Efficient Antibacterial Agent Delivery by Mesoporous Silica Aerogel.

Bacterial infections still cause many health problems for human beings. Silica aerogels with a three-dimensional (3D) porous structure and a large surficial area are promising candidates for drug delivery, but they have rarely been investigated for antibacterial agent delivery. Herein, we study mesoporous silica aerogels as carriers for delivery of three slightly soluble antibacterial agents including cinnamaldehyde (CA, liquid), salicylic acid (SAA, solid), and sorbic acid (SOA, solid) under supercritical fluid carbon dioxide. Notably, all three antibacterial agents form uniform nanocrystals in the mesopores of silica aerogels and the loading efficiency reaches 56 wt %, which assists in overcoming the obstacles of low bioavailability of slightly soluble antibacterial agents. Benefiting from nanocrystallized antibacterial agents, the agent-loaded aerogels exhibit an inhibition rate of 99.99% against Escherichia coli during the initial release; notably, they still have a 95% inhibition rate even after ∼90% of CA is released. Importantly, the agent-loaded silica aerogels demonstrate good biocompatibility in vitro. This work indicates that mesoporous silica aerogels are a promising platform for antibacterial agent delivery.

Introduction

Accompanied by the advancement of medical and health safety in human society, various risks caused by bacterial infections such as inflammation, cholestasis, cysts, and sepsis are earning increasing attention.[1] In coping with bacterial infections, active pharmaceutical ingredients (APIs) are usually used for therapy, but most APIs exhibit low efficiency when used directly and degrade during transportation, thus requiring repeated administration to maintain the therapeutic effect.[2] Alternatively, drug delivery systems that provide APIs locally in a controlled manner could improve the in vivo efficiency of many drugs because they help mitigate side effects, prevent drug degradation, reduce administration frequency, etc.[3,4] Researchers have developed various organic and inorganic materials as carriers for drug delivery, such as liposomes,[5] hydrogels,[6] polymer nanoparticles,[7] phase transition nanoparticles,[8] metal–organic frameworks,[9] mesoporous silica nanoparticles,[10] etc., in which the carriers can improve the safety and effectiveness of the drug by enhancing the stability, solubility, and service life of the packaged drug.[11] Among these materials, mesoporous silica nanoparticles stand out because of their high specific surface area, large porosity, excellent biocompatibility, high drug loading efficiency, and controllable drug release.[12,13] These characteristics enable mesoporous silica nanoparticles to be combined with various antibacterial agents and exhibit excellent endocrine behavior, and these have been approved by the U.S. Food and Drug Administration (FDA) for clinical treatment.[10] However, current drug delivery systems still face many obstacles, mainly due to limited drug delivery efficiency and release efficiency and short lifetime, resulting in undesirable long-lasting therapeutic effects.[14]

As a derivative to mesoporous silica nanoparticles, silica aerogels are a unique class of mesoporous solid materials.[15] Silica aerogels have the potential to be excellent drug delivery carriers for high drug loads and efficient drug release due to such superior characteristics as extremely high specific surface area, wide internal surface area, high porosity, and ultralow density.[16,17] In particular, silica aerogels can improve the drug dispersion in pores and change the crystalline morphology of the drug to cope with the problem of the limited solubility of slightly soluble drugs during direct use, which leads to low bioavailability.[18,19] Meanwhile, silica aerogels can enhance the stability of these drugs, allowing the entire drug delivery system to remain relatively stable under environmental conditions for long hours.[3,20] However, traditional hydrophilic silica aerogels cannot release long-lasting and only apply to a local burst release because the solution quickly penetrates its open pores to destroy the drug-loaded structure.[21] Notably, surface-terminated silica aerogels prepared using hydrophobic groups to replace the hydroxyl groups on the surface of aerogels can solve the problem of liquid entering the internal pores to restrict the entry and exit of liquids, which causes a drop in the release efficiency to be replaced by long-lasting release and higher stability.[22,23] Therefore, surface-terminated silica aerogels would be promising antibacterial agent delivery carriers to provide not only high agent-loading efficiency but also long-lasting bacterial infection treatment effects. However, this has been rarely investigated so far.

