Electrospinning is a highly versatile and robust technique that allows production of fibers with diameters from several nanometers to tens of micrometers [1
]. Drug-loaded electrospun nanofiber mats have been studied intensively and show potential as drug delivery systems (DDSs) [2
] and tissue engineering scaffolds [3
] due to several advantages, such as huge specific surface area, porosity and the possibility to modify the drug release kinetics [5
]. Compared to other delivery systems, nano- and microfiber mats enable control and tuning of the drug release kinetics [8
] and, hence, design the mats with desired properties [9
]. For example, for local antibiotic delivery, the desired drug release needs to follow two steps: initial fast release followed by the slow zero-order kinetics over a longer period of time [11
]. Novel strategies for attaining sustained release have been proposed, for example via the formation of core-shell structures [12
], beads [14
], or modification of nanofiber mat thickness [15
]. The drug release process is affected by several factors, such as the physicochemical properties of the drug and carrier polymer, the structural characteristics of the material system, release environment, and the possible interactions between these factors [16
]. It is known that drug release from electrospun fiber mats may vary depending on the material properties and the structure of the mats [9
Despite the substantial body of literature on electrospun fiber mats and their characterization, there is no standard method for the analysis of drug release from fiber mats. Traditional pharmacopoeial drug dissolution tests have been found useful for the analysis of nanofibers incorporated into capsules or pressed into a tablet [19
]. However, when electrospun nanofiber mats are intended for the local delivery of drug, e.g., wound therapy, the amount of available liquid is low. Thus, mimicking the actual biorelevant conditions in vitro
may be challenging using standardized dissolution testing conditions. Researchers have used methods where the amount of dissolution medium is much reduced and size of the sample is close to the actual size of the nanofiber mat used in vivo
]. Samples are typically collected at predetermined time intervals and analyzed by ultraviolet-visible (UV-VIS) spectrophotometry or high performance liquid chromatography (HPLC).
The biorelevant conditions applied for the drug release studies depend on the exact problem and site of application, and may vary. Hydrogels have been widely used as DDSs for topical applications; however, hydrogels may also provide for a simplistic wound model onto which drug may be released followed by drug diffusion into the hydrogel matrix. Diffusion is one of the major transport mechanisms in the wound [16
], although swelling and erosion may also play a role depending on the formulation. For some electrospun nanofibers, the release rate has also been explained by desorption of the embedded drug from nanopores in the fibers or from the outer surface of the fibers in contact with the water bath [25
]. In addition to the actual testing, simulations have been performed and models proposed that enable prediction of the drug release behavior of the electrospun fiber mats [16
]. These models enable the design of mats with certain structures in order to achieve a desired drug release kinetics [12
Recently, a fully automated fiber-optics based dissolution testing systems for in situ
monitoring of drug release from electrospun fiber mats was proposed [27
]. The direct ultraviolet (UV) measurement of dissolved drug within dissolution medium provided the dissolution profile in real-time. UV imaging technology has emerged in pharmaceutical analysis [28
]. It has found use for the characterization of different pharmaceutical dosage forms, including monitoring drug release from capsules [29
], patches [30
], and hydrogels [31
]. Spatially resolved absorbance values are measured facilitating monitoring of concentrations and concentration gradients by UV imaging, and thereby providing the potential for new insights to the drug dissolution and release processes through real-time monitoring of swelling, precipitation, diffusion, and partitioning phenomena [31
]. Drug release from electrospun fiber mats into hydrogel system has not been investigated before and was of interest within the present study.
In addition to using physical methods in drug release studies, genetically engineered whole-cell bioreporters enable to obtain valuable information during drug release studies. A few of the main advantages of using bioreporters are that they provide physiologically relevant data by measuring only biologically available fraction of the chemical. A typical bioreporter consists of a biological recognition element (i.e., sensor), a transducer, and a reporter protein. Test chemical binds to the sensor element, a transducer initiates the production of the reporter protein, and a signal is produced [35
]. In parallel to the drug release information, such methods give direct information about the bioactivity of the developed antimicrobial DDSs.
