The antimicrobial candidate DD-S052P was produced by Bachem AG (Bubendorf, Switzerland) at the request of HLB Science Inc. using solid-phase peptide synthesis. The peptide was cleaved from the resin with full side-chain deprotection, followed by purification using preparative high-performance liquid chromatography (HPLC), ion exchange treatment, sterile microfiltration, and lyophilization. The final product had a purity of ≥95.0%. Its molecular weight, as determined by HPLC analysis, was 1,645 Da, and its isoelectric point (pI) was 12.59, indicating a highly cationic character.

In addition to DD-S052P (HLB Science), 11 conventional antimicrobial agents were included in broth microdilution tests to determine minimal inhibitory concentrations (MICs): meropenem (Yuhan Co., Seoul, Korea), ampicillin/sulbactam (Samjin Co., Seoul, Korea), ceftolozane/tazobactam (Pfizer, Madison, NJ, United States; imported by: Korea MSD, Seoul, Korea), colistin (SteriMax Inc., ON, Canada), rifampicin (CKD, Seoul, Korea), minocycline (Sigma-Aldrich, St. Louis, MO, United States), polymyxin B (Sigma-Aldrich, St Louis, MO, United States), tigecycline (Pfizer, Seoul, Korea), levofloxacin (Ildong Pharmaceutical Co., Ltd., Seoul, Korea), ceftazidime/avibactam (Pfizer, Seoul, Korea); and piperacillin/tazobactam (CKD, Seoul, Korea).

Clinical isolates of CRAB were collected anonymously from the blood cultures of adult patients (≥18 years of age) diagnosed with CRAB bacteremia at Korea University Anam Hospital (Seoul, Korea; IRB registration number: 2022AN0292; 20 June 2022), with a waiver of informed consent. Non-repetitive CRAB isolates were initially identified in a clinical microbiology laboratory using the MicroScan Pos Combo Panel Type 6 automated system (Baxter Diagnostics, West Sacramento, CA, United States) as part of routine diagnostic procedures. Strain identification was subsequently confirmed using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (Bruker Daltonics, Bremen, Germany).

Antibiotic susceptibility testing of 12 antimicrobial agents was performed using the broth microdilution method to determine their MICs. Each experiment was conducted in triplicate. Testing and interpretation of antibiotic susceptibility breakpoints were performed in accordance with the 2017 Clinical and Laboratory Standards Institute (CLSI) guidelines (Clinical and Laboratory Standards Institute, 2017). Carbapenem resistance was defined as an imipenem MIC of ≥8 mg/L (Clinical and Laboratory Standards Institute, 2017). Two reference bacterial strains, namely, E. coli ATCC® 25922 and Pseudomonas aeruginosa ATCC® 27853, were used as quality controls for antimicrobial susceptibility testing. Multiplex real-time polymerase chain reaction assays were performed to identify antimicrobial resistance genes in the CRAB isolates (Park et al., 2023).

A checkerboard assay was used to evaluate the synergistic activities of two antimicrobial agents. Fifteen CRAB clinical isolates were tested using 96-well plates (Cat. No. 30096; SPL Life Sciences, Pocheon, Korea). The panels of 96-well flat-bottom plates were prepared based on the MIC of each antimicrobial agent, as determined by the broth microdilution method. Dilution intervals were set at 2–8 times and 1/8–1/64 of the MIC values determined in the preliminary analysis. Antibiotic solutions were prepared in cation-adjusted Mueller–Hinton II broth (CA-MHB), with antibiotic-free wells in the lower-left corner of the 96-well plates. After preparing two-fold serial dilutions of both antimicrobial agents, 50 µL of the first antibiotic was dispensed into each row of the plate, and 50 µL of the second antibiotic was added to each column. Consequently, each well contained a 100 µL mixture of the two antimicrobial agents, with varying concentrations across the rows and columns.

The broth-cultured CRAB inoculum in CA-MHB was prepared at twice the desired concentration, based on the 0.5 McFarland standard (100 µL). The final volume in each well was 200 μL, yielding an inoculum concentration of 5 × 105 colony-forming units (CFU)/mL, except in the sterility control well. All plates were then incubated at 37 °C overnight in ambient air. Strains were initially cultured on blood agar plates at 37 °C overnight, followed by a secondary culture in CA-MHB for 2–6 h before use, according to the CLSI guidelines. Bacterial growth was measured using a microplate reader (SpectraMax Plus 384, Molecular Devices, United States) at an optical density of 600 nm (OD600). All assays were performed in triplicate. The fractional inhibitory concentration index (FICI) was calculated as previously described (Ju et al., 2022) and interpreted as follows: FICI ≤ 0.5, synergistic; 0.5 < FICI ≤ 1, additive; 1 < FICI ≤ 2, indifferent; and FICI > 2, antagonistic (Ju et al., 2022). For each well, the reported FICI corresponded to the value observed in at least two of three replicate experiments, and there were no instances in which all three values differed.

