Luria Broth (LB) powder (containing 10 g peptone, 5 g yeast powder and 10 g NaCl) was purchased from Beijing Solarbio Science & Technology Co., Ltd and phosphate buffer saline (PBS) was purchased from Merck KGaA. LB medium was made by LB powder dissolved in 1L sterile PBS and then filtered by 0.22 um, frozen at 4℃ to prepare to use. 0.9% sterile physiological saline was purchased from Baxter Medical Products Co. Ltd and sodium pentobarbital was provided by the Medical Research Center of Beijing Chaoyang Hospital affiliated to Capital Medical University. Gentamicin ELISA kit was purchased from CUSABIO Bioengineering Co. Ltd (website: https://www.cusabio.com/food/Gentamicin-GEN-ELISA-kit-154817.html).
Methicillin sensitive Staphylococcus aureus (MSSA) standard strain and Pseudomonas aeruginosa (P. aeruginosa) were provided by the Guangdong Microbial Species Preservation Center (China, website: http://www.gimcc.net/database1.asp).
Female adult Sprague-Dawley (SD) rats (Charles River Laboratories, weighing 200±20 g) were kept in separate cages under a 12 h light/dark cycle at 23.6 ˚C and 35 % humidity. Animals were fed with sterilized chow diet and water. All procedures were complied with the ARRIVE guidelines and carried out in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The present study was approved by the Capital Medical University Ethics Committee on the use of animals in research and education. At the end of the experiments, rats were euthanized by using an excessive dose of sodium pentobarbital (100 mg/kg, intraperitoneal injection).
femur fractured with wound infection rat model
The specific process was described in our other accepted article (The acceptance letter of the article is in the supplementary materials). Briefly, we built MSSA and P. aeruginosa wound infection rat model with femur fractured. LB medium was used to dissolved bacterial strain powders. Bacteria suspension was cultured at 37˚C, 200 rpm/min overnight. The precipitate was acquired at 10,000 rpm/min, 3 min, then resuspended in LB medium. After 3 times bacterial passage performed, bacterial suspension was diluted 10-fold by sequential transfer of 100 ml into 900 ml PBS. 100 μl of different diluted bacterial suspensions were inoculated in different LB agar media at 37˚C, overnight. The number of colonies yielded between 20 to 200 were counted and multiplied by the corresponding dilution times to obtain the bacterial quantity of the bacterial suspension. Finally, bacterial quantity of the bacterial suspension was adjusted to 106, 107, 108 and 109 CFU/ml by LB medium.
The rat femur fracture model was established on a small animal operating table in a super clean bench. We refer to Husodo, K. et al  for the rat femur fracture model. Briefly, the rats were fasted for 12 h and banned from drinking water for 3 h before anesthesia. Isoflurane (concentration of 3%, oxygen flow 3L/min, small animal anesthesia machine) was used to induce anesthesia and sodium pentobarbital (0.4 ml/100 g, 40 mg/kg, intraperitoneal injection) was used to maintain anesthesia. The right hind limb of rats was shaved and disinfected by 75% alcohol 3 times. Right lateral position was adopted, the right hind limb was draped using a sterile surgical towel. The femoral lateral approach was used. The incision was 1-1.5 cm. Skin, subcutaneous tissue, superficial and deep fascia were opened, and middle part of the femur was revealed from the tensor fascia lata and femoral lateral muscle gap. A bone saw was used to make a femur shaft transverse fracture. Kirschner wire (1.0 mm) was used to fix the fracture as an intramedullary nail. The movement of the hip and knee, the contraposition and alignment of the fracture, as well as the fixation of the fracture, were examined. The method of causing bacterial infection was described as the previously study . Briefly, the bacterial suspension was administered to the rats by gradual dripping onto the surface of the muscle and embrocation with a sterile bacterial inoculation needle. The volume of the bacterial suspension was 0.5 ml at a concentration of 106, 107, 108 and 109 CFU/ml. The muscle the skin incision was sutured using a 3-0 silk suture. Rats were kept in separate cage with sterilized chow diet and water.
The mortality rate and wound tissue sections of rats were obtained following the establishment of the model to evaluate whether the model had succeeded and was stable. Finally, we chose 108 CFU/ml MSSA and 106 CFU/ml P. aeruginosa to establish the rat model. (related results are in the supplementary materials: DOI:10.17632/vpj7vjkz86.1).
