on Raman data.
It should be noted here that in the FTIR spectra of the PVP and CS electrospun fibres, the intensities of high-frequency bands are much lower than the intensities of lowfrequency vibrational bands which is opposite to the case in the Raman spectra discussed in Section 3.2.3 (see also [24,30,35,37]).

Drug Incorporation
5-fluorouracil is an anti-cancer drug commonly used for the treatment of cancer relating to the human colon, stomach, oesophagus, etc. It was chosen as a model therapeutic to load into the CS:PVP system due to its hydrophilicity and the hydrophilicity of the polymers. This should help the miscibility of the blend and thus loading capacity in the fibres. 5-Fu at different concentrations was added to the 4:6 CS:PVP blend and electrospun using instrumentation parameters of; 15 kV, 15 cm tip-collector gap, 0.2 mL/h −1 and an 18-gauge needle. The 4:6 CS:PVP blend was chosen here as it contained the highest concentration of CS. This is critical for improving the biocompatibility and drug loading properties of the composite.
Cell Viability Study A549 human lung epithelial adenocarcinoma cells were chosen as a model cell line to analyse the cytotoxicity of electrospun constructs containing CS/PVP, along with 5-Fu loaded CS/PVP constructs (1, 5 and 10 mg/mL). A WST-8 assay was chosen for cell viability analysis. Following exposure of the cells to electrospun constructs for 24 h, there was no statistical difference between cells alone and the CS/PVP construct, as seen in Figure 9, meaning that the CS/PVP construct did not have a toxic effect on the cells. Research has shown that CS containing nanofibres have "remedying effects" when exposed to normal cells which have been previously exposed to 5-Fu [47]. Additionally, studies have shown that CS containing electrospun nanofibres were non-toxic to cells [48] in agreement to the results we have obtained. Research has also detailed the non-toxic effect of CS within a polymer matrix (PVA and PLA in this case) towards MCF-7 breast cancer cell lines [49]. Statistically, there was a difference between the control groups (cells alone and CS/PVP construct) and both the 5 and 10 mg/mL 5-Fu containing group, which means at both of these concentrations, the drug was efficient at killing cells ( Figure 9). The construct containing 1 mg/mL drug also showed increased cell death compared to controls, but this was not statistically significant-1 mg/mL 5-Fu was not a high enough concentration to kill cells efficiently. Due to the statistical difference in cell death between drug loaded and non-drug loaded groups, and the increase in cell death as the drug concentration is increased (strong negative correlation (no significant) by Pearson's correlation test; r = −0.86, p = 0.06), we can ascertain that the cell death is due to drug release alone. There was also a statistical difference between the 1 and 10 mg/mL 5-Fu containing groups-1 mg/mL 5-Fu was not a high enough concentration to kill cells efficiently. Generally, over 24 h, 5-Fu is released by an initial burst followed by a slower release phase [50]. The release mechanism of 5-Fu from nanofibres is generally through a Fickian diffusion mechanism [18,19]. Release is related to the diffusion or permeation of the drug through nanofibre mat matrices [18,49]. It has been shown that the inclusion of CS and PVP in a 5-Fu delivery system can increase drug exposure time in comparison to 5-Fu only [51] as well as greatly inhibiting tumour growth [51][52][53][54].
Further analysis of the 24 h WST-8 assay shows that the cell viability decreased from 96.93% to; 56% (1 mg/mL), 52.35% (5 mg/mL) and 31.48% (10 mg/mL), as seen in Figure  10. There is a strong negative correlation between drug concentration and cell viability, as identified by Pearson's correlation test [55], but it is not statistically significant owing to the small number of x-variables (r = −0.84, p = 0.07). Overall, 10 mg/mL drug concentration was most efficient at killing cells. It has been reported that CS containing 5-Fu composites decrease cell viability of cancer cells up to ≥50% over 24 h [19]. Other analysis has shown that it may take up to 48 h to efficiently kill ≥50% of liver cancer cells (HepG2) in electrospun CS/5-Fu delivery vehicles [56]. A 30% viability due to 5-Fu release at 24 h, as seen in our study, concurs with other published studies; 5-Fu release from chitosan grafted poly(N-vinylcaprolactam) to PC3, KB and MCF7 cancer cell lines showed similar toxicity levels at 24 h for a 1 mg/mL loaded chitosan sample [54]. The viability of MCF7 breast Figure 9. Cell viability measured using a WST-8 assay after 24 h exposure to constructs (37 • C, 5% CO 2 ). Absorbance was measured at 485 nm; high absorbance indicates high cell viability. Groups containing 1, 5 and 10 mg/mL all contained 5-Fu. Here, * denotes statistical significance between control groups and experimental groups containing 5-Fu. # denotes statistical significance between concentrations of 5-Fu in the CS/PVP groups. Two sample t-test, p < 0.05, n = 3.
