Introduction

The worldwide spread of carbapenem-resistant Klebsiella pneumoniae strains represents a major concern for public health [1]. Carbapenem resistance is mostly due to the production of carbapenemases that hydrolyze all β-lactam antibiotics including carbapenems, which are considered one of the last resorts for treating infections caused by Enterobacteriaceae [2,3,4]. Carbapenemases have been classified into three different molecular classes: Amber class A (such as KPC type), class B (such as VIM type), and class D (such as OXA-48 type). All of these are widely distributed among Enterobacteriaceae, particularly in K. pneumoniae strains associated with systemic infections [5, 6]. Moreover, strains harboring KPC-type genes of class A are often resistant to several other classes of drugs, including fluoroquinolones and aminoglycosides [6, 7]. Therefore, the therapeutic options to treat infections caused by such MDR strains are very limited, and high rates of mortality are recorded [8,9,10]. Among the few drugs that are still effective, colistin, tigecycline, gentamicin, and fosfomycin are often used in combination therapy [11, 12], yet K. pneumoniae strains resistant to these drugs have been recently isolated [13, 14]. It should be noted that development of novel antimicrobial agents against Gram-negative bacteria is also very limited [8, 15].

In this context, cationic antimicrobial peptides (AMP) have gained attention as possible new therapeutic candidates. A generally accepted mechanism of action of antimicrobial peptides involves electrostatic binding between the positively charged peptides and negatively charged structural elements of bacteria, e.g., lipopolysaccharide [16, 17]. Lactoferrin (LF) is a 77 kDa iron-binding multifunctional glycoprotein. It is present in the milk of humans and other mammals, in the blood plasma and neutrophil-specific granules, and is one of the major components of virtually all exocrine secretions of mammals, such as saliva, bile, tears, and pancreatic juice. Its functions include the regulation of innate immunity, iron transfer to cells, control of the level of free iron in blood, interaction with nucleic acids, heparin, and polysaccharides, and pronounced antimicrobial and antiviral activities [18, 19]. Importantly, lactoferrin is a source of peptides with antimicrobial activity. Human LF generates, by acid-pepsinolysis, the antimicrobial peptide lactoferricin H (residues 1 to 47), which contains two cationic domains (residues 2 to 5 and 28 to 31). A synthetic peptide comprising the first cationic domain of lactoferricin H, further referred to as hLF1–11, exerts effective antibacterial and antifungal activities, as demonstrated by in-vitro and in-vivo studies in systemic infections caused by fluconazole-resistant Candida albicans, multidrug-resistant Staphylococcus aureus, and Acinetobacter baumannii strains [20,21,22,23,24,25,26]. Moreover, hLF1–11 exerts modulatory effects on cells of the human immune system [27, 28], which is increasingly recognized as an important contribution to the clearance of infections [29,30,31].

Several AMP have shown synergistic activity with conventional antibacterial and antifungal agents, and combination therapies have also been proposed as a mean to reduce the frequency of emergence of resistant strains [32,33,34,35]. In the present study, we aimed at evaluating the antibacterial activity of hLF1–11 alone or in combination with various antibiotics against carbapenemase-producing K. pneumoniae strains harboring different resistance genes.

Materials and methods

Bacteria

Three K. pneumoniae strains were isolated from positive blood cultures of patients admitted to the Azienda Ospedaliero–Universitaria Pisana (Pisa, Italy) and selected on the basis of their antimicrobial susceptibility profile: a susceptible K. pneumoniae strain, a colistin-sensitive MDR strain, and a colistin-resistant (MIC 16 μg/ml) MDR K. pneumoniae strain. Their antimicrobial susceptibility profile was determined by the Vitek 2 system (bioMeriéux, l’Etoile, France) and confirmed by E-test (bioMeriéux) and Sensititre Aris System (Trek Diagnostic Systems, Thermo Fisher Scientific, OH, USA). The antimicrobials tested were the following: amoxicillin/clavulanic acid, ampicillin, carbapenems (imipenem, meropenem, ertapenem), piperacillin/tazobactam, cephalosporins (cefepime, cefotaxime, ceftazidime), aminoglycosides (amikacin, gentamicin, and tobramycin), fluoroquinolones (levofloxacin and ciprofloxacin), colistin, tigecycline, and trimethoprim/sulfamethoxazole. For both MDR K. pneumoniae strains, phenotypic and genotypic tests were performed to determine the type of carbapenemase. A combination disk test was used for the phenotypic characterization (Neosensitabs®, Rosco, Taastrup, Denmark). For genotypic characterization, after DNA extraction [36], the presence of bla KPC gene was determined using Hy-KPC real-time PCR, kindly provided by Hy-Labs (Hy-Labs, Israel), with the ABI Prism® 7500 instrument (Applied Biosystems, Foster City, CA, USA). Next, the bla KPC gene fragment (1010 bp) was amplified using KPC-fw/KPC-rv primers [37], purified after 1% agarose gel electrophoresis by 5Prime-Agarose GelExtract mini kit (5 Prime GmbH, Hilden, Germany), and sequenced using cycle sequencing technology (dideoxy chain termination/cycle sequencing) on ABI 3730XL sequencing apparatus (Eurofins Genomics Ebersberg, Germany). Sequencing results were analyzed by the sequence alignment Geneious software platform [38], and compared with the reference sequence from NC-021660 (Klebsiella pneumoniae FCF3SP plasmid, bla KPC type 2).