The aim of the current study was to test and compare different drug release model systems for the characterization of electrospun nano- and microfiber antibacterial drug-loaded mats. We studied two different polymeric compositions—polycaprolactone (PCL) alone or in combination with polyethylene oxide (PEO)—with the model antibacterial drug chloramphenicol (CAM). Interestingly, we have previously shown that although these fiber mats with different carrier polymers have different drug release behavior according to the dissolution test results, their antibacterial activity was rather similar in a disc diffusion assay [22
]. Therefore, in order to understand drug release from electrospun polymeric fiber mats better, novel characterization methods were acquired in order to elucidate and rationalize drug release behavior. In the current study, the drug release from electrospun fiber mats into buffer solution and agar hydrogel was investigated using HPLC, UV-VIS spectrophotometry, and bacterial bioreporters responding to the antibacterial drug CAM. UV imaging was used for the first time to monitor real-time the drug release and diffusion from electrospun fiber mats into agarose hydrogel. Antibacterial activity testing of the CAM containing fiber mats was performed using disc diffusion assay in order to shed light on the correlation between the drug release and antibacterial activity and, hence, the intended use of electrospun mats as local antibacterial DDSs for wound infections.
2. Materials and Methods
2.1. Materials, Bacteria and Release Media
Drugs, polymers, supplies. The antibacterial agent chloramphenicol (CAM) was used as a model active pharmaceutical ingredient. CAM, hydrophobic carrier polymer polycaprolactone (PCL, Mw ≈ 80,000), hydrophilic carrier polymer polyethylene oxide (PEO) (Mw ≈ 900,000), and all analytical grade reagents were purchased from Sigma-Aldrich Inc. (Darmstadt, Germany). Ampicillin sodium salt and anhydrous d-glucose used for the bacterial bioreporter preparation were obtained from Carl Roth GmbH + Co. (Karsruhe, Germany) and Fisher Scientific (Waltham, MA, USA), respectively. Type I agarose was purchased from Sigma-Aldrich (St. Louis, MO, USA). Agar hydrogel was prepared using Lennox lysogeny broth (LB) agar (Difco Laboratories, Detroit, MI, USA). FavorPrep Plasmid DNA Extraction Mini Kit and FavorPrep(TM) GEL/PCR Purification Mini Kit was purchased from Favorgen Biotech Corp. (Changzhi Township, Pingtung, Taiwan).
Bacteria (Staphylococcus aureus
DSM No.: 2569) were obtained from Leibniz Institute DSMZ-German Collection of Microorganisms and Cell Cultures. Escherichia coli
MG1655 strain [37
] was used for biosensor construction. Cloning was performed in E. coli
strain DH5α [38
Buffers and agarose hydrogel for drug release testing by UV imaging. For the preparation of the phosphate buffer solution used as a dissolution medium, sodium dihydrogen phosphate dihydrate (NaH2PO4 · 2H2O, Merck, Darmstadt, Germany) was dissolved in distilled water. The pH of the solution (67 mM phosphate buffer) was adjusted to pH 7.40 using 5M sodium hydroxide solution. The 0.5% (w/V) agarose hydrogel was prepared by dissolving type I agarose in an appropriate volume of phosphate buffer kept at 98 °C for 45 min in a water bath. The gels were cast in the quartz cells, and allowed to settle for 30 min prior to commencing the release experiments.
Buffers and agar hydrogels for drug release and antibacterial activity testing
. Drug release studies were conducted using phosphate buffered saline (1x PBS). Lennox lysogeny broth (LB) agar with a concentration of 1.5% (w
) (S. aureous
) and MOPS minimal medium [39
] with 1.5% (w
) agar (E. coli
) were used for drug diffusion and antibacterial activity testing with bacteria.
