Activity of polymyxin B combined with cefepime-avibactam against the biofilms of polymyxin B-resistant Pseudomonas aeruginosa and Klebsiella pneumoniae in in vitro and in vivo models.

The ODc was 0.141, and all strains were identified as strong biofilm producers with the OD595 values ranging from 0.647 to 1.230, and strain P002 demonstrating the strongest biofilm formation ability (Fig.  1 A). Figure  1 B illustrates the OD595 values for all strains at various time points. A gradual increase in OD595 was observed in all strains up to 48 h. The peak OD595 values was reached at 48 h for all strains, followed by a subsequent decline. However, the OD595 values of the experimental strains remained higher than those of the control at 72 h. Fig. 1 A : The mean OD595 values of eight experimental strains. The bars represent standard deviations. B : Biofilm formation at different time points of eight experimental strains. The results are expressed as means± standard deviations A : The mean OD595 values of eight experimental strains. The bars represent standard deviations. B : Biofilm formation at different time points of eight experimental strains. The results are expressed as means± standard deviations Table  1 presents the MICs, MBICs, and MBECs of antibiotics. All strains exhibited resistance to PMB. The MICs of PMB ranged from 4 to 128 µg/mL. The MBIC and MBEC values of PMB and CFP-AVI for biofilms formed by the eight strains were significantly higher than the MIC values of the antibiotics for planktonic cells of all strain, with the MBECs tested exceeded 1,024 µg/mL for all antibiotics. Table 1 The minimal inhibitory concentrations (MIC S ), minimal biofilm inhibitory concentration (MBIC S ), and minimal biofilm eradication concentration (MBEC S ) of polymyxin B (PMB) and cefepime-avibactam (CFP-AVI) Strains PMB CFP-AVI (4:1)** MIC* MBIC MBEC MIC MBIC MBEC Klebsiella pneumoniae  K001 128 512 > 1024 8/2 512/128 > 1024/256  K002 32 256 > 1024 16/4 512/128 > 1024/256  K003 64 1024 > 1024 8/2 512/128 > 1024/256  K004 16 256 > 1024 8/2 128/32 > 1024/256  K005 16 64 > 1024 4/1 128/32 > 1024/256 Pseudomonas aeruginosa  P001 4 128 > 1024 8/2 512/128 > 1024/256  P002 4 > 1024 > 1024 8/2 1024/256 > 1024/256  P003 4 32 > 1024 2/0.5 512/128 > 1024/256 *, The breakpoint values defining the bacterial resistance to PMB is ≥ 4 (mg/L) **, The ratio of CFP to AVI is 4:1 in this study; and an MIC of 8/2 indicates the MIC of 8 µg/ml for CFP and 2 µg/ml for AVI for the combination formulation The minimal inhibitory concentrations (MIC S ), minimal biofilm inhibitory concentration (MBIC S ), and minimal biofilm eradication concentration (MBEC S ) of polymyxin B (PMB) and cefepime-avibactam (CFP-AVI) *, The breakpoint values defining the bacterial resistance to PMB is ≥ 4 (mg/L) **, The ratio of CFP to AVI is 4:1 in this study; and an MIC of 8/2 indicates the MIC of 8 µg/ml for CFP and 2 µg/ml for AVI for the combination formulation As shown in Fig.  2 A, the OD595 values of all strains were significantly reduced in the group treated with PMB-based combinations compared to those of the control and monotherapy groups ( P  < 0.05). Figure  2 B shows that the combination of PMB and CFP-AVI effectively eradicated preformed biofilms in all strains ( P  < 0.05). Fig. 2 Biofilm inhibitory ( A ) and eradication ( B ) effects of polymyxin B (PMB) and cefepime-avibactam (CFP-AVI). The results are expressed as the mean ± standard deviation. *, P  < 0.05, compared with all other groups Biofilm inhibitory ( A ) and eradication ( B ) effects of polymyxin B (PMB) and cefepime-avibactam (CFP-AVI). The results are expressed as the mean ± standard deviation. *, P  < 0.05, compared with all other groups In addition, CLSM revealed significant biofilm formation and eradication effects (Fig.  3 ) . The combination of PMB and CFP-AVI exhibited a notable inhibitory effect on biofilm formation, as indicated by the density of the reduced cell arrangement within the biofilm (Fig.  3 A&B). Furthermore, co-administration of both drugs effectively eliminated mature biofilms, as indicated by the significant reduction in biofilm thickness (Fig.  3 C&D). Fig. 3 CLSM images reveal the