In addition, four multidrug resistant K. pneumoniae strains producing different carbapenemases (OXA-48, KPC-2, KPC-3, and VIM-1) and previously described [39] were used, and a K. pneumoniae ATCC® BAA-1705™ was included as a positive control for KPC production. Bacteria were cultured in Luria Bertani (LB) broth (Sigma–Aldrich, St. Louis, MO, USA) to mid-log phase; aliquots of this culture, containing 108 colony-forming unit (CFU)/ml, were supplemented with 20% (vol/vol) glycerol and stored at −80 °C. For antimicrobial susceptibility tests and synergy studies, each strain was cultured overnight in LB at 37 °C, and sub-cultured for 2 h on a rotary wheel at 37 °C.

Synthetic peptide and antibiotics

The synthetic peptide corresponding to residues 1–11 (GRRRRSVQWCA) of hLF, further referred to as hLF1–11, was purchased from Peptisyntha (Brussels, Belgium). The purity of this peptide exceeded 99%, as determined by reverse-phase high performance liquid chromatography (RP-HPLC). Peptide stocks at a concentration of 10 mM in 0.01% acetic acid (pH 3.7) were stored at −20 °C, and diluted to the desired concentration before use. The following antibiotics were tested in synergy studies: rifampicin, clarithromycin, clindamycin, gentamicin, and tigecycline, all purchased from Sigma–Aldrich (St. Louis, MO, USA). Rifampicin and tigecycline were dissolved in dimethyl sulfoxide (Fluka Chemie GmbH, Sigma–Aldrich Chemie BV, Zwijndrecht, The Netherlands) and stored at −80 °C. The final concentration of DMSO was <0.1%. The other antibiotics were dissolved in sterile distilled water and stored at −20 °C until use. Mueller Hinton (1:16 diluted) and NaPB (10 mM Na-phosphate buffer, pH 7.4) were freshly prepared for each experiment.

Antimicrobial peptide susceptibility assay

To assess the bactericidal activity of hLF1–11, K. pneumoniae cells were harvested in mid-log phase by centrifugation at 4,500 × g for 10 min, washed twice to completely remove traces amounts of culture medium (LB broth), and resuspended at a concentration of 107 CFU/ml in NaPB. Aliquots of this suspension were mixed with equal volumes of various concentrations of hLF1–11, prepared in NaPB, and incubated for 1 h at 37 °C. Thereafter, the number of viable bacterial cells was determined by plating serial dilutions of each sample on blood agar plates (Becton Dickinson & Co, BD; Milan, Italy).

Synergy studies

Synergy analyses of hLF1–11 and antimicrobial agents were carried out by a checkerboard titration method using 96-well round bottom polystyrene microtiter plates. This assay was performed in MH broth (MHB; Oxoid, Milan, Italy) diluted 1/16 in NaPB, since in preliminary experiments hLF1–11 showed no antibacterial activity in full strength medium (data not shown). In fact, similarly to what was observed with other antimicrobial peptides [40,41,42], the high ionic strength of MHB could possibly inhibit the interaction between the positive charges of hLF1–11 and the negatively charged bacterial surface. MIC values of each antibiotic were also assessed for comparison in undiluted MHB (Online supplementary materials, Table S1).

The ranges of concentration of the six antimicrobials tested were as follows: 0.125–32 μg/ml for rifampicin, 0.25–256 μg/ml for clarithromycin, 0.125–64 μg/ml for clindamycin, 0.015–16 μg/ml for gentamicin, and 0.06–16 μg/ml for tigecycline. The range of concentrations of hLF1–11 peptide was 2.7–88 μg/ml.

