Introduction

Multidrug-resistant Gram-negative bacteria (MDR-GNB), such as Klebsiella pneumoniae, are a serious public health concern, particularly infections caused by strains that produce carbapenemase and are only susceptible to a limited number of antimicrobials1,2. Polymyxin antibiotics have historically been used as a last resort to treat infections brought on by Enterobacteriaceae that are resistant to the antibiotic carbapenem.

Polymyxin antibiotics, including colistin (also known as polymyxin E) are cationic antimicrobial peptides that bind to the lipid A phosphate moiety of bacterial lipopolysaccharide (LPS), resulting in leakage of intracellular components from the cell membrane3. However, the emergence of colistin resistance in GNB has been reported in several countries, with resistance mediated via genetic variations represented by chromosomal mutations in genes involved in lipopolysaccharide synthesis, namely phoP/phoQ, pmrA/pmrB or crrA/crrB as well as on the mgrB regulatory gene4,5. K. pneumoniae's acquired colistin resistance has been linked to the inactivation of the mgrB gene, with several genetic events causing changes to these genes in both human and animal isolates6. These mutations lead to overexpression of the genes and an increased synthesis of phospho- ethanolamine (pEtN) and 4-amino-4-deoxy-l-arabinose (LAra4N)7.

In 2016, Yi-Yun Liu et al. discovered a plasmid-mediated colistin resistance gene (mcr-1) encoding a lipid A phosphoethanolamine transferase that confers resistance to colistin by transferring pEtN to lipid A8. Colistin-resistant Enterobacteriaceae, particularly K. pneumoniae, that contain the mcr-1 gene have been reported from humans, animals used for food production, and the environment worldwide, raising the possibility of horizontal transmission of colistin resistance9. This has raised concerns about the potential emergence of pandrug resistance in Enterobacterales. As a result, it is important to continuously and precisely examine how the mcr genes emerged and propagated across bacteria. A systematic review and meta-analysis on the prevalence of colistin resistance of K. pneumoniae isolates in Iran revealed that the pooled prevalence of colistin resistance in clinical isolates was 6.9%10. However, the rate of K. pneumoniae carbapenem resistance was more than 73% in different studies11,12.

The development of reliable techniques for the detection of polymyxin resistance, with low cost and feasibility are necessary. Simner and colleagues described the Colistin Broth Disk Elution (CBDE), which uses colistin disks as a source of these antibiotics13. This study’s objective was to examine the molecular mechanisms of colistin and carbapenem resistance among a collection of extensively drug resistant K. pneumoniae collected from clinical specimens in Tehran, Iran. Also, a comparison was made in susceptibility of K. pneumoniae isolates to colistin using Disk diffusion, Chrome Agar, E-test, and CBDE.

Materials and methods

Bacterial strains

From June 2017 to March 2019, a total of 162 non-duplicate strains of K. pneumoniae were isolated from clinical samples of inpatients and outpatients at Milad Hospital in Tehran, and they were resistant to the colistin in the initial screening. The disc diffusion method and selective agar medium CHROMagar COL-APSE (Paris, France) were used to detect resistance to colistin and finally colistin resistant isolates that were confirmed by both methods were included in the study.

Antimicrobial susceptibility testing

Antimicrobial susceptibility testing was performed on Muller– Hinton agar plates using the standard disk diffusion method according to the Clinical and Laboratory Standards Institute (CLSI, 2022) guidelines14. A total of 10 antibiotics including cefotaxime (CTX: 30 μg), ceftazidime (CAZ: 30 μg), piperacillin/tazobactam (TZP: 100/10 μg), amikacin (AN: 30 μg), gentamicin (GM: 10 μg), fosfomycin (FOX: 200 μg), imipenem (IMP: 10 μg), meropenem (MERO: 10 μg), ciprofloxacin (CIP: 5 μg), and trimethoprim/sulfamethoxazole (SXT: 1.25/23.75 μg) were investigated. Escherichia coli ATCC 25,922 was used as a quality control strain.

Minimum inhibitory concentrations (MICs) of colistin were determined using E-test strips (bioMérieux, Craponne, France) and CBDE13. Isolates with an MIC > 4 µg/L by CBDE were categorized as resistant.

Molecular analysis of carbapenem and colistin resistance-associated genes

DNA extraction was performed commercial High pure PCR template preparation kit (Roche Molecular Biochemicals, IN, USA) according to the manufacturer’s instructions. Carbapenemases-encoding genes such as (blaKPC, blaIMP, blaVIM, blaOXA-48 and blaNDM-1) were detected by singleplex PCR15,16,17,18,19. Singleplex PCRs were also used to test the colistin-resistant isolates for the presence of mcr-1–9 according to previous published papers20,21,22,23,24. In addition, using specific primers that targeted the mgrB coding sequence and several flanking areas, PCR analysis of mgrB was carried out25. The sequence of primers used in this study, which were selected based on common variants, is shown in Table 1. Positive strains for blaNDM, blaVIM, blaKPC and blaOXA-48 that confirmed by sequencing method were used as positive control and E. coli ATCC 25922 was utilized as negative control in the PCR assays.