Statistically, there was a difference between the control groups (cells alone and CS/PVP construct) and both the 5 and 10 mg/mL 5-Fu containing group, which means at both of these concentrations, the drug was efficient at killing cells ( Figure 9). The construct containing 1 mg/mL drug also showed increased cell death compared to controls, but this was not statistically significant-1 mg/mL 5-Fu was not a high enough concentration to kill cells efficiently. Due to the statistical difference in cell death between drug loaded and non-drug loaded groups, and the increase in cell death as the drug concentration is increased (strong negative correlation (no significant) by Pearson's correlation test; r = −0.86, p = 0.06), we can ascertain that the cell death is due to drug release alone. There was also a statistical difference between the 1 and 10 mg/mL 5-Fu containing groups-1 mg/mL 5-Fu was not a high enough concentration to kill cells efficiently. Generally, over 24 h, 5-Fu is released by an initial burst followed by a slower release phase [50]. The release mechanism of 5-Fu from nanofibres is generally through a Fickian diffusion mechanism [18,19]. Release is related to the diffusion or permeation of the drug through nanofibre mat matrices [18,49]. It has been shown that the inclusion of CS and PVP in a 5-Fu delivery system can increase drug exposure time in comparison to 5-Fu only [51] as well as greatly inhibiting tumour growth [51][52][53][54].
Further analysis of the 24 h WST-8 assay shows that the cell viability decreased from 96.93% to; 56% (1 mg/mL), 52.35% (5 mg/mL) and 31.48% (10 mg/mL), as seen in Figure 10. There is a strong negative correlation between drug concentration and cell viability, as identified by Pearson's correlation test [55], but it is not statistically significant owing to the small number of x-variables (r = −0.84, p = 0.07). Overall, 10 mg/mL drug concentration was most efficient at killing cells. It has been reported that CS containing 5-Fu composites decrease cell viability of cancer cells up to ≥50% over 24 h [19]. Other analysis has shown that it may take up to 48 h to efficiently kill ≥50% of liver cancer cells (HepG2) in electrospun CS/5-Fu delivery vehicles [56]. A 30% viability due to 5-Fu release at 24 h, as seen in our study, concurs with other published studies; 5-Fu release from chitosan grafted poly(N-vinylcaprolactam) to PC3, KB and MCF7 cancer cell lines showed similar toxicity levels at 24 h for a 1 mg/mL loaded chitosan sample [54]. The viability of MCF7 breast cancer cells exposed to electrospun chitosan composites containing 5-Fu showed 60% viability after 24 h, but viability decreased further over time and was dependent on the composite materials and processing parameters [50].
Chemosensors 2021, 9, x FOR PEER REVIEW 16 of 19 cancer cells exposed to electrospun chitosan composites containing 5-Fu showed 60% viability after 24 h, but viability decreased further over time and was dependent on the composite materials and processing parameters [50].