Briefly, the two-fold dilutions of each agent were set up in 100 μl of 1/8 strength Mueller–Hinton broth, and then an equal volume of the mid-log phase bacterial suspension in NaPB was inoculated into each well of the plate at a final concentration of approximately 5 × 105 CFU/ml. Sterility control wells, containing the medium alone, were included in each plate.

After 18–24 h incubation at 37 °C, the MIC of both the peptide and antibiotics were defined on the basis of the turbidity of the wells as the lowest concentration of the agent that produced the complete inhibition of visible growth. A variability of one dilution was considered acceptable to determine the MIC of hLF1–11 and antibiotics for each strain. The fractional inhibitory concentration (FIC) index for the combinations was calculated using the following formula: FIC index = (MIC drug A in combination)/(MIC drug A alone) + (MIC drug B in combination)/(MIC drug B alone). The FIC indices were interpreted as follows: ≤0.5, synergy, >0.5 to ≤4, indifference, and >4, antagonism [43]. FIC index was reported in this study as the mean of the lowest FIC indices of at least three independent experiments.

In addition, for antimicrobial combinations resulting in indifference by the checkerboard method, bactericidal kinetics synergy studies were performed at sub-inhibitory concentrations of each antimicrobial agent (1/2 MIC of hLF1–11, and 1/4 MIC of tigecycline), diluted in 1:16 MHB immediately before use. At 0, 1, 2, 3, 6, and 24 h, serial dilutions of cultures were plated onto solid media to determine the number of CFU per milliliter. Synergy was defined as a decrease in CFU/ml of ≥2 Log of the combination of hLF1–11 and drug, in comparison with its most active constituent [44]. All tests were performed in triplicate.

Hemolysis assay

A hemolysis assay was used to evaluate the cytotoxicity of the peptide and antibiotics alone and in combination. Briefly, blood from three healthy individuals was collected in vacuum tubes containing citrate (Becton Dickinson & Co, BD; Milan, Italy) as anticoagulant. Red blood cells (RBCs) were harvested by centrifugation at 1,600 × g for 5 min at room temperature, washed three times with phosphate-buffered saline (PBS, pH 7.4) and resuspended in PBS to a concentration of 8% (v/v). An aliquot (100 μl) of this suspension was transferred into each well of a 96-well microtiter plate and mixed with 100 μl of peptide or antibiotic solution at twice the desired concentration or 50 μl of 4× the peptide and antibiotic for synergy combinations.

After incubation for 1 h at 37 °C, the microtiter plate was centrifuged (1,600 × g, 5 min) and 100 μl of the supernatants were transferred to a flat-bottom 96-well plate for measurement of the hemoglobin release by reading the absorbance at 450 nm. Data were normalized between the 0% hemolysis of RBCs in PBS and 100% hemolysis of RBCs in 1% Triton X-100. The percentage of hemolysis was calculated by the following formula: (A peptide/antibiotic─ A PBS)/(A Triton X-100─A PBS) × 100%.

Statistical analysis

Data were expressed as means ± standard deviation of the mean (SD). Results obtained by the antimicrobial peptide susceptibility test and bactericidal kinetics synergy studies were analyzed by one-way ANOVA test after logarithmic transformation, using GraphPad Instat software (version 6.05 for Windows, La Jolla, CA, USA). Comparisons between the means of the untreated control and treated groups were made applying the Bonferroni post-hoc test. Comparisons between three strains for each peptide concentration were made applying the Tukey post-hoc test. The level of significance was set at a P value of <0.05.

Results

Antimicrobial peptide susceptibility assay

The antimicrobial activity of hLF1–11 was evaluated against three K. pneumoniae strains isolated from positive blood cultures. One strain was susceptible to all the tested antimicrobials with the exception of levofloxacin and ciprofloxacin, in addition to ampicillin to which K. pneumoniae is naturally resistant. Another strain, further referred to as 1R, showed a multidrug-resistant profile (including colistin resistance), and a third strain had a multidrug-resistant profile but was susceptible to colistin. The complete antimicrobial susceptibility profile of these K. pneumoniae strains is reported in Table S2 (Online supplementary materials). The characterization of the MDR strains by the combined disk method revealed that both K. pneumoniae strains were KPC-producing. Genotypic characterization by Hy-KPC real-time PCR indicated the presence of bla KPC gene in both the MDR K. pneumoniae strains. For the colistin-resistant 1R strain, which was further investigated in synergy studies, the variant of bla KPC gene was evaluated by sequencing, and the results revealed the presence of the variant type 3 of the bla KPC gene (Table 1).