inhibitory ( A & B ) and eradication ( C & D ) effects of polymyxin B (PMB), cefepime- avibactam (CFP-AVI) on biofilm formation by Klebsiella pneumoniae K002 (A&C) and Pseudomonas aeruginosa P002 ( B & D ). Scale bar, 100 μm CLSM images reveal the inhibitory ( A & B ) and eradication ( C & D ) effects of polymyxin B (PMB), cefepime- avibactam (CFP-AVI) on biofilm formation by Klebsiella pneumoniae K002 (A&C) and Pseudomonas aeruginosa P002 ( B & D ). Scale bar, 100 μm PK/PD of PMB and CFP-AVI detected in the modified in vitro biofilm PK/PD model The targeted and measured concentrations of PMB and CFP-AVI in the in vitro biofilm PK/PD model are shown in Fig.  4 . The measured and targeted drug time curves were nearly identical (within ± 20% of the targeted value). Figure  5 shows the bacterial counts in the biofilms with or without antibiotics for all strains. After 48 h of pre-culturing the biofilm, the initial bacterial load of all strains on the sample coupons ranged from 8.17 ± 0.13 to 9.56 ± 0.12 log10 cfu/mL. PMB monotherapy did not effectively kill all strains after 72 h, as bacterial counts in biofilms remained similar to those in the control. In contrast, CFP-AVI monotherapy exhibited effective activity against most strains (Δlog10 cfu/mL = -2.560 to -1.245), except for P002. The combination of PMB and CFP-AVI demonstrated significant bactericidal activity against all biofilm-producing strains at 72 h (Δlog10 cfu/mL = -5.345 to -3.243). Compared to the most active single agent, the combination of PMB and CFP-AVI showed additivity (K003, Δlog10 cfu/mL = -1.621; K005, Δlog10 cfu/mL = -1.676) or synergy (in all strains except K003 and K005, with Δlog10 cfu/mL ranging from − 5.089 to -2.019). Fig. 4 The targeted and measured concentrations of polymyxin B (PMB) (A), cefepime (CFP) (B) andavibactam (AVI) (C) in the in vitro biofilm PK/PD model. The targeted and measured concentrations of polymyxin B (PMB) (A), cefepime (CFP) (B) andavibactam (AVI) (C) in the in vitro biofilm PK/PD model. Fig. 5 Bacterial killing by different treatments against biofilm-embedded bacteria of K. pneumoniae and P. aeruginosa in the in vitro biofilm PK/PD model after 72 h treatment Bacterial killing by different treatments against biofilm-embedded bacteria of K. pneumoniae and P. aeruginosa in the in vitro biofilm PK/PD model after 72 h treatment The LD 90 values for the strains tested in the larval infection model ranged from 10 7 to 10 9 cfu/mL for K. pneumoniae and from 10 3 to 10 4 cfu/mL for P. aeruginosa (Table  2 ). As shown in Fig.  6 , there was no signigfcant difference in the 7-day survival rates between larvae injected with PMB or CFP-AVI only and those in the control group ( P  > 0.05). The survival curves of the infected larvae with biofilms are shown in Fig.  7 . PMB monotherapy was not effective in rescuing biofilm-infected larvae for all strains, with a 7-day survival rate ranging from 0 to 6.7%. In the groups treated with CFP-AVI monotherapy, the 7-day survival rates for larvae infected with biofilm-forming strains were between 40.0% and 63.3%. However, in the groups receiving combined PMB and CFP-AVI treatment, the larvae survival rate increased between 46.7% and 90.0%. Compared to the CFP-AVI monotherapy groups, the combination treatment groups showed significantly increased survival rates ( P  < 0.05, n  = 6). Table 2 The lethal doses for 90% mortality (LD 90 ) of the eight strains of Klebsiella pneumoniae and Pseudomonas aeruginosa for Galleria mellonella larvae Strains LD 90 K. pneumoniae  K001 ~ 10 7 cfu/mL  K002 ~ 10 9 cfu/mL  K003 ~ 10 9 cfu/mL  K004 ~ 10 9 cfu/mL  K005 ~ 10 7 cfu/mL P. aeruginosa  P001 ~ 10 3 cfu/mL  P002 ~ 10 3 cfu/mL  P003 ~ 10 4 cfu/mL The lethal doses for 90% mortality (LD 90 ) of the eight strains of Klebsiella pneumoniae and Pseudomonas aeruginosa for Galleria mellonella larvae Fig. 6 Survival of G. mellonella larvae treated only with polymyxin B (PMB) and cefepime-avibactam (CFP-AVI) without infection. ns, no statistical differences, compared between normal saline and antibiotics Survival of G. mellonella larvae treated only with polymyxin B (PMB) and cefepime-avibactam (CFP-AVI) without