Evaluation of Patterns of Susceptibility of Clinical Isolates of Acinetobacter baumannii Towards Polymyxins and Different Cationic Antimicrobials

1. Introduction

The Acinetobacter genus is diverse, comprising oxidase-positive and -negative, nonpigmented, Gram-negative coccobacilli. With over 50 species identified [1] most are environmental and non-pathogenic. The species most frequently linked to infections are A. baumannii, followed by A. calcoaceticus and A. lwoffii [2], then other species including A. haemolyticus, A. johnsonii, A. junii, and A. ursingii [3,4,5,6,7,8,9]. A. baumannii is the most virulent species among them [10].

Infections caused by Acinetobacter spp. became more widespread in the 1960s and 1970s, coinciding with the growing use of intensive care techniques [11,12], mechanical ventilation, central venous and urinary catheterization, and associated antibacterial therapies [13,14,15,16]. Analysis of data from 2018-2020 found Acinetobacter spp. accounted for 1.2% of all healthcare-associated infections in USA [17]. There is an estimated 1 million cases globally each year [18]. Similar infection rates are observed in ICUs across Europe and Latin America [19,20,21,22,23]. In some Asian and Latin American countries, Acinetobacter ranks among the top three causes of bacteraemia and nosocomial pneumonia [24,25,26,27,28].

Acinetobacter baumannii is ranked as a top priority pathogen of critical importance by the World Health Organisation for which new antimicrobials are urgently required [29]. It has an increasing resistance to all traditional antibiotics, with up to 90% of strains being resistant to cephalosporins, carbapenem, fluoroquinolones, and β-lactams [30]. Furthermore, multidrug resistance has been reported in up to 99.8% of carbapenem-resistant A. baumannii [31,32]. Infection with carbapenem-resistant strains results in higher in-hospital mortality [17]. Polymyxins are becoming more widely used to treat Acinetobacter infections due to this increase in resistance to traditional antibiotics. However, resistance to polymyxins is also increasing, [33,34] with 21% of strains being resistant to colistin in one study [30]. Acinetobacter baumannii can become resistant to polymyxins via a variety of mechanisms including addition of phosphoethanolamine to its outer membrane, loss of lipopolysaccharide from its outer membrane, loss of outer membrane proteins such as OmpA, and expression of efflux pumps [35,36].

Due to the potential for cross contamination, disinfection of hospital surfaces and appliances is common. Cationic disinfectants such as chlorhexidine (CHX), polyhexamethylene biguanide (PHMB) and polyquaternium-1 (PQ-1) are widely used for disinfection purposes in health care settings [37,38,39]. However, excessive or improper use of disinfectants can contribute to bacterial resistance and promote cross-resistance to antibiotics, thereby worsening the global antimicrobial resistance crisis [40]. Common disinfectants such as CHX may trigger adaptive bacterial mechanisms such as efflux pump activation, biofilm formation, membrane alterations, stress-response pathways, and horizontal gene transfer [41]. These adaptations can increase bacterial tolerance not only to disinfectants but also to multiple antibiotics, leading to multidrug-resistant pathogens [41]. There is the potential for resistance to cationic disinfectants to affect the susceptibility to polymyxins [42]. Cationic antimicrobial peptides are possible alternative new antibiotics or disinfecting agents that are being developed. They typically consist of short chains of amino acids and their net positive charge at physiological pH is usually due to the presence of lysine or arginine [43]. This positive charge is critical for their interaction with microbial cell membranes, which are often negatively charged [44]. The naturally occurring human cathelicidin LL-37 and the chimeric AMP melimine have been studied for their antimicrobial activity [45,46,47]. Use of LL-37 can improve polymicrobial wound infection in clinical trials [48,49], and a modified version of melimine, Mel4, successfully completed a Phase III clinical trial as an antimicrobial contact lens coating [50]. The changes that occur in membranes to develop polymyxins and cationic disinfectants, can also affect sensitivity to antimicrobial peptides [51], although this is not always the case [52,53].

This research was designed to determine the susceptibility pattern of clinical isolates of Acinetobacter baumannii against disinfectants, polymyxins and cationic antimicrobial peptides, and to evaluate whether resistance to disinfectants correlated to resistance to polymyxins, and whether resistance to polymyxins was correlated with resistance to antimicrobial peptides. The hypotheses being tested were (1) resistance to disinfectants would correlate with resistance to polymyxin antibiotics, and (2) resistance to polymyxin antibiotics would correlate with resistance to naturally occurring and synthetic antimicrobial peptides.