There are two main ways to load drugs into silica aerogels when preparing a drug delivery system: (1) loading in solvent conditions during sol–gelation or aging[24] and (2) loading in supercritical fluids during or after drying.[25] Supercritical drug impregnation technology based on supercritical carbon dioxide (scCO2) fluids has attracted increasing attention in the pharmaceutical field because of its inherent characteristics such as low critical temperature, mild critical pressure, simple work-up, environment-friendliness, nontoxicity, and recyclability.[26,27] However, conventional supercritical drug impregnation technology lacks controlled regulation of the loaded drug form, limiting its application in drug delivery.[28] Crystallization from supercritical solutions (CSS) is a process for preparing tiny particles from supercritical carbon dioxide for slightly soluble drugs, which can be controlled by modulating the pressure conditions of the system to form smaller crystals in a single nucleation process with a high crystal formation rate.[29] As the reduction of drug particles to the nanoscale can increase the surface area of pure drug particles will obtain higher solubility and dissolution rates and protection against degradation,[30] the CSS method can find novel applications in the carriers of silica aerogels, which integrate highly efficient agent delivery and release efficiency and long-term therapeutic efficacy, to combat bacterial infections.

In this contribution, we used the CSS method to fabricate silica aerogels loaded with antibacterial agents and investigated their antibacterial performance. Three slightly soluble antibacterial agents, including cinnamaldehyde (CA, liquid), salicylic acid (SAA, solid), and sorbic acid (SOA, solid), were loaded in commercial hydrophobic silica aerogels by the CSS method in the medium of supercritical carbon dioxide to build an antibacterial agent delivery system. To demonstrate the mechanism of loading antibacterial agents, we attempted to reveal changes in the chemical properties and crystal structure of antibacterial agents during the agent-loading process through a series of physical and chemical characterizations such as scanning electron microscopy (SEM), transmission electron microscopy (TEM), Fourier transform infrared (FT-IR), X-ray diffraction (XRD), and thermogravimetric analysis (TGA). The antibacterial effects of agent-loaded aerogels were evaluated by agent release curve and in vitro antibacterial experiment. Moreover, the cytotoxicity of agent-loaded aerogels to normal cells was studied to examine its potential in clinical applications.

Results and Discussion

Morphology of Silica Aerogels

The basic physical and chemical properties of the commercial silica aerogels were investigated to examine their potential as antibacterial agent delivery carriers. Since the surface of the silica aerogels was methylated, the silica aerogels exhibited hydrophobic properties, which could be better applied to the loading of organic antibacterial agents. The microscopic morphology of silica aerogels observed by SEM and TEM showed a loose amorphous state with no obvious boundary features and a porous structure uniformly dispersed in the skeleton (Figure a,b). The nitrogen adsorption/desorption isotherm analysis further revealed the pore structure of silica aerogels, including the Brunauer–Emmett–Teller (BET) surface area and pore size distributions. As shown in Figure c, silica aerogels exhibited type IV isotherms and H3 hysteresis at P/P0 = 0.8–1.1 according to the IUPAC classification, demonstrating the existence of abundant mesopores. The specific surface area and pore volume of silica aerogels were as high as 714 m2 g–1 and 4.05 cm3 g–1, respectively. The pore size distribution plots based on the Barrett–Joyner–Halenda (BJH) model indicated that the diameter of mesopores generally was around 20 nm and the mean pore size was about 22.7 nm (Figure d). The pore structure information of silica aerogels, which was obtained by nitrogen adsorption/desorption isotherm analysis, was in good agreement with the results in TEM. In addition, the bulk density of silica aerogels measured by the true density analyzer was 3.62 g cm–3. These results indicate that silica aerogels have a uniform mesoporous structure and a large specific surface area, allowing silica aerogels to significantly improve the loading efficiency and controlled release of antibacterial agents as carriers.

Figure 1

(a) SEM image, (b)TEM image, (c) N2 adsorption–desorption isotherms, and (d) pore size distributions based on the BJH model of silica aerogels.

