The cationic triblock copolymers, PEG-PGC20-PLLA20 (L2) and PEG-PGC20-PDLA20 (D2), were synthesized according to our previous protocols [22]. The scheme of monomer structure is shown in Supplementary Fig. 1.

Twenty clinically-isolated multidrug-resistant A. baumannii strains were gotten from blood and phlegm samples of the patients hospitalized in The First Affiliated Hospital of Medical College, Zhejiang University (Hangzhou, China). All isolates were identified and stored in 20% (v/v) glycerol at –80 ${}^{\circ }$C prior to use.

After grown overnight in Mueller-Hinton (MH) agar plate at 37 ${}^{\circ }$C, the A. baumannii isolates were harvested in mid-exponential growth phase. The MICs of the L/D2, imipenem and ceftazidime were tested using the broth microdilution method. In Brief, the three agents were diluted serially with MH broth (MHB) to various concentrations. The L/D2 was diluted serially from 256 to 0.5 mg/L, imipenem from 512 to 0.5 mg/L, and ceftazidime from 2048 to 8 mg/L. The bacterial suspension was diluted with 0.45% NaCl solution to adjust the turbidity approximately to the standard McFarland 0.5, which corresponds to the concentration of 1 $\times$ 10${}^8$ CFU/mL. Subsequently, bacterial suspension was further diluted 100-fold in MHB to a final concentration of 10${}^6$ CFU/mL. Equal volumes (100 $\mu$L) of microbial suspension and agent solution with varied concentrations were mixed in each well of a 96-well plate. Finally, the final concentration of bacteria in each well was 5 $\times$ 10${}^5$ CFU/mL and those of L/D2, imipenem and ceftazidime were 128 to 0.25 mg/L, 256 to 0.25 mg/L and 1024 to 4 mg/L, respectively. After incubated for 18 h at 37 ${}^{\circ }$C, MIC values were read with unaided eyes as the concentration of the agents, at which no microbial growth was observed. Broth containing microbial cells alone was considered as the negative control, and each test was performed in 3 replicates.

A. baumannii 10086 isolate was grown overnight in a MH agar plate at 37 ${}^{\circ }$C and then suspended in 0.45% NaCl solution to adjust the turbidity to that of a McFarland 0.5 standard (1 $\times$ 10${}^8$ CFU/mL). The bacterial suspension was further diluted 100-fold in MHB to a final concentration of 10${}^6$ CFU/mL. Various concentrations of L/D2, ceftazidime and imipenem were added to bacterial suspension to achieve final concentrations corresponding to 1 $\times$ MIC, 2 $\times$ MIC and 4 $\times$ MIC. Bacterial suspension without any treatment was used as the negative control. Samples were incubated at 37 ${}^{\circ }$C for 2 h. After each 10 min of incubation, samples were pipetted out and then diluted serially in 10-fold to various concentrations. The diluted bacterial solution (50 $\mu$L) was plated on MH agar plate and incubated at 37 ${}^{\circ }$C for 24 h. Subsequently, the number of viable colonies was counted. The experiments were performed in triplicate and the results are shown as mean l g (CFU/mL) $\pm$ SD.

Antibiofilm activity of the L/D2 was evaluated according the protocols reported previously [26]. A. baumannii 10086 biofilm was formed after 7 days of culture and treated with the peptide at 1 $\times$ MIC, 2 $\times$ MIC, 4 $\times$ MIC, and 8 $\times$ MIC. Viability of the cells in the biofilm and biomass were tested with or without the peptide treatment.

A. baumannii 10086 in mid-exponential growth phase was acquired as described in Section 2.3. The bacterial suspension was incubated with L/D2 at a concentration of 2 $\times$ MIC and 16 $\times$ MIC for 1 h. The bacterial suspension treated with PBS was used as the negative control. The solution was then centrifuged (5000 rpm, 10 min), and the supernatant was subsequently removed. PBS containing 2.5% glutaraldehyde was added to the bacterial suspension, and incubated overnight at 4 ${}^{\circ }$C for fixation. After washing with PBS three times, the bacterial suspension was subject to a post-fixing procedure for 1 h using 1% OsO${}_4$ in PBS. The fixed samples were then washed three times with PBS, followed by dehydration using a graded ethanol series. The samples were placed on copper tape, air-dried and coated with platinum prior to observation under a field emission SEM (JSM-7400F, JEOL, Tokyo, Japan).