infection. ns, no statistical differences, compared between normal saline and antibiotics Fig. 7 Survival of G. mellonella larvae infected with biofilm implants, with or without polymyxin B (PMB) and cefepime-avibactam (CFP-AVI) treatment. Experiment was conducted with 30 larvae per group. *, P  < 0.05, compared between CFP-AVI monotherapy and PMB combined with CFP-AVI; **, P  < 0.001, compared between positive controls and PMB combined with CFP-AVI; ns, no statistical differences, compared between CFP-AVI monotherapy and PMB combined with CFP-AVI Survival of G. mellonella larvae infected with biofilm implants, with or without polymyxin B (PMB) and cefepime-avibactam (CFP-AVI) treatment. Experiment was conducted with 30 larvae per group. *, P  < 0.05, compared between CFP-AVI monotherapy and PMB combined with CFP-AVI; **, P  < 0.001, compared between positive controls and PMB combined with CFP-AVI; ns, no statistical differences, compared between CFP-AVI monotherapy and PMB combined with CFP-AVI

The present study investigated the anti-biofilm effects of PMB combined with CFP-AVI on PMB-resistant biofilms using biofilm formation inhibition and eradication assays, modified in vitro biofilm PK/PD model and in vivo Galleria mellonella larvae biofilm implant infection model. Our results revealed, for the first time, that the combination of PMB and CFP-AVI exhibited a significant synergistic effect against PMB-resistant K. pneumoniae and P. aeruginosa biofilms, both in vitro and in vivo. In the antimicrobial susceptibility assays, PMB-resistant K. pneumoniae and P. aeruginosa biofilms demonstrated considerably higher resistance to PMB and CFP-AVI compared to planktonic bacterial cells. The MBICs and MBECs of all antibiotics were 4 to over 256 and 16 to over 1,024 times higher than their MICs, respectively, suggesting that conventional antibiotic concentrations are insufficient to eradicate the biofilms. The effective steady-state concentration of PMB in clinical practice is 2–4 mg/L [ 13 ]. In our biofilm formation inhibition and eradication assays, the combination of PMB with CFP-AVI demonstrated a synergistic anti-biofilm effect, with PMB concentrations (1/4 MIC or 1/2 MIC) in the combination ≤ 4 mg/L for three strains of P. aerugino sa. These findings further substantiate the potential clinical applicability of the combination. The biofilm development mainly includes four stages: initial adhesion of bacteria, formation of microcolonies, maturation of biofilm, and dispersion of biofilm. There are mainly two anti-biofilm strategies: inhibiting the formation of biofilm and eradicating mature biofilm. Sub-MIC PMB has an inhibitory effect on the formation of biofilm but could not eradicate the already formed biofilm. This may be due to the fact that in the early stages of biofilm formation, the biofilm structure is not yet mature, and bacteria have relatively weak resistance to antibiotics. Mature biofilms can enhance bacterial tolerance due to their structural integrity and the permeability barrier effect of extracellular polysaccharides. Previous studies have shown that sub-MIC doses of cefepime can upregulate the expression of biofilm-related genes and enhance biofilm by P. aeruginosa clinical isolates [ 40 ]. The present study also showed similar results, that is, sub-MIC CFP-AVI increased biofilm formation in monotherapy in the case of several P. aeruginosa. However, whether it also upregulates the expression of related genes needs to be further investigated. CFP-AVI is a new BL-BLI antibiotic with high susceptibility rates against Gram-negative bacteria, specifically targeting extended-spectrum β-lactamase-, AmpC-, KPC-, and OXA-48-producing Enterobacteriaceae , and multidrug-resistant P. aeruginosa [ 41 , 42 ]. Our previous static in vitro time-kill study demonstrated that CFP-AVI alone exhibited bacteriostatic activities against PMB-resistant P. aeruginosa and K. pneumoniae [ 25 ]. However, its efficacy against biofilm-related infections has not been well-established. In the present study, the modified in vitro PK/PD biofilm model showed that CFP-AVI had bacteriostatic or effective activities against the majority of biofilms