Abstract
Objectives:
We aimed to describe the mechanisms of colistin resistance in Acinetobacter baumannii.
Materials and Methods:
Twenty-nine patients diagnosed with colistin-resistant A. baumannii infection were included to the study. The mutations in pmrCAB, lpxA, lpxC, and lpxD genes, expression of pmrCAB, carbapenemases, and mcr-1 positivity were studied.
Results:
Twenty-seven (93%) of the patients received IV colistin therapy during their stay, and the case fatality rate was 45%. All mutations in pmrC and pmrB were found to be accompanied with a mutation in lpxD. The most common mutations were I42V and L150F in pmrC (65%), E117K in lpxD (65%), and A138T in pmrB (58.6%). The colistin minimum inhibitory concentrations (MICs) of the isolates having any of these four mutations were higher than the isolates with no mutations (p < 0.001). The two most common mutations in pmrC (I42V and L150F) were found to be associated with higher expressions of pmrA and pmrC and higher colistin MIC values (p = 0.010 and 0.031). All isolates were blaOXA-23 positive.
Conclusion:
Coexistence of the lpxD mutation along with mutations in pmrCAB indicates synergistic function of these genes in development of colistin resistance in A. baumannii.
Introduction
Acinetobacter baumannii is an important hospital-acquired pathogen associated with severe respiratory tract, bloodstream, urinary tract, surgical site, and wound infections. In recent years, a substantial increase in infections caused by multidrug-resistant A. baumannii strains was reported in several countries.1–3
Colistin is a member of polymyxin group of antimicrobials, which is used as a last treatment option for infections caused by carbapenem-resistant bacteria. Unfortunately, colistin resistance among A. baumannii strains has been increasing over time mainly due to the increasing usage of colistin to treat patients with carbapenem-resistant bacterial infections.4,5 A study reported the global colistin resistance rate below 7% in most of the countries, however, some countries exhibit high resistance rates; 40.6% and 30.6%, in Spain and Korea, respectively. 5 In Turkey, a multicentered surveillance study reported 6% of colistin resistance in A. baumannii. 6
There are two main mechanisms inducing colistin resistance in A. baumannii. The first mechanism is the reduction of the binding affinity of colistin to lipopolysaccharide (LPS) due to the cationic modifications on LPS as a result of mutations in pmrAB two-component system and pmrC.7,8 The second mechanism is the complete loss of LPS as a result of mutations or insertions in lpxA, lpxC, and lpxD genes, which are involved in lipid A biosynthesis pathway. 6 In 2016, plasmid-mediated colistin resistance gene named mcr-1 was reported for the first time, which raised a global concern.9,10 Until now, there is no report of mcr-1 presence in A. baumannii. 11
Despite the global influence of the problem, the clinical, microbiological, and molecular characteristics of colistin-resistant A. baumannii infections need to be detailed. In this study, we aimed to identify the clinical and molecular factors that might induce colistin resistance in A. baumannii.
Materials and Methods
Study population and data collection
A total of 29 patients with colistin-resistant A. baumannii infection from four different hospitals located in Turkey (12 from Menemen Hospital [Izmir], 14 from Baskent University Hospital [Ankara], 2 from Kosuyolu State Hospital [Istanbul], and 1 from American Hospital [Istanbul]) between April 2015 and June 2016, were included. A study protocol in accordance with the criteria suggested by Strengthening The Reporting Of Molecular Epidemiology for Infectious Diseases was designed. 12 Patients' demographic data, underlying diseases, type of infection, isolation site, predisposing factors such as having operation within last one month, intensive care unit admission, type of antimicrobial agents used for empirical and agent-specific therapy, duration of colistin therapy before isolation of colistin-resistant isolates, carbapenem resistance, and 30-day fatality were recorded on a standardized case form. Exclusion criteria were missing key data, colonization, and subsequent episodes of the same patient. The study was approved by Koç University Institutional Review Board by the protocol number of 2015.048.IRB1.008.
Antimicrobial susceptibility testing and detection of carbapenemase encoding genes
The resistance to colistin was diagnosed by E-test (BioMérieux, France) in routine laboratories, and colistin minimum inhibitory concentrations (MICs) were determined by broth microdilution in reference laboratory at Koc University. The resistance breakpoint was set as >2 mg/L. 13 MICs of carbapenems were tested by broth microdilution. 13 Carbapenamase type was identified by multiplex PCR using blaOXA-23, blaOXA-24, and blaOXA-58 primers described previously. 14
Genotyping of the isolates
The genetic relatedness of colistin-resistant isolates was determined by repetitive extragenic palindromic PCR (Biomerieux, France). Isolates more than 95% similarity were considered as clonally related.
