Background

Pseudomonas aeruginosa is a Gram-negative opportunistic pathogen commonly found in hospital environments, particularly in intensive care units (ICUs). It is responsible for various nosocomial infections, including pulmonary, bloodstream, urinary tract, surgical site, and skin and soft tissue infections [1,2,3,4]. The treatment of P. aeruginosa infections typically involves the use of β-lactam, fluoroquinolone, and aminoglycoside antibiotics either alone or in combination [5, 6]. However, the misuse and overuse of antibiotics have led to the emergence of antibiotic resistance in P. aeruginosa strains, making it difficult to effectively treat these infections. The rise in antibiotic resistance among Gram-negative bacteria, which began in the 1970s, continues to be a significant challenge [7,8,9]. In 2017, the World Health Organization (WHO) declared antimicrobial resistance a global crisis. The organization listed 12 bacterial families as the greatest threats to human health, with carbapenem-resistant P. aeruginosa strains being one of the most important [10]. Carbapenem antibiotics are commonly recommended for the treatment of multidrug-resistant (MDR) P. aeruginosa strains [8]. However, in line with the WHO report, our previous studies have confirmed the high prevalence of carbapenem-resistant P. aeruginosa isolates in various cities of Iran, particularly in Ardabil in the northwest of the country [3, 11]. This challenging situation has increased the reliance on polymyxins, such as polymyxin B and colistin, which are cationic lipopeptide antibiotics and considered as the last resort for treating Gram-negative bacteria, including carbapenem-resistant P. aeruginosa strains that are resistant to all other available antibiotics [12]. While colistin (also known as polymyxin E) resistance has not been observed in clinical isolates of P. aeruginosa in Ardabil city thus far, colistin-resistant P. aeruginosa has been reported in other cities in Iran and worldwide. Therefore, one of the aims of this study was to investigate the resistance of P. aeruginosa clinical strains to colistin in Ardabil, as well as their genetic linkage. The mechanisms of polymyxin resistance in Gram-negative bacteria are not fully understood. However, they can be mediated through plasmid-encoded genes, such as the acquisition of the mobilized colistin resistance-1 (mcr-1) gene, or through chromosomally encoded genes, including 1) cationic modification of lipid A in lipopolysaccharides (LPS) via two-component systems (TCSs: PhoPQ and PmrAB), 2) loss of LPS, 3) overexpression of efflux pump systems and capsular polysaccharides, 4) down-regulation of porin (OprD), and 5) enzymatic inactivation of colistin [7]. Therefore, another objective of this study was to determine the most common mechanisms of colistin resistance among drug-resistant P. aeruginosa clinical isolates collected from patients referred to hospitals in Ardabil, Iran.

Materials and methods

P. aeruginosa clinical isolates, materials, and equipment

In this cross-sectional study conducted in Iran, a country located southwest of the Asian continent, a total of 200 P. aeruginosa clinical isolates were utilized. These strains were collected from various specimens, including urine (n = 90), sputum (n = 55), wound (n = 28), blood (n = 26), and cerebrospinal fluid (CSF) (n = 1), obtained from patients referred to hospitals in Ardabil city, northwest of Iran. The hospitals included Imam Khomeini (n = 105), Alavi (n = 55), Imam Reza (n = 25), Bu-Ali (n = 6), Sabalan (n = 6), Fatemi (n = 2), and Ghaem (n = 1). The data collection period spanned from June 2019 to May 2023.

The initial identification of P. aeruginosa clinical isolates was performed using phenotypic standard laboratory tests, which included assessments of pigment production, colony morphology, oxidase, catalase, IMViC pattern, and Gram staining. Confirmation of the species was subsequently achieved using the polymerase chain reaction (PCR) with a specific species primer [3].

The most important materials used in our study were the Master Mix for PCR/ERIC-PCR and real-time PCR (Ampliqon, Denmark), primers (Metabion, Germany), TRIzol™ Reagent (Bio Basic, Ontario, Canada), cDNA synthesis kit (Yekta Tajhiz Azma, Tehran, Iran), Mueller Hinton agar (Conda, Pronasida, Spain), colistin sulfate salt powder (Sigma-Aldrich co, St. Louis, MO, ≥ 15,000 U/mg), antibiotic disks (Padtan Teb, Iran), Cetrimide agar (Conda, Pronasida, Spain), and Trypticase Soy Broth (TSB) (QUELAB/UK). Additionally, the following equipment was used: Eppendorf thermal cycler (Germany), LightCycler® System (Roche Diagnostics), NanoDrop 2000c Spectrophotometer (Thermo Scientific, USA), Agarose Gel Electrophoresis (Padideh Nojen Pars, Iran), and ELISA microplate reader (BioTek, USA).

