SiNPs were made by reacting 7.5 mL of cyclohexane, 1.8 mL of hexanol-1, and 1.7 g of Triton X-100 and stirring them at 800 rpm for 1.0 h. This solution was then mixed with 500 µL of dH₂0 and 100 µL of ammonia >99%. The particles were allowed to sit for a period of 1 h while stirring, and tetraethyl orthosilicate (TEOS) > 98% was added at a volume of 100 µL to the solution. This mixture was allowed to stabilize, and then 100 µL of TEOS was added to increase the particle size. The particles were left spinning, covered in aluminum foil, for 3 days. The final volume added was ≈300 µL. The solution was opened, and the excess of ammonia was allowed to evaporate. The solution was centrifuged at 10,000 rpm for 10 min to remove unreacted reactants. The supernatant was removed from the pellet. The pellets were washed with 0.1 M HEPES buffer and centrifuged three more times at 10,000 rpm until the particles were removed from excess organic solvents. The pellets were resuspended in HEPES buffer and then stored in the fridge at 4 °C for a period of 8 h. SiNPs were always made fresh before experiments were conducted (Brito-Silva et al., 2013; Sobral-Filho et al., 2017). Transmission electron microscopy (TEM) was used to image the particles, and these images were acquired in a Jeol 1011 transmission electron microscope.

The SiNPs were removed from the 4 °C fridge and centrifuged at 8,000 rpm for 10 min. An aliquot of 1.0 mL of freshly prepared 0.1 M HEPES was added to the centrifuged tube, and the tubes were suspended on a floating foam raft in a sonication water bath for 30 min at 80 kHz in ice water at a temperature of ≈4 °C–8 °C. The tubes were then removed, and the particles were centrifuged at 5,000 rpm for 2 min with the supernatant discarded. A solution was made using two different kinds of antibiotics: clindamycin hydrochloride and tetracycline hydrochloride, both with purities greater than >99%. The antibiotic removed from −20 °C freezer was dissolved in 0.1 M HEPES buffer (1.14 g in 40 mL dH₂O), and the antibiotic was kept on ice for the duration of the incubation with the SiNPs. The concentration of 200 mg/mL for both antibiotics was added at a 1:1 volume with the SiNPs to create a final concentration of 100 mg/mL for this specific concentration gradient. The SiNPs were kept at 4 °C on a rotary shaker at 300 rpm for 24 h after incubation with the antibiotic. The SiNPs were used immediately. Every experiment was conducted using freshly prepared particles produced in batches of >100 mL (Gounani et al., 2018). The particles were initially stored in PBS buffer, but this caused the particles to disintegrate. The SiNPs were put into microtubes and then centrifuged at 10,000 rpm for 5 min to spin down the particles, and then the supernatant was removed and discarded. Various concentrations of antibiotics were then tested as the optimal loading capacity was determined, ranging from 500 mg to 8 µg of clindamycin hydrochloride (HCl) and tetracycline HCl. The amounts of antibiotic were weighed on a top-loading scale, the material was transferred to a clean microtube, and 1.0 mL of HEPES buffer was added to the vial. Serial dilutions were performed if needed. The dose range was selected to mimic an oral dosage of the antibiotics.

Staphylococcus aureus ATCC 27217 (ATCC, Virginia, US) and Staphylococcus epidermidis (obtained from the Biochemistry and Microbiology Department, University of Victoria) were cultured in Luria–Bertani (LB) broth (Thermo Fisher Scientific, Massachusetts, US). Bacteria were mixed with 2.0 mL of LB broth in a clean glass test tube and placed in a 37 °C biological incubator for an 18 h period to create an overnight broth culture. The bacteria were then expanded by taking 1.0 mL and adding it to 9.0 mL of fresh LB broth that had been previously autoclaved. The broth tube was placed in the incubator at 37 °C at 300 rpm for 2–4 h to obtain the exponential phase of bacterial growth of OD₆₀₀ 0.2–0.4, determined after growth cultures were taken at different ratios. Aliquots (1.5 mL) were removed and placed in a spectrophotometer blanked with the sterile LB broth. The aliquot of bacteria was placed in the clean cuvette, and the OD₆₀₀ value was given as a value between 0 and 2. The OD₆₀₀ was determined in the exponential phase of growth of the bacteria by comparison to literature values of Staphylococcus aureus OD₆₀₀ over a period of 24 h.

