**1. Introduction**

Preventing deep bone infections after surgery and enhancing bone regeneration reduces the time required for patient recovery and the costs associated with long hospital stays. Deep bone infections are very challenging to treat once established due to the di fficulty of achieving a suitable antibiotic concentration in the a ffected area and ensuring all bacteria have been removed. Bacteria such as *Staphylococcus aureus* (*S. aureus*) and *Escherichia coli* (*E. coli*) are responsible for causing bone infections post-surgery [1,2], and the application of excessive antibiotics can lead to antimicrobial resistance. Antibacterial poly(methyl methacrylate) beads have been used routinely for treating infections in a reactive manner with additional disadvantages of requiring surgical removal after the treatment ends [3]. Silver has often been used to successfully treat bacterial and fungal infections [4,5] and can be delivered in a localized area. Metallic implants coated with silver have been investigated to prevent infections [6]. However, silver by itself does not enhance wound healing and can have a toxic e ffect on mammalian cells at higher concentrations [7,8].

To facilitate bone healing, surgeons currently employ hydroxyapatite in granular or paste form to stimulate bone regrowth. Hydroxyapatite is the mineral component of bone and can be manufactured synthetically and has been demonstrated to aid in bone regeneration [9]. Nanoscale hydroxyapatite (nHA) is an increasingly popular biomaterial due to its similarity to the hydroxyapatite in human bone and tooth enamel. Our group recently produced a bioinspired nHA material through a rapid mix preparation, which is slightly calcium deficient to better mimic biological apatite [10]. Using nHA-based medical devices to increase the regeneration rate of the bone after surgery could greatly decrease patient healing times. Unfortunately, despite its success in aiding bone regeneration, hydroxyapatite has no innate antimicrobial properties to prevent infection. Moreover, the implantation of medical devices provides an increased risk of bacterial colonization and eventual biofilm formation if left untreated [11]. Once a biofilm has been established it is very di fficult to treat using antibiotics and even surgical intervention involves a high risk of reoccurrence [12].

There is growing interest in using a combination of silver and hydroxyapatite to produce a material with both antimicrobial/antifungal [13,14] and bone-regenerating properties [15–18]. Jiang et al. [19] produced foamed polyurethane sca ffolds containing up to 40 wt.% silver phosphate particles (Ag3PO4) and nHA. These mold-formed sca ffolds were porous and eluted silver for up to 22 days without toxic e ffects on the human osteosarcoma cell line MG63. Saravanan et al. [20] produced sca ffolds using the freeze-drying technique incorporating nano-hydroxyapatite with nanoparticles of silver. Wilcock et al. [21] developed a novel, rapid mixing, wet precipitation method of producing bioinspired, nanoscale hydroxyapatite crystals doped with silver for antimicrobial action. A degradable sca ffold could be placed against existing implants or used to hold packed bone substitutes (e.g., bone chips) into a bone defect.

The biomaterials' research community has been keenly interested in using electrospinning as a containment system for ceramic-derived materials including hydroxyapatite [22]. Electrospinning is a technique for producing fibrous sca ffolds that roughly resemble the body's extracellular matrix. These can be made porous to allow cell thoroughfare or nanoscale to act as a barrier to cells, fulfilling a wide variety of applications. Ciraldo et al. [23] investigated three di fferent manufacturing approaches using mesoporous glasses containing silver. These were glass coatings prepared using electrospraying, sol-gel methods, and directly spinning with polycaprolactone (PCL). For electrospinning, the authors were limited to 5% mesoporous glasses in PCL but did not investigate antimicrobial action. However, a recent paper by the same group demonstrated that the mesoporous glasses loaded with silver were antimicrobial [24]. Anjaneyulu et al. [25] manufactured doped hydroxyapatite through the sol-gel method and electrospun with poly(vinyl alcohol). Schneider et al. [26] electrospun tricalcium phosphate nanoparticles into poly(lactic-co-glycolic acid) (PLGA) sca ffolds, looking at a single concentration of silver, but did not investigate mammalian toxicity of the sca ffold. Furtos et al. [27] electrospun a PCL sca ffold to contain both nHA and amoxicillin, which they tested against four bacterial strains and found that increasing the nHA content reduced the loading e fficiency of the antibiotic.