Molecular mechanisms of colistin resistance
The mutations on pmrCAB, lpxA, lpxC, and lpxD genes were studied by Sanger sequencing (Applied ABI3500; Applied Biosystems) using primers described previously.6,15 The sequence reads of 1,179 bp lpxA, 1,164 bp lpxC, 1,502 bp lpxD, and 3,699 bp pmrCAB were aligned with corresponding genes of ATCC 19606, ATCC 17978 standard strains, and colistin susceptible two clinical isolates by Applied Maths BioNumerics (version 7.5). Expression studies of pmrCAB were carried out by quantitative real-time (qRT)-PCR using previously described primers (LightCycler 480, Roche, Germany).6,15 For housekeeping gene 16S rRNA was used. A. baumannii ATCC 17978 strain was selected for normalization of Ct values. The presence of plasmid encoded mcr-1 was examined by PCR using primers described by Liu et al. 10
Statistical analysis
Statistical analysis was performed using the R. 16 In the heatmap, normalized −ΔΔ signal values of probes and hierarchical clustering with complete linkage to obtain dendrograms of rows and column were used. Wilcoxon rank-sum test (also known as the Wilcoxon–Mann–Whitney test or the Mann–Whitney U test) was used to make statistical comparisons between two samples. All results of statistical analysis are available at https://midaslab.shinyapps.io/colistin_resistant_acinetobacter_analysis/.
Results
A total of 29 patients infected with colistin-resistant A. baumannii were included in this study (Table 1). Twenty-seven (93%) of the patients received IV colistin therapy during their stay in the indicated hospitals with a median duration of 7 days. Seventeen of the 29 patients (59%) had diagnosis of pneumonia, and 5 (17%) patients had bacteremia. The 11 patients (38%) had a history of operation within the last month, 6 had (21%) diabetes, and 7 (24%) patients had cardiovascular diseases. The fatality rate was detected as 45% (Table 1). All isolates were found to be resistant to carbapenems and were blaOXA-23 positive. The MIC50 of colistin was 32 μg/mL (range: 2–256 μg/mL). None of the isolates was positive for mcr-1 gene.
Characteristics of the Patients Infected with Colistin-Resistant Acinetobacter baumanni
In analysis of pmrCAB, nine nonsynonymous mutations were found in pmrB (A138T, R263H, H89L, I163F, I164L, A224V, E229D, P233S, and A408E), four in pmrC (I42V, R109H, I115N, and L150F), and one in pmrA (M12I). In the lpxD, three nonsynonymous mutations (Q4K, V63I, and E117K) were detected (Table 2). No mutation was detected in lpxA and lpxC genes. The most common amino acid variations were I42V and L150F in pmrC (65%), E117K in lpxD (65%), and A138T in pmrB (58.6%).
Mutations in the pmrC, pmrA, pmrB, and lpxD
In all isolates, any of the mutations in pmrCAB region was accompanied with a lpxD mutation (Fig. 1). The combination of pmrC I42V, pmrC L150F, and lpxD E117K was detected in 19 isolates, and 16 of them (55.2%) were positive for A138T in pmrB as well (Fig. 1). The colistin MICs of the isolates having any of these four mutations were significantly higher than the isolates without mutations (median of 12 μg/mL vs. 3 μg/mL, p < 0.001) (Supplementary Data).

Heatmap of pmrCAB and lpxD mutations in colistin-resistant Acinetobacter baumannii. Color images are available online.
The relative expressions of pmrC, pmrA, and pmrB were found to be increased in 28 (97%), 25 (87%), and 26 (90%) of the 29 isolates with mean values of 35.3-, 12.3-, and 8.2-fold, respectively. Mutations in pmrC strongly stimulated expression of pmrCAB genes. The two most common mutations in pmrC (I42V and L150F) were found to be associated with higher expressions of pmrA and pmrC and higher colistin MIC values (p = 0.010 and 0.031, respectively). Furthermore, I115N in pmrC was associated with overexpression of pmrA, pmrB, and pmrC genes (p = 0.003, 0.022, and 0.008, respectively). The most common mutations in pmrB (A138T) and lpxD (E117K) were associated with increased expression of pmrA gene (p = 0.04 and p = 0.01).
