Characterization of trimethyl chitosan/polyethylene glycol derivatized chitosan blend as an injectable and degradable antimicrobial delivery system

Abstract

Advanced local delivery systems are needed as adjunctive treatments for severe injuries with high infection rates, such as open fractures. Chitosan systems have been investigated as antimicrobial local delivery systems for orthopaedic infection but possess mismatches between elution and degradation properties. Derivatives of chitosan were chosen that have enhanced swelling ratios or tailorable degradation properties. A combination of trimethyl chitosan and poly(ethylene glycol) diacrylate chitosan was developed as an injectable local delivery system. Research objectives were elution of antimicrobials for 7 days, degradation as open fractures heal, and cytocompatibility. The derivative combination eluted increased active concentrations of vancomycin and amikacin compared to the non-derivatized chitosan paste, 6 vs. 5 days and 5 vs4 days, respectively. The derivative combination degraded slower than non-derivatized paste in an enzymatic degradation study, 14 vs. 3 days, which increased antimicrobial delivery duration. Cytocompatibility of the combination with fibroblast and pre-osteoblast cells exceed the cell viability standard set in ISO 10993–5. Combination paste requires an increased ejection force of 9.40N (vs. 0.64N), but this force was within an acceptable injection force threshold, 80N. These preliminary results indicate combination paste should be further developed into a clinically useful adjunctive local delivery system for infection prevention.

1. Introduction

Approximately 100,000 open fractures occur each year in the United States [1]. Open fractures are traumatic injuries that engage both soft and hard tissues. Treatment of open fractures is exceedingly difficult due to compromised vasculature and irregular geometries, and may be further complicated by antimicrobial resistant bacteria and the presence of biofilm [2–4]. Necrotic tissue, devascularized bone, and fracture fixation devices are ideal substrates for bacterial adherence and biofilm formation. Pathogenic bacteria such as Staphylococcus aureus (S. aureus), Staphylococcus epidermidis (S. epidermidis), and Pseudomonas aeruginosa (P. aeruginosa) account for nearly 75% of biofilm-based infections in medical devices [5]. Despite following established treatment protocols, up to 50% of patients sustaining these injuries will become infected [6–8]. Therefore, infection prevention is crucial in the management of patients with open fractures.

Local delivery of antimicrobials has become a popular adjunctive therapy in combination with systemic antimicrobials for infection prevention [7]. Local administration of therapeutic agents reduces systemic toxicity and maximizes their concentration at the injury site. Several materials have been fabricated into local delivery systems for the treatment of orthopaedic infections over the past several decades, including poly(methyl methacrylate) (PMMA), calcium sulfate (CaSO₄), and collagen [7, 9]. These materials have been used clinically with some success but possess drawbacks that limit treatment efficacy. PMMA is incompatible with temperature-sensitive antimicrobials due to heat released during polymerization and requires explantation [7]. Unless removed in a timely manner, antimicrobials are eluted at subtherapeutic dosages that may promote antimicrobial resistance [10]. CaSO₄ is a resorbable biomaterial that eliminates the need for removal surgeries. However, rapid degradation may lead to production of sterile wound drainage [11, 12]. Collagen is a resorbable biopolymer, but its main drawback is the rapid release of antimicrobials does not correspond to its rate of degradation within the body [7, 9].

Chitosan, a natural glycomaterial, is a polysaccharide that is principally derived from the exoskeleton of crustaceans. This abundant biopolymer has been developed into a wide variety of therapeutic delivery systems due its biocompatibility and ability to degrade within the body [13, 14]. Other advantages include its mucoadhesivity and intrinsic antimicrobial properties [15]. Release kinetics of antimicrobials from chitosan systems typically exhibit a bolus release and show similar drawbacks as collagen systems [11]. However, the properties of chitosan can be modified due to the presence of reactive functional groups, and derivatives of chitosan can be specifically produced to enhance desirable characteristics while retaining its biocompatibility and degradation properties [15–17]. Quaternized chitosans are water soluble, have enhanced antimicrobial properties, and have been shown to degrade at the same rate or faster than unmodified chitosan [18, 19]. Cross-linked chitosans have been shown to extend release kinetics and enhance swelling ratio, but they also slow the degradation rate [20, 21].