A variety of polymers and calcium compounds have been used for producing these antimicrobial sca ffolds. Di fferent manufacturing technologies have also been employed to produce these as sca ffolds. A relatively small number of the above publications have reported testing their material with primary mesenchymal stem cells (MSCs), which are key to understanding bone-healing mechanisms. Few of the published reports have investigated microbiological testing beyond agar di ffusion assays with no time course elution antimicrobial testing.

Here we present the fabrication of novel, dual-function sca ffolds that combine silver to impart antimicrobial properties and hydroxyapatite to enhance bone regeneration. Our laboratory previously developed a silver-doped nano-hydroxyapatite material formed by the wet precipitation method [21]. Here we show the incorporation of this innovative material into an electrospun sca ffold alongside a thorough investigation of metal ion release, antimicrobial activity against clinically relevant bacterial strains, and in vitro biocompatibility using cultured mammalian MSCs. This research has provided an important proof-of-concept study that has the potential to underpin the development of a new generation of dual-action medical devices to address the growing challenges of deep bone infection and antimicrobial resistance.

#### **2. Materials and Methods**

#### *2.1. Silver Nano-Hydroxyapatite (Ca10(PO4)6(OH)2) Manufacture via Wet Precipitation*

This was based on our previously published methods [10,21]. In brief, calcium hydroxide (Sigma Aldrich, Dorset, UK) (50 mmol) was suspended in 500 mL deionized water (dH2O) into which 0, 1, 2.5, or 5 mmol silver nitrate (Sigma Aldrich, UK) (corresponding to 0, 2, 5, or 10 mol%, respectively) was added and stirred at 400 rpm for 1 h on a hotplate (Fisher Scientific, Loughborough, UK) at 90 ◦C. Phosphoric acid (Sigma Aldrich, UK) (30 mmol) was dissolved in 250 mL dH2O, poured into the calcium and silver preparation, and stirred for a further hour. The suspension was left to settle overnight, after which the supernatant was poured o ff and the silver-doped nHA suspension was washed with dH2O (3 × 500 mL). The silver-doped nHA suspensions were dried at 60 ◦C and ground in an agate mortar and pestle.

#### *2.2. Imaging the Silver Nano-Hydroxyapatite and Electrospun Fibers Using SEM*

Dried nHA powder or a 5 × 5 mm section of electrospun sca ffold was attached to a carbon tab and sputter gold coated and then imaged using a TESCAN VEGA3 SEM (Tescan, Brno, Czech Republic) with an accelerating voltage of 15 kV.

#### *2.3. Imaging the Silver Nano-Hydroxyapatite Using Transmission Electron Microscopy (TEM) and Energy-Dispersive X-Ray Spectroscopy (EDX)*

The nHA powder (0.1 g) was dispersed in ethanol in an ultrasonic bath (ThermoFisher, Shanghai, China) for 1 h and a drop was placed onto a copper mounting grid with carbon film and allowed to dry under ambient conditions. Samples were then imaged on a transmission electron microscope (Tecnai G2 F30, ThermoFisher, Shanghai, China) under standard imaging and EDX detecting elements Ca, P, Ag, and O. EDX settings used were 20 kV, takeo ff 14.8, amp time 7 μs, and 126.5 eV resolution.

#### *2.4. Analyzing Crystal Phases Present Using X-Ray Di*ff*raction (XRD)*

Phase analysis was performed on powdered samples using XRD (STOE IP, Darmstadt, Germany) with Cu Kα λ = 1.5418 Å radiation to determine the crystal structure of the materials. The di ffractometer was operated at 40 kV and 35 mA, with a 2θ range of 20–50◦. The following The International Centre for Di ffraction Data Powder Di ffraction (ICDD-PDF) cards were used for phase identification: 9-432 hydroxyapatite, 6-505 silver phosphate.