formed by PMB-resistant P. aeruginosa and K. pneumoniae despite no obvious bactericidal effect. Additionally, in the Galleria mellonella larvae biofilm implant infection model, the survival rate of larvae treated with CFP-AVI ranged from 40.0 to 63.3%, which is a significant improvement compared to the positive control, where no larvae survived after 7 days. Our results suggest that CFP-AVI has potential value in treating biofilm-related infections. In contrast, one other study [ 43 ] found that the new BL-BLI ceftazidime-avibactam did not exhibit reliable anti-biofilm activity. Thus, further investigation is required to assess the anti-biofilm activities of BL-BLI antibiotics. It is acknowledged that a nylon toothbrush bristle segment was inserted in the last left proleg of each larva to allow biofilm formation but the effect of the antibiotic treatments on the in vivo biofilms was assessed without consideration of the bristle segment, which is a limitation of this study design. We will keep this issue in mind and test the biofilm on the implanted bristles in the future study, which will provide a more intuitive conclusion. In the present study, the LD 90 values were significantly different between K. pneumoniae (10 7 -10 9 cfu/mL) and P. aeruginosa (10 3 -10 4 cfu/mL), a finding consistent with other studies [ 44 ]. The high virulence of P. aeruginosa may be attributed to its secretion of various virulence factors, such as outer membrane proteins, alginate, flagellum, type IV pili, and protein secretion systems [ 45 ]. In our modified in vitro biofilm PK/PD model, PMB and CFP-AVI exhibited synergistic or additive effects when administered at clinical doses (PMB at a 2.5 mg/kg loading dose, followed by 1.5 mg/kg every 12 h; CFP-AVI at 2 g every 8 h through a 2-h infusion). However, the in vitro biofilm PK/PD model cannot fully simulate the in vivo environment to reflect the human immune response. Therefore, we applied the established Galleria mellonella larvae biofilm implant infection model as Galleria mellonella has both cellular and humoral defenses that are analogous to the human immune system, making it widely used in models of biofilm-related infections. These characteristics would compensate for the limitations of the in vitro dynamic PK/PD model. Previous studies found consistent results between larvae and mammals [ 46 , 47 ]. We observed that the PMB and CFP-AVI combination treatment significantly increased the survival rate in group, which confirmed the findings in the in vitro biofilm PK/PD model. It should be mentioned that although the humoral immunity of larvae is similar to that of humans, the adaptive immune system of larvae is different from that of mammals, and thus the results obtained in the present study need to be further verified in clinical practice. The synergistic effects produced by the combination may be based on the following two possible mechanisms. First, subpopulations in biofilms have different metabolic activities. PMB could kill the subpopulation exhibiting low metabolic activity in the inner layers of biofilm, whereas the β-lactam antibiotics could specifically kill the metabolically active subpopulation in biofilm because the activity of these antibiotics is highly dependent on the rate of bacterial growth [ 48 ]. Therefore, the rationale for the combination of these two types of drugs is based on the potential for subpopulation synergy, with each drug in combination targeting a different subpopulation. Second, it is well known that poor antibiotic penetration into biofilms plays an important role in biofilm resistance. PMB can bind to the lipopolysaccharides (LPS) of Gram-negative bacteria, which may disrupt the outer membrane and facilitate the penetration of other antibiotics in the biofilm [ 49 , 50 ]. In conclusion, the combination of PMB and CFP-AVI exhibits effective anti-biofilm activities against PMB-resistant K. pneumoniae and P. aeruginosa both in vitro and in vivo. This study provides valuable preclinical evidence supporting the use of PMB and CFP-AVI in combination as a promising strategic approach to overcome the challenges associated with treating biofilm-related infections.