Table 1 MIC values of the hLF1–11 peptide or antibiotics against K. pneumoniae strains used in this study
Full size table

In-vitro killing assays revealed that the hLF1–11-induced bactericidal activity was significantly (P < 0.05) higher against the non-carbapenemase-producing than both carbapenemase-producing K. pneumoniae strains (colistin-sensitive and colistin-resistant) at all the concentrations tested of hLF1–11 ≥ 88 μΜ, as shown in Fig. 1.

Fig. 1
figure 1

Antibacterial activity of hLF1–11 at different concentrations against three K. pneumoniae strains: a clinical susceptible K. pneumoniae strain (S, square), a colistin-resistant KPC K. pneumoniae strain (KPC R, circle, a), and a colistin-sensitive KPC K. pneumoniae strain (KPC S, triangle, b). Data are means ± SD of at least three independent experiments. K. pneumoniae cells (106 CFU/ml) were incubated with hLF1–11 for 1 h at 37 °C. The number of viable bacterial cells was determined by plating serial dilutions of each sample onto blood agar plates. * Significantly different (P < 0.05, ANOVA, Bonferroni test) as compared to untreated K. pneumoniae cells; # Significantly different (P < 0.05, ANOVA, Tukey test) from values obtained with the susceptible K. pneumoniae strain

Full size image

Synergistic effect of hLF1–11 with antibiotics

Synergy studies combining various concentrations of hLF1–11 with antibiotics were performed by the checkerboard method against the 1R K. pneumoniae strain and other strains harboring different bla KPC genes and producing different types of carbapenemases. The MIC values of hLF1–11 and the various antibiotics tested are reported in Table 1. All K. pneumoniae strains were inhibited by hLF1–11, with MIC values ranging from 22 to 88 μg/ml.

The results obtained by the combination of hLF1–11 with the various antibiotics, expressed as FIC index, are shown in Table 2. The results revealed that hLF1–11 showed synergism with all antibiotics against at least some K. pneumoniae strains. A synergistic effect was observed with rifampicin, clarithromycin, and clindamycin (hydrophobic antibiotics) against all tested strains. The FIC index ranged from 0.22 to 0.47 in the combination of hLF1–11 with rifampicin, from 0.15 to 0.5 and from 0.19 to 0.5 in the combinations with clarithromycin and clindamycin respectively. In the presence of hLF1–11, the MIC of rifampicin was reduced by 64-fold (from 8 μg/ml to 0.125 μg/ml) for OXA-48 and VIM-1 K. pneumoniae strains, by 32-fold (from 16 μg/ml to 0.5 μg/ml and from 8 μg/ml to 0.25 μg/ml, respectively) for K. pneumoniae ATCC® BAA-1705™ and KPC-3 (colistin-sensitive), and by 8- to 16-fold for KPC 1R and KPC-2 strains.

Table 2 Effect of the combination of the hLF1-11 peptide and antibiotics, expressed as FIC index, against carbapenemase-producing K. pneumoniae strains harboring different resistance genes
Full size table

In the peptide–clarithromycin combination, the MIC of the antibiotic was decreased from 4- to 64-fold, with the highest reductions observed against KPC-2 (from 32- to 64-fold), ATCC® BAA-1705™, and VIM-1 (both from 16- to 32-fold) strains. Among all strains, KPC 1R and KPC-3 showed the highest values of FIC indices.

In the peptide–clindamycin combination, the MIC of the antibiotic was decreased from 4- to 32-fold, with the highest reductions observed against VIM-1 (from 8- to 32-fold), and ATCC® BAA-1705™ (from 8- to 16-fold) strains. Among all strains, KPC 1R and KPC-3 showed the highest values of FIC indices.

Synergy was also obtained when hLF1–11 was tested in combination with gentamicin, as evidenced by FIC indices ranging from 0.21 to 0.5, and with tigecycline with FIC indices of 0.38–0.5. In contrast, no synergistic effect was observed with the hLF1–11 peptide-tigecycline combination against the KPC 1R strain and K. pneumoniae harboring the bla KPC-3 gene, with FIC indices of 0.84 and 0.75 respectively. Antibacterial kinetics experiments also showed no synergism in the peptide–tigecycline combination against the 1R strain at 24 h (Fig. 2), but revealed a synergistic effect at earlier time points. In fact, as shown in Fig. 2, the results of antibacterial kinetics experiments revealed a 3 Log CFU reduction at 6 h by the peptide–antibiotic combination in comparison to its most active constituent [44].