Evaluation of the gentamicin-loading of sponge
Gentamicin powder was dissolved into 0.9% physiological saline and we adjusted the concentration of gentamicin were 40mg/ml, 16mg/ml, 8mg/ml, 4mg/ml, 1.6mg/ml, 0.8mg/ml or 0mg/ml. Gelatin sponge was cut into 1×1×0.5cm size and sterilized with ethylene oxide (Fig.1). According to the concentration of gentamicin solution, gentamicin sponges were divided into 7 groups. The sponge was immersed in gentamicin solution of different concentrations for 12h, 24h, 48h, 96h or 120h. Then, sponge was air dried until the quality no longer changed (about 48h). The weight of sponge was measured by electronic balance to evaluate the gentamicin-loading of sponge.
Preparation of air-dried sponge
We used the similar methods to prepare the air-dried sponge. Briefly, gentamicin powder was dissolved into 0.9% physiological saline and the concentration of gentamicin were adjust to 40mg/ml, 16mg/ml, 8mg/ml, 4mg/ml, 1.6mg/ml, 0.8mg/ml or 0mg/ml. Gelatin sponge was cut into 1×1×0.5cm size and sterilized with ethylene oxide. According to the concentration of gentamicin solution, gentamicin sponges were divided into 7 groups. The sponge was immersed in gentamicin solution of different concentrations for 48h. Then, sponge was air dried until the quality no longer changed (about 48h). Finally, sponges were sterilized with ethylene oxide and kept in -20℃ until used.
Evaluation of the release of air-dried gentamicin-saturated sponge
The air-dried sponge was immersed into 10ml 0.9% physiological saline 30min,2h,6h, 12h,24h,48h and 96h. Gentamicin ELISA kit was used to test the concentration of gentamicin in 10ml 0.9% physiological saline. The specific operation steps are in accordance with the manual.
Evaluation of the antibacterial effect of gentamicin sponge in vitro
After 3 times bacterial passage performed as described before, 100ul bacterial suspension was inoculated in LB agar media at 37˚C, overnight to make the surface of medium be full of colonies. The air-dried sponge was put on the center of the surface of the LB agar media at 37˚C, 48h. Then, the sponge was removed and the antibacterial area (area that bacteria do not grow) was measured to evaluate the antibacterial effect of gentamicin sponge.
Evaluation of the effects of gentamicin sponge on femur fractured with wound infection rat model
70 rats were divided into 2 groups as MSSA infection (35 rats) and P. aeruginosa infection (35 rats) group. Then, each group was divided into 7 subgroups as: 40mg/ml air-dried sponge (5 rats), 16mg/ml air-dried sponge (5 rats), 8mg/ml air-dried sponge (5 rats), 4mg/ml air-dried sponge (5 rats), 1.6mg/ml air-dried sponge (5 rats), 0.8mg/ml air-dried sponge (5 rats) and 0mg/ml air-dried sponge (5 rats). As described before, we chose 107 CFU/ml MRSA and 106 CFU/ml P. aeruginosa to establish the femur fractured with wound infection rat model. After the bacterial suspension was administered to the rats, air-dried sponge was put in the bacteria inoculated muscle gap. Then, the muscle the skin incision was sutured using a 3-0 silk suture. Rats were kept in separate cage with sterilized chow diet and water. We recorded the mortality and wound suppuration rate of the rats at day 7.
SPSS 23.0 was used to analyze the data. Measurement data were reported as median (interquartile), Kruskal-Wallis H test was used to compare multiple groups and Mann-Whitney U test was used to compare two groups. Enumeration data were as a percentage, in addition, χ2 test and Fisher definite probability methods was used to compare groups. P<0.05 was considered as statistically significant.
The equations of gentamicin-loading of sponge and release of air-dried gentamicin-saturated sponge was constructed by curve estimation of SPSS software. R2 was used to evaluate the matching degree of the constructed equation. We used the median of each group at each timepoint to construct the surface estimation in Origin2019b software. The maximum number of iterations was set at 400, tolerance was 1e-9 and iterative algorithm was levenberg-Marquardt algorithm. Surface function was set to Poly2D and the coefficients of the surface equation are expressed by mean± standard deviation. R2 was also used to evaluate the matching degree of the constructed surface equation. Microsoft Excel 2016 was used to make the data visualization of wound suppuration rate of rats.