The bacteria were inoculated by transferring aliquots with serological pipettes into autoclaved media bottles (Thermo Fisher Scientific, Massachusetts, US). Clindamycin hydrochloride (Hello Bio^®, New Jersey, US) and tetracycline hydrochloride (Sigma-Aldrich, Missouri, US) were the two antibiotics tested in this experiment. Distilled water was used to make up the LB and toxin superantigen (TSA) (Thermo Fisher Scientific, Massachusetts, US) media. Bacteria were grown at 37 °C with shaking at 300 rpm for 4 h, then read at an OD₆₀₀ of ≈0.70 (0.71 was the actual recorded value). Clindamycin HCl was diluted to a concentration of 8.0 mg/mL in HEPES buffer 0.1 M and kept on ice by weighing 8 mg of −20 °C clindamycin antibiotic and adding 1.0 mL of HEPES buffer 0.1 M. A 1.0 mL aliquot of the broth culture was placed in each well of a 12-well plate, and 1.0 mL of the clindamycin sample was added to the first well. 1.0 mL was removed and mixed into the second well using a clean sterile filter tip for each well. The serial dilution was finished by discarding the remaining contents after the last well was completed. The well plate was wrapped in parafilm and incubated at 37 °C in the bacteria incubator. The 12-well plate was then incubated for 24 h to allow the culture to grow, and the visual determination of the MIC was considered based on the qualitative appearance of growth/no growth that was present in each well.

Staphylococcus aureus was incubated as described above, and 1.0 mL of the culture was plated into each well of a 12-well plate. Different amounts of clindamycin HCl (500 mg, 250 mg, 100 mg, and 62.5 mg) were diluted in 1.0 mL of 0.1 M HEPES buffer. SiNPs were previously sonicated at 80 Hz in a water bath sonicator filled with ice, ≈4 °C, for 30 min. The SiNP particles were then removed and incubated with 1.0 mL of the antibiotic solution and placed in a rotary incubator shaker set at 4 °C and 300 rpm for 24 h. After incubation, 1.0 mL of the antibiotic solution was added to the culture broth and incubated in the bacterial incubator for 24 h at 37 °C, with shaking at 300 rpm. The plate was placed in the Cytation 5 plate reader, and the absorbance was read at OD₆₀₀ to determine whether the bacterial concentration had decreased after the incubation with the antibiotics.

Fibroblast cells (Primary Dermal Fibroblast Normal; Human, Neonatal (HDFn), ATCC^® PCS-201-010™—Neonatal foreskin fibroblasts, male donor) were used. Cells were transferred into a T75 flask (Thermo Fisher Scientific, Massachusetts, US) containing 15 mL of fibroblast growth media (DMEM, high glucose, no glutamine, no phenol red (Thermo Fisher Scientific, Massachusetts, US) + fetal Bovine Serum (FBS) certified 15% + 1% GlutaMAX (Thermo Fisher Scientific, Massachusetts, US), and Hanks Balanced Salt Solution (HBSS) (Thermo Fisher Scientific, Massachusetts, US) in a biological safety cabinet (BSC). The T75 flask containing cells was placed in a 37 °C, 5% CO₂ incubator overnight. Cell cultures were monitored using phase-contrast microscopy. After initial cell culture, 1.0 mL of trypsin was added for cell passage. Cells were maintained by replacing the culture media every 3 days.

The fibrin-based bioink was prepared based on previously published articles (Abelseth et al., 2019; Benwood et al., 2023; de la Vega et al., 2018; Sharma et al., 2020). The mechanical and rheological properties of this bioink were reported previously (Sharma et al., 2021). Fibroblasts were resuspended in a fibrin-based bioink at a concentration of 1.0 × 10⁶ cells/mL. Dome-shaped constructs were bioprinted based on the specifications detailed in the relevant CAD file (Figure 1) generated using Aspect’s studio software (V1.2.59.0, Aspect Biosystems, Vancouver, BC, Canada) using a rectilinear infill pattern in a repeated layer-by-layer fashion. The constructs consisted of six deposited layers of cell-laden bioink, according to previous studies (Sharma et al., 2020). Specific pressures are applied to each channel to monitor the flow rate to provide sufficient time for the crosslinking reaction to occur. The bioprinting speed was 25 mm/s, and the pressures for bioink, crosslinker, and buffer channels were 50 mbar, 60 mbar, and 100 mbar, respectively. The final structure comprised a 1.0-cm-diameter dome with six fiber layers, with an average width of ∼1.1 cm and a height of ∼0.7 cm.