Characterization of Agent-Loaded Aerogels

Benefiting from the excellent drug delivery potential of these silica aerogels, three typical antibacterial agents that are trapped in low water solubility were selected to be loaded into silica aerogels. These antibacterial agents, CA (liquid),[31] SAA (solid),[32] and SOA (solid),[33] whose structures and properties are shown in Table , were selected to be loaded into silica aerogels via the CSS method under the action of scCO2 (50 °C, 15 MPa). Silica aerogels, after loading with antibacterial agents, all showed a powder form (Figure a). Among them, CA@SiO2 gave off a stronger aroma than CA and retained the yellow color characteristic of CA. In addition, the appearance and odor of SAA@SiO2 and SOA@SiO2 were not significantly different from those of the pure silica aerogels.

Table 1

Properties of Antibacterial Agentsab[43,44]

Water solubility is regulated by European Pharmacopoeia 6.0.

y = mole fraction solubility. Condition: T = 35 °C and P = 10 MPa.

Figure 2

(a) Physical appearance, (b) calcination curves, (c) XRD spectrum, and (d) FT-IR spectrum of silica aerogels, CA@SiO2, SAA@SiO2, and SOA@SiO2.

To examine the loading efficiency of agent-loaded aerogels, we investigated the calcination curves of silica aerogels and agent-loaded aerogels by a thermogravimetric analyzer (TGA) (Figure b). The sample was gradually heated to 1200 °C at a constant heating rate of 10 °C min–1. Silica aerogels would yield a total weight loss of ∼20 wt % at ∼200 and ∼500 °C due to the collapse of micropores. Among them, most CA and SOA in silica aerogels were lost rapidly at ∼230 and ∼150 °C, respectively. In addition, SAA was slowly lost during the heating process. According to the weight loss of agent-loaded aerogels during the heating process, the loading efficiencies of CA@SiO2, SAA@SiO2, and SOA@SiO2 were obtained as 56, 35, and 17 wt %, respectively. These results suggest that the loading efficiency was related to the structure of agents and the different solubility of agents in scCO2. The low loading efficiency of SAA@SiO2 might be caused by the poor solubility of SAA in scCO2 and the poor compatibility between the aromatic ring structure and siloxane chains.[22,34] Particularly, as the chemical structure of CA and SOA had a certain flexible aliphatic chain, it facilitated the formation of hydrogen bonds between the agents and the silica aerogels, thereby increasing the loading efficiency.[35]

The internal structure of the agent-loaded aerogels was analyzed by Fourier infrared spectrum (FT-IR) and X-ray diffraction (XRD) to determine whether the chemical properties of the antibacterial agents changed during the agent-loaded process, As illustrated in Figure d, silica aerogels had a strong absorption at 700–1200 cm–1, attributed to the stretching vibration of Si–O bond in its skeleton. The sharp peak at 1350 cm–1 was due to the stretching vibration of the C–O bond in the skeleton of the aerogels. In the FT-IR spectrum of CA@SiO2, the stretching vibration peak of the benzene ring at 1400–1600 cm–1, the stretching vibration peak of the C=C bond at 1600 cm–1, and the stretching vibration peak of the C=O bond at 1780 cm–1 could be detected. For CA@SiO2, the peaks at 1400–1600, 1600, and 1780 cm–1 were attributed to the stretching vibrations of the benzene ring, C=C bond, and C=O bond, respectively. For SAA@SiO2, the peak of the benzene ring and the C=O bond by stretching vibration could be identified and the broad peak at 3400 cm–1 was caused by the stretching vibration of the O–H bond. Similarly, the characteristic peaks of the original agents could be detected in the FT-IR spectrum of SOA@SiO2. This evidence indicated that the chemical properties of the agents were better preserved in the silica aerogels by the CSS method.