2. Results

2.1. Conformation of Strains Identity

All strains were confirmed as Acinetobacter baumannii (Table 1). Apart from strains ATCC 19606, ATCC 19606 (R), Ab08, Ab39 and Ab61, other strains were not selected on the basis of their resistance to polymyxins or any of the other antimicrobials being studied.

2.2. Minimum Inhibitory Concentration of Cationic Antimicrobials

The results demonstrate diversified patterns of susceptibility toward different tested cationic antimicrobials. Except four strains, (ATCC 19606R, Ab08, Ab39 and Ab 61) all strains were susceptible to polymyxin B (1-2 μg/ml) and colistin (0.25-2 μg/ml). The MICs to polymyxin B for the resistant strains ranged from 16 μg/ml for Ab08 to 1250 μg/ml for ATCC 19606 (R) (Table 2). For the disinfectants, PQ-1 was the most effective with a median MIC of 18.75 μg/ml, ranging from 6.3–50 μg/ml (Table 2). The median MICs for PHMB and CHX were the same, at 25 μg/ml, although there was a slightly greater range for PHMB of 3.13 to 50 μg/ml (Table 2).

For the antimicrobial peptides, LL-37 gave the most consistently low MICs, with a median of 62.5 μg/ml, ranging from 31.3 to 125 μg/ml (Table 2). This was followed by Mel4 and melimine, with median MICs of 125 μg/ml, but with a slightly greater range from 15.6 to 250 μg/ml for Mel4 (Table 2). Lactoferricin was virtually inactive against the strains tested, with MICs >500 μg/ml (Table 2).

Overall, Australian and Pakistani isolates had very similar median MICs to each antimicrobial, with the exception of Mel4, where the median MIC for Pakistani isolates was 250 μg/ml compared to 62.5 μg/ml for Australian isolates.

When the MICs to Mel4, Melimine, LL-37, PQ-1, PMHB, CHX of the polymyxin resistance isolates (ATCC 19606 (R), Ab08, Ab39, Ab61) were compared with the polymyxin susceptible strains, there were markedly lower MICs to PQ-1 (median 6.3 vs. 25; p = 0.01) and PHMB (median 3.1 vs. 25; p = 0.003), with a trend for the same difference occurring with CHX (15.6 vs. 25; p = 0.08) suggesting collateral sensitivity to membrane-active antiseptic agents. In contrast, susceptibility to peptide antimicrobials (Mel4, Melimine, LL-37) showed no consistent directional change.

2.3. Correlation Analysis of MICs

Within the antimicrobial groups, polymyxin B and colistin showed strong positive correlation (r = 0.51; p = 0.0041) (Figure 1). Mel4 and colistin demonstrated the next strongest correlation (r = 0.45; p = 0.0127), followed by Mel4 and melimine (r = 0.41; p = 0.0258). On the other hand, colistin was significantly negatively correlated with PQ-1 (r = -0.67; 0.0001), and PHMB (r = -0.44; p = 0.0151). Similarly, susceptibility was significantly negatively correlated with PHMB susceptibility (r = -0.48; 0.0074). In contrast, LL-37 susceptibility was not significantly correlated with any of the other antimicrobials tested.

3. Discussion

The present study was designed to evaluate the pattern of susceptibility among Acinetobacter strains against different cationic antimicrobials and to evaluate whether high MICs to disinfectants correlated to resistance to polymyxins or MICs to cationic antimicrobial peptides. The impetus for this research was the use of disinfectants to control spread of Acinetobacter and the increasing isolation of polymyxin resistant strains. The hypothesis was that resistance to disinfectants might cross-react to resistance to polymyxins, as these can act in similar ways on the membrane of bacteria. Furthermore, mechanisms of resistance to disinfectants such as efflux pump activation and membrane alterations [41] are also used be Acinetobacter to become resistant to polymyxins [33].

