Colistin and capsaicin were obtained from Selleck (Huston, TX, United States) and Tocris Bioscience (Bristol, United Kingdom), respectively. For in vitro studies, capsaicin was dissolved in dimethyl sulfoxide (DMSO; Sigma-Aldrich, St. Louis, MO, United States), and then diluted with culture medium to the desired concentrations. For in vivo studies, capsaicin was dissolved in ethanol, and then diluted with normal saline to the required concentration, with a final ethanol concentration of 0.2% (Zhang et al., 2018).

Two non-clonal MDR clinical A. baumannii strains Ab13 and Ab156 isolated at the Affiliated Hospital of Yangzhou University and the type strain ATCC 19606 were used in this study. All of these strains are sensitive to colistin. The distinct antibiotic resistance profiles of Ab13 and Ab156 are shown in Table 1. The minimum inhibitory concentrations (MICs) of colistin and capsaicin against A. baumannii were evaluated in triplicate, based on the standard broth microdilution method recommended by the Clinical and Laboratory Standards Institute (CLSI) (Matuschek et al., 2018). The criteria for the interpretation of the susceptibility tests were based on the CLSI guidelines.

A. baumannii strains Ab13, Ab156, and ATCC 19606 were subcultured in 5 mL of Mueller–Hinton broth (MHB; Oxoid Ltd., Cambridge, United Kingdom) containing colistin (Sigma-Aldrich) at concentrations ranging from 0 to 32 μg/mL for 24 h at 37°C with shaking at 200 rpm. Resistant mutants were selected from MH agar plates containing 10 μg/mL of colistin and were then stored as glycerol stocks at −80°C. The obtained colistin-resistant A. baumannii were designated Ab13-R, Ab156-R, and ATCC 19606-R, respectively.

Checkboard assays were carried out to test the MIC of the combination of colistin and capsaicin. In brief, increasing concentrations of capsaicin (0–256 μg/mL) in each column and increasing concentrations of colistin (0–64 μg/mL) in each row were set up in 96-well microtiter plates. Each well was inoculated with 100 μL of suspension containing 5 × 10⁵ CFU/mL of the test strain, and then the required concentration of drugs was added to a final volume of 200 μL. The plates were incubated at 37°C for 24 h. Data were obtained from at least three independent experiments.

The test strains were subjected to time-kill assays using colistin alone, capsaicin alone or colistin combination with capsaicin. Assays were performed in duplicate. Tubes containing MHB, with or without antibiotic, were inoculated with approximately 5 × 10⁶ CFU/mL log-phase inoculum to a final volume of 5 mL. Then, the tubes were cultured in an ambient atmosphere at 37°C with shaking at 200 rpm. The number of bacteria was determined at time points of 0, 2, 4, 6, 8, and 12 h by enumerating colonies in 10-fold serially-diluted specimens of 100 mL aliquots plated on Mueller–Hinton agar at 37°C. Synergism was defined as a >2 log10 decrease in CFU/mL after 12 h when the combination treatment was used compared with colistin or capsaicin alone.

Hemolytic activity of colistin in the presence of capsaicin was evaluated as follows. 8% sheep blood cells that prepared from fresh sterile defibrinated sheep blood, were treated with various concentrations of colistin (8–32 μg/mL) alone or in combination with capsaicin (8–128 μg/mL) at 37°C. After incubation for 1h, the absorption of supernatant was measured at 576 nm by Infinite M200 Microplate reader (Tecan). PBS (0.01M, pH = 7.4) was used as a negative control, and PBS with Triton X-100 (0.2%) was used as a positive control. Comparing the absorbance of the sample with the positive control is the hemolysis rate.