Figure c exhibits the XRD results of the crystal lattice structure of silica aerogels and agent-loaded aerogels. For the reference sample, silica aerogels showed an amorphous state without the crystal lattice; the solid antibacterial agents, including SAA and SOA, possessed the obvious crystal lattice; the liquid antibacterial agent CA clearly did not possess the crystal lattice. Notably, there were distinct diffraction peaks in the XRD spectrum of CA@SiO2, indicating that crystallization behavior occurred when CA was loaded onto the silica aerogels. We adjusted the pressure change in the system more gently by rate-limiting the decompression phase during the preparation. On this condition, the pressure changed in the system relatively significantly and the temperature changed relatively slowly, which might have led to the formation of smaller crystals during the one-time nucleation process by the CSS method.[26,36] Also, the rich mesoporous structure of silica aerogels provided sufficient sites for the formation of these crystals to provide higher crystal formation rates.[37] For both SAA@SiO2 and SOA@SiO2, the appearance of new diffraction peaks in their XRD spectrum accompanied by the disappearance of the original diffraction peaks suggested that the solid agent might also exhibit a crystallization behavior similar to that of the liquid agent during the supercritical process, which formed smaller crystals in the mesoporous structure of the silica aerogels.[38] Probably, this crystallization behavior would precipitate the antibacterial agents in the mesoporous structure of the silica aerogels in the form of nanocrystals, thereby affecting the agent release characteristics of the agent-loaded aerogels.

Considering the special crystallization behavior of agents in the mesoporous structure of silica aerogels, we performed TEM, SEM, and nitrogen adsorption/desorption isotherm analysis observations of agent-loaded aerogels to determine their microscopic morphology and pore structure further. The antibacterial agents were seen in high-resolution transmission electron microscopy (HRTEM) images of the agent-loaded aerogels as formed nanosized microcrystals with clearly visible lattice stripes (Figure a–c). Combined with the XRD spectrum of the agent-loaded aerogels, it was suggested that the antibacterial agents might be loaded into the pore structure of silica aerogels partially in the form of nanocrystals, which would affect the agent release pattern and the performance of the agent delivery system. By comparing the SEM image of CA@SiO2, SOA@SiO2, and silica aerogels (Figure d–f), it was found that CA and SOA had sufficiently penetrated and entered the pores of silica aerogels, turning the original loose structure into certainly lumpy, which indicated that CA@SiO2 and SOA@SiO2 were effectively and fully loaded on the silica aerogels. The structure of SAA@SiO2 still retained part of the sparsely porous structure. This indicates that the loading of SAA was not adequate, which was reflected in its low loading efficiency. The pore structure of agent-loaded aerogels was further revealed by the nitrogen adsorption/desorption isotherm analysis, as shown in Figure . The agent-loaded aerogels displayed type IV isotherms and H3 hysteresis similar to the original silica aerogels, indicating that the pore structure of the silica aerogels carrier did not change much during the loading process. Furthermore, their specific surface area, pore volume, and mean pore size based on the BJH model (Table ) decreased with the increase in loading efficiency, which suggests that the antibacterial agents penetrated into the pores of the silica aerogels by the supercritical process.

Figure 4

N2 adsorption–desorption isotherms of agent-loaded aerogels; (a) CA@SiO2, (b) SAA@SiO2, and (c) SOA@SiO2. The pore size distributions of agent-loaded aerogels; (d) CA@SiO2, (e) SAA@SiO2, and (f) SOA@SiO2.

Figure 3

TEM image of silica aerogels loaded with antibacterial agents: (a) CA@SiO2, (b) SAA@SiO2, and (c) SOA@SiO2 (inset in (a–c), the HRTEM images of CA@SiO2, SAA@SiO2, and SOA@SiO2, respectively). SEM image of silica aerogels loaded with antibacterial agents: (d) CA@SiO2, (e) SAA@SiO2, and (f) SOA@SiO2.

Table 2

Pore Structure of Silica Aerogels and Agent-Loaded Aerogels

simple	specific surface area (m2 g–1)	pore volume (cm3 g–1)	mean pore size (nm)
silica aerogels	714	4.05	16.5
CA@SiO2	22.7	5.53 × 10–2	7.22
SAA@SiO2	620	3.62	12.96
SOA@SiO2	342	1.84	9.43