#### *2.5. Electrospinning Polycaprolactone (PCL) and Incorporating the Ag–nHA*

Polycaprolactone (average 80,000 Mn, Sigma, UK) was dissolved in a specific solvent and then propelled by an electric field onto a collector where polymer fibers formed as the solvent evaporated. Adding both the silver-doped and undoped nHA powder to the solution at 20 wt.% allowed it to be incorporated into the electrospun sca ffold. PCL (1.5 g) was dissolved into a solvent mixture (90% Dichloromethane [DCM], 10% Dimethylformamide [DMF]) (Sigma Aldrich, UK) and stirred using a magnetic stirrer bar (5 mm bar, Sigma Aldrich, UK) for 2 h. The nHA powders were incorporated at 20 wt.% of the PCL component and added into the dissolved PCL solution. The polymer-ceramic-solvent mixtures were placed in an ultrasonic bath for 30 min, before stirring using a magnetic stirrer bar for 30 min to ensure even distribution of powder in the solution. A 1-mL Luer-Lok syringe (BD, EU) was

loaded with 1 mL of the PCL/nHA solution and this was electrospun using a custom electrospinning rig described previously [28]. In total, 2 mL of solution at 1 mL/hr was used to fabricate each sca ffold at 17 kV with a spinning gap between the needle and collector of 20 cm.

#### *2.6. Accelerated Degradation Study of Electrospun Sca*ff*olds and Release Profile*

To investigate the rate at which nHA was released during sca ffold degradation, an accelerated degradation study was performed over 30 days. Samples were weighed and added to a 5-mL solution of 0.1 M NaOH (Sigma Aldrich, UK) in dH2O and placed in an oven at 37 ◦C. At each time point (days 0, 1, 7, 14, 21, and 28) material was also weighed and 0.5 mL of solution was removed, which was replaced with fresh solution. The solutions obtained were diluted to 5 mL using dH2O.

#### *2.7. Inductively Coupled Plasma (ICP) Measurement of Material and Degradation Product*

To ascertain the presence and abundance of various elements within the sample, ICP was utilized. Solid samples (e.g., 0.25 g–0.5 g) were weighed into a glass tube and 9 mL of concentrated hydrochloric acid (Sigma Aldrich, UK) added and 3 mL of concentrated nitric acid (Sigma Aldrich, UK) added. The samples were then placed into the heating block and the temperature was gradually raised to 150 ◦C and maintained at this temperature for 30 min. After removal from the block and cooling, the samples were diluted to 50 mL with 1% nitric acid. Samples were diluted down to 1 in 100 for analysis. Samples were analyzed using an ICP-OES (Spectro-Ciros-Vision Optical Emission Spectrometer (Spectro, Kleve, Germany) using a procedure with a detection limit of at least 10 ppm.

#### *2.8. Sca*ff*old Preparation for Cell Culture*

Disks of each electrospun sca ffold were prepared and sterilized prior to culture with cells. Using a 13 mm cork borer, disks (13 mm) of each sca ffold were punched out of the primary collector sheet. On the day of use, the culture sca ffolds were disinfected with 70% ethanol (Sigma Aldrich, UK) in dH2O for 30 min and then washed three times in dH2O for 5 min.

#### *2.9. Contact Bacterial Toxicity Test*

Antimicrobial activity was investigated using a classical di ffusion agar assay. Isolates of *E. coli* and *S. aureus* were purchased from the China Center of Industrial Culture Collection (CICC, Beijing, China): *E. coli*, CICC23657 and *S. aureus*, CICC10384. The bacterial suspensions at stationary phase were firstly obtained by culture in LB (Luria-Bertani) medium overnight at 37 ◦C in a 120-rpm shaker. Then 1 mL of the *E. coli* and *S. aureus* suspensions were pipetted from the overnight phase into another 100-mL flask containing 50 mL of fresh LB. The mixture was then cultured at 37 ◦C for another 6 h to obtain bacterial suspensions at exponential growth phase. After that, the bacterial suspension was centrifuged at 8000 rpm, washed, and resuspended in sterilized physiological saline to reach a concentration of 1 × 10<sup>7</sup> colony forming units (CFU)/mL. Then, 0.1 mL of the bacteria suspension with 1 × 10<sup>7</sup> CFU/mL was pipetted and plated on the agar plate uniformly. Three disks of the electrospun material (12 mm diameter) were placed on each Petri dish. These were incubated at 37 ◦C for 24 h. The zone of inhibition ring around the electrospun sca ffolds was measured.