Cytotoxicity on mammalian cells was performed on A549 cells (adenocarcinomic human alveolar basal epithelial cells)by Cell Counting Kit-8 (CCK-8) assay according to the manufacturer’s instructions (Beyotime, Shanghai, China). A549 cells were seeded in a 96-well plate at about 2 × 10³ cells/well and cultured for 20 h. These cells were then treated with capsaicin (8–128 μg/mL) for 24 h, with three replicates for each concentration. After 24 h of culture, 10 μl of CCK-8 solution per well was added. Absorbance was measured at 450 nm to calculate the cytotoxicity of capsaicin.

For the in vivo experiments, the efficacy of colistin and capsaicin for the treatment of A. baumannii infection was determined using the murine bacteremia infection model. Four-week-old BALB/C male mice of approximately 20 g, purchased from the Experimental Animal Center of Yangzhou University (Yangzhou, China), were infected intraperitoneally (i.p.) with a washed overnight culture of A. baumannii ATCC19606-R (1 × 10⁷ CFU). After 24 h, the mice were randomly divided into four groups (n = 6 per group) and received daily treatment via i.p. injection: the control group received sterile PBS, the colistin group received 5 mg/kg colistin, the capsaicin group received 3 mg/kg capsaicin, and the combination group received 5 mg/kg colistin and 3 mg/kg capsaicin. Treatment lasted for 3 days. Twenty-4 hours after the last therapeutic dose, mice were euthanized and the tissues (e.g., spleen, kidney, liver, and lungs) were removed, weighed, homogenized, and serially diluted in PBS. The bacterial counts in the target tissues of each group were computed and presented as the mean (±SD) log10 CFU/g tissue or log10 CFU/mL blood.

Mice were handled following the Guidelines for the Care and Use of Laboratory Animals and procedures were approved by the Ethical Committee of Yangzhou University and the Laboratory Animal Management Committee of Yangzhou Province. All animals were acclimated to the controlled environment for 1 week prior to the experiment.

The effects of capsaicin on A. baumannii biofilm formation were analyzed using a 96-well tissue culture-treated polystyrene plate. The bacterial cell suspension (1 × 10⁶ CFU/mL) was cultured in wells supplemented with capsaicin (0–128 μg/mL). Negative controls containing MH only were included. The inoculated plates were incubated aerobically at 37°C for 24 h. Then, the planktonic bacteria and remaining capsaicin were discarded and the wells were washed twice with phosphate-buffered saline (PBS) and fixed with paraformaldehyde (4%) for 15 min. After removing the paraformaldehyde and staining with 0.4% (w/v) crystal violet (Sinopharm Chemical Reagent Co. Ltd., Shanghai, China) for 30 min, the cells were finally washed with ddH₂O until no color was visible in the control well. Aliquots of 95% ethanol were added to each well to dissolve the crystals. The average optical density values at 595 nm (OD₅₉₅) were calculated for the samples. The rate of biofilm inhibition was calculated via the following equation (Wei et al., 2017):

[1-(OD₅₉₅ of cells treated with test agent/OD₅₉₅ of untreated control)]× 100.

The combination of colistin (2 μg/mL) and capsaicin (32 μg/mL) or capsaicin (32 μg/mL) alone on the biofilm formation of colistin-resistant A. baumannii were done similar to the above study. Minimum biofilm inhibitory concentration (MBIC) is defined as the lowest concentration of an antimicrobial agent that can inhibit the formation of biofilms, and minimal biofilm eradication concentration (MBEC) is defined as the lowest concentration at which bacteria biofilms failed to regrow (Mi et al., 2016). The MBIC values of capsaicin on A. baumannii were done as the following steps: 100 μL of bacterial suspensions (1 × 10⁶ CFU/mL) were transferred to the wells of a flat-bottom 96-well microtiter plate. Capsaicin at different concentrations (0–512 μg/mL) were then added and the plates were incubated at 37°C for 24 h. Then, the consequent steps were done similar to the above study. The MBEC values of capsaicin on A. baumannii were the lowest concentration that can prevent visible growth in the biofilm recovery medium after 18 h incubation (Ceri et al., 1999; Lin et al., 2020).

