Introduction

Pseudomonas aeruginosa is a ubiquitous microorganism capable of adapting to several environments. It is an opportunistic pathogen associated with airway infections, which, when well established, is practically impossible to eradicate, especially the chronicity. This morbidity can culminate in pulmonary exacerbations and irreversible lung damage in patients with cystic fibrosis (CF) [1].

CF is a multisystemic, autosomal recessive disease caused by cystic fibrosis transmembrane conductance regulator (CFTR) gene mutations located on the long arm of chromosome 7, involving mucus and sweat producer cells. The lungs are the most severely affected, increasing mucus viscosity and inability to clear bacteria, resulting in chronic airway inflammation. Airway damage mainly occurs due to bacterial infections caused by typical microorganisms, which can lead to reduced lung function, leading to decreased quality of life in patients with CF [2,3,4].

P. aeruginosa possesses a diverse pool of resistance mechanisms, such as porins, efflux pumps and enzymes, affecting its susceptibility to several antimicrobial classes, such as β-lactams, carbapenems, aminoglycosides, fluoroquinolones and polymyxins [5, 6]. Resistance to carbapenems, mediated by permeability alterations and carbapenemases, is a significant concern as carbapenems are considered first-line drugs in the treatment of airway infections caused by P. aeruginosa in CF [5, 7, 8]. Similarly, resistance to polymyxins, facilitated by alterations in two-component system (TCS) and acquisition of exogenous DNA (plasmid mcr gene) is also a cause for concern since polymyxins are often considered a last resort treatment for MDR P. aeruginosa [9,10,11].

One of the challenges encountered in P. aeruginosa infection is the existence of multidrug-resistant strains and their spread among patients with CF [12], such as the Liverpool, Manchester, Midlands, Australia, Denmark, and Canada strains [13, 14]. Phylogenetic analysis aims to identify common bacterial clones that can be used as strains with important resistance profiles against CF [3, 15].

Whole-genome sequencing (WGS) is a standard technique for investigating outbreaks and bacterial typing. In addition to the rapid identification of agents that cause infections with high accuracy, it allows the detection of genes associated with antimicrobial resistance and relevant virulence factors. Despite the high cost and requirement for specialized professionals, it has been adopted by several laboratories and hospital institutions. In silico analysis of bacterial data obtained from WGS guarantees speed in the characterization and management of outbreaks, personalized therapy, and facilitates approaches in the control of nosocomial infections [16].

In recent years, an increase in P. aeruginosa isolates obtained from individuals with CF, followed up in two centers in Rio de Janeiro, showed decreased susceptibility to carbapenems, polymyxin B, and other antimicrobials [17]. This results in difficulties in the treatment and eradication of the pulmonary environment; therefore, it is necessary to analyze the phenotypic and genotypic characteristics of these isolates. The aim of this study was to perform a genetic analysis of P. aeruginosa resistance to polymyxin B, which may contribute to understanding the mechanisms of antimicrobial resistance, its phylogeny, and other important genomic aspects.

Materials and methods

Clinical specimens

Isolation and identification

We selected ten P. aeruginosa isolates from microbiological cultures of the respiratory secretions of five patients with CF (1 to 4 isolates per individual), obtained between 2010 and 2014 (Table 1), and stored in the bacteriological collection at the State University of Rio de Janeiro. Isolates from archived strains were used and no patient data were consulted for this study, therefore, informed consent was not obtained. All procedures performed in this study were in accordance with the ethical standards of the institutional research committee (CAAE: 79547616.1.0000.5259), and the approval was waived by the local Ethics Committee of Universidade do Estado do Rio de Janeiro.

Table 1 Distribution of patients, isolates and colonial morphology of Pseudomonas aeruginosa strains from cystic fibrosis patients
Full size table

Biochemistry test evaluations of the oxidative metabolism of glucose (non-fermentative), oxidase production (positive), arginine decarboxylation (positive), and physiological tests, such as motility and growth at 42ºC were conducted as previously described [18]. The sample inclusion criteria was a non-susceptibility profile to polymyxin B using the supplemental colistin agar test [19].