In Vitro Release Kinetics

The in vitro release kinetics of the three-agent-loaded aerogels and the original agents were further investigated in vitro (Figure ). For the silica aerogels loaded with the liquid agent, the in vitro release kinetic curves were determined by the mass loss when slowly released in a constant environment at 37 °C. CA in the agent-loaded aerogels had a fast release rate in the early stage; however, as the release time increased, the release rate gradually slowed down. Notably, on the eighth day, ∼13% of the agents were still unreleased. For silica aerogels loaded with solid agents, the in vitro release kinetics curves were determined by analyzing the changes in absorbance by ultraviolet–visible (UV–vis) spectrophotometer, and these curves showed that the agents were slowly released in a dissolving medium at 37 °C (Figure S1). Silica aerogels carriers provided burst release for the solid agents in the early stage, and the release amount of SAA@SiO2 and SOA@SiO2 in the first 60 min reached 88 and 91%, respectively. It was worth noting that after the initial release of the agents in the agent-loaded aerogels, the release rate of the agents was significantly slowed down. Even after another 480 min, there was still some portion of the agents that had not been released. Additionally, silica aerogels exhibited good chemical stability in an aqueous buffer (Figure S2). The rapid and long-lasting agent release performance of the agent-loaded aerogels might be related to the crystallization behavior of antibacterial agents and the structural properties of silica aerogels. Since the hydrophobic surface of silica aerogels limited the penetration rate of liquid in the early release and affected the release efficiency of agent-loaded aerogels, the nanosized agent crystals significantly increased the effective surface area of the diffusion layer, which in turn increased the agent release rate. From the results, the acceleration effect provided by the nanocrystalline nature of the antibacterial agents dominated this process, thereby increasing the dissolution rate of the poorly soluble drugs. Meanwhile, due to the vast space in the pores of the silica aerogels to store the antibacterial agent and the local uneven penetration during release, the antibacterial agents continuously released from the pores of the silica aerogels come out.[26,38,39]

Figure 5

Release kinetics of (a) CA and CA@SiO2, (b) SAA and SAA@SiO2, and (c) SOA and SOA@SiO2.

In Vitro Antibacterial Study and Biocompatibility Test

Inspired by the advantage of in vitro release kinetics of agent-loaded aerogels, in vitro antibacterial experiments were further carried out using Escherichia coli as a bacterial model (Figure a,b). The antibacterial efficacy of the agent-loaded aerogels against E. coli at the initial release and after ∼90% agent release was evaluated by the conventional plate counting method for the rapid release and long-lasting release performance of the agent-loaded aerogels, respectively. First, in the absence of silica aerogels or agent-loaded aerogels treatment, E. coli grew and reproduced healthily. Second, when treated with silica aerogels for the first time, the colony-forming unit (CFU) of E. coli slightly decreased with the increase in its concentration, which indicated that silica aerogels would not significantly affect the normal growth of E. coli. Then, when the agent-loaded aerogels were used by treating E. coli, the CFU of E. coli was greatly reduced. The CFU of E. coli treated with 1000 ppm CA@SiO2 and SOA@SiO2 were reduced by 99.99%, which suggests that they had excellent antibacterial properties against E. coli. However, the antibacterial effect of SAA@SiO2 was not so good, which might be due to the low loading efficiency of SAA@SiO2. After that, to verify the long-lasting therapeutic effect of agent-loaded aerogels, we further conducted in vitro antibacterial experiments on agent-loaded aerogels, which had released ∼90% of the agents, at a concentration of 1000 ppm. The results indicated that both CA@SiO2 and SOA@SiO2 could still kill 95% of E. coli even under this condition, which suggests the long-term antibacterial performance for the agent-loaded silica aerogels. Compared with the existing antibacterial systems, CA@SiO2 showed efficient antibacterial application where it exhibited high loading efficiency for agents and a high killing rate for bacteria (Table S1). Taken together, silica aerogels loaded with antibacterial agents had a burst release efficiency to inactivate bacteria and a long-lasting antibacterial effect.

Figure 6

(a) Plate photographs and (b) survival of E. coli treated with silica aerogels, CA@SiO2, SAA@SiO2, and SOA@SiO2 at a concentration range of 0–1000 ppm and at a concentration of 1000 ppm after ∼90% of the agent has been released. (c) Cell viability of silica aerogels, CA@SiO2, SAA@SiO2, and SOA@SiO2 at 1000 ppm labeled of HUVEC after 24 h.