Colistin agar test

The susceptibility pattern of P. aeruginosa clinical isolates to colistin was determined based on the agar dilution MIC (minimum inhibitory concentration) method on Mueller–Hinton agar as suggested by the Clinical and Laboratory Standards Institute (CLSI) [13]. For this purpose, 3–5 fresh P. aeruginosa colonies from Mueller Hinton agar plates were picked and transferred to 4–5 mL sterile saline to prepare 0.5 McFarland turbidity standards. Bacterial standard suspensions were diluted in saline (1:10). A 10 μL of each diluted bacterial suspension was poured onto a colistin agar plate. Colistin agar plates were prepared in required dilutions, i.e., 0.5–16 μg/mL. Incubation condition and length were maintained at 37 °C for 16–18 h. P. aeruginosa clinical isolates with MIC values ≥ 4 μg/mL were considered as resistant strains. A colistin-resistant Acinetobacter baumannii clinical isolate was used as the positive control (MIC = 16 μg/mL) (Ethics ID: IR.ARUMS.REC.1400.071).

The disk diffusion method was used to determine multiple drug resistance patterns (multidrug-resistant (MDR), extremely drug-resistant (XDR), and pandrug-resistant (PDR) strains) among colistin-resistant P. aeruginosa clinical isolates [14]. Furthermore, this method was employed to assess the resistance rates of colistin-resistant P. aeruginosa isolates against various antibiotics, including piperacillin (100 μg), piperacillin-tazobactam (100/10 μg), ceftazidime (30 μg), cefepime (30 μg), aztreonam (30 μg), imipenem (10 μg), meropenem (10 μg), gentamicin (10 μg), tobramycin (10 μg), amikacin (30 μg), ciprofloxacin (5 μg), levofloxacin (5 μg), norfloxacin (10 μg), and ofloxacin (5 μg), as per our previous study [3]. P. aeruginosa ATCC 27853 was used as a reference strain.

Biofilm formation assay

Evaluation of biofilm production among colistin-resistant P. aeruginosa isolates was performed by a colorimetric assay [15]. For this aim, of the 1:100 diluted suspensions of P. aeruginosa isolates which had grown in TSB mediums, 150 μL were inoculated into a sterile 96-well flat bottom plate and incubated at 37 °C for 24 h. The plate was washed with 200 μL of phosphate-buffered saline (PBS) (pH ~ 7.4) three times. Biofilm fixation was done with 100 μL of methanol (99%) for 15 min and then the wells were air-dried. 150 μL of crystal violet stain (1%) was added to wells for 20 min, unbound stain washed with water, and then bound stain released through 150 μL of acetic acid (33%). ELISA microplate reader was used to measure the optical density (OD) of wells at 590 nm. Colistin-resistant P. aeruginosa isolates were divided into four categories including no biofilm producer if the OD of a strain (ODs) was less or equal to the OD negative control (ODc), weak biofilm producer if ODc < ODs < 2 × ODc, moderate biofilm producer if 2 × ODc < ODs < 4 × ODc, and strong biofilm producer if 4 × ODc < ODs. P. aeruginosa ATCC 27853 and sterile TSB medium were used as positive and negative controls, respectively. All experiments are performed in triplicate.