Table 1 Primers used in this study.
Full size table

Genomic sequencing

Draft genome sequences were created using genomic sequencing on an Oxford Nanopore GridION in National Reference Laboratory.

Statistical analysis

Data analysis was performed using IBM SPSS Statistics v.22.0 (IBM Corp., Armonk, NY). The frequency of isolates between groups was compared by χ2 test. A P-value of ≤ 0.05 was considered statistically significant.

Results

Epidemiology of colistin-resistant K. pneumoniae isolates

During the study periods, 162 colistin resistant strains of K. pneumoniae belonging to 94 (58%) men and 68 (42%) women were studied. The average age of the patients was 67.25 ± 1.2 (between 19 and 98 years) and most of the studied patients 103 (63.6%) were older than 65 years. Out of 162 patients, 155 were inpatients and only 7 were outpatients. The largest number of studied samples included trachea (41, 25.3%), urine (37, 22.8%) and sputum (35, 21.6%) samples, more than half of which (95, 58.6%) were received from the intensive care unit (ICU). The demographic characteristics of patients, sample types and distribution of study isolates among the different hospital wards are shown in Table 2.

Table 2 Sources of strains isolation.
Full size table

Antimicrobial susceptibility testing

79.6% of the tested strains were resistant to cotrimoxazole, and except for amikacin, resistance to other antibiotics was observed above 90% (Fig. 1). Out of 162 K. pneumoniae isolates, 161 (99.3%) were resistant to ciprofloxacin. The resistance rate of the third generation of cephalosporins including ceftriaxone and ceftazidime was 98.8%. Resistance to carbapenems, which are used as alternative antibiotics in the treatment of strains resistant to cephalosporins, was high, so that 94.4% of the tested strains were resistant to imipenem and 96.3% were resistant to meropenem. From aminoglycoside antibiotics, gentamycin and amikacin were tested, and the resistance to gentamycin was about 90%, while the resistance to amikacin was about 44%, and 12% of the strains had intermediate resistance to this antibiotic. Resistance of K. pneumoniae strains to tazobactam-piperacillin was 92%. Colistin resistance (MIC > 4 µg/L) was observed in 161 isolates (99.4%) by CBDE (Table 3). To compare the antimicrobial sensitivity test using chrome agar, E-test and CBDE methods in colistin resistant K. pneumoniae isolates, 95.6% of the methods used in this study overlapped to detect colistin sensitivity. Meanwhile, a big difference was observed with the disc diffusion method. 50 (30.8%) of the isolates that showed resistance to colistin by other studied methods, showed a sensitive phenotype by disc diffusion method.

Figure 1
figure 1

Results of Antimicrobial susceptibility testing for different antibiotics in colistin resistant K. pneumoniae isolates. SXT Trimethoprim/Sulfamethoxazole, FOX Fosfomycin, CTX cefotaxime, CAZ Ceftazidime, IMP Imipenem, GM Gentamicin, MERO Meropenem, CIP Ciprofloxacin, AN Amikacin, TZP Piperacillin-Tazobactam.

Full size image
Table 3 Comparison of colistin resistance results obtained with different methods in K. pneumoniae isolates.
Full size table

Molecular analysis of carbapenem and colistin resistance-associated genes

The gene encoding the KPC enzyme was the most common and was identified in 95 strains (58.6%), followed by IMP and VIM in 47 (29%) and 23 (14.2%) strains, respectively. The gene encoding the OXA-48 enzyme was found in 12 strains (7.4%), while the NDM-1 gene was not detected.

PCR screening for plasmid-mediated colistin resistance genes (mcr-1 to mcr-9) were examined on tested strains. According to the findings, no positive mcr isolates were found. Therefore, all the strains were examined for the presence of the mgrB gene, and 152 (92.6%) strains harboured the gene. Genomic sequencing of strains harboured mgrB gene revealed that only the conversion of codon 39 from TCA to TGA was present in two isolates, since the TGA codon is a stop codon, which causes the termination of transcription and the loss of protein function (Accession number KF852760: m colistin-susceptible K. pneumoniae strain; coverage:53%).

Discussion

The global increase in multidrug-resistant K. pneumoniae strains has increased the use of colistin to treat these infections, resulting in the emergence of colistin resistance worldwide26,27. The important concern that must be considered in the management of nosocomial infections caused by K. pneumoniae are periodic surveillance to identify the resistant strains for optimizing available infection control policies and treatment options in different parts of the hospitals28. This study investigated the molecular mechanisms of colistin and carbapenem resistance among a collection of extensively drug resistant K. pneumoniae clinical isolates in Tehran, Iran. Also, different phenotypic methods including disc diffusion, E-test, chrome agar and CBDE were compared to determine the susceptibility of K. pneumoniae isolates to colistin.