Outer membrane permeability:1-N-Phenyl-naphthylamine (NPN) assay was used to assess the outer membrane permeability according to a previously study (Jangra et al., 2018). Due to NPN is a hydrophobic fluorescent probe, the fluorescence of NPN increases when it is transferred from aqueous to hydrophobic environment. The permeability of NPN were tested using 96 well plates (black with clear bottom). After treating the bacteria with different concentrations of capsaicin (0–128 μg/mL) or colistin (0–32 μg/mL) without or with capsaicin (32 μg/mL) for 1 h, 1-N-phenylnaphthylamine (NPN) was added with a final concentration of 10 μM, and the fluorescence intensity were measured with the excitation wavelength of 350 nm and emission wavelength of 420 nm.

Cell membrane integrity:Cell membrane integrity experiments were performed as previously described with some modifications (Jangra et al., 2019). Propidium iodide (PI, Thermo Fisher Scientific), a membrane-impermeable fluorescence dye, was used to detect the cell membrane integrity. After washing bacterial cells three times with normal saline (0.9% NaCl), the cell concentration was normalized to an OD₆₀₀ of 0.1. Then, treat the bacteria cells with different concentrations of capsaicin (0–128 μg/mL) or colistin (0–32 μg/mL) without or with capsaicin (32 μg/mL) for 1 h, the supernatants were collected and PI was added to a final concentration of 10 nM. The fluorescence intensity was measured with the excitation wavelength of 535 nm and emission wavelength of 615 nm.

Total ROS: The levels of reactive oxygen species (ROS) in A. baumannii were measured using 2′,7′–dichlorodihydrofluorescein diacetate (DCFH-DA) according to the manufacturer’s instructions (Beyotime, Shanghai, China). With the entering of DCFH-DA into A. baumannii cells, DCFH-DA was switched to a membrane-impermeable homologous by cellular esterases or oxidized to a fluorescent form by hydroxyl radical, superoxide, orhydrogen peroxide (H₂O₂) (Chen et al., 2020). Resistant strains were loaded with DCFH-DA at a final concentration of 10 μM, and then treated by different concentrations of capsaicin (0–128 μg/mL). Fluorescence intensity was measured with the excitation wavelength of 488 nm and emission wavelength of 525 nm.

Intracellular ATP: Intracellular ATP levels of A. baumannii were determined using an Enhanced ATP Assay Kit (Beyotime, Shanghai, China). The A. baumannii were cultured overnight and resuspended in PBS, then treated with different concentrations of capsaicin (0–128 μg/mL) for 1 h. After centrifugation, the bacterial precipitates were treated with lysozyme and the supernatants were collected for subsequent determination. The test solutions were added into the 96-well plates and incubated at room temperature for 5 min. The luminescence was detected by multi-function microplate reader (Biotek, United States).

ATCC19606-R were treated with colistin (2 μg/mL) or capsaicin (32 μg/mL)or their combination for 4 h and were fixed with 2.5% glutaraldehyde at 4°C for 24 h. Then the bacteria were gradual dehydration with ethanol (30, 50, 70, 80, 90, 95, and 100%). The processed samples were dried using a critical point dryer and coated with a layer of gold-palladium using an ion sprayer, and finally observed with SEM (GeminiSEM 300).

ATCC19606-R were grown in MHB to the early exponential phase, and treated with colistin (2 μg/mL) alone or in combination of capsaicin (32 μg/mL) for 4 h. After quality inspection of RNA samples, cDNA synthesis, cDNA fragmentation modification and sorting, and library amplification were processed, then, the constructed cDNA library was sequenced using Illumina HiSeqTM. The technical services of this process and subsequent bioinformatics analysis were provided by Nanjing Motif Zhigu Biotechnology Co., Ltd.