Detection of colistin resistance genes

Molecular identification of the genes encoding efflux pumps (i.e., mexA, mexC, mexE, and mexY genes), TCSs (i.e., phoP, phoQ, pmrA, and pmrB genes), outer membrane porin (oprD gene), and mcr-1 gene, as well as genes involved in biofilm formation of P. aeruginosa (i.e., algD, pslD, pelF, Ppgl, and PAPI-1 genes) were performed by the PCR method. Used primers along with the PCR program for the detection of each gene were listed in Table 1. In brief, genomic DNA was extracted from the 200 P. aeruginosa clinical isolates by the boiling method [3] and confirmed by a spectrophotometer. Amplification of the genes was performed in a final volume of 15 μL (10 μL of Master Mix, with 3 μL of template DNA (50 ng/µL), and 2 μL of primers (10 μmol/L)) and then their presence was confirmed using the agarose gel electrophoresis and sequencing (Sanger method, Pishgam, Iran) techniques. It is worth mentioning that for some genes with non-specific bands on agarose gel, the values mentioned above along with the PCR conditions were changed. Finally, we employed the enterobacterial repetitive intergenic consensus (ERIC)-PCR method to assess the genetic relatedness among colistin-resistant P. aeruginosa isolates. In pursuit of this, amplification reactions were conducted in a final volume of 50 μL using the primers and ERIC-PCR program outlined in Table 1. Subsequently, the ERIC-PCR products were electrophoresed on a 2% agarose gel, and the resulting band patterns were analyzed using the Dice distance method for similarity and the UPGMA analysis method for clustering (GelQuest software version 3.3.5.0). ERIC-PCR band patterns exhibiting > 80% similarity were categorized as belonging to the same cluster [5].

Table 1 Used primers along with PCR/ERIC-PCR and qRT-PCR programs
Full size table

Mutational analysis of the PhoPQ and PmrAB

Detection of colistin resistance-associated mutations among P. aeruginosa clinical isolates with multiple drug resistance was performed using the sequencing method. The PCR products of the phoQ and pmrB genes from P. aeruginosa clinical isolates MDR, XDR, and resistant to colistin were sent for sequencing. The nucleotide sequences were compared with colistin-susceptible P. aeruginosa reference strain ATCC 27853 using the BioEdit software (version 7.2.5). Additionally, an online data analysis service (available at https://web.expasy.org/translate/) was utilized to assess amino acid alterations.

Expression of the genes encoding efflux pumps, TCSs, and outer membrane porin

Expression levels of the mexA, mexC, mexE, mexY, phoP, pmrA, and oprD genes were determined among resistant isolates using the quantitative reverse transcription PCR (qRT-PCR) and specific sets of primers (Table 1). In brief, the total RNA of colistin-resistant P. aeruginosa isolates was extracted using the TRIzol™ Reagent. After confirming the quality and quantity of extracted RNAs (1 μg), cDNA synthesis was done according to the manufacturer’s instructions. The qRT-PCR of the genes was carried out under conditions presented in Table 1 and in a final volume of 15 μL (SYBR Green PCR Master Mix (7 μL), primers (2 μL, 10 μmol/L), cDNA (1 μg/μL), and DEPC-treated water (5 μL)). The 30S ribosomal rpsL gene was used as the normalizing gene. The expression levels of genes in colistin-resistant P. aeruginosa isolates were determined relative to their expression levels in P. aeruginosa ATCC 27853 using the 2−ΔΔCt method. Expression for each gene was assessed in duplicate.

Interpretation of the results of qRT-PCR was performed as follows: for the mexA and mexC genes; twofold, for the mexE gene; tenfold, and for the mexY gene; fourfold expression rates compared with the reference strain of P. aeruginosa ATCC 27853 were considered as overexpression [16]. For the oprD gene, the expression rate ≤ 30% relative to the reference strain was considered as down-regulation [17]. In addition, for the phoP and pmrA genes, expression levels higher than those of P. aeruginosa ATCC 27853 were considered as increased gene expression [12].

Result

Among 200 P. aeruginosa clinical isolates obtained from hospitalized patients, 18 isolates (9%) were resistant to colistin antibiotic according to the agar dilution method. Characteristics of these 18 colistin-resistant P. aeruginosa isolates were presented in Table 2. In addition, MIC values of colistin for clinical isolates of P. aeruginosa are reported in Fig. 1. Among 200 P. aeruginosa strains, MIC values of 0.5, 1, 2, and 4 μg/mL were found in 2 (1%), 125 (62.5%), 55 (27.5%), and 18 (9%) isolates, respectively. All colistin-resistant isolates showed MIC = 4 μg/mL.