Antimicrobial susceptibility test (AST)

Disk-diffusion test (DDT) was performed for the following antimicrobials: piperacillin-tazobactam (PPT) (100/10 µg), ceftazidime (CAZ) (30 µg), cefepime (FEP) (30 µg), meropenem (MER) (10 µg), imipenem (IMP) (10 µg), doripenem (DOR) (10 µg), aztreonam (ATM) (30 µg), ciprofloxacin (CIP) (5 µg), amikacin (AMI) (30 µg) and tobramycin (TOB) (10 µg) (Becton, Dickinson and Company, BD, Sparks, NV, USA) as described in Clinical and Laboratory Standards Institute (CLSI) [19]. The P. aeruginosa ATCC 27,853 strain was used as a quality control.

Minimum inhibitory concentration

Minimum inhibitory concentration (MIC) was determined by broth microdilution technique [20] only to polymyxin B (10.000 µg/mL, Sigma, St. Louis, USA) using the quality control strain P. aeruginosa ATCC 27,853, and the breakpoints of ≤ 2 µg/mL as intermediate and ≥ 4 µg/mL as resistant [19].

Bacterial resistance to multiple antimicrobial agents

P. aeruginosa isolates were classified as multidrug-resistant (MDR) when they were resistant to ≥ 1 agent in ≥ 3 antimicrobial categories tested; and as extensively drug-resistant (XDR) when they were resistant to ≥ 1 agent in all but ≤ 2 categories tested as it was described by Magiorakos and collaborators [21].

Genomic analysis and bioinformatic tools (in silico analysis)

Bacterial DNA were extracted and purified from recent bacterial cultures using a QIAmp DNA Mini Kit (QIAGEN, Hilden, Germany). DNA was quantified by the QuantiFluor® (Promega, Madison, Wisconsin, USA) system, genomic libraries were created using Nextera XT DNA Library Preparation (Illumina Inc, California, USA) kit, and then sequenced on the Illumina MiSeq System (Illumina Inc, California, USA).

Reads quality control were carried out in FastQC v.0.11.9 (usergalaxy.org.au) program and then assembled in contigs using Unicycler v.0.4.8 [22] in Bacterial and Viral Bioinformatics Resource Center (BV-BRC) (bv-brc.org) platform. Annotations were performed using Rapid Annotation Subsystem Technology (rast.nmpdr.org/rast.cgi) and BV-BRC (bv-brc.org). The Basic Local Alignment Search Tool (blast.ncbi.nlm.nih.gov/Blast.cgi) database was used to ensure the integrity of the sequences using BLASTn and BLASTp tools. Extrinsic resistance genes were identified using the Center for Genomic Epidemiology tool, ResFinder (cge. food. dtu. dk/services/Resfinder/), and intrinsic resistance genes were identified using the BV-BRC databases. The sequences were aligned using the P. aeruginosa PAO1 reference strain in Bioedit (bioedit.software.informer.com/7.2/). Plasmids were assembled and annotated using Galaxy Australia (usergalaxy.org.au) tools, plasmidSPAdes v.3.9.0 [23] and Bakta v.1.5.0 [24]. The online tool PHAge Search Tool Enhanced Release was used for phages [25]. Molecular typing was performed using the Public Database for Molecular Typing and Microbial Genome Diversity (pubmlst. org/), and the ST relation of the samples was compared with epidemic CF clones using the Grapetree tool [26].

The WGS Shotgun projects have been deposited at the National Center for Biotechnology Information (NCBI) under the following accession numbers: JAPTHH000000000 (9876, BioProject PRJNA890687), JAOYMC000000000 (10,705, BioProject PRJNA890688), JAOYMD000000000 (11,227, BioProject PRJNA890692), JAOYME000000000 (14,297, BioProject PRJNA890694), JAOYMF000000000 (14,339, BioProject PRJNA890695), JAOYMG000000000 (17,138, BioProject PRJNA890696), JAOYMG000000000 (17,749, BioProject PRJNA890697), JAOYMI000000000 (17,801, BioProject PRJNA890698), JAOYMJ000000000 (17,828, BioProject PRJNA890700) and JAOYMK000000000 (17,973, BioProject PRJNA890701).