In this study, we developed a blend of trimethyl chitosan (TMC) and poly(ethylene glycol) diacrylate chitosan (PEGDAc) to combine the water solubility and degradation properties of TMC with the enhanced swelling ratio of PEGDAc. When the composite is hydrated with antimicrobial solution, a paste is formed that can be used as an injectable delivery system for preventing infection in open, complex musculoskeletal trauma. Toward demonstrating feasibility for this combination as an injectable infection prevention biomaterial, elution kinetics of vancomycin and amikacin, activity of antimicrobials, and degradation of the composite were compared to non-derivatized chitosan controls. Additionally, biocompatibility and injectability were evaluated against clinically relevant standards of 70% viability and 80N of force to assess feasibility of delivery for the complex geometry of open fractures [22, 23].

2. Materials and Methods

2.1. Synthesis and fabrication

Two chitosan derivatives were prepared from chitosan (Chitinor AS, Norway) with a molecular weight of 250.6 kDa and degree of deacetylation of 82.46%. The first derivative was N-trimethyl chitosan (TMC), a quaternized chitosan derivative, and the other derivative was poly(ethylene glycol) diacrylate chitosan (PEGDAc), a graft copolymer of chitosan. Unmodified chitosan delivery systems previously developed in this lab were prepared using the same batch of chitosan and used as controls [24, 25]. Previous evaluations demonstrated that chitosan with this molecular weight and degree of deacetylation has favorable biocompatibility and degradation properties for local delivery of antimicrobials [25, 26].

2.1.1. Preparation of trimethyl chitosan

TMC was synthesized according to a method adapted from Verheul et al. with minor modification [27]. This method was chosen to avoid side reactions that are present in other methods for preparing TMC [28]. Ten grams of chitosan was dissolved at a concentration of 4% (w./v.) in 30 mL formic acid (Fisher Chemical, USA), 40 mL formaldehyde (Fisher Chemical, USA), and 180 mL deionized (DI) water and modified using an Eschweiler-Clarke reaction. The solution was heated to 70°C and allowed to stir for 120 hours (Fig. 1). The resulting solution containing the intermediate product, dimethyl chitosan (DMC), was evaporated under vacuum. This solution was gelled using 1 M NaOH and washed with DI water. DMC was dissolved in DI water, adjusted to pH = 5 with 1 M HCl, and dialyzed for 3 days against DI water. The purified solution was frozen at −80°C and lyophilized for 3 days. To remove particulates, DMC was dissolved in DI water, gelled using 1 M NaOH, and washed with DI water and acetone. The approximate yield for the conversion of chitosan to DMC was 90%. Two grams of DMC were suspended in N-methyl-2-pyrrolidone (NMP, EMD Millipore, USA), and excess iodomethane (Alfa Aesar, USA) was added to the solution. This solution was heated to 40°C and allowed to stir for 72 hours (Fig. 1). The solution was dropped in a 50:50 mixture of ethanol (Fisher Chemical, USA) and diethyl ether (Acros Organics, USA) to precipitate the final product. TMC was separated from the solution using centrifugation and dissolved in a 5% (w/v.) solution of NaCl. This solution was dialyzed against DI water for 3 days, frozen at −80°C, and lyophilized. The yield for conversion of DMC to TMC was approximately 65%.

Figure 1.

Schematic representation of two step synthesis of N-TMC from chitosan.

2.1.2. Preparation of poly(ethylene glycol diacrylate) chitosan

PEGDAc was synthesized according to the method described in Shitrit et al. with minor modification [29]. Two grams of chitosan was dissolved at 1% (w/v.) in a 1% (v./v.) solution of blended lactic and acetic acid (3:1 ratio). This solution was stirred for 24 hours to allow the chitosan to completely dissolve. Then, 2 g of poly(ethylene glycol) diacrylate (PEGDA) with a molecular weight of 8000 (Alfa Aesar, Massachusetts) was added to the solution to also achieve 1% (w./v.) concentration. The mixed solution was stirred for 15 minutes to allow the polymer to dissolve and heated to 60°C for 3 hours (Fig. 2). A Michael addition reaction between the amine groups on chitosan and the acrylate groups on PEGDA would result in the partially cross-linked structure of PEGDAc. The reaction mixture was dialyzed against DI water for 3 days to remove unreacted PEGDA molecules, frozen at −80°C, and lyophilized for 3 days. The freeze-dried product was treated using 0.25M NaOH, and the pH was reduced to neutral using copious DI water. Finally, the neutralized product was frozen at −80°C and lyophilized for 3 days. Final yield for conversion of chitosan to PEGDAc was approximately 50%.

Figure 2.

Schematic representation of one step synthesis of PEGDAc from chitosan.