The RX1 was used to 3D bioprint skin constructs containing fibroblasts in each well of a 12-well plate, and then 2.0 mL of fibroblast media was added to each well. Fibroblast media was made by adding 34 mL of phenol-free DMEM to a sterile conical and adding 15 mL of FBS and 1.0 mL of 1% GlutaMAX. Fresh media was prepared for each experiment. The constructs were stained with the CellMask™ Deep Red Actin Tracking Stain (Thermo Fisher Scientific, Massachusetts, US) according to the manufacturer’s instructions after printing. This process allowed for visualization under the confocal microscope using the appropriate excitation and emission.

Staphylococcus aureus was grown to an OD₆₀₀ as described previously. SiNPs loaded with either clindamycin or tetracycline were spun down in a centrifuge at 10,000 rpm and washed three times with equal amounts of 0.1 M HEPES buffer (Figure 2).

The constructs were moved with a spatula to three separate plates: one for clindamycin, one for tetracycline, and one for no treatment. The 12-well plate was opened, and 1.0 mL of the overnight broth culture of S. aureus at OD₆₀₀ ≈ 0.70 was added to each well and placed at the side of the culture to allow the bacteria to disperse over the construct. The construct was moved to the Cytation 5 plate reader for incubation of the bacteria/fibroblast mixture with 5% CO₂ and was left for a period of 4 h at 37 °C. After the incubation time, the 12-well plate was taken into the BSC, 1.0 mL of each antibiotic-loaded-SiNP in 0.1 M HEPES buffer was added to the wells, and the plate was placed back in the incubator for 6 h. The BacLight™ LIVE/DEAD^® Bacterial Viability Kit Protocol (kit L13152, Thermo Fisher Scientific, Massachusetts, US) was freshly prepared according to the manufacturer’s instructions. A 1.0 mL aliquot of the final solution was added to each well. Fluorescence levels were determined using the Cytation 5 plate reader using the Texas Red filter and the GFP filter after 15 min of incubation wrapped in aluminum foil. The plate was then fixed with 1:1 methanol-free 4% formaldehyde in pre-sterile PBS and filtered. The well plate was then read on the confocal microscope by analyzing the excitation at 485 nm and 530 nm for the emission of the green fluorescence from Syto9, and 485 nm excitation and 630 nm emission of the red fluorescence from the propidium iodide. Zen Black software was used to image the bacterial constructs, and Zen Blue software was used to process the images and interpret the data.

One-way ANOVA (analysis of variance) and Tukey’s post-hoc test were performed for colony-forming units (CFUs) counting using GraphPad Prism 6.0 (GraphPad Software, Inc., La Jolla, CA, USA), which was also used to depict all graphs. A p-value <0.05 was considered statistically significant. Data are presented as the mean ± standard error of the mean.

SiNPs were synthesized and imaged using TEM by diluting them in ethanol and placing them on a TEM grid (Figures 3a,b). The images were then analyzed on ImageJ software to determine the distribution size histogram, as shown in Figure 3e.

The size distribution of the particles mainly fell in the 40–50 nm target after preparation (Figure 3e). The size was controlled by the injection of TEOS, which increased the amount of orthosilicic acid in the system. Increasing size leads to decreased toxicity, as detailed by Mannerström et al. (2016). Figures 3c,d show the greater polydispersity of the particles.

The buffer of choice for the nanoparticles needed to have no ultraviolet light absorbance, as HPLC was later used to detect the antibiotics of interest, which have a peak at ≈220 nm. There has been some analysis of the cytotoxic effects of light-exposed HEPES buffer due to the formation of hydrogen peroxide (Devlin et al., 2021; Liu et al., 2022; Rahaman et al., 2023). This limitation was overcome during the experimentation by creating a fresh solution before the experimentation. The solution was kept in a conical flask wrapped in aluminum foil to reduce the light able to penetrate the buffer. The buffer has membrane impermeability that is necessary because it will not penetrate the cells or the bacterial membrane (Hu et al., 2017). The particles were washed 3× in an aqueous solution of pH 7.0 adjusted with a HEPES buffer balanced with sodium hydroxide.

The bacterial cultures were streaked onto Vogel and Johnson agar (VJA), which is classically used in diagnostics using plating methods rather than genomics to ensure species typing. The presence of black colonies after incubation for a maximum of 24 h at 37 °C is a positive indication of S. aureus, along with a yellow border. Figure 4a illustrates the plating technique that showed the red agar turning black/yellow. S. epidermidis was also cultured in VJA, and the presence of pink media was seen behind the black colonies, which can qualitatively distinguish the species.