Severe infections caused by multidrug-resistant Gram-negative bacteria (MDR-GNB) are on the rise globally and pose a significant threat to public health. 62% of the causative agents of septic shock include MDR-GNB bacterial strains [ 1 – 3 ]. Patients with severe infections often require the insertion of biocompatible materials, such as central venous catheters, closed thoracic drainage tubes, and tracheal intubation. Certain MDR-GNBs, including P. aeruginosa and K. pneumoniae , readily adhere to the surfaces of these biomaterials, forming complex biofilms and leading to biomaterial-related infections [ 4 – 7 ]. Compared to planktonic cells, bacteria within biofilms demonstrate stronger adhesion and fertility, with antibiotic resistance that is 10-1000 times higher, and thus can evade clearance by both the host immune system and antibiotics. Therefore, biofilm-related infections are characterized by low cure rates and high mortality, especially when caused by multidrug-resistant bacteria [ 8 – 12 ]. Polymyxin B (PMB), a polypeptide antibiotic known as the “last line of defense” against MDR-GNB, is widely used in clinical practice for severe infections of MDR-GNB [ 13 ]. PMB has been shown to inhibit the biofilm of PMB-sensitive MDR Acinetobacter baumannii [ 14 ]. However, with the widespread use of PMB, bacteria resistant to PMB have proliferated globally. For example, a retrospective hospital-based study indicated that the resistance rate to PMB in P. aeruginosa was 25.1% in Nepal [ 15 ], and the rate of polymyxin E-resistant K. pneumoniae increased from 36 to 50% from 2011 to 2015 in Italy [ 16 ]. In addition, it has been reported that 5-40% of carbapenem-resistant Enterobacteriaceae are also resistant to PMB [ 17 ]. Treating biofilm infections caused by PMB-resistant bacteria presents more significant challenges than those caused by PMB-sensitive bacteria. Some studies have demonstrated the effectiveness of colistin (a PMB congener) in combination with other antimicrobial agents, such as fosfomycin, gentamycin, mefloquine, and several new compounds, including PFK-158, ethylenediaminetetraacetic acid, or furanone C-30, against biofilms formed by colistin-resistant Gram-negative bacteria [ 18 – 23 ]. However, studies evaluating on PMB-based combinations with other antimicrobial agents for combatting PMB-resistant biofilms remains limited. An in vitro study reported that PMB, when used in combination with meropenem, was effective against the biofilm of PMB-resistant P. aeruginosa [ 24 ]. Our previous study demonstrated that PMB plus a new β-lactam/β-lactamase inhibitor (BL/BLI) combination restored PMB activity against PMB-resistant K. pneumoniae and P. aeruginosa [ 25 ]. Moreover, we further observed that among various BL/BLI combinations, CFP-AVI exhibited a highest synergistic rate (52.8% in the checkerboard assay), and obvious synergistic effects for all eight tested PMB-resistant clinical isolates [ 25 ]. These preliminary findings suggest that PMB plus CFP-AVI is potentially effective against the biofilm of PMB-resistant K. pneumoniae and P. aeruginosa . However, the impact of this combination on the biofilm has not yet been elucidated. Therefore, the present study was performed to evaluate the activities of PMB in combination with CFP-AVI against the biofilms formed by PMB-resistant K. pneumoniae and P. aeruginosa .