2.2. Preparation and Characterization Methods
Preparation of electrospinning solutions and fiber mats
. Fiber mats were prepared using an ESR200RD robotized electrospinning system (NanoNC, Seoul, Republic of Korea). Fiber mats with different compositions were prepared in order to provide different drug release kinetics. The exact compositions of the solutions used for electrospinning and electrospinning conditions are shown in Table 1
A mixture of chloroform:methanol (3:1) (V/V) was used as a solvent system for the preparation of PCL and PCL/PEO systems, and a total of 10 mL was electrospun using rotation (20 rpm) and moving stage (speed 25 mm/min, distance 140 mm). For the preparation of the electrospinning solution, the polymers were dissolved in the solvent system under stirring overnight. The desired CAM concentration in the fibers was 4% (dry solid state %) and CAM was added together with the polymer into the solvent system immediately after the preparation. The electrospun fiber mats were collected onto aluminum foil and put into ziploc bags. The samples were kept in a desiccator at 0% relative humidity above silica gel to avoid humidity induced changes in the mats.
Morphology and solid state characterization of electrospun fiber mats.
Morphology and diameter of electrospun fiber mats were investigated using scanning electron microscopy (SEM). Samples were mounted on aluminum stubs and magnetron-sputter coated with 3 nm gold layer in argon atmosphere prior to microscopy. Solid state characterization of the electrospun fiber mats and drug-loaded fiber mats was performed as described previously using attenuated total reflection-Fourier transform infrared (ATR-FTIR) spectroscopy (IRPrestige-21 spectrophotometer (Shimadzu Corp., Kyoto, Japan) with Specac Golden Gate Single Reflection ATR crystal (Specac Ltd., Orpington, UK) and verified with X-ray diffraction (XRD) (D8 Advance, Bruker AXS GmbH, Karlsruhe, Germany) [22
]. The thickness of the fiber mats was verified using a Precision-Micrometer 533.501 (Scala Messzeuge GmbH, Dettingen, Germany) with the resolution of 0.01 mm. The thickness of the mats was 0.07 ± 0.01 mm for the PCL fiber (0.05 ± 0.01 mm with CAM) and 0.08 ± 0.01 mm for the PCL/PEO fiber (0.08 ± 0.01 mm with CAM) mats.
Drug loading and distribution in fiber mats. High performance liquid chromatography (HPLC) (Shimadzu Prominence HPLC with LC20, PDA detector SPD-M2QA, controlled by LC Solution software (1.21 SP1 Shimadzu); Shimadzu Europa GmbH, Duisburg, Germany) was used to determine the CAM concentration in the electrospun fiber mats and to evaluate its distribution uniformity throughout the fiber mats. Analyses were performed according to the official European Pharmacopoeia method for a related substance CAM sodium succinate. Briefly, CAM-loaded fiber samples were cut into 1 cm2 pieces, weighed, and dissolved in chloroform and methanol (3:1 V/V). The HPLC measurements were performed using an octadecylsilyl column (Phenomenex, Luna C18(2), 250 × 4.6 mm, 5 μm). The flow rate was 1.0 mL/min, and injection volume was 20 μL. The mobile phase consisted of 2% phosphoric acid R, methanol R and water R in the volume ratio 5:40:55. A wavelength of 275 nm was used.
2.3. Drug Release Studies
2.3.1. Drug Release Testing into Buffer Solution and Agar Hydrogel by UV-VIS Spectrophotometry and HPLC
Drug release to buffer solution and agar hydrogel was investigated using UV-VIS spectrophotometry and HPLC, respectively.
Release to phosphate buffer solution:
The in vitro
drug release of CAM from electrospun PCL and PCL/PEO fiber mats was carried out as described previously [22
]; however, a more frequent sampling protocol was used. Briefly, 4 cm2
= 3) cut from the mats were weighed and placed into 20 mL of 1× PBS (pH 7.4) at 37 °C in 50 mL plastic tubes. The tubes were put into a dissolution apparatus vessel (Dissolution system 2100, Distek Inc., North Brunswick, NJ, USA) containing water maintained at 37 °C using rotation (paddle system, 100 rpm). Aliquots of 2 mL were removed and replaced with the same amount of 1× PBS at set time points. The aliquots were analyzed using UV-spectrophotometry (Shimadzu UV-1800, Shimadzu Europa GmbH, Duisburg, Germany) at 278 nm.