Data were presented as mean ± standard deviation (SD) of three replications. Mean values between groups were compared using the two-tailed Student’s t-test. All statistical analyses were performed using GraphPad Prism version 5.04 (GraphPad Software Inc., San Diego, CA, United States) with a p value < 0.05 considered significant.

The potency of capsaicin as an antibacterial was initially evaluated against colistin-resistant strains using standard broth dilution assays to determine MICs. The MICs of capsaicin were >512 μg/mL for all of the strains tested, revealing the lack of direct antimicrobial activity of capsaicin alone against the ATCC 19606-R, Ab13-R, and Ab156-R colistin-resistant strains.

To explore the interaction of capsaicin and colistin, a checkerboard assay was performed to determine the FIC indices for the laboratory-generated colistin-resistant A. baumannii. Checkerboard assays revealed the obvious dose-dependent synergistic activity of capsaicin with colistin on the ATCC 19606-R (FICI = 0.03), Ab13-R (FICI = 0.06), and Ab156-R (FICI = 0.04) colistin-resistant strains (Table 2). Capsaicin could decrease the colistin MICs to values lower than the susceptibility breakpoint (2 μg/mL) in most cases (Table 2). These results suggested that colistin-resistant A. baumannii strains have recovered or enhanced sensitivity to colistin when combined with different concentrations of capsaicin.

To further confirm the synergism of capsaicin with colistin, time-kill assays were carried out with the three colistin-resistant A. baumannii strains. The colistin concentration used corresponded to the clinical breakpoint for susceptibility (2 μg/mL), and the capsaicin concentration used for Ab13-R was 256 μg/mL, while that used for Ab156-R and ATCC19606-R was 32 μg/mL, which could inhibit the growth of most bacteria when combined with colistin (Table 2). As shown in Figure 1 A, B, and C, three colistin-resistant strains grew normally within 12 h of using the drug alone. Neither capsaicin nor colistin monotreatment killed colistin-resistant A. baumannii. However, the combination led to a reduction of bacterial load approximate >2 log10 (Figure 1). The above results showed that capsaicin indeed drastically enhanced the effect of colistin against colistin-resistant A. baumannii.

To evaluate the safety of capsaicin, we performed toxicity tests in red blood cells (RBCs) and mammalian cells. As shown in Figure 2 A, we found that capsaicin (≤128 μg/mL) had negligible effects on the hemolysis of colistin to red blood cells (RBCs). The results of CCK-8 assay shown that capsaicin (≤32 μg/mL) had no cytotoxicity to A549 cells (Figure 2B). 128 μg/mL capsaicin has a lower cell viability of 82% (Figure 2B), which is different with the results of hemolytic activity. Thus, the use of capsaicin is relatively safe under the concentrations of 128 μg/mL.

Lethal bacteremia is a common outcome of A. baumannii infection so to determine the potential use of combination therapy in clinical situations, we investigated the in vivo efficacy of capsaicin combined with colistin in a bacteremia model infected with colistin-resistant A. baumannii. In preliminary experiments, we found that, by 24 h, all organs in the mice were heavily infected with A. baumannii, suggesting that the infection was systemic. Furthermore, the combined treatment group showed reduced A. baumannii densities in the target tissues compared with the untreated or COL/CAP alone groups (Figure 3).

All of the strains used in this study have the ability to form biofilm. Considering the synergistic effects of capsaicin on colistin observed above, we hypothesized that capsaicin might inhibit biofilm formation of A. baumannii and this was tested using 8, 32 and 128 μg/mL of capsaicin. As shown in Figure 4, capsaicin could suppress the biofilm formation of all test strains. Interestingly, the amount of biofilms formed by colistin-resistant strain ATCC19606-R, Ab13-R, and Ab156-R were significantly increased compared with corresponding sensitive strains, suggesting that colistin resistance was positively correlated with biofilm formation in this strain. This provides insight into the potential coevolution of antimicrobial resistance and biofilm formation in Gram-negative pathogens. The MBIC values of capsaicin on ATCC19606, ATCC19606-R, Ab13 and Ab156 were 256 μg/mL. The MBIC values of capsaicin on Ab13-R and Ab156-R were 512 μg/mL. The MBEC values of capsaicin on ATCC19606, ATCC19606-R, Ab13 and Ab156 were 512 μg/mL, and the MBEC values of capsaicin on Ab13-R and Ab156-R were higher than 512 μg/mL. The MBIC and MBEC values of ATCC19606, ATCC19606-R, Ab13 and Ab156 were lower than Ab13-R and Ab156-R. This could be due to the differences in the ability to form biofilms of the bacterial strains tested.