First, to screen for synergistic effects of various antibiotic combinations, a time-kill assay was performed using a randomly selected CRAB isolate. Subsequently, time-kill assays were conducted for the three combinations expected to produce reproducible synergistic effects across the remaining 14 CRAB isolates. In brief, tubes containing freshly prepared CA-MHB supplemented with antimicrobial agents, alone or in combination, were inoculated with CRAB isolates at a concentration of 5 × 10⁵ CFU/mL. The inoculum was prepared using the McFarland 0.5 standard (Schofield, 2012). Each tube containing 30 mL of suspension was incubated at 37 °C with shaking at 200 rpm under ambient conditions. Next, 100 μL aliquots were collected from each tube at specified time points (0, 2, 4, 8, 12, and 24 h of incubation) and serially diluted in saline to measure bacterial growth. Following a 1:10 serial dilution, 100 µL was plated onto blood agar for colony counting and incubated overnight at 37 °C. Colony counts were used to estimate the initial bacterial density of the original sample based on the dilution factor. The lower limit of detection was 1 log₁₀ CFU/mL, as defined by the detection capability of the method. The antibiotic concentrations for monotherapy were set at 1 × MIC for each antimicrobial agent. For combination therapy, concentrations of 0.5 × MIC, 1 × MIC, and 2 × MIC were used, as determined based on the individual MIC values of each antibiotic in the combination. For example, the 0.5 × MIC condition for the DD-S052P + SAM combination consisted of 0.5 × MIC of DD-S052P with 0.5 × MIC of SAM. The antibiotic combinations tested were identical to those used in the checkerboard assay.

For both single antimicrobial agents and their combinations, bactericidal activity was defined as ≥ 3 log₁₀ reduction in viable cell count within 24 h compared with that at the initial time point. A synergistic effect was defined as a ≥2 log₁₀ CFU/mL reduction within 24 h for the combination compared with the most active individual agent at different time points (Wences et al., 2022). An increase of >2 log₁₀ was considered indicative of antagonism (Pankey et al., 2014). Indifference was defined as any outcome that did not meet the criteria for synergy or antagonism (Pankey et al., 2014).

Immunocompetent specific pathogen-free CD-1 (Institute of Cancer Research; ICR) young male mice (average weight 35 ± 3 g; 8 weeks old) were supplied by Orient Bio Inc. (Seongnam, Korea). Prior to experimentation, mice were housed under specific-pathogen-free conditions for 1 week to allow acclimatization. All mice were maintained in the infectious animal facility of Korea University, which operates under ABSL-2 conditions.

A CRAB isolate was inoculated into MHB and incubated for 6 h, followed by cell harvesting via centrifugation. The supernatant was discarded, and the bacterial pellet was resuspended in saline to adjust the concentration to 2 × 10⁸ CFU/mL using a McFarland densitometer (DEN-1B) (Harris et al., 2013). For the CRAB challenge, a 500 µL suspension (yielding approximately 20% survival) was injected into the central abdomen of healthy mice using an insulin syringe at a 90° angle to a depth of approximately 4 mm. To evaluate the efficacy of each antimicrobial agent as monotherapy, the agents were administered 30 min after bacterial inoculation to allow for bacterial proliferation and spread (Kleinhenz et al., 2022). Assuming a mouse body weight of 40 g, seven antimicrobial agents were prepared at five concentrations each using two-fold serial dilutions. Each antibiotic solution was dissolved in saline, and a 200 µL solution was prepared for administration. Drugs were administered in 50 µL injections at four separate sites (Hoyle et al., 2020). Two concentrations of each drug were used: the lowest concentration (H) that achieved the highest survival rate as a single dose and one-quarter of that concentration (L). The same combinations tested in the checkerboard assays were used, with antibiotics prepared at twice the concentrations employed under single-agent conditions. A 100 µL injection of each drug, prepared at twice the final concentration, was administered, resulting in a total injection volume of 200 µL. In brief, combinations of drugs A and B—A(H)–B(H), A(H)–B(L), A(L)–B(H), and A(L)–B(L)—were prepared. The A(H)-B(H) combination was administered as four 50 µL injections (A at 2 × H, B at 2 × H, A at 2 × H, and B at 2 × H) at 30 min intervals, for a total of 200 µL. Saline was administered as a control following the same injection schedule. Following bacterial inoculation into the abdominal umbilical region of the mice, antibiotics were administered to all four abdominal quadrants: upper right, lower right, upper left, and lower left. The antibiotic injection sites were located approximately 1.5 cm from the bacterial injection site. The study endpoint was set at 48 h. Each experimental group included five mice, and each antibiotic combination set included five experimental groups: control, H–H, H–L, L–H, and L–L. A total of 15 antibiotic combination sets were tested, resulting in 75 experimental groups; the experiment was performed once. All animal experiments were approved by the Institutional Animal Care and Use Committee of Korea University (Approval No. KOREA-2025-0035) and conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. During the experiments, mice were maintained under standard laboratory conditions with ad libitum access to food and water.