2.1.3. Grinding and reconstitution

Final lyophilized products were ground into small flakes and stored in a desiccator. Chitosan derivatives were combined in a 3:1 weight ratio of PEGDAc to TMC due to preliminary evaluations showing this combination to have favorable cytocompatibility and the longest elution profile compared to other ratios. To form the injectable delivery system, the ground products were weighed directly into a Luer-Lock syringe. In a separate Luer-lock syringe, 4.5 mL of antimicrobial solution or PBS per gram of ground chitosan derivatives was drawn into the second syringe. Loss of product was minimized by coupling the two syringes, hydrating the powder by mixing the liquid solution into the powder, and moving the components between the syringes to form a paste.

2.1.4. Preparation of controls

Two controls were prepared for comparison to the developed system: an injectable paste and lyophilized sponges. Berretta et al. previously developed an injectable system and was prepared according to their methods [25]. Blended chitosan control paste was prepared by dissolving 1% chitosan (w./v.) in 0.85% (v./v.) acetic acid solution and adding 1% (w./v.) poly(ethylene glycol) with a molecular weight of 8000 (PEG, Alfa Aesar, USA) to the solution . This solution was frozen at −80°C, lyophilized for 3 days, and not neutralized. The PEG blended chitosan control was not mixed with TMC or PEGDAc. Chitosan sponge controls were prepared according to the methods in Noel et al. with minor modification [24]. Lyophilized sponges were prepared by dissolving 1% chitosan (w./v.) in a 1% (v./v.) blended lactic and acetic acid solution (3:1 ratio). This solution was frozen at −80°C, lyophilized for 3 days, and treated using 1.0M NaOH. The pH of the sponges was brought to neutral using DI water, frozen at −80°C, and lyophilized. Control sponges were hydrated by placing them in petri dishes filled with solution and allowed to passively absorb the solution for two minutes prior to experimentation.

2.2. Fourier transform infrared spectroscopy

A Nicolet iS10 (Thermo Scientific, USA) using attenuated total reflectance (ATR) FTIR was used to collect spectra for chitosan, DMC, TMC, PEGDA, and PEGDAc. Samples were prepared by vacuum drying at 40°C for 3 days. Data was collected using a deuterated tryglycine sulfate (DTGs) potassium bromide detector over the range of 525–4000 cm⁻¹ with 64 scans and a resolution of 2 cm⁻¹. Ambient air was used to apply baseline corrections. Spectra were analyzed using Thermo Scientific OMNIC Software.

2.3. Nuclear magnetic resonance spectroscopy

The ¹H NMR spectra were collected using a JEOL Resonance 400 MHz NMR spectrometer at 25°C. TMC and PEGDAc samples were dispersed at 1% (w/v.) concentration in 1.0 mL of D₂O (Acros Organics, USA) with 0.01 mL of DCl (Acros Organics, USA) added to dissolve the polymers. Solutions were vortexed overnight to ensure complete dissolution.

2.4. Intrinsic viscosity

Viscosity of TMC and Chitosan materials were determined using an Ubbelohde viscometer. In this method, intrinsic viscosity values are obtained from flow times of solvent and material recorded at 25°C. First, flow times were recorded for the solvent (0.3M Acetic Acid/0.2M Sodium Acetate) in triplicate. Next, flow times of 4 different concentrations of TMC and chitosan were recorded, also in triplicate. Recorded flow times were used to calculate the relative viscosity, specific viscosity, and reduced viscosity of each sample. A graph of reduced viscosity vs concentration was plotted and the intrinsic viscosity obtained using data linear regression [30, 31].

$${\eta }_r=\frac{T_i}{T_0}\text{Relative Viscosity}$$

$${\eta }_{sp}={\eta }_r-1\text{Specific Viscosity}$$

$${\eta }_{red}=\frac{{\eta }_{sp}}{c}\text{Reduced Viscosity}$$

2.5. Antimicrobial elution

Combinations of TMC and PEGDAc and non-derivatized paste were hydrated using a 5 mg/mL combination solution of vancomycin and amikacin (MP Biomedicals, USA). Approximately 0.6 mL of hydrated paste (n = 4) was injected into cell crowns (Scaffdex, Finland) with nylon filters (pore size = 41 μm) attached (See Supplementary Material for visual representation). Each sample was placed in 5 mL of phosphate buffered saline (PBS), incubated at 37°C, and sampled daily. Upon sampling, each sample was completely refreshed with PBS. Vancomycin was detected and quantified using a high performance liquid chromatography (Dionex UltiMate 3000 HPLC, Thermo Scientific, Waltham, MA) system interfaced with a UV/Vis spectrophotometer at 209 nm [32]. Amikacin was quantified using a previously described method of pre-column derivatization with an o-phthaldialdehyde reagent (AdipoGen Life Sciences, USA) and subsequent detection with an HPLC system using a fluorescence detector (Excitation = 340 nm, Emission = 455 nm) [33]. Both detections utilized reverse-phase columns, C18 150 × 4.6 mm (Hypersil Gold, Thermo Scientific) for vancomycin and C8 100 × 4.6 mm (Hypersil BDS, Thermo Scientific) for amikacin, with mobile phase consisting of 85% phosphate buffer at pH = 7.4 and 15% acetonitrile.