Table 2 Mechanisms of colistin resistance among 18 colistin-resistant P. aeruginosa clinical isolates
Full size table
Fig. 1
figure 1

MIC values of colistin for 200 clinical isolates of P. aeruginosa

Full size image

Of the 18 colistin-resistant P. aeruginosa strains, 6 strains showed multiple drug resistance patterns (4 MDR and 2 XDR) in the disk diffusion method. The susceptibility patterns of colistin-resistant P. aeruginosa strains to various antibiotics along with their virulence gene profiles are presented in Table 3. The resistance rates of 18 colistin-resistant P. aeruginosa isolates to different antibiotics were as follows: piperacillin 11.1%, piperacillin-tazobactam 5.5%, ceftazidime 11.1%, cefepime 11.1%, aztreonam 0%, imipenem 38.9%, meropenem 11.1%, gentamicin 11.1%, tobramycin 11.1%, amikacin 5.5%, ciprofloxacin 50%, levofloxacin 50%, norfloxacin 50%, and ofloxacin 55.5%.

Table 3 Profiles of antibiotic resistance and virulence genes of 18 colistin-resistant P. aeruginosa clinical isolates
Full size table

A high genetic diversity was observed among the 18 colistin-resistant P. aeruginosa clinical isolates by detecting 15 different ERIC-PCR band patterns (Fig. 2). Among tested strains, 15 colistin-resistant P. aeruginosa exhibited unique genotypes (subgroup), while genotype subgroup 6 comprised three isolates. Details of band patterns for each species are depicted in Supplementary Figure S1.

Fig. 2
figure 2

Dendrogram of colistin-resistant P. aeruginosa clinical isolates based on the ERIC-PCR band patterns

Full size image

As shown in Table 3, colistin-resistant P. aeruginosa strains were carrying the genes encoding resistance to other antibiotics such as metallo-β-lactamase (IMP gene), AmpC cephalosporinase, extended-spectrum β-lactamase (TEM and PSE genes), oxacillinase (oxa-2 and oxa-23 genes), and efflux pumps (qacEΔ1, qacE, qacG, and cepA genes). These genes are involved in the emergence of MDR and XDR P. aeruginosa strains.

Biofilm formation was identified in all 18 colistin-resistant P. aeruginosa isolates in the colorimetric assay. Among them, 7 (38.9%) isolates were weak biofilm producers, 8 (44.4%) isolates were moderate biofilm producers, and 3 (16.7%) isolates were strong biofilm producers. In addition, the presence of genes encoding biofilm (i.e., algD, pslD, pelF, Ppgl, and PAPI-1 genes) was detected in all 18 colistin-resistant P. aeruginosa isolates in the PCR and confirmed by sequencing. The GenBank accession numbers for our nucleotide sequences of detected genes in this study are OR855380 to OR855387.

The role of the phoQ and pmrB gene mutations in the emergence of colistin-resistant P. aeruginosa was evaluated in 6 MDR and XDR clinical isolates. Analysis of the phoQ gene revealed the following nucleotide substitutions: at positions 473 C → T, 495 G → A, 552 T → C, 583 T → C, 675 T → G, 702 A → G, 1110 C → T, 1146 C → T, and 1155 C → T. None of these nucleotide substitutions led to amino acid alterations. Furthermore, similar results were observed for the pmrB gene and nucleotide changes did not result in amino acid substitutions except for Tyr345His. Nucleotide substitutions of the pmrB gene were as follows: G at position 507 to A, G at position 639 to A, A at position 645 to G, G at position 690 to C, T at position 696 to C, A at position 750 to G, A at position 762 to G, C at position 891 to T, and T at position 1033 to C.

Expression levels of the mexA, mexC, mexE, mexY, phoP, and pmrA genes in 18 colistin-resistant P. aeruginosa strains were as follows: 88.8% (n = 16), 94.4% (n = 17), 11.1% (n = 2), 83.3% (n = 15), 83.3% (n = 15), and 38.8% (n = 7), respectively. In addition, down-regulation of the oprD gene was observed in 44.4% (n = 8) of colistin-resistant P. aeruginosa strains. The presence of a plasmid-borne mcr-1 gene and its association with colistin resistance in P. aeruginosa strains was not confirmed in this study.