The clinical and microbiological characteristics of the CRAB isolates included in this study are presented in Table 1. All patients from whom CRAB isolates were isolated from blood samples succumbed to CRAB infection. All 15 clinical isolates were resistant to meropenem, with an MIC of ≥16 mg/L. All isolates carried the OXA-51-type and OXA-23-type carbapenem-hydrolyzing oxacillinases but did not harbor class B metallo-carbapenemases or other class D carbapenemases. The susceptibility rates, MIC₅₀, MIC₉₀, and MIC range associated with each conventional antibiotic are summarized in Supplementary Table S1.

Checkerboard assays performed against the 15 CRAB isolates revealed that DD-S052P exhibited in vitro synergistic effects when combined with select conventional antibiotics (Table 2). Synergistic effects were confirmed when observed in at least two of three replicate experiments. Among DD-S052P-based combinations, DD-S052P plus meropenem exhibited the highest in vitro synergistic activity against CRAB isolates (100%), followed by DD-S052P plus minocycline (93%), DD-S052P plus ampicillin/sulbactam (93%), DD-S052P plus ceftolozane/tazobactam (80%), and DD-S052P plus colistin (80%). In contrast, DD-S052P plus tigecycline showed a considerably lower synergistic rate of 13% (Table 2). Overall, in the checkerboard assay, combination therapies excluding DD-S052P exhibited lower synergistic activity than DD-S052P-based combinations. Notably, tigecycline-based combinations did not demonstrate significant synergistic effects, whereas the colistin-tigecycline combination showed antagonism in 80% of the 15 CRAB isolates (Table 2). The representative raw OD₆₀₀ values from the checkerboard assay are provided in Supplementary Table S2. The FICI values calculated from these data are summarized in Supplementary Table S3.

To evaluate the time-dependent effects of the antibiotic combinations, a time-kill assay was performed on a randomly selected CRAB isolate (No. 3) using the same combination regimens tested in the checkerboard assay (Supplementary Table S4).

The bactericidal effects of antibiotic monotherapy (1 × MIC) and combination therapy (0.5 × MIC, 1 × MIC, and 2 × MIC) against the CRAB isolate were evaluated. Meropenem monotherapy exhibited bactericidal activity, whereas ceftolozane/tazobactam, tigecycline, and minocycline showed bacteriostatic effects. Notably, monotherapy with ampicillin/sulbactam, colistin, and DD-S052P did not exhibit the expected antibacterial activity at 24 h (1 × MIC; Supplementary Table S5). A ceiling effect was observed with meropenem monotherapy, precluding the assessment of its synergistic effects when combined with conventional antibiotics. No antagonism was observed in the time-kill assay. At an antibiotic concentration of 0.5 × MIC, DD-S052P combined with ampicillin/sulbactam, ceftolozane/tazobactam, tigecycline, and colistin exhibited both in vitro synergistic and bactericidal effects against the CRAB isolate at 24 h in the time-kill assay (Table 3). Among combination therapies excluding DD-S052P, colistin plus minocycline, colistin plus tigecycline, and tigecycline plus ceftolozane/tazobactam demonstrated both bactericidal and synergistic effects at 24 h (Table 3).

When combined with DD-S052P, ampicillin/sulbactam, ceftolozane/tazobactam, and colistin exhibited in vitro synergistic effects at 24 h in both the checkerboard and time-kill assays (0.5 × MIC). Furthermore, all combinations exhibited bactericidal effects against the CRAB isolates in the time-kill assay. When combined with DD-S052P, ampicillin/sulbactam, ceftolozane/tazobactam, tigecycline, and colistin also exhibited synergistic and bactericidal effects at all tested concentrations (0.5 × MIC, 1 × MIC, and 2 × MIC). Notably, DD-S052P combined with colistin, ampicillin/sulbactam, or tigecycline demonstrated bactericidal activity at all examined time points (Supplementary Figure S2). DD-S052P combined with colistin achieved bacterial killing at a lower colistin concentration (4 mg/L) than colistin alone (8 mg/L). At a minimal media concentration of 0.5 × MIC, the combination was bactericidal, with bacterial killing occurring up to 2 h faster than with colistin alone.

When combined with DD-S052P, ampicillin/sulbactam, ceftolozane/tazobactam, and colistin demonstrated synergistic effects against all 15 CRAB strains (1.0 × MIC). Excluding strains exhibiting a ceiling effect, bactericidal activity was observed in 86.7%, 86.7%, and 100% of isolates for DD-S052P combined with ampicillin/sulbactam, ceftolozane/tazobactam, and colistin, respectively (Table 4).

Table 5 summarizes the 48 h mortality rates in mouse models treated with various regimens. Antibiotic concentrations used in the animal experiments are provided in Supplementary Table S6. In CRAB-infected mice, low-dose administration of two-drug combinations resulted in 100% survival for the following combinations: DD-S052P plus minocycline, DD-S052P plus colistin, colistin plus ceftolozane/tazobactam, colistin plus minocycline, and tigecycline plus minocycline. Notably, the DD-S052P plus colistin combination was the only regimen that consistently demonstrated synergistic effects across the checkerboard, time-kill, and animal experiments. The Kaplan–Meier survival curve for the DD-S052P plus colistin combination is shown in Figure 1 as a representative curve. Survival curves for other experimental groups are presented in Supplementary Figure S3.