2.6. Antimicrobial activity

Activity of the eluates against relevant orthopaedic pathogens S. aureus (UAMS-1; ATCC 49230) and P. aeruginosa (PA; ATCC 27317) was evaluated using zone of inhibition (ZOI) assays. Bacteria were grown separately overnight at 37°C in tryptic soy broth (TSB, MP Biomedicals, USA). Overnight growth was diluted, 1:10 for S. aureus and 1:50 for P. aeruginosa, in TSB, and 100.0 μL of these solutions was spread on tryptic soy agar plates. Blank paper discs (diameter = 6.0 mm) were hydrated with 30.0 μL of eluate and placed on the plates. These were incubated for 24 hours at 37°C and photographed. ImageJ software was used to determine the ZOI for each disc.

2.7. Enzymatic degradation

Approximately 0.3 mL (n = 3) of each system was hydrated using PBS and placed in petri dishes. Non-derivatized neutral sponges were cut into equal parts (n = 3), hydrated with PBS, and placed in petri dishes. Degradation solution was prepared by dissolving 1 mg/mL lysozyme type VI (MP Biomedicals) and 100 μg/mL Normocin antibiotic/antimycotic (Invivogen, USA) in PBS. Then, 5 mL of degradation solution was added to each petri dish, and samples were placed in the incubator at 37°C. At days 1, 3, 5, 7, and 14 samples were taken, aspirated through a nylon filter (pore size = 41 μm), and placed in an oven at 45°C. Fresh degradation solution was replaced every other day by aspirating the old solution through a nylon filter (pore size = 41 μm) and adding 5 mL of fresh solution. After drying, samples were weighed and compared to their initial weight to determine degradation rate. Different samples were used each day due to the destructive nature of the test.

$$Percentremaining=\frac{Massafterdegradation}{Originaldrymass}\times 100\%$$

2.8. Cytocompatibility

NIH3T3 fibroblast and MC3T3 pre-osteoblast cells were seeded at 1 × 10⁴ cells/cm² in 24-well plates and grown in Dulbecco’s Modified Eagle’s Medium (DMEM, HyClone, USA) supplemented with 10% fetal bovine serum (Corning, USA) and 100 μg/mL Normocin for 24 hours at 37°C and 5% CO₂. Approximately 0.3 mL (n = 3) of each paste was hydrated using sterilized PBS and injected into cell culture inserts (Falcon, pore size = 8 μm) and placed in each well. Cells were exposed to eluates from the paste for 24 and 72 hours and quantified using a Cell-Titer Glo (Promega) assay. Results were normalized as a percent viability of cells grown on blank tissue culture plastic. Samples were sterilized with ethylene oxide gas (EtO) prior to testing.

2.9. Injectability

Injectability of the pastes was evaluated by measuring the maximum ejection force from a standard 1 mL syringe. Approximately 0.9 mL (n = 3) of paste was hydrated using PBS and ejected. Each syringe was placed in an Instron Universal Testing Machine (Instron) with a 500 N load cell and programmed to compress the syringe at a constant rate of 1 mm/second to fully eject the pastes. Force required to fully eject the pastes was measured, and maximum force required was recorded. Syringes filled with air were used for baseline corrections.

2.10. Statistical analysis

Statistical analysis of the results was performed using Sigma Plot 14 (Systat Software). Two way analysis of variance (ANOVA) with group and time point as the two factors and Student-Newman-Keuls post hoc analysis were used to determine statistical differences between groups for degradation rate, antimicrobial elution, and relative cell viability. T-tests were used to assess differences for injectability of the pastes. P values < 0.05 were considered statistically significant.

3. Results and Discussion

3.1. Fourier transform infrared spectroscopy

The ATR-IR spectrum of PEGDAc clearly show the characteristic peaks attributed to PEG. They are, for example, at 1105, 1464, and 2877 cm⁻¹ corresponding to C-O stretching, C-H bending due to methylene groups, and C-H stretching, respectively (Fig. 3). These peaks are absent in the non-derivatized chitosan sample. This result proved the existence of PEGDA within modified chitosan, PEGDAc, however the detailed reaction between PEGDA and chitosan was not able to be confirmed by IR. Noticeable differences were not observed between the spectra of DMC, TMC, and chitosan.