Release to agar hydrogel:
The amount of drug released into agar plates was investigated by sampling different zones of the agar (illustrated with a figure in Section 3.4
). Pieces of fiber mat (PCL/CAM and PCL/PEO/CAM fiber mat discs, with a diameter of 1 cm) were weighed, put onto pre-warmed LB agar plates, kept at 37 °C, and removed at set time points. Zones of the agar were cut out, the agar sample was put into ethanol (96%) and sonicated for 15 min. This extraction process was repeated twice and the obtained ethanol solutions were combined. The vials with ethanol solutions were left under a fume hood without caps, for the ethanol to evaporate. The residues left in the vials were dissolved in 1.5 mL of ethanol (96%) and the amount of CAM analyzed with HPLC. In the present study, the limit of detection for CAM was 1 µg/mL. Triplicate measurements were performed. The extraction efficacy was tested separately confirming that two times extraction resulted in 100% efficacy (Appendix A
, Table A1
2.3.2. UV Imaging for Drug Release Monitoring in Hydrogel
Complementary to the traditional HPLC method, an Actipix D200 Large Area Imager (Paraytec Ltd., York, England) controlled by Actipix D200 acquisition software ver. 18.104.22.168 was used to image the release of CAM from PCL and PCL/PEO fiber mats. These experiments were performed in a heating cabinet from Edmund Bühler TH30 (Bodelshausen, Germany) set to 37 °C. Imaging was performed at four alternating wavelengths: 525 nm, 280 nm, 255 nm, and 214 nm. Images for each wavelength were recorded at a frequency of 0.125 s−1 for the release experiments as well as the standard curve. The imaging area (28 × 28 mm2; pixel size 13.8 µm2) encompassed three quartz cells (Pion Inc., UK; 62 mm × 4 mm × 7 mm (L × H × W)), allowing three measurements to be performed simultaneously. The fibers were cut to fit the inner dimensions of the quartz cells (7.0 mm in width and 4.0 mm in height). The fibers were positioned perpendicular to the imaging direction in contact with the agarose gel. The fibers were backed by silicone plugs to ensure good contact with the gel and correct alignment. Parafilm was used to seal the quartz cells preventing evaporation of water from the gels. The release of CAM from the fibers was imaged for 3 h at 37 °C. Each imaging experiment allowed measurements of two CAM-containing fibers and one blank fiber (control). The positioning of the fibers in the imaging system (top, middle or bottom row) was randomized.
Standard Curve for quantification by UV imaging. A CAM stock solution (5 mM) in phosphate buffer was used to make the dilutions for the standard curve in 0.5% (w/V) agarose gel. These were made by mixing 1.5 mL 1% (w/V) agarose in phosphate buffer with a defined volume of CAM solution and phosphate buffer to obtain 3 mL of the mixture. The 1% (w/V) agarose solution and the phosphate buffer were both heated in water bath (98 °C) to facilitate mixing leading to a homogeneous mixture. The gels were cast in the quartz cells, and allowed to settle for 30 min prior to the experiments. As reference, a 0.5% (w/V) agarose gel in phosphate buffer without CAM was used.
2.4. Antibacterial Activity Studies—Drug Release and Effect on Bacterial Growth
2.4.1. Antibacterial Activity Testing
The antibacterial activity of released CAM on agar plates was investigated at different time points mimicking the drug diffusion tests into agarose hydrogel during UV imaging studies. Overnight culture (20 h) of S. aureus
DSM No.: 2569 was grown from DMSO stock (100 uL to 3 mL of LB). Preparation of all bacterial DMSO stocks used in the present study is described in Appendix B
. The culture was diluted to optical density (OD) 0.05 in LB and 100 μL was plated onto pre-warmed LB agar plates (1.5% (w
)). PCL/CAM and PCL/PEO/CAM fiber discs, and a positive CAM filter paper control were applied onto each plate. At specific time points, the discs were removed and the LB plates were incubated at 37 °C for 24 h prior to measurement of the inhibition zones.
2.4.2. Bioreporter Plasmid and Strain Preparation
All cloning was performed using CPEC cloning method [40
]. Plasmid vector backbone was low-copy pSC101 plasmid. Timer
reporter gene in plasmid pSC101-Ptet-Timer [41
] was replaced with two fluorescent reporter genes GFPmut2
] and mScarlet-I
]. In order to increase the expression of the green fluorescence protein (GFP) during antibiotic stress additional stress-inducible dnaK1 promoter (PdnaK1) originating from E. coli
MG1655 genomic DNA was added upstream of the tet-promoter (Ptet). In addition, kanamycin resistance gene kanR
was replaced with ampicillin resistance gene ampR.