The results of this study show that the hypothesis was not proven, and resistance to polymyxin in A. baumannii negatively correlated with resistance to disinfectants (especially PQ-1 and PHMB). Whilst the antibacterial mechanism of action of PHMB has been traditionally thought to be via membrane disruption, recent studies have shown that whilst PHMB can attach to Gram-negative bacterial membranes, this does not result in pore formation (often a key mechanism for antibacterial activity) but it then penetrates through the membranes and interacts with intracellular DNA, which is likely the antibacterial mode of action [55,56]. Similarly, the quaternary amine compound benzalkonium chloride can have two modes of action against A. baumannii [57]. At high concentrations benzalkonium chloride acts on the bacterial membrane, but at lower concentrations it can induce protein aggregation and this can lead to cell death [57]. This may also occur for polyquats such as PQ-1. The polymyxin antibiotics appear to have predominant modes of action on the outer membrane of A. baumannii and other gram-negative bacteria, although they may also act intercellularly by inhibiting respiratory enzymes and triggering production of reactive oxygen species [58,59]. These differences may at least partially account for the lack of correlation between the MICs for the polymyxins and disinfectants. Reinforcing this, the polymyxin resistant mutant A. baumannii ATCC (R)19606 with complete loss of outer membrane [54] with MIC 1250 μg/ml and >2500 μg/ml against polymyxin B and colistin, respectively, showed high sensitively to the disinfectants. A previous study found that loss of LPS caused a 4 fold reduction in MIC of CHX [60].

This study also evaluated whether high MICs to the disinfectants or resistance to polymyxins resulted in high MICs to antimicrobial peptides [61]. The AMPs chosen for study were two naturally occurring AMPs, LL-37 (the only human cathelicidin) and lactoferricin (a small peptide from lactoferrin), as well as the synthetic AMPs melimine and Mel4. Should there have been a correlation, it would mean that resistance to polymyxins might have large impacts on the pathogenesis of disease caused by A. baumannii. Epithelial cells upregulate AMPs during A. baumannii infection [62] and AMPs such as LL-37 are key components of the innate defence system [63]. Furthermore, the AMPs melimine and Mel4 are being examined as potential new therapeutic or prophylactic agents [64,65,66], and so cross-resistance might curtail this research and development. Previous studies have determined that polymyxin resistance in A. baumannii is mediated by several factors including phosphoethanolamine addition to lipopolysaccharides in the outer membrane, loss of lipopolysaccharide from the outer membrane, and overexpression of efflux pumps [33,35,54]. These same mechanisms can confer resistance to antimicrobial peptides such as LL-37 [45]. Indeed, clinical use of polymyxin can induce cross-resistance to host antimicrobial peptides in A. baumannii [42].

However, the current study found no statistically significance evidence for cross-resistance between the naturally occurring LL-37 and polymyxins. This implies a lack of effect of polymyxin resistance, especially loss of the outer membrane which is the known mechanism of resistance of strain ATCC 19600 (R), on susceptibility to LL-37. There was a significant correlation between colistin resistance and MIC to Mel4, implying the possibility of similar mechanisms being involved. Although, this correlation did not hold for polymyxin B, which is structurally similar to colistin. Further research is needed to identify if any known polymyxin resistance mechanisms can confer cross-resistance to Mel4, and why these do not correlate with resistance to polymyxin B or MICs to melimine (the parent peptide of Mel4). The correlation between MIC to Mel4 and PQ-1 is intriguing, suggesting some cross-resistance, and this should also be followed up in future experiments. These future experiments could mutate susceptible strains with genes known to be involved polymyxin, disinfectant and antimicrobial peptide and determine which ones confer cross resistance.

4. Materials and Methods

4.1. Collection and Preservation of Bacterial Strains

Clinical isolates of Acinetobacter were collected from Australia (Liping Li, Macquarie University, NSW, Australia) Pakistan (Bushra Jamil, Sir Syed CASE University, Islamabad, Pakistan) and Switzerland (Stefano Mancini, Institute of Medical Microbiology, University Zurich, Switzerland). This study included the polymyxin-resistant mutant A. baumannii ATCC 19606R, which has a complete loss of its outer membrane due to mutation in lpx genes [54]. All the strains were preserved in 25% glycerol and stored at −80 °C at the microbial culture collection of School of Optometry and Vision Science, Faculty of Medicine and Health, UNSW Sydney.