Staphylococcus aureus can be diagnosed in several ways, depending on the type of infection and location. The initial step involved is the Gram stain test, which will indicate a Gram-positive species of bacteria that will appear in clustered cocci form (Prasad et al., 2012). The next plating technique was done on MacConkey agar to examine the lactose-fermenting abilities of bacteria. It is categorized as a selective and differential medium (Sigma-Aldrich). The agar was used in this study to examine the difference between Staphylococcus aureus and Staphylococcus epidermidis, the two bacterial species used. Staphylococcus aureus can ferment lactose, causing the pink plate to turn yellow. In contrast, Staphylococcus epidermidis cannot ferment lactose, and the plate will stay pink. This test was completed for both species grown on the same plate at 37 °C for 24 h to confirm the identity of both species.

The VJA agar and MacConkey agar were subcultured from the original TSA plates. Other testing, such as the coagulase test, was done on both species, and S. aureus was found to be positive, while S. epidermidis was negative. Figure 4b shows that S. aureus grown on TSA had a characteristic yellow color when streaked (top), and S. epidermidis had white colonies (bottom). This was further accompanied by a literature study that illustrated the name “aureus,” meaning gold, characterized the 2–3 mm diameter circle of gold pigment produced by the colonies of S. aureus (Pal et al., 2023). This contrasted with the small white colonies 1–2 mm in diameter. This identification was found to be consistent with all TSA plating. This analysis confirmed the species of bacteria as the species of interest and, for the purposes of this study, that they could be compared in the analysis. Antibiotic sensitivity testing was conducted on the primary species of interest, S. aureus, grown in a confluent layer with two antibiotic discs, tetracycline and vancomycin, respectively. The disks were placed on TSA agar plates and then positioned in a bacterial incubator for 24 h (Supplementary Material S1) to confirm susceptibility.

This study selected two antibiotics of interest due to their molecular structure and current treatment of S. aureus bacterial infections. The samples were run on HPLC to determine if the antibiotic was observed after the washing stages. Antibiotic standards were run in buffer to determine the peaks of interest for clindamycin HCl and tetracycline HCl. The HEPES buffer was also run. Samples were run in triplicate (clindamycin-loaded SiNP supernatant, tetracycline-loaded SiNP supernatant, and bare SiNPs not loaded with antibiotics) (Supplementary Figure S2). The HPLC detection of clindamycin post incubation and washing in the supernatant of SiNP data is presented in Table 1.

The presence of the antibiotic could be observed in the data sets obtained by HPLC analysis. As a control, an initial weight sample of clindamycin HCl in 500 mg/mL in water was used to determine whether any peaks were present and to determine the mobile phase. Clindamycin HCl was also tested in ethanol and in buffer. Due to the polarity and hydrophobicity of the molecule, it was difficult to determine the optimal pH range, which was achieved by adding 1–2 sodium hydroxide pellets into the buffer to neutralize the solution to ≈ pH 7.5, which was confirmed using pH strips. Further analysis was completed to create a standard curve using five standards. This was challenging, as saturation of the antibiotic in the buffer solution was difficult to overcome. Clindamycin was determined to be optimal in the HEPES buffer at 14.2 mg/mL, and this was then serially diluted by adding 500 μL of this original standard to 500 μL of fresh buffer that was prepared on the day of the experiment.

The presence of clindamycin HCl was verified using the HPLC method due to the lack of specificity and accuracy of UV-Vis spectrophotometry (Devlin et al., 2021; Liu et al., 2022; Rahaman et al., 2023). The amount of antibiotic was quantified using a calibration curve to detect the presence of clindamycin HCl recovered in a sample. This method of recovery extraction is needed to determine the amount of loss of antibiotic due to washing and to determine the loading capacity and saturation levels. The loading of clindamycin HCl into the SiNPs was consistent across all batches, as confirmed by reproducible concentration profiles obtained in triplicate measurements over multiple iterations. The antibiotic peak was clearly identified within the expected range, confirming successful loading and detection. Quantitative HPLC analysis verified that the concentration of clindamycin HCl in the particles was stable and measurable within the linear range of the calibration curve. The indication that the particles were exhibiting the same levels of antibiotic as 200 mg/mL in buffer suggests that the particles had reached their loading capacity. While this method indicates that antibiotics were successfully loaded into the SiNPs, it does not provide insight into the time course of release.