PMB-resistant clinical strains, comprising five strains of K. pneumoniae (K001-K005) and three strains of P. aeruginosa (P001-003), were obtained from patients at the Third Hospital of Changsha, Changsha, Chain and Sir Run Run Shaw Hospital, Zhejiang, China. These strains were identified using the automated Vitek 2 system (bioMérieux, Lyon, France). Reference strains, K. pneumoniae ATCC700603 and P. aeruginosa ATCC27853 were used for routine quality control. Reference substances of PMB (lot number: 130313–201310) and cefepime (lot number: 130524–202005) were acquired from the China National Institutes of Food and Drug Control. Avibactam, which was used as another reference substance, was generously provided by Nanjing Yoko Bio-Pharma Co., Ltd. (lot number: 1192491-61-4). These substances were used in antimicrobial susceptibility assays. For PK/PD experiments, the active pharmaceutical ingredients of PMB (no. A1411171, Xellia Pharmaceuticals), CFP (no. 0074MJ86D8, Nanjing Yoko Bio-Pharma Co., Ltd.), and AVI (no. 1192491-61-4, Nanjing Yoko Bio-Pharma Co., Ltd.) were used. All antimicrobial stock solutions were freshly prepared according to the Clinical and Laboratory Standards Institute (CLSI) guidelines and filtered using a 0.22 μm sterile filter membrane (Millex-GP 33 mm PES 0.22 μm Sterile; Millipore, USA) prior to use. Overnight cultures of the tested strains were adjusted to a 0.5 McFarland using sterile saline and then diluted 1:100 in fresh tryptic soy broth (TSB). A standardized inoculum of 200 µL, equivalent to 10 6 cfu/mL, was added to each well of a 96-well plate (flat bottom, Corning, USA) and incubated at 37 °C for various time points (6, 12, 24, 48, 56, and 72 h). After incubation, the wells were rinsed three times with sterilized water to remove planktonic bacteria, and then 200 µL of 1% crystal violet (AnYan, China) was added into each well for 15 min, followed by two washes with sterilized water. After air drying at room temperature, 200 µL of 95% ethanol (KESHI, China) was added into each well to solubilize the bound crystal violet. The formation of biofilm biomass was detected and quantified with a microplate reader at an optical density of 595 nm (OD595). Negative controls consisted of wells without any inoculum. The mean OD595 values of the biofilms at different time points were used to assess changes in biofilm biomass over time, while the mean OD595 values of the biofilms at 24 h were used to evaluate the biofilm-formation capacity of the strains. The optical density cut-off (ODc) was defined as three standard deviations above the mean OD of the negative controls. Accordingly, the strains were categorized according to their biofilm formation capacity as non-biofilm producers (OD ≤ ODc), weak biofilm producers (ODc < OD ≤ 2ODc), moderate biofilm producers (2 ODc  4 ODc). The minimal inhibitory concentrations (MICs) of PMB and CFP-AVI were determined by using the broth microdilution method, following the CLSI guidelines [ 26 ]. Antimicrobial susceptibility tests were conducted in 96-well microplates, with each well containing a final bacterial concentration of approximately 10 5 cfu/mL in Mueller-Hinton broth medium (Oxoid, Hampshire, England). All antibiotics were prepared using the serial two-fold dilution method, with the concentrations tested ranging from 0.25 to 128 µg/mL for PMB and 0.0625–256 µg/mL for CFP-AVI. The results were recorded after an incubation period of 18–22 h at 37 °C. Biofilm antimicrobial susceptibility tests were performed by using the Calgary Biofilm Device (Innovotech Inc. Canada), which consists of a standard 96-well plate and a plastic lid with 96 pegs [ 27 ]. Biofilms were formed on the pegs by culturing bacterial suspension with a final concentration of 10 6 cfu/mL at 37℃ for 24 h. After incubation, the lid was washed three times with sterilized water. The lid pegs were then transferred to a new sterile 96-well plate with TSB containing different concentrations of antibiotics (PMB or CFP-AVI alone) and incubated at 37 °C for 18–22 h. The minimal biofilm inhibitory concentration (MBIC) was identified as the lowest concentration of antibiotics that did not show visible growth or turbidity [ 28 ]. Subsequently, the lid was rinsed three times with sterilized water and placed in another sterile 96-well plate containing 200 µL of fresh TSB without any antibiotics, and then incubated at 37 °C for 24 h. The minimal biofilm