Results

The 10 P. aeruginosa isolates were resistant to polymyxin, as noticed by screening tests, with MIC ranging from 4 to 8 µg/mL, highlighting that almost all had an MIC of 4 µg/mL. All isolates were tested against 10 different antibiotics using DDT. Full or intermediate resistance was observed for all antimicrobials evaluated, with the highest marker resistance observed for carbapenems (IMP and MER) and the lowest for CAZ. Five isolates were categorized as MDR, and four as XDR (Table 2).

Table 2 Antimicrobial susceptibility in Pseudomonas aeruginosa isolates
Full size table

After annotation of the genomes, the contigs per isolate ranged from to 45–120 with an average genome size of 6.434.232 kpb (6.250.921–6.738.939 kpb) and 72–58 RNA genes (Table 3).

Table 3 Genomic composition data of Pseudomonas aeruginosa isolates
Full size table

According to molecular typing, the isolates were classified as ST252 (isolate 9876), ST865 (10,705), ST871 (14,297), ST2211 (17,138 and 17,749) and ST1560 (17,801). The three new STs were identified as ST4051 (11,227), ST4052 (14,339), and ST4053 (17,828 and 17,973) (Table 4). Using a phylogenetic tree, the following CF-transmissible strains of P. aeruginosa were compared with clinical isolates: Liverpool, Manchester, Midlands, Praire, Australian 1, 2, and 3, Denmark, Clone C, and Dutch strains. Through phylogenetic analysis based on the distances calculated from the known STs and the strains included in this study, it was not possible to establish phylogenetic correlation between the described isolates and the typical transmissible strains observed around the world (Fig. 1).

Table 4 Sequence type identification and allele profile of Pseudomonas aeruginosa isolates from this study and from cystic fibrosis epidemic clones
Full size table
Fig. 1
figure 1

– Spanning tree of Pseudomonas aeruginosa isolates from this study and from cystic fibrosis epidemic clones. (Figure 1: Spanning tree of Pseudomonas aeruginosa isolates comparing isolates in this study and cystic fibrosis epidemic clones; Figure was generated in Public Database for Molecular Typing and Microbial Genome Diversity (pubmlst. org/) platform with Grapetree [26] tool in 2023; ST: sequence type; MLST: Multi-locus Sequence Typing; LES: Liverpool epidemic strain; MES: Manchester epidemic strain; PES: Praire epidemic strain; AES: Australian epidemic strain; DK: Denmark strain; DES: Dutch epidemic strain)

Full size image

The genes associated with intrinsic antimicrobial resistance were also investigated. Isolates 17,138 and 17,749, both ST2211 strains, exhibited the highest number of mutations (substitutions, “s”) in the efflux pump regulators mexR (5 s), nalC (17 s), nalD (4 s), nfxB (26 s), mexS (6 s), and mexZ (18 s). The remaining isolates showed variations of 1–3 s for the same genes. In isolate 14,339, a premature stop codon (sc) was detected in the mexZ amino acid chain at position 134 (Supplementary Material 1).

Each isolate exhibited substitutions in at least one of the studied porin genes, with isolate 11,227 displaying the highest number of alterations in oprD. When compared with other porin genes, oprD showed the greatest number of modifications (deletions “Δ”, insertions “ins” and substitutions) among the isolates, being the only gene among porins that showed deletions (Table 5). In oprD, isolates 9876, 10,705, 11,227 and 14,339 presented a frameshift deletion in codons 372, 373 and 383; 17,138, 17,749 and 17,801, exhibited a premature stop codon in the codon 277. Additionally, 9876, 10,705, 11,227, and 14,339 exhibited a sequence of mutations ranging from codon 375 to 382 (Supplementary Material 1).