Outer membrane permeability of L/D2 was tested by means of uptaking 1-N-phenylnaphthylamine (NPN). A. baumannii 10086 was suspended in PBS to a concentration of 2 $\times$ 10${}^9$ CFU/mL, and was incubated with NPN (8 $\mu$L from a 500 $\mu$M stock in acetone) for 30 min at 25 ${}^{\circ }$C. After transferred to cuvettes, fluorescence was tested using an F-2500 fluorescence spectrophotometer (Hitachi, Tokyo, Japan) at an excitation wavelength of 350 nm and an emission wavelength of 420 nm.

Similarly, L/D2 was added to the bacteria suspension to a concentration of 2 $\times$ MIC and 8 $\times$ MIC. The supernatant was harvested at 0.5, 2 and 6 h by means of centrifugating and filtering through a 0.22 $\mu$m filter. The release of LPS was tested using a chromogenic limulus amebocyte lysate assay (Xiamen Bioendo Technology, Co., Ltd., Xiamen, China).

A. baumannii 10086 were suspended in PBS to a concentration of 2 $\times$ 10${}^9$ CFU/mL. L/D2 was added to the bacteria suspension to a concentration of 1 $\times$ MIC, 2 $\times$ MIC, 4 $\times$ MIC, and 16 $\times$ MIC. After incubated at 37 ${}^{\circ }$C for 2 h, bacteria suspension was then filtered with a 0.22 $\mu$m filter to harvest the supernatant. The supernatant was subsequently tested for its absorbance using the Thermo Scientific NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) based on UV absorption at 260 nm. The data were normalized against the absorbance of supernatant of the untreated cells in PBS. Each assay was performed in triplicates.

A. baumannii 10086 was exposed to L/D2, ceftazidime, and imipenem for MIC determination as previously described. Bacteria were then harvested from wells of 0.5 $\times$ MIC, washed, and grown overnight in MH agar, and then subjected to MIC determination for up to 10 similar serial passages. The development of drug resistance in the bacteria was monitored by recording the changes in the MIC.

Checkerboard assays were used to evaluate the synergistic effect in vitro. As described in Section 2.3, the plates were set up with serial doubling dilutions of L/D2, imipenem or meropenem at various concentrations. After incubated for 18 h at 37 ${}^{\circ }$C, the synergistic or additive effect was evaluated by calculating the FIC indices (FICI) based on the formula: FICI = (MIC of A in combination/MIC of A) + (MIC of B in combination/MIC of B). The synergy or additive was defined based on the standard criteria (FICI $\le$ 0.5 was defined as synergistic; 0.5 FICI $\le$ 1 was defined as additive; 1 FICI $\le$ 4 was defined as indifference; FICI $>$ 4 was defined as antagonism).

ICR mice (female, 8 weeks old) were used in the following in vivo studies. Immunosuppression was induced by intraperitoneal injection of 200 mg cyclophosphamide (Hengrui Corp, Lianyungang, China) per kg of bodyweight 4 days before the injection of bacteria. All animal studies were conducted in accordance with protocols approved by the Animal Studies Committee, China.

The in vivo synergistic efficacy of L/D2 and imipenem was evaluated using a peritonitis mouse model. Overnight cultures of A. baumannii 10086 were harvested and suspended in PBS. Each of the cyclophosphamide-pretreated mice was injected intraperitoneally with 0.5 mL bacterial suspension at doses of 2.0 $\times$ 10${}^8$ CFU/mL. With this does, mice were diagnosed to be peritonitis at which numbers of bacterial colonies were found through method of anatomy. Mice were then randomly divided into four subgroups (PBS group, L/D2 group, imipenem group, L/D2 + imipenem group). Subsequently, L/D2 and imipenem were administered through intraperitoneally at 1h and 6 h after infection at designated dose (0.5 mg/kg, and 3.0 mg/kg for the L/D2 and imipenem respectively, 0.2 mL per injection, 15 mice per group). At 24 h post infection, five mice in each group were sacrificed to obtain peritoneal fluid, blood, and organ samples. In order to obtain peritoneal fluid sample, 3.0 mL of PBS was injected into the peritoneal cavity, and the abdomen area was gently massaged. After that, the abdomen of each mouse was opened, and the peritoneal fluid (2.0 mL per mouse) was recovered from the peritoneum. Peritoneal fluid and blood samples were diluted and plated on MH agar plates. Spleen, liver, and kidneys were then removed and homogenized in 2.0 mL of PBS. The homogenate was diluted and plated on MH agar plates. After overnight incubation at 37 ${}^{\circ }$C, the number of bacterial colonies was counted. The data are presented as l g (CFU/mL of homogenate). Survival of the rest of mice was followed up for 7 days for each treatment group.