#### *2.10. Antimicrobial Di*ff*usion Toxicity Test*

A disk of the electrospun sca ffold (12 mm diameter) was immersed into the bacteria suspension (10 mL) containing 1 mL 10<sup>7</sup> CFU/mL and 9 mL sterile physiological saline in a 15-mL falcon tube. The falcon tubes were placed on a carousel (IKA, Staufen, Germany) and rotated 360◦ in the z axis at 30 rpm in an incubator at 37 ◦C. At the time points 3, 18, 24, 48, 72, and 96 h, 200 μL of phosphate bu ffered saline (PBS) was removed and serially diluted to 10−<sup>7</sup> before 100 μL of diluted solutions were added to an agar plate and incubated for 24 h before imaging. The CFUs on plates were counted using the dilution, which was the highest with all bacterial colonies visible as single entities.

#### *2.11. Rat Mesenchymal Stem Cell (MSC) Isolation and Passage*

MSCs were isolated from four 4–5-week-old male Wistar rat femurs following the method previously described in literature [29,30]. Briefly, the femur was removed from the rats, opened at both ends, and flushed with media. This was then cultured to collect adherent cells, cells from all rats were intermixed, and cells were passaged twice before use. Cells were cultured in minimum essential medium (αMEM) supplemented with 10 U/mL penicillin (Sigma Aldrich, UK), 0.1 mg/mL streptomycin (Sigma Aldrich, UK), 20 mM alanyl-glutamine (Sigma Aldrich, UK), and 10% v/v fetal bovine serum (FBS; Biosera, Heathfield, UK). Cells were passaged 1:5 using trypsin-EDTA (ethylenediaminetetraacetic acid, Sigma Aldrich) and used below passage 10 for toxicity studies and below passage 5 for differentiation studies.

#### *2.12. Noncontact Mammalian Cell Toxicity Study*

Toxicity of elution from the electrospun scaffolds was measured by suspending the material over on cultured cells for 24 h. Both primary rat MSCs and the established fibroblast 3T3 cell line were used to test this material. Both cell lines were treated the same in the experiment apart from the rat MSCs were cultured in αMEM as described above and the 3T3 cells were cultured in high-glucose Dulbecco's Modified Eagle Medium (DMEM) with 10% FBS, both supplemented with 20 mM alanyl-glutamine (Sigma Aldrich, UK) and 0.1 mg/mL of streptomycin (Sigma Aldrich, UK). Cells were trypsinized and added to a 24-well plate (ThermoFisher, Loughborough, UK) at 30,000 cells per well in 1 mL of respective media. The cells were cultured for 24 h in an incubator at 37 ◦C and 5% CO2. Media was replaced and a Transwell® insert (Millicell Hanging Cell Culture Insert, polyethylene terephthalate (PET) 0.4 μm, 24-well) was added to each well containing cells. In triplicate, the following were added to the Transwell® inserts: No scaffold (positive), 1 mL 70% industrial methylated spirit (IMS) (negative), 0% Ag nHA scaffold, 5% Ag nHA scaffold, and 10% Ag nHA scaffold. One mL of media was added to each Transwell® insert and plates were incubated for 48 h at 37 ◦C and 5% CO2. After 48 h the cells within the wells were imaged and metabolic activity was assessed using PrestoBlue™ (ThermoFisher, UK). Media was removed and 700 μL of PrestoBlue™ (10% in complete media) was added to each well and the well plate was incubated for two hours. From the wells, 200 μL of PrestoBlue™ was added in triplicate to a 96-well plate. The fluorescence of the solutions was measured using a plate reader (FLx 800 Bio-Tek Instruments; Thermo Fisher Scientific, UK) with an excitation wavelength of 535 nm and an emission wavelength of 590 nm. The experiment was repeated four times (N = 4, n = 3).