Table 5 Alterations in the genes of interest in Pseudomonas aeruginosa isolates
Full size table

Regarding TCS genes, in connection with their capacity to induce polymyxin resistance through genetic mutations, isolates 17,138 and 17,749 displayed the highest number of substitutions among all isolates. This emphasizes the involvement of pmrB (28 s and 1sc), parS (32 s), and cprS (25 s) genes. Furthermore, both isolates exhibited identical alterations and were the only isolates that manifested mutations in all TCS genes. Mutations in colR were unique to these two isolates (3 s). The other isolates displayed a variable range of alterations, ranging from one to five substitutions (Table 5).

Similar to that observed for TSC, four genes (gyrA, gyrB, parC, and parE) related to the DNA gyrase and Topoisomerase IV subunits showed the same mutations in 17,138 and 1749 isolates (Table 5).

Among the extrinsic resistance genes, oxacillinases OXA-4 (17,828 and 17,973), OXA-50 (14,339), OXA-395 (14,297, 17,801, 17,828 and 17,973), and OXA-486 (9876, 10,705, 11,227 and 14,297) were found (Table 6).

Table 6 Analysis of extrinsic resistance genes found in the investigated samples of Pseudomonas aeruginosa using Resfinder
Full size table

The aph(3’)-llb and aadA24 genes, related to aminoglycoside resistance, were exclusively detected in isolates 17,828 and 17,973, both ST4053. crpP, which is associated with CIP resistance, was detected in isolates 17,138, 17,749, and 17,801. The sul1 gene, associated with sulfamethoxazole resistance, was only detected in isolate 17,801. catB7 and fosA, related to chloramphenicol and fosfomycin resistance respectively, were present in all isolates, except for 17,138 and 17,749 (ST2211). Notably, isolates 17,828 and 17,973 (ST4053) had the highest number of extrinsic resistance genes (Table 6).

Broad variations were observed in the size and presence of genes in the annotated plasmids. The smallest plasmid (15,805 bp) was found in isolate 9876, whereas the largest (414,162 bp) was found in isolate 17801. According to the Bakta database, each isolate carried one plasmid. Resistance genes were identified in the plasmid of isolates 17138 (crpP), 17801 (OXA-395), 17828, and 17973 (OXA-1like and aaC(6’)lb).

Intact phages found were Pseudo_YMC11/02/R656 (9876 and 10,705 isolates), Pseudo_H66 (11,227 isolate), Pseudo_JD024 (14,297 isolate), Escher_vB_EcoM_ECO1230_10 (17,749 isolate), Pseudo_Dobby (17,828 and 17,973 isolates), Pseudo_ F10 (17,828 and 17,973 isolates).

Discussion

Pseudomonas aeruginosa is one of the main pathogens associated with CF, severely affecting the lungs and contributing to the worsening of disease prognosis through inflammation and airway damage. It is difficult to eradicate due to its ability to withstand antimicrobial agents and establish chronicity [27]. Several mechanisms of antimicrobial resistance, biofilm formation, and persistence of multidrug-tolerant cells are strongly associated with chronic infections [8, 28].

In the present study, P. aeruginosa polymyxin-resistant isolates demonstrated reduced susceptibility to at least one carbapenem, a concern in patients with CF, owing to the use of these first-line drugs in the treatment of chronically infected patients. This makes it even more difficult to achieve a positive therapy response and eradicate pulmonary infection [29].

Studies have highlighted the emergence of carbapenem-resistant P. aeruginosa in CF in several regions of the world [29,30,31,32]. However, in Brazil, carbapenem-resistant P. aeruginosa reports on CF remain scarce [17, 33, 34].

One treatment option for carbapenem-resistant P. aeruginosa in CF is the use of polymyxins; however, monotherapy may lead to resistance [35, 36]. Variations in polymyxin susceptibility among patients with CF have also been observed in Europe. A study conducted in Italy reported an increase in polymyxin susceptibility, from 93 to 98% [37]. However, an investigation in Denmark from 2008 to 2016 identified an increase in polymyxin resistance from 7 to 13% among individuals with CF colonized with P. aeruginosa [38]. In Brazil, data on P. aeruginosa polymyxin resistance in individuals with CF are limited. In a recent study of 179 isolates from individuals with CF, 7.2% exhibited polymyxin resistance [17].