Discussion

In recent years, there has been a significant reduction in the susceptibility of P. aeruginosa to polymyxins, despite its inherent sensitivity to these antibiotics [21]. The current study reported, for the first time, a prevalence of 9% for colistin-resistant P. aeruginosa in Ardabil city. Other studies conducted in different regions of Iran and abroad have reported varying rates of colistin resistance among P. aeruginosa isolates. These include studies by Abd El-Baky et al. in Egypt (21.3%) [22], Rossi et al. in Brazil (6.3%) [23], Wi et al. in South Korea (7.4%) [24], Zarate et al. in Peru (7.2%) [25], Farajzadeh Sheikh et al. in Iran (Ahvaz, Tehran, and Isfahan) (1.3%) [26], Goli et al. in Iran (Tabriz) (2%) [27], Heidari et al. in Iran (Isfahan and Shiraz) (7%) [28], Malekzadegan et al. in Iran (Shiraz) (0%) [29], Tahmasebi et al. in Iran (Hamadan) (3.9%) [30], and Talebi et al. in Iran (Tehran) (0%) [19].

The variations in colistin resistance rates among P. aeruginosa isolates in different regions can be attributed to several factors. These include differences in sample size, methods used for antimicrobial susceptibility testing, and specific local factors. In the case of Ardabil, the higher rate of colistin resistance compared to other cities in Iran may be attributed to three main factors. Firstly, the use of colistin in veterinary medicine for the growth promotion of food-producing animals, as Ardabil is an agricultural and animal husbandry province [31]. This practice can contribute to the selection and spread of colistin-resistant strains. Secondly, the misuse of antibiotics in intensive care units, which can lead to the emergence of resistant strains [22]. It's worth noting that Iran has been reported to have the second-highest antibiotic consumption rate in the world, with antibiotic consumption 16 times higher than the global standard (mehrnews.com/xYQDx), as stated by the secretary of the Infectious Diseases Association of Iran. Thirdly, while the transmission of colistin-resistant P. aeruginosa within hospitals or across the entire healthcare system has been a concern [32], our investigation revealed no genetic correlation among colistin-resistant strains collected from different hospitals in Ardabil city. However, we did observe three strains in genotype subgroup 6 that were shared between two distinct hospitals (Fig. 2). A similar finding was reported by Khosravi et al. for MDR P. aeruginosa isolates in Ahvaz [33]. In contrast, Zarei et al. identified clonal relatedness between clinical and environmental P. aeruginosa isolates [20].

Regarding the emergence of colistin-resistant P. aeruginosa strains in Ardabil hospitals, various bacterial factors may be involved. One such factor is biofilm formation. A biofilm is a bacterial population encased in an outer polymer layer, consisting of host immune system products or bacterial secreted polymers like exopolysaccharides (EPS), extracellular DNA, and proteins. Biofilms play a role in acute burn wound infections [1, 34]. Studies have shown that extracellular DNA within the biofilm, along with factors such as magnesium or calcium starvation, low pH, and antimicrobial peptides (including colistin), can contribute to polymyxin resistance through modifications in LPS via the activation of TCSs like PmrAB and PhoPQ [34, 35]. Our study's findings are consistent with the previous research discussed above. We observed that all colistin-resistant strains of P. aeruginosa in our study were capable of producing biofilms, as indicated in Table 2. This ability poses a significant challenge in the treatment of P. aeruginosa infections, particularly those affecting wounds. Furthermore, the presence and excessive production of alginate EPS within the biofilm provide P. aeruginosa with protection against phagocytic cells and antibiotic treatments [34]. Apart from the algD gene, which encodes alginate, we identified the presence of other genes associated with biofilm formation (pslD, pelF, Ppgl, and PAPI-1) in all colistin-resistant P. aeruginosa isolates. However, the prevalence of these chromosomal genes differed from a study conducted by Rajabi in Iran, where the respective prevalence rates were as follows: algD 78.6%, pelF 70.5%, pslD 36.6%, Ppgl 0%, and PAPI-1 77.6% [18]. Notably, the presence of the PAPI-1 gene in all P. aeruginosa isolates in our study supports the findings of Qiu et al., which suggest that this large pathogenicity island can be transmitted between P. aeruginosa strains [36].