We further measured the changes of cell membrane permeability after capsaicin treatment. NPN (1-N-phenylnaphthylamine) was used to study the outer membrane permeability of three resistant strains after capsaicin treatment. Due to the experiments calibrated in the same way, the fluorescence intensities showed in Figure 5 A and B were comparable. Figures 5A,B showed that the addition of capsaicin significantly increased outer membrane permeability in a concentration dependent manner. PI (propidium iodide) is often used to evaluate the integrity of inner membrane. As shown in Figures 5C,D, there was no significant changes in fluorescence intensity among the three resistant strains exposed to capsaicin alone or combination with colistin. Similarly, the fluorescence intensities showed in Figures 5C,D were also comparable. These results indicated that capsaicin destroyed the integrity of outer membrane, thus increasing the membrane-damaging ability of colistin, but had no effect on the integrity of inner membrane.

The effect of capsaicin on intracellular ATP level was determined by luciferase bioluminescence method. As shown in Figure 6A, capsaicin significantly decreased intracellular ATP level in a dose-dependent manner. We speculated that the decrease of intracellular ATP level is related to the increase bactericidal effect of colistin.

The membrane damage is related to the production of ROS and affects the content of ATP, and the bacteria will produce oxidative damage under antibiotic stress. Therefore, we speculate that capsaicin may increase oxidative damages. As shown in Figure 6B, capsaicin decreased the ROS levels of ATCC19606-R and Ab13-R, while in Ab156-R strain, capsaicin had no effect on ROS levels. It shows that capsaicin mainly plays its antioxidant activity here, though there are differences in different strains.

The resistance of colistin is mainly mediated by the decrease of affinity between colistin and LPS. Therefore, we speculate that capsaicin can restore the ability of colistin to destroy the outer membrane. In order to verify this hypothesis, scanning electron microscopy (SEM) was used to observe the morphological changes of ATCC19606-R after capsaicin treatment. As shown in Figure 7, the surface of cells treated colistin combination with capsaicin showed depression, shrinkage and even collapse and lysis compared with monotreatment, indicating that the bactericidal effect of capsaicin may be related to the rapid destruction of cell surface structure.

To gain a deeper understanding of the synergy mechanisms of capsaicin, we performed transcription analysis of ATCC19606-R after exposure to colistin or colistin-capsaicin combination for 4 h. The results showed that compared with colistin alone group, the expression of 271 genes in ATCC19606-R were up-regulated and 327 genes were down regulated after 4 h treatment with capsaicin and colistin (Figure 8A). KEGG enrichment analysis showed that the up-regulated genes were significantly enriched in ribosome synthesis (Figure 8B), while the down regulated genes were involved in microbial metabolism (Figure 8C). In addition, the 30S and 50S subunit synthesis-related genes were up-regulated under the combined treatment, which may be due to the inhibition of protein synthesis by capsaicin that was compensated by the up-regulation of ribosomal synthesis genes. The expression of multidrug efflux related genes were significantly down regulated, which indicated that the efflux pumps function of bacteria were inhibited under combined treatment. In addition, the expression of outer membrane protein related-genes also showed decreased, and these proteins were related to the ability of adhesion, invasion, colonization and proliferation (Figure 8D).