The primary function of fibroblasts is to maintain the structural integrity of connective tissues by supporting the extracellular matrix. This cell line was chosen because it mimics the natural infection site of the S. aureus. 3D-bioprinted fibroblast constructs were used to illustrate an in vitro analysis to show Staphylococcal infection on the constructs. Once infected, the constructs can show the effectiveness of the treatment in patients with serious skin infections (Dokoshi et al., 2025; Kendall and Feghali-Bostwick, 2014). Figure 5 shows the green fluorescence signal, indicating live bacteria (Figures 5a,c,e), and the red fluorescence signal, indicating dead bacteria (Figures 5b,d,f). The comparison demonstrates the ability of the SiNPs loaded with clindamycin 500 μg/mL to induce bacterial death. Figures 5c,d show that the bare SiNPs did not have an impact on bacteria death, indicating the particles were non-toxic drug carriers. Figure 6 show S. aureus bacterium incubated without the use of antibiotics or particles, which gave a green fluorescence signal, indicating viability. The use of this control was significant because the fluorescence kit that was used is specific for bacteria and illustrated that the bioprinted construct did not have any natural antimicrobial properties, as was anticipated.

The constructs inoculated with bacterial cultures were labeled with green fluorescent protein (GFP) and then were treated with antibiotic-loaded SiNPs. The bacteria were transformed to express GFP, as done previously, to enable easy visualization (Devlin et al., 2021; Liu et al., 2022; Rahaman et al., 2023; Yang et al., 2022). GFP labeling enabled us to track the presence of the bacteria. Different conditions were investigated for the bioprinted constructs, including pre-infection and pre-treatment. The pre-infection group corresponded to a control group of bioprinted fibroblast constructs with no bacteria. The post-infection and pre-treatment group corresponded to the bacterial inoculation in the 3D-bioprinted construct prior to treatment with SiNPs, and the post-infection and post-treatment group corresponded to the bacterial inoculation in the 3D-bioprinted construct with SiNP-loaded clindamycin at a treatment of 500 mg/mL. As expected, the third group had fewer bacteria, as indicated by GFP expression, due to the effective action of the SiNP-loaded clindamycin treatment in the “dermis” model of 3D-bioprinted construct against S. epidermidis (Figure 6).

Staphylococcus epidermidis was used as a model organism for a live, real-time analysis that would allow for experimentation of the in vitro modeling to determine if saturation times >24 h would improve the SiNP release, in contrast to the 4–6 h that was completed in this study.

Experiments were also performed to quantify the effectiveness of the previous experiment in the “dermis” model of 3D-bioprinted construct inoculated with S. epidermidis and treated with SiNP-loaded clindamycin 500 mg/mL, mimicking the real infectious skin disease. Figure 7a showed that the CFU counts in the pre-infection and pre-treatment conditions had higher S. epidermidis CFUs than in the HDFn media. In this test, the results were regarding the timeframe before the bacteria were inoculated in the 3D-bioprinted construct, as the bacteria were separated from the fibroblast media, which confirms a lack of contamination in the 3D-bioprinted construct that could change the results. Figure 7b shows the CFU counts in the post-infection and pre-treatment condition, indicating higher S. epidermidis growth on day 2 post-infection. The growth was uncontrolled, as it was untreated. Figure 7c shows the CFU counts in the post-infection and post-treatment conditions, highlighting the effectiveness of SiNP-loaded clindamycin 500 mg/mL in killing S. epidermidis in the 3D-bioprinted construct. It was similar to the treatment with clindamycin at 500 mg/mL with no SiNP delivery, which was also favorable. However, the treatment with clindamycin at 100 mg/mL was not effective, showing considerable CFUs. Also, SiNPs alone showed some S. epidermidis growth. Figure 7d shows the growth curve of the S. epidermidis over time. Figure 7e shows bacterial imaging of S. epidermidis using BacLight^® fluorescence detection. Bacteria were imaged in LB broth after the addition of SiNP-loaded clindamycin 500 mg/mL. The GFP-stained live bacteria are the green dots in the image, and the Texas Red-stained dead bacteria are the red dots in the image. The absence of red dots indicates that the SiNP-loaded clindamycin 500 mg/mL prevented the S. epidermidis from forming colonies or biofilms in the 3D-bioprinted construct.