0 mg/mL 1 mg/mL 5 mg/mL 10 mg/mL The inclusion of a higher CS content in a polymeric blend has been shown to improve drug loading and release properties. This is because of interactions between CS and 5-Fu [18]. Following the above fibre production protocol that we developed, we achieved fibres containing a high ratio of CS to PVP (4:6) which shows good potential for use in a drug delivery system. Moving forward, a concentration between 5 and 10 mg/mL 5-Fu will be chosen to exploit the cytotoxic properties of the material for cancer research. The biggest limitation to the work is in converting our in vitro results to the in vivo scenario. To move forward, our electrospun constructs will be made using our chosen 'tuned' parameters to the relevant anatomical size and re-assessed for drug release and cell viability in a scaledup culture system. An in vivo study investigating 5-Fu release from electrospun PLLA mats used tumour-bearing mice; drug loaded electrospun mats were surgically implanted onto the tumour site and the tumour size assessed daily. The drug loaded mats showed increased tumour suppression and necrosis and increased survival time in treatment groups compared to control groups and results correlated with the preceding in vitro work performed [3]. Therefore, it is encouraging that our in vitro work outlined here will translate to a clinical application as we advance our research.
Note, that the additional information on drug release profile and degradation rate study is provided in the Supporting Information Section.

Conclusions
Overall, sample E, containing 4% chitosan and 6% PVP, proved to be the best performing fibres and presented characteristics of unaltered chitosan and PVP. This sample was achieved under the following electrospinning protocol; voltage set at 15 kV, tip-collector gap set at 15 cm with a flow rate of 0.1 mL h −1 . The content of chitosan incorporated in the fibres is very significant. This concentration of chitosan is higher than anything that has been reported in the previous literature when forming a blend with another polymer [13,57]. Chitosan and PVP presented un-altered features post-electrospinning.
The SEM results show that the morphological structure of sample E fully homogenous with no heterogeneities (i.e., beads) observed within the investigated structure. XPS Figure 10. % Cell viability measured using a WST-8 assay after 24 h exposure to constructs (37 • C, 5% CO 2 ). Viability was calculated using the following formula: [A549 (sample)/A549 (control)] × 100.
The inclusion of a higher CS content in a polymeric blend has been shown to improve drug loading and release properties. This is because of interactions between CS and 5-Fu [18]. Following the above fibre production protocol that we developed, we achieved fibres containing a high ratio of CS to PVP (4:6) which shows good potential for use in a drug delivery system. Moving forward, a concentration between 5 and 10 mg/mL 5-Fu will be chosen to exploit the cytotoxic properties of the material for cancer research. The biggest limitation to the work is in converting our in vitro results to the in vivo scenario. To move forward, our electrospun constructs will be made using our chosen 'tuned' parameters to the relevant anatomical size and re-assessed for drug release and cell viability in a scaled-up culture system. An in vivo study investigating 5-Fu release from electrospun PLLA mats used tumour-bearing mice; drug loaded electrospun mats were surgically implanted onto the tumour site and the tumour size assessed daily. The drug loaded mats showed increased tumour suppression and necrosis and increased survival time in treatment groups compared to control groups and results correlated with the preceding in vitro work performed [3]. Therefore, it is encouraging that our in vitro work outlined here will translate to a clinical application as we advance our research.
Note, that the additional information on drug release profile and degradation rate study is provided in the Supplementary Materials Section.

Conclusions
Overall, sample E, containing 4% chitosan and 6% PVP, proved to be the best performing fibres and presented characteristics of unaltered chitosan and PVP. This sample was achieved under the following electrospinning protocol; voltage set at 15 kV, tip-collector gap set at 15 cm with a flow rate of 0.1 mL h −1 . The content of chitosan incorporated in the fibres is very significant. This concentration of chitosan is higher than anything that has been reported in the previous literature when forming a blend with another polymer [13,57]. Chitosan and PVP presented un-altered features post-electrospinning.
The SEM results show that the morphological structure of sample E fully homogenous with no heterogeneities (i.e., beads) observed within the investigated structure. XPS showed that the surface composition of the materials contained PVP. However, intermolecular interactions (via H-bonding) between PVP and CS molecules analysed in Raman and FTIR data indicate the presence of CS within the samples. In terms of cell viability studies, constructs containing 10 mg/mL of the drug were the most efficient at killing cells over 24 h. The constructs containing 1 mg/mL of the drug were not efficient at killing cells. The study concludes that chitosan-based electrospun constructs have potential in the use of drug delivery systems. Further research would be performed to optimise drug delivery of chitosan/PVP/5-Fu constructs to human lung cancer cells.