Figure 3.

ATR-FTIR absorbance spectra for I: PEGDA 8000, II: PEGDAc, and III: unmodified chitosan. Peaks of interest are indicated, and peak at 2350 cm⁻¹ is due to CO₂.

3.2. Nuclear magnetic resonance spectroscopy and Intrinsic Viscosity

The ¹NMR spectra for PEGDAc and TMC disclosed their chemical structures and compositions. An NMR spectrum was also collected for DMC and demonstrated a quantitative dimethylation of amino groups (See Supplement 1). For PEGDAc, the sharp peak at 3.5 ppm corresponds to the PEG portion of PEGDA, and the peak at 1.9 ppm corresponds to the acetyl moiety (Fig.4).The existence of the vinyl end group of PEGDA was examined with ×100 enlarged spectrum between 5.9–6.5 ppm (inserted spectrum in Fig. 4) to evaluate the crosslinking reaction. No noticeable peaks were found. This may indicate that the PEGDA molecules present are cross-linked to the chitosan backbones. Combining NMR results with the FTIR spectra provides evidence for the synthesis of PEGDAc.

Figure 4.

NMR spectrum of 10 mg/mL PEGDAc dissolved in D₂O/DCl at room temperature.

For TMC, the peak at 2.9 ppm corresponds to the N-dimethyl group, and the peak at 3.2 ppm corresponds to the N-trimethyl group. The peak at 1.9 ppm (singlet) is attributed to the hydrogens of the acetyl moiety. Peaks 3.25, 4.95, and between 3.5–4.2 ppm are attributed to the hydrogens present on the chitosan backbone (Fig. 5). NMR results provided evidence that TMC was successfully synthesized. Small peaks not belonging to TMC can be noted at 3.32, 2.6, 2.3, and 1.85 ppm; these were attributed to residual NMP that had not removed by diethyl ether washes and extensive dialysis. Due to this residual impurity, additional washing and dialysis steps were added to the synthesis protocol. A sharp peak at approximately 2.1 ppm was present in both spectra, and this was attributed to residual acetone from cleaning of the NMR tubes.

Figure 5.

NMR spectrum of 10 mg/mL TMC dissolved in D₂O/DCl at room temperature.

The NMR spectra were also used to calculate the grafting ratio for PEGDAc and degree of quaternization (DQ) of TMC. Although the accurate content of PEGDA was difficult to calculate due to the overlap with a broad chitosan peak, the integration ratio of separated peaks revealed that the PEG content in PEGDAc was only 3–6 weight %. It was assumed that PEGDA was completely removed by dialysis, but some PEGDA chains potentially remained in the solution. However, assuming the most of PEGDA chains are crosslinked between chitosan, the 3–6% content will still improve the stability of the chitosan paste to slow the drug release and the degradation. For the TMC composition, the ratio of N-trimethyl and N-dimethyl of the final product was calculated to be 12 and 88 mol %, respectively. The relatively low DQ obtained in this study, 12%, could be a result of different chitosan being used in the study that the synthesis was adapted from [27]. Unlike previous results with this synthesis, viscometry measurements showed a lower viscosity for TMC (1.37 dL/g) compared to the parent chitosan polymer (6.07 dL/g)[27]. This may indicate that degradation or chain scissions occurred during the synthesis, similar to findings by Jintapattanakit et al. that increasing number of reaction steps led to decreased molecular weight and intrinsic viscosity [34]. Due to the larger molecular weight in the present study, iodomethane may not have been added in enough excess to get a DQ higher than 50%. However, even with this relatively low DQ, one of the major shortcomings of nonderivatized chitosan, solubility at neutral pH, can be overcome by TMC [28]. Some intrinsic properties of chitosan, mucoadhesive and antimicrobial effects, have been attributed to the positive charge on the amine group. TMC has been reported to have superior characteristics compared to non-derivatized chitosan, even with a low DQ [35], and increasing the DQ of TMC has been shown to enhance its antimicrobial properties [36]. Future studies will investigate alternative synthetic strategies that incorporate a base catalyst or protecting groups that would likely result in a higher DQ [37, 38].

3.3. Antimicrobial elution

Similar release kinetics were observed for both antimicrobials and delivery systems (Figs. 6a, 6b, 7a, and 7b). A burst release of antimicrobials was observed on day 1 for both systems that tapered below detectable levels after day 7. Concentrations of vancomycin dipped below the minimum inhibitory concentration (MIC) of S. aureus after day 6, and amikacin concentrations dropped below the MIC of P. aeruginosa by day 5. Cumulative release of vancomycin from the combination paste was 76.5 ± 13.5% and 76.4 ± 10.6% from the non-derivatized paste. Similar cumulative release profiles were observed for amikacin with the combination eluting 81.7 ± 13.2%, and the non-derivatized paste eluting 81.6 ± 10.3%. No significant differences were observed between the groups, but the combination paste eluted vancomycin at day 7 while the non-derivatized paste did not.