#### *2.13. Contact Biocompatibility Study on Rat MSCs*

A direct-contact biocompatibility study was used to assess the toxicity of the scaffolds on cells over 21 days. Rat MSCs were used for this study. Disks of each scaffold were added in triplicate to a 24-well plate and 1 mL of media was added to each for 30 min (including a tissue culture plastic [TCP] control). Cells were trypsinized, suspended in media, and then 30,000 cells were added to each well. Well plates were agitated to aid even cell-seeding across the scaffold. Well plates were incubated for 24 h before cell metabolic activity was measured using PrestoBlue™. The scaffolds were transferred to a new well plate before PrestoBlue™ to avoid interference of cells growing on the tissue culture plastic surrounding the scaffold. PrestoBlue™ was measured at 4, 7, 14, and 21 days in culture. N = 2, n = 3.

#### *2.14. Live and Dead Cell Staining*

Cells were stained with two dyes to identify if the cell was alive or if the cell scaffold was disrupted (dead). Cells were stained using the LIVE/DEAD™ Viability/Cytotoxicity Kit (Thermofisher, UK) as per the instructions. In brief, cells were washed in PBS and then Calcien and Ethidium homodimer-1 (both at 5 μM concentration) were incubated with cells in media for 30 min. Samples were imaged using a fluorescent microscope (Axioplan 2, Zeiss, Jena, Germany) using the built-in filters for FITC (calcein) and TRITC (Ethidium homodimer-1).

#### *2.15. Total Sample DNA and Alkaline Phosphatase Production Measurements*

Alkaline phosphatase (ALP) is an early indicator of MSC di fferentiation to osteoblasts and can be measured using an ALP detection test. ALP was normalized to DNA to give a reading of alkaline phosphatase activity per cell regardless of cell number. Media was removed and samples washed three times with PBS. Cell digestion bu ffer was made up using deionized water and 1% triton X-100 (Sigma, UK). Samples were incubated at 4 ◦C for 1 h in the cell digestion bu ffer. The samples were then subjected to two freeze–thaw cycles (–20 ◦C to room temperature), which included solution agitation by pipetting when thawed to ensure thorough lysis of cells. After the second freeze–thaw cycle, an alkaline phosphatase diethanolamine detection kit was used to assess the alkaline phosphatase activity of the cultures. Then, 40 μL of cell culture lysate solution was placed in a 96-well plate in triplicate with 210 μL alkaline phosphatase reaction bu ffer. A standard curve of para-nitrophenyl (pNP) product was prepared with concentrations of 250, 100, 50, 10, and 0 nmole.ml−<sup>1</sup> pNP in the alkaline phosphate reaction bu ffer. Then, 250 μL of the standard solutions were added in triplicate to the 96-well plate. The para-nitrophenylphosphate (pNPP) substrate solution was prepared by dissolving 6.6 mg pNPP per 1 mL of dH2O. Then, 10 μL substrate was then added to each well using a multichannel pipette and the plate was incubated in the dark at 37 ◦C. The plate was monitored for color change and, at an appropriate time point, the absorbance at 405 nm was read, ensuring that the readings were within the range provided by the standard solutions. The amount of pNP in the sample wells was then calculated using the standard curve and used as a measure of ALP activity.

After a further freeze–thaw cycle, Quant-iT ™ PicoGreen ® dsDNA assay (ThermoFisher, UK) was used in accordance with the manufacturer's instructions to quantify the DNA content of the cultures. In detail, 50 μL of cell culture lysate was placed in a 96-well plate in triplicate along with 50 μL dH2O. A standard curve of DNA was also prepared in triplicate with final concentrations of 1000, 500, 100, 50, 10, 5, and 0 ng.mL−<sup>1</sup> when 100 μL of PicoGreen reagen<sup>t</sup> was added to the 100 μL of standard solution in each well. Then, 100 μL of freshly made PicoGreen reagen<sup>t</sup> was added to all the wells using a multichannel pipette. The plate was incubated in the dark for 4 min, after which the fluorescence was read using an excitation wavelength of 485 nm and an emission wavelength of 528 nm. The standard curve was plotted and used to convert the fluorescence units of the wells into the DNA content of the original cultures.

#### *2.16. Statistical Methodology*

Statistical analysis was conducted using ANOVA run with multiple comparisons using GraphPad Prism 8 (GraphPad Software, Inc., San Diego, CA, USA) with significance indicated with \* when *p* < 0.05. No demarcation of lower *p* values was used as all were tested with *p* = 0.05.