To assess the toxicity of L/D2 toward the major organs in peritonitis mouse model, blood samples were extracted from the periorbital plexus of anesthetized mice at 72 h after the treatment of PBS and L/D2. Each mouse in the L/D2 group was injected intraperitoneally with L/D2 at designated dose (4 mg/kg of body weight, once daily i.p. injection for 3 consecutive days). Analysis of aspartate transaminase (AST), alanine transaminase (ALT), creatinine, ureanitrogen, sodiumions and potassium ions was made through blood biochemistry.

Analyses for difference between the control and treatment arms were conducted using one-way analysis of variance (ANOVA) and Student’s t-test. Statistically significant differences were set as a p value 0.05.

To assess the antimicrobial ability of L/D2, the MICs of L/D2 against 20 clinical isolates of MDR A. baumannii were tested firstly. Simultaneously, in order to compare with the first-line antimicrobial agents in clinic, the MIC values of ceftazidime and imipenem against these A. baumannii strains were also determined (Table 1). All of the A. baumannii strains could be inhibited by L/D2 and imipenem, with MICs ranging from 16 to 64 mg/L and 16 to 128 mg/L, respectively. The bactericidal effect of L/D2 was convincingly superior to ceftazidime, with 100% of strains inhibited at a much lower concentration. Moreover, against a particular MDR strain A. baumannii 10086, L/D2 exhibited superior efficacy with dramatical lower MIC values (16 mg/L) in comparison with both imipenem (128 mg/L) and ceftazidime (256 mg/L) (Table 2).

The time-killing curves of L/D2 were subsequently plotted against A. baumannii 10086, in comparison with antibiotics imipenem and ceftazidime. As shown in Fig. 1, L/D2 exhibited a more rapid bactericidal kinetics against A. baumannii 10086 than imipenem and ceftazidime. At 1 $\times$ MIC concentration, L/D2 could completely eradicate all bacteria within 40 min, while viable bacteria were still observed with imipenem or ceftazidime-treated samples within the same time frame. Importantly, even at higher concentrations of up to 4 $\times$ MIC, ceftazidime or imipenem couldn’t obviously eradicate A. baumannii 10086 within 1 h. Additionally, the killing kinetic of L/D2 against A. baumannii 10086 was also showed to be dose dependent. It was noted that the killing time was halved (40 to 20 min) when L/D2 concentration quadrupled from 1 $\times$ to 4 $\times$ MIC (Fig. 1D). Taken together, these results demonstrated that L/D2 was able to eradicate the A. baumannii rapidly and effectively.

Fig. 1.

In vitro killing kinetics of L/D2, ceftazidime, and imipenem. CFUs of A. baumannii 10086 from the different treatment arms were compared at varying concentrations (A) 1 $\times$ MIC, (B) 2 $\times$ MIC, and (C) 4 $\times$ MIC. (D) Effect of varying concentrations of L/D2 on killing kinetics. Error bars indicate respective standard deviations.

To study the antibiofilm activity of L/D2, A. baumannii 10086 biofilm was formed after 7 days of culture and treated with L/D2 for 24 h at different concentrations. As a result, the L/D2 showed a dose-dependent antibiofilm efficacy (Fig. 2). The viability of A. baumannii 10086 in the biofilms decreased to ~20%, and the amount of biomass reduced to ~32% after a single treatment at 8 $\times$ MIC. These results showed that L/D2 was able to effectively eradicate the bacteria in the biofilm and disperse the biofilm matrix.

Fig. 2.

Antibiofilm activity of L/D2. (A) Cell viability and (B) biomass of A. baumannii 10086 biofilm after L/D2 treatment for 24 h at various concentrations. NS, not significant. *p 0.05; **p 0.01; ***p 0.001.

In order to ascertain if membrane-lysis of action was operative for L/D2, A. baumannii 10086 was treated with L/D2 at concentration of 2 $\times$ MIC or 16 $\times$ MIC for 1 h while phosphate-buffered saline (PBS, pH 7.4) was utilized as a negative control. Pretreated samples were subsequently observed under scanning electron microscope (SEM). No obvious membrane lysis was seen after treatment of L/D2 at 2 $\times$ MIC, as compared with the smooth and integrated surface of control (Fig. 3A). However, after increasing the treatment concentration from 2 to 16 $\times$ MIC, membrane corrugation and lysis were observed on the surface of bacteria. As shown in Fig. 3B, a significant increase in the fluorescence intensity was observed 30 min post incubation, which indicate that outer membrane integrity was damaged by L/D2 in a time-dependent manner. Moreover, the LPS release was tested to further study the effect of L/D2 on the outer membrane of A. baumannii 10086. As a result, a significant increase of LPS was detected in the supernatant at 2 h post incubation at dose of 2 $\times$ MIC (Fig. 3C). In addition, a rapid release of LPS could also be detected when L/D2 concentration quadrupled from 2 $\times$ MIC to 8 $\times$ MIC. Subsequently, the bacterial membrane integrity was investigated after treatment with L/D2 by detecting leakage of cytoplasmic materials. As shown in Fig. 3D, the release of the cytoplasmic materials was L/D2 concentration-dependent, and it increased rapidly after L/D2 treatment at 4 $\times$ MIC. Taken together, these results convincingly indicated that L/D2 served membrane-disruption antibacterial mechanism.