In order to reduce the expression of ampR
resulting from reverse direction transcription initiation from PdnaK1, additional rrnB T2 terminator was added between ampR
In order to construct the ribosomal stalling reporter plasmid pSC101-CAM-bioreporter transcription attenuation-based regulatory trpL2Ala
region together with a terminator and constitutive T5 promoter from plasmid, pRFPCER-TrpL2A [44
] was inserted between GFPmut2
ribosomal binding site. mRNA from the reporter gene mScarlet-I,
and therefore, red fluorescence is only produced when ribosomal stalling occurs, e.g., due to CAM presence (Appendix C
CPEC products were transformed into E. coli
DH5α and plasmids were purified using FavorPrep Plasmid DNA Extraction Mini Kit. All plasmids were verified by sequencing. Purified bioreporter plasmid pSC101-CAM-bioreporter was transformed into E. coli
MG1655 chemical competent cells via heat shock. The transformants were selected on ampicillin (100 μg/mL) containing LB-agar plates after overnight incubation. The full nucleotide sequence of the pSC101-CAM-bioreporter plasmid is provided as Supplementary Materials
in GeneBank (gb) file format.
2.4.3. Bioreporter Disc Diffusion Assay
For bacterial bioreporter disc diffusion assay agar plates with defined MOPS minimal medium [39
] supplemented with 0.4% (w
) glucose as the carbon source and 1.5% (w
) agar were prepared in sterile conditions by measuring 20 mL of warm agar medium per plate, and plates were dried for 30 min under laminar flow hood.
Bioreporter strain DMSO stock was thawed, diluted 20× into sterile 1× PBS and 75 μL was plated on each minimal plate. A sterile cotton bud dipped into 1× PBS was used to spread the cells evenly. Plates were left to incubate at 37 °C for 10 h. After incubation, the weighted fiber mats (PCL/PEO/CAM and PCL/CAM) were added to each plate. Individual plates were first scanned with the Amersham Typhoon scanner (GE Healthcare Europe GmbH, Freiburg, Germany) (pixel size 100 μm; green fluorescence: 488 nm laser, 525BP20 filter, PMT voltage 352V; red fluorescence: 532 nm laser, 570BP20 filter, PMT voltage 621V) after adding the mats and re-scanned every hour for 6 h. Scan time for each plate was approximately 6 min. The plates were incubated at 37 °C between the scans.
2.5. Data Analysis
Data are given as average ± standard deviation (SD), unless stated otherwise. Data were analyzed and figures plotted using MS Excel 2017 and/or 2016, GraphPad Prism 7 ver. 7.04 or OriginPro 8.5. Statistical analysis was performed by two-tailed Student’s t-test assuming unequal variances (p < 0.05) where applicable.
The SDI Data analysis software ver. 2.0.60624 (Paraytec Ltd., York, England) was used to analyze the recordings made on the Actipix D200 system. The CAM molar absorption coefficient determined from the standard curve was used to calculate CAM concentrations in the gels. The analysis was performed at 255 nm, because the linear range of the standard curve covered the absorbance values encountered in the release experiments in contrast to 280 nm. A 6.25 mm wide zone starting from the fiber-gel interface protruding into the gel in the CAM transport direction was defined. For each pixel column, the absorbance was averaged, converted to concentration and plotted as a function of distance from the fiber mat to attain concentration-distance profiles. From the concentration-distance profiles, the area under curve (AUC) was calculated to determine the total amount of CAM released from the fibers.
The inhibition zones free of bacterial growth (diameters, mm) were determined using ImageJ software [45
] program version 1.52n. Tests were run at least in triplicate. ImageJ software was also used for obtaining numerical values of fluorescent zones from disc diffusion assay images. Green and red fluorescence images were analyzed separately. 0.8 mm wide lines were chosen as regions of interest for analysis. The 1.5 cm long line was drawn starting from the fiber mat and plot profiles of grey values for these regions were recorded.