eradication concentration (MBEC) was determined by visually identifying the lowest concentration of antibiotics with no observable bacterial growth [ 28 ]. Briefly, bacterial suspensions were prepared at a final concentration of approximately 10 6 cfu/mL in a 96-well plate with 1/4 MICs of PMB and CFP-AVI, alone or in combination. Negative controls consisted of wells containing the inoculum without any antibiotics. After incubation at 37 °C for 24 h, 1% crystal violet was added into each well for 15 min, and the formation of biofilm biomass was detected and quantified as described above. For biofilm eradication experiment, the inoculum was prepared in a 96-well plate as described above. Following a 24-h incubation at 37 °C, the unattached bacteria were removed by washing twice with sterilized water. The mature biofilms preformed were then exposed to 1/2 MIC of all antibiotics alone or in combination at 37 °C for another 24 h. Negative controls did not contain any antibiotics. The quantity of biofilm biomass was assessed as described above. Two representative strains, K. pneumoniae K002 and P. aeruginosa P002, were selected for confocal laser scanning microscopy (CLSM) to further observe the biofilm formation inhibition and eradication effects of PMB and CFP-AVI, alone or in combination. Briefly, biofilms of the two strains treated with the antibiotics in the above experiments were cultured in a 24-well plate (Jet Bio-Filtration Co., Ltd, China) with each well containing one cell climbing slice. Subsequently, each cell climbing slice was thoroughly rinsed with sterilized water and stained with the fluorescent dye 4′, 6′-diamidino-2-phenylindole for 15 min in the dark. Biofilms without the treatment (as negative controls). CLSM was used to observe the biofilms on the cell climbing slices. An in vitro biofilm PK/PD model was used to determine the activities of PMB in combination with CFP-AVI against the biofilms formed by PMB-resistant K. pneumoniae and P. aeruginosa . Briefly, during the biofilm culture phase, 1 mL of bacterial suspension (10 8 cfu/mL) was inoculated into a CDC biofilm reactor (BioSurface Technologies, Bozeman, MT), which contained eight rods with each rod housing three removable coupons where bacterial biofilms would form, followed by a 24-h static batch culture at 37ºC [ 29 ]. Subsequently, fresh sterile TSB was continuously pumped into the reactor at a constant flow rate, calculated by dividing the reactor volume (315 mL) by the residence time [ 30 ]. Upon the completion of the biofilm culture phase, the therapeutic phase began. For monotherapy, an antibiotic solution was introduced into the reactor using an infusion pump (B. Braun Medical (Shanghai) International Trading Co., Ltd.), while a peristaltic pump (Baoding Lead Fluid Technology Co., Ltd.) was used to pump fresh medium, simulating the antibiotic half-life. For combination therapy, we innovatively incorporated a compensation chamber to compensate for drugs with longer half-lives, deviating from conventional methods. This chamber facilitated the simulation of the simultaneous half-lives of both drugs, thereby aligning the concentration changes in the reactor more closely with the real concentration changes following combination therapy in the human body. A total of four regimens were evaluated on each of the eight clinical strains over a 72-h treatment period. These included: (i) PMB alone (2.5 mg/kg load, 1.5 mg/kg q12h, 2-h infusion) with an assumed f Cmax of 4.1 mg/L, 50% protein binding, and targeting an adult half-life of 11.9 h; (ii) CFP-AVI alone (2 g–0.5 g q8h, 2-h infusion), with CFP having an f Cmax of 68.73 mg/L, 16% protein binding, and targeting an average adult half-life of 1.98 h, and AVI has an f Cmax of 22.19 mg/L, 6.95% protein binding, and the same half-life; (iii) a combination of PMB (2.5 mg/kg load, 1.5 mg/kg q12h, 2-h infusion) and CFP-AVI (2 g–0.5 g q8h, 2-h infusion); and (iv) a drug-free group. The doses and PK parameters of PMB and CFP-AVI were determined based on the findings of our previous research, particularly in the context of severe infections in critically ill individuals [ 31 – 35 ]. PK samples (1 mL) of PMB and CFP-AVI were collected from the reactor exhaust port at various time points (0, 1, 2, 3, 4, 6, 8, 10, 12 h after the first dose, and before and after each subsequent dose). All samples were filtered through a 0.22 μm sterile filter membrane and stored at -80 °C until analysis. The concentrations of PMB and AVI were measured by liquid chromatography-tandem mass spectrometry as our previously described [ 36 , 37 ]. The linearity ranges for PMB1 and PMB2 were 0.1–10.0 mg/L (y = 1.74123x-0.390872, R² =0.9992) and 0.01-1.00 mg/L (y = 4.45587x + 0.00447435, R² =0.9991), respectively, with lower quantification limits of 0.10 and 0.01 mg/L. Intra-day and inter-day assay precisions for PMB1 and PMB2 ranged from 1.23 to 4.82% and 2.57–7.00%, respectively. For AVI, the linearity range was 0.1–25 mg/L (y = 0.322873x + 0.0435643, R² =0.9991), with a lower quantification limit of 0.08 mg/L and intra-day and inter-day assay precision ranging from 1.50 to 3.71%. CFP concentration was measured using high performance liquid chromatography with a linearity range of 0.5–120.0 mg/L (y = 0.114973x + 0.054603, R² =0.9992), a lower quantification limit of 0.50 mg/mL, and assay precision ranging from 0.21 to 5.04%. One rod was aseptically removed from the CDC biofilm reactor at 0, 8, 24, 32, 48, 56, and 72 h. The coupons were then rinsed twice with sterile saline to remove unbound bacteria. Each coupon was transferred into 10 mL of normal saline, and biofilm bacteria were recovered by sonication, followed by serial dilution in saline. A 100 µL aliquot of the bacterial suspensions was seeded in Mueller-Hinton agar plates using an automatic spiral plater (easySpiral; Interscience, Cantal, France). The bacterial colonies were counted after incubation at 37 °C for 18–22 h. Each experiment was performed three times, with a reliable detection limit of 50 cfu/mL. Time-kill curves were constructed by plotting log10 cfu/mL against time (h) to determine the anti-biofilm activity of the different dose regimens. Bactericidal and effective activities were, respectively, defined as a ≥ 3-log, and ≥ 1-log and < 3-log reduction in bacterial density at 72 h compared to the initial inoculum. Synergy and additivity in the treatment were defined as a ≥ 2-log reduction and a ≥ 1-log to < 2-log reduction, respectively, in bacterial density caused by combination therapy compared to the most active antimicrobial alone at 72 h [ 30 ]. A Galleria mellonella larvae infection model was also used to determine the activities of PMB in combination with CFP-AVI against the biofilms formed by PMB-resistant K. pneumoniae and P. aeruginosa . Galleria mellonella larvae were obtained from Huiyude Biological Technology (Tianjin, China), specifically selected for their weight, approximately 300 mg per larva. The lethal dose for 90% mortality (LD 90 ) was first determined for each strain in the larvae, and subsequently used as the initial inoculum for the Galleria mellonella larvae infection model. Then, the Galleria mellonella larvae implant biofilm infection model was established as previously described [ 38 ]. Briefly, commercially available hard nylon toothbrush bristles were cut into 1 cm segments and sterilized in an autoclave before use. A single bristle segment was carefully inserted into the last left proleg of each larva, while a 10 µL LD90 bacterial suspension was injected into the last right proleg using a Hamilton syringe. To evaluate the effect of PMB combined with CFP-AVI on the survival of the larvae with biofilm infection, three groups of larvae with biofilm infection were administered with the following three dose regimens 1 h after bristle insertion and infection: (i) a single dose of PMB (2.5 mg/kg), according to the loading dose in clinical settings, producing in vivo levels of 3.53 mg/L; (ii) a single dose of CFP-AVI (21 mg/g), based on a previous in vivo study [ 39 ], producing in vivo levels of 42.3 mg/L; and (iii) a combination of PMB and CFP-AVI (2.5 mg/kg + 21 mg/g). In addition, larvae with and those without biofilm infection received normal saline without any antibiotic treatment served as positive and negative controls, respectively. Each group included 30 larvae. The larvae were incubated at 37 °C, and their survival rates were observed every 24 h for 7 days. Thirty larvae were used in each experimental group. Phoenix WinNonlin 6.0 software (Pharsight Co. Ltd., Missouri, USA) was utilized for the PK analysis. Differences in the OD595 values between single-drug and combination groups in biofilm inhibition and eradication experiments were assessed using a one-way analysis of variance (ANOVA). A P value of less than 0.05 was considered statistically significant.