Modifications of the negatively charged phosphate groups of lipid A through adding phosphoethanolamine mediated by the mcr gene can lead to polymyxins resistance [35]. However, our study did not find evidence of the involvement of this plasmid-borne gene in the emergence of colistin-resistant P. aeruginosa strains. Similar results were obtained in Tabriz, Iran, where all colistin-resistant Gram-negative isolates, including P. aeruginosa strains, tested negative for mcr genes [37]. Additionally, another study conducted in Ardabil on clinical isolates of colistin-resistant A. baumannii also demonstrated the absence of this gene (data not published). One possible explanation for the 0% prevalence of the mcr-1 gene in our study is its association with Enterobacteriaceae, particularly Escherichia coli, which are resistant to colistin and commonly isolated from animal sources [37].

Multiple studies have confirmed that mutations in the components of TCSs, namely PmrB and PhoQ proteins, play a significant role in the development of polymyxin resistance in P. aeruginosa strains. These mutations result in the upregulation of the arnBCADTEF operon, leading to the substitution of phosphate groups of lipid A with the cationic 4-amino-4-deoxy-L-arabinose in the LPS structure [38]. Several different mutations have been reported in various studies; however, in our study, we observed only the amino acid alteration Tyr345His in the sensor kinase protein PmrB, a component of the PmrAB TCS. This amino acid alteration has also been reported in other studies conducted by Barrow et al. [39], Sellera et al. [40], Lee et al. [41], and Schurek et al. [42]. It appears that the Tyr345His substitution in the PmrB protein is not involved in the activation of the response regulator PmrA through phosphorylation in colistin-resistant P. aeruginosa strains isolated from hospitals in Ardabil. As shown in Table 2, except for strain 141, colistin-resistant P. aeruginosa strains containing the Tyr345His mutation did not exhibit overproduction of the pmrA gene. In our study, nucleotide substitutions in the phoQ gene did not result in amino acid alterations. Therefore, similar to the pmrB gene, there is no association between mutations in the phoQ gene and subsequent overproduction of the phoP gene in the emergence of colistin-resistant P. aeruginosa strains. The overproduction of PmrB and PhoP genes among colistin-resistant P. aeruginosa strains may be attributed to factors other than mutations, such as low levels of magnesium or calcium, low pH, and antimicrobial peptides [34, 35].

Table 2 provides evidence that the development of colistin resistance in P. aeruginosa strains in Ardabil hospitals is a result of multiple factors. Previous reports have suggested that the ParRS TCS in P. aeruginosa is also involved in the emergence of polymyxin-resistant strains by down-regulating the expression of the porin protein OprD [43]. In this study, this resistance mechanism was confirmed in 44.4% of colistin-resistant P. aeruginosa strains. Furthermore, mutations in the ParRS TCS lead to low to moderate levels of resistance to polymyxins by enhancing the production of the MexXY/OprM efflux pump [44]. It is noteworthy that a significant production of efflux pumps, compared to other resistance mechanisms, was observed in 18 colistin-resistant P. aeruginosa strains in the current study: MexAB-OprM 88.8%, MexCD-OprJ 94.4%, MexEF-OprN 11.1%, and MexXY-OprM 83.3%. Goli et al. also demonstrated increased expression of genes encoding the MexAB-OprM and MexXY-OprM efflux pumps in two colistin-resistant P. aeruginosa strains [27].

Limitation of the study

In the current research, there were the following limitations due to insufficient resources: 1) the mutations in the pmrB and phoQ genes were not assessed in all colistin-resistant P. aeruginosa isolates, 2) the role of other TCSs (such as ParRS) in the emergence of polymyxin-resistant strains was not studied. And, 3) all variants of mcr (including mcr-2 to -9) gene were not investigated.

Conclusion

The detection of colistin resistance among clinical isolates of P. aeruginosa in Ardabil hospitals, higher than in other cities in Iran, is a significant finding. Our study suggests that this resistance can be attributed to various mechanisms, including amino acid alterations in TCSs, overproduction of TCSs, down-regulation of porin, and overproduction of efflux pumps. These results indicate that there may be insufficient infection control measures in Ardabil hospitals and a potential issue with the indiscriminate use of colistin in both humans and animals, which can complicate the treatment of P. aeruginosa infections. Therefore, it is recommended to avoid the unnecessary use of this antibiotic to prevent the potential increase in colistin resistance in the future.