Figure 6.

In vitro elution of vancomycin. (a) Full elution profile for days 1–7. (b) Shows zoomed in elution profile for days 3–7. Data are presented as mean ± standard deviation (n = 4). * denotes significant differences from control (p < 0.05).

Figure 7.

In vitro elution of amikacin. (a) Full elution profile for days 1–7. (b) Shows zoomed in elution profile for days 3–7. Data are presented as mean ± standard deviation (n = 4).

A primary goal of this research was to reduce the mismatch between elution kinetics and degradation rates that affects chitosan-based delivery systems [7]. Elution kinetics for these systems typically exhibit an initial bolus release that is followed by low levels of release until the system is degraded [25, 39]. PEGDA has previously been investigated as a cross-linked hydrogel and a cross-linking agent with a variety of systems to sustain drug delivery [21, 40, 41]. In preliminary work not contained in this manuscript, our group showed that PEGDAc had twice the swelling ratio of unmodified chitosan, and other groups have produced similar results [21, 40]. This could help explain the extended release profile observed compared to unmodified chitosan and other injectable chitosan paste studies [25, 26]. The extended release profile observed in Chen et al.’s study of a thiolated chitosan, PEGDA, and β-glycerophosphate complex is similar to the current study [21]. However, their study showed a much longer release profile that was likely due to the size of the protein which would have a considerably lower diffusion coefficient [40]. Similar to the current study, Liang et al. observed sustained release of amoxicillin, a common antimicrobial agent, and doxorubicin from an injectable, cross-linked chitosan hydrogel [42]. Their study investigated release for the first 60 hours, which makes direct comparison to this study difficult, but it appeared that elution would continue after that point.

3.4. Antimicrobial activity

Large zones (> 10mm) were observed for both systems on day 1 that tapered off to clinically irrelevant zones (< 1mm). Combination paste inhibited growth of S. aureus for 6 days while the non-derivatized paste only prevented growth for 5 days (Table 1). Similar to the previous results, combination paste produced zones against P. aeruginosa for longer (5 days) compared to the non-derivatized paste (4 days) (Table 2).

Table 1:

Zone of inhibition results for S. aureus. Zones greater than 10mm are indicated with +++, larger than 6mm are indicated with ++, larger than 1mm are indicated with +, and smaller than 1mm are indicated with −.

Formulation	Eluate Sample Time (Day)
	1	2	3	4	5	6	7
25TMC : 75 PEGDAC	10.10 ± 0.76 +++	8.27 ± 0.49 ++	6.71 ± 0.47 ++	4.37 ± 0.38 +	2.25 ± 0.46 +	1.40 ± 0.40 +	0.00 ± 0.00 −
Control	10.01 ± 0.59 +++	8.11 ± 0.68 ++	5.06 ± 0.53 +	3.56 ± 0.75 +	2.30 ± 1.11 +	0.45 ± 0.90 −	0.00 ± 0.00 −

Table 2:

Zone of inhibition results for P. aeruginosa. Zones greater than 10mm are indicated with +++, larger than 6mm are indicated with ++, larger than 1mm are indicated with +, and smaller than 1mm are indicated with −.

Formulation	Eluate Sample Time (Day)
	1	2	3	4	5	6	7
25TMC : 75 PEGDAC	13.09 ± 0.49 +++	9.85 ± 0.84 ++	7.38 ± 0.77 ++	3.45 ± 0.73 +	1.50 ± 0.30 +	0.00 ± 0.00 −	0.00 ± 0.00 −
Control	13.99 ± 0.43 +++	10.51 ± 0.94 +++	5.98 ± 1.87 +	2.87 ± 1.14 +	0.55 ± 1.10 −	0.00 ± 0.00 −	0.00 ± 0.00 −

MIC values reported in this study were lab-generated values produced by exposing the bacteria to serial dilutions of the antimicrobials used. The antimicrobials used in this study, vancomycin and amikacin, were chosen due to their effectiveness against orthopaedic pathogens, complementary spectrum of activity against Gram-positive and Gram-negative bacteria, and synergistic effects [43]. Activity observed for eluate samples indicate that incorporation of antimicrobials into the system did not inactivate them. Chitosan has been reported to possess antimicrobial properties, and quaternized chitosan derivatives, such as TMC, and chitosan that has been grafted with antimicrobial moieties, such as aniline tetramer, have been shown to have enhanced antimicrobial effects compared to chitosan [13, 19, 44–46]. The boost in activity observed against P. aeruginosa could possibly be due to an additive antimicrobial effect between amikacin and TMC. Future studies will investigate the antimicrobial activity of the delivery system against these pathogens to elucidate this additional activity.