4.2. Confirmation of Strain Identities by MALDI-TOF MS

Identity of strains was confirmed by MALDI-TOF MS (Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry). Briefly, strains were cultured on blood agar overnight and protein extraction was done according to MALDI Biotyper® Protocol Guideline [67,68] with 300μl of sterile MilliQ water. After air-drying the pellet at room temperature, 25 μL of 70% aqueous formic acid was added and mixed, followed by addition of 25 μL 100% acetonitrile and mixing. The mixture was then centrifuged for 2 minutes at 13,000 rpm and 1 μL of the supernatant was deposited onto a MALDI target plate and allowed to dry at room temperature. A 1μl aliquot of Bruker Bacterial Test Standard (Bruker, Billerica, MA, USA) was transferred to a MALDI target plate and allowed to dry. After drying. 1 μl of HCCA matrix solution was added to each sample and standard. Identification was done by using Bruker ultrafleXtreme MALDI-TOF/TOF biotyper system.

4.3. Determination of Minimum Inhibitory Concentrations of the Antimicrobials

Two cationic polypeptide antibiotics, polymyxin B (PMB) (Fluka Chemie GmbH, Buchs, Switzerland) and colistin sulfate (Polymyxin E; Sigma-Aldrich, Massachusetts, United States), were used. Four cationic antimicrobial peptides (LL-37, Mel4, melimine, and lactoferricin) were synthesised by solid state synthesis and purchased from AusPep Peptide Company (Tullamarine, VIC, Australia) with a purity of ≥90%. The cationic disinfectants PQ-1 (Sigma Aldrich), PHMB (Biosynth, Gardner, MA USA) and CHX (Sigma Aldrich) were also used.

Minimum inhibitory concentration (MIC) of the cationic antimicrobials was determined by broth micro-dilution assay [69]. Standard solutions of 128μg/mL for polymyxin B, colistin, PQ-1, PHMB and CHX were made sterile distilled water. For the AMPs, standard solutions of 5mg/mL were made in sterile distilled water. Cation adjusted Muller Hinton broth (Ca-MHB; Oxoid, Basingstoke, Hampshire, UK) was used as described previously [70]. Briefly, 100μl of each antimicrobial was diluted two fold in 100μl Ca-MHB, then serially diluted in Ca-MHB. A standard bacterial inoculum was prepared from overnight culture by adjusting to 0.5 McFarland standard (~1.5 × 10⁸ colony forming units (CFU)/mL) followed by further dilution to achieve a final concentration of approximately 1.5 × 10⁶ CFU/mL and 100μl of bacterial cells were added to each well and plates were incubated for 24 hours and MIC was determined as inhibition of visible growth.

4.4. Statistical Analysis

Correlation analysis among MICs of different cationic antimicrobials was determined by GraphPad Prism software (version 11.02.2. GraphPad Software, San Diego, CA, USA) by determining Spearman correlation coefficient from MIC data. Mann-Whitney U test was performed to determine whether differences in the median MICs of polymyxin resistant and susceptible strains to Mel4, Melimine, LL-37, PQ-1, PMHB, CHX were statistically different.

5. Conclusions

Acinetobacter baumannii strains demonstrate diverse patterns of susceptibility to cationic antimicrobial peptides and disinfectant. One hypothesis being tested was to determine whether high MICs to disinfectants that can be used to clean surfaces to reduce the spread of Acinetobacter spp. were correlated to resistance (high MICs) to polymyxin antibiotics which are antibiotics of last resort to treat Acinetobacter infections. The study found an inverse correlation, implying that disinfectant resistance does not result in resistance to polymyxins. The second hypothesis being tested was whether resistance to polymyxins was correlated to resistance to antimicrobial peptides involved in defence against Acinetobacter or being developed as new therapeutics or prophylactic agents. The lack of correlation between MIC for LL-37 and polymyxins implies no cross resistance.

This study has important translational implications for both infection control and therapeutic development. The finding that polymyxin-resistant A. baumannii remains susceptible, or even shows collateral sensitivity, to commonly used disinfectants suggests that current hospital disinfection strategies are unlikely to drive resistance to last-line antibiotics, and may in fact help control resistant strains. Furthermore, the absence of cross-resistance between polymyxins and key antimicrobial peptides such as LL-37 highlights the potential for host defence–based or AMP-derived therapies to remain effective against highly drug-resistant A. baumannii, supporting their ongoing clinical development as alternative treatments.