Fig. 2
figure 2

Kinetics of the antibacterial activity of hLF1–11 and tigecycline against KPC 1R K. pneumoniae strain. K. pneumoniae cells (106 CFU/ml) were incubated with 22 μg/ml hLF1–11 (diamond), 0.25 μg/ml tigecycline (circle), the combination of the same concentrations of hLF1–11 and tigecycline (square), or no treatment (triangle). The number of viable bacterial cells was determined at 0, 1, 2, 3, 6, and 24 h by plating serial dilutions of each sample onto blood agar plates. Data are means ± SD of at least three independent experiments. * Synergistic effect of the combination of hLF1–11 and tigecycline in comparison to its most active constituent (P < 0.05, ANOVA, Bonferroni test)

Full size image

Hemolysis assay

A hemolysis assay was performed to evaluate possible toxic effects of the hLF1–11 peptide and antibiotics on human red blood cells (Fig. 3). The results revealed that hLF1–11 alone had no hemolytic activity (<1%) even at 10× the highest MIC value (880 μg/ml). In addition, all the tested antibiotics alone showed no hemolysis (≤1%), with the exception of rifampicin, which exhibited 1–5% hemolysis at 8–32 μg/ml and 37% hemolysis at 10× the highest MIC value (320 μg/ml). All combinations of the peptide with antibiotics showed no hemolysis (<1%) even at 10× MIC for both constituents (Fig. 4), with the exception of the combination of hLF1–11 with rifampicin, which exhibited no hemolysis at MIC (<1%) but 3% hemolysis at 10× MIC.

Fig. 3
figure 3

Hemolytic activity of peptide and antibiotics alone at various MICs. The values above the bars indicate the tested concentrations (μg/ml). The antibiotic or the peptide was incubated with 8% RBC suspension. The results are expressed as mean of percentage of hemolysis. RBCs incubated with 1% Triton X-100 and PBS (untreated) were considered as 100% and 0% hemolysis respectively. The percentage of hemolysis was calculated as follows: (A peptide/antibiotic─ A PBS)/(A Triton X-100─A PBS) × 100% ± SD from three independent experiments. RIF, rifampicin; CLR, clarithromycin; CLI, clindamycin; GEN, gentamicin; TGC, tigecycline

Full size image
Fig. 4
figure 4

Hemolytic activity of the combination of the hLF1–11 peptide and antibiotics at MICs (open bars) and 10× MICs (gray bars). The values above the open bars indicate the tested concentrations (μg/ml). The tested concentrations are reported as x/y, where x and y represent the concentrations of the antibiotic and peptide respectively. The antibiotic and peptide were incubated with 8% RBC suspension. The results are expressed as mean of percentage of hemolysis. RBCs incubated with 1% Triton X-100 and PBS (untreated) were considered as 100% and 0% hemolysis respectively. The percentage of hemolysis was calculated as follows: (A peptide/antibiotic─ A PBS)/(A Triton X-100─A PBS) × 100% ± SD from three independent experiments. RIF, rifampicin; CLR, clarithromycin; CLI, clindamycin; GEN, gentamicin; TGC, tigecycline

Full size image

Discussion

Carbapenem-resistant K. pneumoniae strains are often resistant to multiple classes of antibiotics, including fluoroquinolones, aminoglycosides, and β-lactams [45] and, therefore, treatment of infections caused by such strains often meets with failure, and is associated with high mortality rates. Combination therapies have been shown to significantly improve the therapeutic efficacy in the treatment of bacterial infections [46], and there is evidence suggesting that emergence of antimicrobial resistance may be reduced by using combination therapy regimens [47, 48]. However, this is a controversial issue and some studies indicate that, unless highly inhibitory antibiotic doses are maintained until the pathogen is eradicated, combination therapy can have the opposite effect, i.e., to accelerate the emergence and spread of MDR bacterial strains [49, 50]. Another, recently proposed approach to combat infections caused by MDR bacterial strains consists in combining antibiotics and antimicrobial peptides [51]. These latter show broad-spectrum antimicrobial activity, frequently show strong synergism with conventional antibiotics [52], and have been suggested to help in preventing or delaying the emergence of antibiotic resistance [51, 52].