The detection of P. aeruginosa polymyxin resistance in individuals with CF is uncommon; however, resistance to polymyxins has been increasingly reported in non-CF cases [39,40,41,42]. This could be related to mutations in the TCS genes or the presence of plasmid genes that alter the polarity of lipid A, such as mcr [43,44,45,46]. To date, P. aeruginosa polymyxin resistance harboring mcr has not yet been described in individuals with CF, including those in our study.

The emergence of MDR P. aeruginosa isolates from patients with CF has been described for more than two decades [47,48,49,50]. Nevertheless, cases of MDR isolates from CF centers in Brazil have been reported over the last decade [17, 33, 34, 51].

In the present study, most isolates were classified as either MDR or XDR strains. This result was expected because of the study’s inclusion criteria, which required the isolates to exhibit resistance to polymyxin B. Based on the criteria established by Magiorakos and collaborators [21], isolate 10,705 was not classified as MDR or XDR despite exhibiting resistance to IMP and polymyxin B. This still draws our attention as isolate 10,705 exhibited mutations in mexS, oprD, opdP, opdD, pmrA, pmrB, parS and parC genes. All these genes, except for parC, can mediate resistance to carbapenems or polymyxins through mechanisms such as efflux pumps (overexpression of MexEF), porins (OprD, OpdP and OpdD) and lipid A modification (PmrAB and ParRS). In the case of parC, which can confer resistance to fluoroquinolones through change in the target site [52], our study only tested ciprofloxacin. This may present a methodology gap, as we cannot conclusively determine the isolate’s resistance to other fluoroquinolones, which could classify this isolate as MDR. Nevertheless, numerous definitions outlining the phenomenon of MDR in Gram-negative organisms are prevalent worldwide. These definitions exhibit variations contingent on their intended applications and originating country or institution. The choice of specific definitions of MDR should facilitate uniformity in epidemiological surveillance practices [53,54,55].

Through molecular typing, the isolates were categorized into eight ST, among which three new STs were identified. Notably, among the 10 isolates analyzed, 17,138 and 17,749, both recovered from the same individual with a one-year interval between isolations, shared the same ST (2211), and exhibited resistance to all tested antimicrobials. This is an important issue because multidrug resistance is often induced by the exposure of individuals with CF to multiple antimicrobials, thereby reducing therapeutic options and life expectancy [56].

Using a comparative analysis to assess the relationship between these STs and ST related to CF-transmissible strains, no similarity was observed, similar to previous studies conducted by our group [34, 57]. The existence of these transmissible strains has initiated debates concerning infection control protocols and the management of patients with CF. Continuous surveillance is imperative due to their considerable potential for global dissemination and their ability to harbor multiple resistance and virulence genes.

Genes associated with intrinsic and extrinsic antimicrobial resistance were investigated by WGS. Notably, isolates 17,138 and 17,749, which were observed in the same patient, exhibited the highest number of mutations in the same gene (Supplementary Material 1). The mutations observed in these isolates, such as the efflux pump regulators MexAB (mexR, nalC and nalD), MexCD (nfxB), MexEF (mexS), porins (opdP, opdD, opdH, and oprE), TCS (phoP, phoQ, pmrA, parR, cprR, cprS, colR, and colS), and the enzymatic DNA gyrase and Topoisomerase IV subunits (gyrB and parE), did not correlate with previously described mutations in MDR strains.

The oprH porin gene, involved in polymyxin resistance [8], was searched in all isolates in this study. The gene sequences were aligned with the reference strain P. aeruginosa PAO1, revealing that, with exception of isolates 17,138 and 17,749, all isolates exhibited 100% similarity with the oprH from P. aeruginosa PAO1. Blast comparison showed that oprH from isolates 17,138 and 17,749, shared 96.52% similarity with the oprH from P. aeruginosa PAO1, indicating the presence of the same protein function, with only allele variations.

mexZ, a MexXY efflux pump regulator, displayed a premature stop codon at codon 134 in isolate 14,339. Although alterations at this position have not been published, loss of protein function and impairment of repressor function have been reported when mutations occur at other positions in the amino acid chain [58, 59]. This may be related to the overexpression of the efflux pump, a determinant of resistance to aminoglycosides in the CF P. aeruginosa [60], similar to the resistance profile observed in AST.