Fig. 3.

Membrane-lytic mechanism of L/D2. (A) SEM images of A.baumannii 10086. (B) LPS levels in the supernatants was detected at 0.5, 2 and 6 h after 2 $\times$ MIC or 8 $\times$ MIC L/D2 treatment. (C) Outer membrane permeabilization of A. baumannii 10086 after L/D2 treatment was measured by detecting the fluorescence intensity of NPN. (D) Concentration of nucleic acids in the supernatants was detected after L/D2 treatment at doses of 1 $\times$ MIC, 2 $\times$ MIC, 4 $\times$ MIC and 16 $\times$ MIC. (E) In vitro evolution of antimicrobial resistance in A. baumannii 10086. NS, not significant. *p 0.05; **p 0.01; ***p 0.001.

As known commonly, prolonged repeated exposure of bacteria to non-lethal doses of antibiotic could induce acquisition of drug resistance. To determine if L/D2 could inhibit drug resistance development of MDR A. baumannii, the representative strain A. baumannii 10086 was serially passaged in the presence of L/D2, imipenem and ceftazidime at sub-lethal doses for 10 passages. As shown in Fig. 3E, MIC value of L/D2 stayed unchanged till the last 10 passage. Nevertheless, MIC value of imipenem started to increase by the 6th passage, and increased by 16 times. Astonishingly, by the 10 passage, MIC value of ceftazidime increased by 128 times.

With the promising bactericidal phenomenon in vitro presented above, we proceed to assess the synergistic effect of L/D2 and carbapenems by using the checkerboard method. All fractional inhibitory concentration (FIC) indices were $\le$1 for six representative A. baumannii strains, indicating that L/D2 served synergistic or additive effect with imipenem or meropenem (Fig. 4A). The Fig. 4B,C indicated that, the antimicrobial activity of imipenem and meropenem against A. baumannii 10086 was restored in the presence of sub-MIC L/D2, with the MIC value decreased from 128 to 16 mg/L. Combined usage of L/D2 and imipenem/meropenem exhibited dramatical bactericidal capacities against A. baumannii 10086 at concentration much lower than MIC when mono-administered (Fig. 4D, E). The synergistic antibacterial activity of L/D2 and imipenem/meropenem was further investigated in A. baumannii 10086-induced peritonitis mouse models. As shown in Fig. 5A, combined usage of L/D2 and imipenem was able to effectively remove bacteria from the peritoneal fluid, blood, and organs with $>$99.0% efficiency as compared with the control as well as L/D2/imipenem alone. Furthermore, combined usage of L/D2 and imipenem increased the survival rates from 0% for the control group and 20% for the imipenem or L/D2 group to 80% at 7 days post-treatment (Fig. 5B).

Fig. 4.

Synergistic effect between L/D2 and carbapenems in vitro. (A) The fractional inhibitory concentration index (FICI) of the L/D2/carbapenems combination against the various A. baumannii strain. (B) Microdilution checkerboard analysis of the combined effect of L/D2 and imipenem against A. baumannii 10086. (C) Microdilution checkerboard analysis of the combined effect of L/D2 and meropenem against A. baumannii 10086. (D) CFUs of A. baumannii 10086 after combined use of L/D2 and imipenem as well as its monotherapy. (E) CFUs of A. baumannii 10086 after combined use of L/D2 and meropenem as well as its monotherapy. *p 0.05; **p 0.01; ***p 0.001.

Fig. 5.

Synergistic effect between L/D2 and imipenem in vivo. (A) CFUs of A. baumannii 10086 in blood, peritoneal cavity, spleen, liver, and kidney at 24 h post infection. (B) Mice survival. Tracked for up to 7 days post infection. *p 0.05; **p 0.01; ***p 0.001.

To further evaluate whether the L/D2 might induce any side effect toward the major organs in the peritonitis mouse models, the liver and kidney functions, and the balance of electrolytes in blood were tested through blood biochemistry. As shown in Table 3, the levels of ALT, AST, urea nitrogen, creatinine, and sodium ion in the blood samples of the L/D2 treatment group exhibited no significantly changes compared with the control group treated with PBS. Taken these results together, it can be concluded that L/D2 showed synergistic effects with imipenem against MDR A. baumanniiin vitro and in vivo with negligible toxicity.