3.5. Enzymatic degradation

Combination paste was almost completely dissolved/degraded by day 14 with 5.55 ± 1.93% remaining of the original mass (Fig. 8). A decrease in mass was observed for each day which is in contrast to the non-derivatized paste. Non-derivatized paste was degraded to 10% of its original mass by day 5 and remained there for the duration. Degradation of the paste delivery systems was likely due to the combination of enzymatic cleavage by lysozyme followed by dissolution. Neither paste dissolved upon initial placement into the solution, and the nylon filter would help prevent bulk material from being aspirated. The non-derivatized sponge did not experience any observable degradation. By day 14, it weighed 138.26 ± 2.07% of its original mass, and this is likely due to the salts in PBS being retained and crystallizing in the sponge.

Figure 8.

In vitro enzymatic degradation of chitosan-based delivery systems. Results are presented as mean ± standard deviation (n = 3). * denotes significant differences from control (p < 0.05).

Previous studies demonstrated release of vancomycin and amikacin from unmodified chitosan pastes for 3 days, but the observed degradation rates did not correspond to their elution kinetics. Berretta et al. showed almost complete degradation after day 2, and Rhodes et al. showed degradation for 10 days without complete degradation [25, 26]. Qu et al. and Zhao et al. developed chitosan-based hydrogels that were susceptible to hydrolytic degradation, and they showed approximately 50% degradation in PBS after 4 and 6 weeks, respectively [47, 48]. Rapid degradation of a previously developed chitosan paste for antimicrobial delivery led to exploration of methods to slow degradation [25]. Degradation of chitosan-based systems is influenced by the DDA and MW of the starting material. High MW and DDA chitosans have slow degradation rates but lowering the MW or reducing DDA closer to 50% accelerates degradation time. Adjusting the DDA to 50% or incorporating pendant groups into the backbone lowers the crystallinity of the material.

Preliminary evaluations demonstrated that increasing reaction time of TMC from 24 to 72 hours slowed its degradation rate. Extended degradation of the system under investigation could be due to a combination of the increased molecular weight of TMC and the cross-linking of PEGDAc [49]. The 14 day degradation was within the time frame of initial healing, but this may mean that material would remain after antimicrobials dropped below effective levels. Concentrations of lysozyme used in this study were higher than physiological levels, and this might indicate that degradation would take longer than 14 days [50]. However, Stinner et al. demonstrated that chitosan sponges were almost completely degraded in an in vivo model of infection in 42 hours. Whereas, in this study, there was no measurable degradation after two weeks. In vitro evaluations lack the complexity to mimic clinical scenarios accurately, and this makes translation of in vitro degradation results difficult to predict actual in vivo degradation time.

3.6. Cytocompatibility

Each delivery system evaluated was cytocompatible according the criteria established in ISO 10993–5 [22]. Cellular viability was normalized to cells not exposed to chitosan-based biomaterials. Viability of fibroblasts cells was reduced by approximately 12 and 13% after being exposed to combination paste for 24 and 72 hours, respectively (Fig. 9). There was no reduction in viability for the non-derivatized sponge and paste after 24 hours, but there was approximately a 25 and 30% reduction after 72 hours. Similar results were observed with pre-osteoblast cells with combination paste producing a 4 and 17% reduction in viability on days 1 and 3, respectively (Fig. 10). Again, there was not a reduction in viability for the non-derivatized sponge and paste after 24 hours, but there was a reduction of 6 and 16% after 72 hours.

Figure 9.

In vitro cytocompatibility evaluation with NIH3T3 fibroblast cells. Results are presented as mean ± standard deviation normalized to cells grown on blank tissue culture plastic (n = 3). * denotes significant differences from blank tissue culture plastic (p < 0.05). The black bar shows the threshold value outlined in ISO 10993-5 for cytocompatibility.

Figure 10.

In vitro cytocompatibility evaluation with MC3T3 pre-osteoblast cells. Results are presented as mean ± standard deviation normalized to cells grown on blank tissue culture plastic (n = 3). * denotes significant differences from blank tissue culture plastic (p < 0.05). The black bar shows the threshold value outlined in ISO 10993-5 for cytocompatibility.