The present study aimed at evaluating the antibacterial activity of hLF1–11, alone or in combination with various antimicrobial drugs, against different strains of K. pneumoniae. The mechanism of action of cationic AMP is commonly related to their ability to interact with negatively charged components of cell membranes, and a correlation between colistin resistance and cross-resistance to host antimicrobial peptides has been reported, thus suggesting the existence of shared mechanisms of action for these compounds [53]. Therefore, the hLF1–11-induced bactericidal activity was evaluated against two MDR K. pneumoniae strains differing for colistin susceptibility and a K. pneumoniae strain susceptible to most of the antibiotics tested. The results revealed that the hLF1–11-induced bactericidal activity was significantly (P < 0.05) higher against a non-carbapenemase-producing K. pneumoniae than a colistin-sensitive and a colistin-resistant carbapenemase-producing K. pneumoniae strain in an in-vitro killing assay. This evidence indicates that resistance to colistin and to the hLF1–11 peptide are not strictly associated. Therefore, other distinct mechanisms of resistance to the hLF1–11 peptide and to colistin should also be considered [54].

The main conclusion from the present data is that hLF1–11 is a promising candidate for combination therapies with various antibiotics in the treatment of infections caused by MDR K. pneumoniae strains. This conclusion is based on the following findings. First, the hLF1–11 peptide exerted synergistic effects in combination with most of the antibiotics tested by the checkerboard assay. Rifampicin, clarithromycin, or clindamycin, which are hydrophobic antimicrobial drugs, are normally not able to permeate through the outer membrane of Gram-negative bacteria, thereby being ineffective against these microorganisms [55]. Indeed, hLF1–11 induced up to a 64-fold reduction in the MIC of these hydrophobic antibiotics and a 4- to 16-fold reduction in the MIC of gentamicin or tigecycline against the tested K. pneumoniae strains. These results suggest a hLF1–11-induced sensitizing effect on KPC K. pneumoniae strains to antibiotics. The mechanism underlying the synergistic effect between hLF1–11 and otherwise impermeable hydrophobic antibiotics might be the result of a transient loss of membrane potential induced by hLF1–11 and subsequent increase in cell membrane permeabilization. Further studies will help to elucidate the mechanism of action underlying these synergistic effects. Second, the highest FIC indices by the checkerboard assay, indicating indifference, were observed in the combinations of hLF1–11 and tigecycline against the two K. pneumoniae strains harboring the bla KPC-3 gene. However, antibacterial kinetics studies, performed against K. pneumoniae 1R showed a synergistic effect exerted by the combination of hLF1–11 and tigecycline at 6 h. In agreement with the results obtained by the checkerboard assay, no synergistic effect was observed at 24 h. Due to the chemical nature of the peptide, it is possible that hLF1–11 was progressively inactivated by bacterial cell components, such as proteases, released by dead cells after incubation with the antimicrobial peptide. A multiple daily administration regimen for the combination hLF1–11/tigecycline as well as for other antimicrobial peptide/antibiotic combinations might be considered in order to achieve complete eradication of bacteria. Chemical modification of the peptide enhancing its stability in biological fluids might also be attempted [56]. Alternatively, resumption of bacterial growth might be due to rapid evolutionary adaptation of K. pneumoniae persistence, though this phenomenon is usually associated with repeated antibiotic application rather than with a single-dose exposure [57]. Further studies will be needed to shed light on this issue. Third, all combinations of the peptide with antibiotics showed no hemolytic activity (<1%) even at 10× MIC, with the only exception of rifampicin, which exhibited 3% hemolysis at 10× MIC in combination with hLF1–11. The latter observation is not surprising, since the rifampicin-induced hemolytic activity by eryptosis has already been described [58]. These results indicate that hLF1–11 might be safe to be used in combination with antibiotics in the treatment of infections caused by KPC K. pneumoniae strains.

Overall, the results of the present study indicate that a combination therapy consisting of hLF1–11 and conventional antibiotics may be considered, and might be helpful as a last resort to treat infections sustained by antibiotic resistant bacterial strains. Such combination therapies represent a promising approach to treat infections caused by MDR K. pneumoniae strains for which conventional antibiotics are no longer effective and hydrophobic antibiotics are not indicated. Since the ionic strength in biological fluids may affect the antibacterial activity of hLF1–11, further studies will be needed to assess the in-vivo efficacy of hLF1–11, alone or in combination with antibiotics, against MDR K. pneumoniae strains, and to elucidate the mechanisms of action underlying the sensitizing effect of hLF1–11 to antibiotics.