Analyses of oprD, related to carbapenem resistance in P. aeruginosa, showed frameshift deletions and premature stop codons, previously reported as modifications promoting resistance to these antimicrobials and porin suppressors [61, 62].

In previous studies, defective or absent oprD was reported in non-CF P. aeruginosa, and some of these alterations were similar to those observed in the present study [63, 64]. Isolates 9876, 10,705, 11,227, and 14,339 showed a sequence of mutations (372 M-382Y) with alternating deletions and substitutions, and some similar previously described cases demonstrated that changes in these amino acids were associated with increased susceptibility to MER [62, 65]. Notably, this was true for two isolates (10,705 and 14,339) in this study.

Substitutions in pmrB (15 V-I, 68G-S, and 343T-A), observed in isolates 11,227, 17,138, 17,749, and 17,801, were thought to be associated with colistin non-susceptibility in non-CF P. aeruginosa, as these are common mutations described solely in non-susceptible isolates [66]. Additionally, a substitution in parS (398 H-R), detected in all isolates in this study, has been linked to colistin resistance, as it occurs under external stress conditions, thereby affecting colistin resistance [67]. Both genes are known to be involved in lipid A polarity modifications which can cause polymyxin resistance.

Substitutions in gyrA (83T-I) and parC (87 S-L) were identified in isolate 17,801, resistant to CIP, already well established as causing high-level resistance to fluoroquinolones in non-CF P. aeruginosa [52, 68, 69].

crpP, which encodes extrinsic resistance to CIP, was identified in three phenotypically resistant isolates (17,138, 17,749, and 17,801), which are not always correlated [70]. Corroborating the findings of the authors, two isolates were resistant to CIP, but did not have crpP, which can be explained by another mechanism of resistance, such as the overexpression of efflux pumps [71].

A correlation between the results of phenotypic resistance obtained in the AST and the presence of OXA-4, OXA-50, OXA-486, and crpP was observed. However, for OXA-395, aph(3’)-llb, aadA24 it was not possible to establish the same association, possibly due to not being expressed, expressed at a very low level, or another mechanism [72]. Sulfamethoxazole, chloramphenicol, and fosfomycin were not tested in this study; therefore, the relationship between sul1, catB7, and fosA cannot be reported.

In Brazil, the presence of carbapenemases in P. aeruginosa strains isolated from patients with CF is rare. This was highlighted in a report of a single isolate carrying the blaSPM−1 gene [33]. Furthermore, the presence of blaKPC has been documented in Klebsiella pneumoniae and Enterobacter cloacae among individuals with CF under surveillance in facilities within the same city. This observation has raised concerns regarding the potential transmission of these genes, including those within P. aeruginosa [73, 74].

In the plasmid analysis, the presence of mcr was not observed in any of the isolates, which may suggest a different origin of resistance to polymyxin B. Nonetheless, the crpP, OXA-395, and aaC(6’)lb plasmid genes were identified in different isolates of P. aeruginosa polymyxin resistance in CF; to our knowledge, this study is the first time this has been described in Brazil.

Most isolates showed at least one intact phage. Notably, YMC11/02/R656 was present in two isolates (9876 and 10,705) from individuals with CF treated at the same follow-up center. The same phage was described in a study by our group with Achromobacter ruhlandii in a different CF center [75]. This is concerning since lung infections via CF are polymicrobial, and the potential of P. aeruginosa to acquire transmissible genetic information allows the dissemination of resistance factors, including those from different bacterial species [76, 77].

Conclusion

Among the isolates, a high resistance rate was observed, with many multi-resistance profiles. WGS is fundamental for the identification of mutations related to permeability and alterations in the outer membrane and target site, which could influence the resistance of the isolates. Additionally, it allows the localization of plasmid resistance genes and bacteriophages, which may spread and perform horizontal gene transfer between other bacteria in the same center. Furthermore, molecular typing described new STs but did not correlate with previously described epidemic clones.