A goal of this research was to maintain the biocompatibility of chitosan after derivatization. Early studies of TMC reported that this material was cytotoxic, and that increasing DQ enhanced the cytotoxic effects [19, 51,52]. Synthetic methods used to prepare TMC for the studies that showed cytotoxicity produced methylation at the amino and hydroxyl groups. More recent evaluations use synthetic procedures that specifically target the amino group and report minimal cytotoxic effects [53–55]. Also, among the N-alkylated chitosan derivatives, TMC has been shown to exhibit the least cytotoxicity [56]. Mazzoccoli et al. reported that concentrations of PEGDA between 20–40% weight percent reduce cellular viability by 20–64% as the concentration increased [57]. PEGDAc produced in this study was manufactured using PEGDA at a 1% weight percentage and showed no cytotoxicity in preliminary studies. Similar to the current study, Zhao et al. showed excellent cytocompatibility with L929 cells and minimal hemolysis of red bloods cells with a chitosan-based hydrogel that incorporated quaternized moieties and cross-linked using PEG derivatives [58, 59].

The apparent reduction in viability for each group and between days was higher than expected. It was noted during the experiment that cell culture media turned pinker for the nonderivatized sponge and yellower for non-derivatized paste, which indicated increased alkalinity and acidity, respectively. In addition, combination paste appeared to absorb some of the media and could have sequestered growth factors from the cell culture media, or residual solvent could have been leached from the device. Another consideration is that these evaluations used a static fluid flow model that would result in extended exposure of the cells to the materials, resulting in reduced cellular viability. Cytocompatibility considerations in this study do not account for the influence of large doses of antimicrobials being delivered to the cells. Antimicrobial levels attained in this study should not be cytotoxic to osteogenic cells present in open, complex musculoskeletal trauma according to Rathbone et al [60]. Their study showed that concentrations of vancomycin and amikacin up to 2000 μg/mL are not toxic to human osteoblasts [60].

3.7. Injectability

After applying baseline corrections, non-derivatized paste required 0.64 ± 0.22N to eject, and the combination paste required 9.40 ± 0.83N (Fig. 11). Both injectable delivery systems are below clinically relevant values for ejection force.

Figure 11.

Injectability assessment of combination paste. Results are presented as mean ± standard deviation (n = 3). * denotes significant differences from control (p < 0.05).

Preparation of an injectable delivery system offers distinct advantages over other systems currently used for the treatment of open fractures. Hydration immediately prior to application allows clinicians to choose which antimicrobials are incorporated, allowing adaptation of therapy to the clinical scenario. Also, the hydration ratio can be tailored to modify the injectability properties. By hydrating paste with more aqueous antimicrobial solution, the injection force may be lowered, and more antimicrobials can be delivered. MacDonald et al. determined the maximum ejection force of female healthcare workers using a standard chuck grip to be 79.5N [23]. Force values reported in this study are almost 10 times lower than this threshold value. Previous studies that have investigated injectable chitosan paste reported values for ejection force ranging from 30–150N [25, 26]. Injectability of the system allows for complete coverage of open fractures that may possess complex geometries. This is an advantage over other delivery systems that are not able to conform to these wounds.

4. Conclusions

This study sought to determine elution, degradation, and injectability of a derivatized chitosan paste combination containing TMC and PEGDAc to assess feasibility as an injectable local delivery system for antimicrobials. The combination exhibited an extended elution profile of active antimicrobials and increased degradation time compared to the non-derivatized paste without sacrificing the injectability or cytocompatibility of the delivery system. These enhanced properties may be due to the characteristics of the selected chitosan derivatives [28, 40].

TMC and PEGDAc were synthesized, characterized, and developed into an injectable paste. In vitro evaluations indicate that the combination of TMC and PEGDAc may be developed into an improved local delivery system for infection prevention in severe injuries with high infection rates. The derivative combination’s ability to elute active concentrations of vancomycin and amikacin, cytocompatibility with representative cell lines, and injectability through standard 1 mL syringes demonstrates its feasibility as an injectable antimicrobial delivery system. Capability to elute antimicrobials beyond 3 days make the system applicable for infection prevention in open fractures, and degradability within 14 days facilitates removal in a similar time frame to initial healing of these injuries. Injectability and degradability of the system provides a distinct advantage over other systems such as PMMA or CaSO₄ currently used for open fractures. Future studies will further characterize the materials properties by using more sophisticated methods to measure the molecular weight and viscosity of the derivatives, swelling ratio of the developed system, and microscopy to investigate the structure and morphology. Clinical potential of this system will be further investigated by using in vivo assessments for biocompatibility and infection prevention efficacy.