Based on the excellent antibacterial effect of agent-loaded aerogels, the biocompatibility of the agent-loaded aerogels was explored by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) method examined in human umbilical vein endothelial cell (HUVEC) (Figure c). It is noted that the cytotoxicity of silica aerogels to HUVEC was negligible. Among the three-agent-loaded aerogels, CA@SiO2 was slightly toxic to HUVEC due to its excessively high loading efficiency and initial release efficiency; SAA@SiO2 had negligible toxicity to HUVEC; the trace amount of SOA released by SOA@SiO2 also had the ability of cell proliferation to promote the proliferation of HUVEC.[40] Combined with the in vitro antibacterial experiment, agent-loaded aerogels had a prominent antibacterial performance and biocompatibility, suggesting they have excellent potentials for combating bacterial infections.

Conclusions

We demonstrate a simple agent delivery strategy with a high loading efficiency, rapid removal of bacteria, and long-lasting therapeutic effect. A series of slightly soluble antibacterial agents, including liquid CA and solid SAA, and SOA, were successfully loaded into hydrophobic silica aerogels via supercritical carbon dioxide fluid and used for bacterial ablation. Agent-loaded aerogels exhibited significantly enhanced agent delivery performance and retained their chemical properties by forming nanocrystals in the mesoporous structure of silica aerogels by the CSS method of antibacterial agents, which contributed to the fast release rate in the early stage and sustained long-lasting release in the later stage. The porous structure of silica aerogels provided a large number of sites for the crystallization of antibacterial agents under suitable supercritical process conditions, which improved the loading efficiency of CA@SiO2 to achieve 56 wt %. Importantly, CA@SiO2 exhibited excellent antibacterial properties at the beginning of its release, and even though it has been released, 90% of CA can inactivate 99.99% of E. coil in the in vitro antibacterial experiment. Moreover, the biocompatibility test indicated that silica aerogels had excellent biocompatibility and can be used as excellent carriers for antibacterial agents. In general, this work offers a facile and green pharmaceutical preparation engineering based on silica aerogels for various slightly soluble antibacterial agents, including solid and liquid, to achieve the advantages of high loading, long-lasting bacterial treatment, and excellent biocompatibility.

Experimental Section

Materials

Silica aerogels were purchased from Aerogel Technology Co., Ltd, China. Hydrochloric acid (HCl, 35%), cinnamaldehyde, salicylic acid, sorbic acid, sodium dodecyl sulfate, and phosphate-buffered saline (PBS) were obtained from Macklin. Carbon dioxide of 99.99% purity was supplied by Huatepeng Special Gas Co., Ltd., China, and used without further purification unless otherwise stated. Luria-Bertani (LB) broth and LB agar were from USB Co. E. coli was obtained from China General Microbiological Culture Collection Center. Human umbilical vein endothelial cells (HUVECs) were supplied by Kang Lang Biological Technology Co., Ltd., China.

Characterization

The morphology of silica aerogels and agent-loaded aerogels was analyzed using scanning electron microscopy (SEM, JEOL JSM-7800F) and transmission electron microscopy (TEM, JEOL JSM-6700F). The internal structure of silica aerogels was obtained by the Brunauer–Emmett–Teller (BET, Microtrac BELSORP-max) analysis. Fourier transform infrared (FT-IR) spectra were recorded on a Shimadzu IR Affinity-1 spectrophotometer using KBr tablets. The crystal structure of silica aerogels and agent-loaded aerogels was determined by X-ray diffraction (XRD, PANalytical B.V. Empyrean) analysis. The loading efficiency of the agent on the silica aerogels was determined using a thermogravimetric analyzer (TGA, NETZSCH STA409PC). UV–vis absorption (UV) spectra were recorded on a Shimadzu UV-2550 spectrophotometer.

Loading of Antibacterial Agents

The loading of antibacterial agents into silica aerogels using scCO2 was carried out in the apparatus as shown in Scheme . Silica aerogels were dried in a 333 K oven for 12 h to remove moisture before the loading experiment. After the dried silica aerogels were wrapped in filter paper, they were hung in an autoclave to avoid direct contact with the agents. Excessive antibacterial agents were placed in a container at the bottom of the autoclave to ensure that the antibacterial agent reached the solubility limit in scCO2. The autoclave was flushed using carbon dioxide for 2 min to remove air. The autoclave was heated to 50 °C and slowly pressurized to 15 MPa. The system was stored for 36 h under these conditions to ensure the adsorption equilibrium of the agents between the silica aerogels and scCO2. Afterward, the vessel was depressurized to ambient pressure by discharging carbon dioxide at a constant flow rate of 0.5 MPa s–1. The sample was taken out of the autoclave to offer the agent-loaded aerogels. The loading efficiency of the agents was calculated by the mass changes before and after the loading process.