In the comparison between the phenotypic methods, the highest number of colistin-resistant strains was identified with the CBDE method. So that out of 162 tested strains, except for one strain, the other isolates were resistant to colistin by the CBDE method, then the E-test and chrome agar methods showed the highest resistance (95.6% overlap to detect colistin resistance), and finally, the disc diffusion method showed the lowest resistance. Some studies have compared agar dilution, E-test and disk diffusion methods to measure colistin resistance in Gram-negative bacilli29,30. According to CLSI and EUCAST guidelines, they confirmed that the use of diffusion-based methods for antimicrobial susceptibility testing against colistin was unreliable, which is due to the large size of the colistin, prevents its uniform diffusion in agar-containing media14. Despite this recommendation, the results of this study showed that the disk diffusion method performed with both types of commercial disks containing colistin was able to successfully differentiate resistant strains. According to this result, it seems that revisions should be made about the diagnostic value of the disc diffusion method to determine the susceptibility to colistin and this test should not be completely abandoned. Because disc diffusion method is able to identify at least strains with high level of resistance to colistin and can be an important tool in the direction of rapid screening of resistant strains with high resistance level. CBDE is a simple and low-cost phenotypic method to test antimicrobial sensitivity against colistin in Gram-negative bacilli, including Enterobacteriaceae, which has been approved by CLSI and EUCAST. Studies have shown that CBDE method was comparable with the broth microdilution method as a reference method with a 100% correlation13,31.

The colistin-resistant rate in Iran were reported about 11.6%32. Nevertheless, data from reports of neighboring countries showed the resistance to colistin are ranging from 0 to 31.7%33. As expected for colistin-resistant isolates in our study, most were also resistant to other clinically useful antimicrobial agents. In the present study, 96% of isolates showed phenotypic resistance to meropenem and/or imipenem, which similar to many other studies, 95 of them harboured blaKPC gene34. We found that none of the colistin-resistant K. pneumoniae isolates had plasmid-encoded mcr-genes, suggesting that the resistance is mediated by chromosomally encoded mechanisms. MCR-1 is still extremely uncommon in clinical isolates worldwide35. In previous investigations, mcr-1 prevalence in Enterobacteriaceae was reported to range between 0.1 and 1%36,37. The literature shows the prevalence of the mcr-1 gene was lower in Enterobacteriaceae strains isolated from human sources than in strains isolated from animal and food samples38. This assumes that their reservoir is at least in animals and the environment, following the important use of colistin in animal production and in general agriculture. In Iran, the massive use of colistin in clinical practice following the spread of carbapenemase-producing Enterobacteriaceae has led to the selection of multidrug-resistant bacteria in hospital settings39.

Interestingly, the results of the current study showed that mcr-1- negative K. pneumoniae isolates had high level of colistin resistance. Consistent with this findings, previous studies reported that K. pneumoniae with chromosomal mutations in mgrB exhibited a high level of colistin resistance40,41. The mechanisms of colistin resistance other than those attributed to the mcr-1 gene in the K. pneumoniae isolates in this study are being further studied to define the precise molecular mechanism of resistance. Of 162 colistin resistant K. pneumoniae isolates, 150 (92.6%) isolates had mgrB. Critical changes in mgrB, such as disruption of the promoter or coding sequence, are thought to cause the gene to be silenced or lead to the generation of shortened mgrB40. In reality, PhoP/PhoQ activation follows mgrB inactivation by any of these occurrences, which in turn activates the PmrA response regulator responsible for modification of the lipopolysaccharide polymyxin target25. So mentioned mutation in the present study may causes the termination of transcription and the loss of protein function. Avgoulea et al. reported that insertional inactivation of the mgrB gene conferred resistance to colistin in all isolates tested42.

There were some limitations related to the present study. Firstly, it should be noted that this study was performed using extracted DNA samples from K. pneumoniae that were selected based on the results of initial screening for colistin resistance. The lack of strains carrying mcr genes except mcr-1 as a positive control is another limitation. From an epidemiological standpoint, the monocentric nature of the study is a significant limitation.

Conclusion

Our findings indicate that the colistin resistance of K. pneumoniae isolates may be associated with mgrB gene mutation. These data provide added insight into the mechanism of colistin. Colistin resistance developed with a number of new mutations among highly resistant populations, limiting the availability of further antimicrobial medicines and resulting in pandrug resistance. The prevalence of resistance to carbapenems and colistin in Iran should be surveyed and new therapeutic strategies including old drugs should be evaluated and used in Iran.