Analyses regarding clinical factors showed that isolates with the group of four mutations were significantly more common among the cases, who had operation within last month (p = 0.003). Operation within last month was also found to be associated with overexpression of pmrB and pmrC genes (p = 0.039 and 0.044, respectively). There was no association between any of mutations and colistin use, bacteremia, pneumonia, and fatality (Supplementary Data).
The dendogram showed that 12 isolates from three different hospitals were clustered with >95% similarity. However, pmrC, pmrA, and pmrB expression patterns of these clonal isolates were diverse in expression dendogram (Supplementary Data).
Discussion
We present the genetic mechanisms of colistin resistance in A. baumannii infection in correlation with clinical factors. The possible genetic mechanisms of colistin resistance in A. baumannii have been reported by previous studies with limited number of clinical isolates or laboratory-derived mutants.8,15,17,18
In A. baumannii, mutations in pmrCAB operon, especially in pmrB region are known as the major cause of colistin resistance. 6 Various mutations and small in-dels in different domains of the pmrB gene were reported as the most common cause of colistin resistance in A. baumannii.16,19 In this study, the number of mutations were the highest in pmrB region. The most common pmrB mutation was A138T, but it was in the site of unknown function. The active site mutations in pmrB, including P233S, were seen only in 12 (41%) of the isolates. pmrC encodes an inner membrane protein responsible for the incorporation of phosphoethanolamine into lipid A. Mutation of pmrC results in synthesis of lipid A that lacks phosphoethanolamine. 20 In our study, 65% of the colistin resistant isolates had I42V and L150F and 28% had R109H mutations in the pmrC. All three mutations were on the transmembrane domain of the protein, and they were found to be significantly associated with higher colistin MIC values (p < 0.01). In a study from Greece, L150F and R109H in pmrC were detected both in colistin susceptible and resistant A. baumannii isolates. 21 In Salmonella, mutations in pmrC were shown to increase the MICs of colistin. 20 These findings suggested that additional supplementary variations on the target of colistin may lead to development of resistance. The additional outer membrane modifications such as vacJ and pldA mutations were shown to have a role on colistin resistance among laboratory-derived colistin-resistant strains. 22
In our study, E117K mutation of lpxD was found in 65% of the isolates, and it was significantly associated with high colistin MIC values (p = 0.02). Mutations and/or insertions in any of lpx genes may cause termination of Lipid A synthesis and LPS formation.8,23 In a previous study, E117K was found in both colistin susceptible and resistant Acinetobacter isolates. 21 All these results suggested us that the mutations in pmrCAB or lpxD genes alone are not sufficient to induce colistin resistance in A. baumanni. Occurrence of mutations in both of the regions, which encodes different part of cell wall lipids, might have synergistic effect in promoting colistin resistance.
Overexpression of pmrC was considered as a major cause of colistin resistance because of impairment of the binding of polymyxin to the outer membrane,6,15,18 and mutations in pmrCAB genes were shown to be responsible for the overexpression of pmrC gene. 19 In the expression analyses, we found that overexpression of pmrC was positively correlated with higher MIC values (Rho = 1.41) (Supplementary Data) and associated with the mutations of I42V, I115N, and L150F in the pmrC gene (p = 0.031, 0.0085, and 0.031, respectively).
The comparison of the DNA and expression dendograms showed that isolates of the same clone have different expression patterns of pmrC, pmrA, and pmrB genes (Supplementary Data). This finding verifies the importance of environmental and host factors on gene expression behaviors. Studies so far demonstrated that colistin use is the major factor for emergence of chromosomal colistin resistance. 24 Our patients, except two, were treated with colistin, and so, we could not analyze the effect of colistin exposure on resistance. In our study, the case fatality rate was detected as 45%, and it was reported to be between 30% and 69% by other studies.25,26
In conclusion, mutations in pmrC or pmrB were found to be associated with colistin resistance in A. baumannii. The presence of a lpxD mutation along with mutations in pmrCAB indicates synergistic function of these genes in development of colistin resistance.
Ethical Approval
This study was approved by Koç University Ethics Committee, and the protocol number was 2015.048.IRB1.008.
Footnotes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