Scheme 1

Schematic Diagram of the Experimental Setup for Loading Antibacterial Agents into Silica Aerogels Using scCO2

(a) CO2 cylinder; (b) purification filter; (c) CO2 cooling system; (d) manual gas pressurizing device; and (e) high-pressure vessel and heating control system.

In Vitro Release Kinetics of Antibacterial Agents

The release kinetic curve of CA@SiO2 was measured by the following method. A quantitative amount of CA@SiO2 was put into a filter bag and then placed in an environment without the influence of an ambient airflow at 37 °C to make sure it is released slowly. The amount of cinnamaldehyde released at each moment was calculated by measuring the sample mass loss at a predetermined time interval. The release kinetic curve of pure cinnamaldehyde was tested in the same method.

The release kinetic curve of SAA@SiO2 and SOA@SiO2 was measured by the following method. The agent-loaded aerogels were added to the dissolving medium, which was a 0.1 M HCl solution containing 1% sodium lauryl sulfate.[41,42] The agents were uniformly distributed in the container at a constant temperature of 37 (±0.5) °C and a stirring speed of 100 rpm. At predetermined time intervals, 2 mL of the sample solution was taken from the container each time, and 2 mL of fresh culture medium solution was added at the same time. Since the cumulative loss of the extracted agent had been calculated with the release rate, this method could maintain a constant dissolved volume of the agent during the release period and ignore the dilution effect. The calculated amount of the medicine released at each time included the sum of the amount of medicine released in the container each time and the cumulative loss of all of the samples before it. The contents of salicylic acid and sorbic acid in the samples were analyzed by UV–vis spectrophotometers with wavelengths of 305 and 257 nm, respectively. At each sampling, the absorbance data of the sample was read three times, and then the average value and variance were calculated. Within the absolute average relative deviation (AARD) of 3.5% concentration, the absorbance/concentration standard curve of each agent was further drawn to determine the unknown concentration of the sample at any time during the release process. The net released at any time was calculated by the formula of the previous researcher.[42] The release kinetic curve of pure salicylic acid and sorbic acid was tested in the same way.

In Vitro Antibacterial Experiments

The antibacterial performance before the release and after ∼90% release of CA@SiO2, SAA@SiO2, and SOA@SiO2 was evaluated via the colony-forming unit (CFU) counting method. Taking before the release of CA@SiO2 as an example, E. coli suspension (OD600 = 1.0) and silica aerogels with concentration gradient from 0 to 1000 ppm were added to a 96-well plate. Bacteria suspension without CA@SiO2 and with silica aerogels was used as the control group. After that, the bacterial suspension was diluted to 1 × 105 times with PBS and inoculated on an LB agar plate for 16 h of incubation. The diameter of the solid agar plates was 90 mm. The bacterial inhibition ratio (IR) was calculated under light and dark conditions according to the following equationwhere C is the CFU of the experimental group treated by CA@SiO2 and C0 is the CFU of the control group without incubation with CA@SiO2. The experimental conditions and operations of the remaining samples toward microbes were totally the same as that before the release of CA@SiO2.

Biocompatibility Test

First, human umbilical vein endothelial cells (HUVECs) were seeded on a 96-well plate with a density of 8000 cells/well and cultured for 24 h. Afterward, HUVECs were treated with silica aerogels, CA@SiO2, SAA@SiO2, and SOA@SiO2 at a concentration of 1000 ppm and hatched at 37 °C for 24 h. After gently washing each well with PBS three times, the culture medium containing MTT was joined and continued to be cultured at 37 °C for 3 h. After removing the MTT culture medium, dimethyl sulfoxide (DMSO) was joined to dissolve the formazan crystals. At last, the microplate was shaken for 5 min and the absorbance value at 450 nm was detected on the microplate reader to measure cell viability.