Detection of Plasmid Mediated Colistin Resistant mcr-1 Gene Among Escherichia Coli and Klebsiella Pneumoniae Isolated From Clinical Samples

Background: The prevalence of antimicrobial resistance (AMR) among Gram-negative bacteria is alarmingly high. Reintroduction of colistin as last resort treatment in the infections caused by drug-resistant Gram-negative bacteria has led to the emergence and spread of colistin resistance. This study was designed to determine the prevalence of drug-resistance among beta-lactamase-producing strains of Escherichia coli and K. pneumoniae, isolated from the clinical specimens received at a tertiary care centre of Kathmandu, Nepal during the period of March to August, 2019. Methods: A total of 3216 different clinical samples were processed in the Microbiology laboratory of Kathmandu Model Hospital. Gram-negative isolates (E. coli and K. pneumoniae) were processed for antimicrobial susceptibility test (AST) by using modied Kirby-Bauer disc diffusion method. Drug-resistant isolates were further screened for extended-spectrum beta-lactamase (ESBL), metallo-beta-lactamase (MBL), carbapenemase and Klebsiella pneumoniae carbapenemase (KPC) production tests. All the suspected enzyme producers were processed for phenotypic conrmatory tests. Colistin resistance was determined by minimum inhibitory concentration (MIC) using agar dilution method. Colistin resistant strains were further screened for plasmid-mediated mcr-1 gene using conventional polymerase chain reaction (PCR). Results: Among the total samples processed, 16.4% (529/3216) samples had bacterial growth. A total of 583 bacterial isolates were recovered from 529 clinical samples. Among the total isolates, 78.0% (455/583) isolates were Gram-negative bacteria. The most predominant isolate among Gram-negatives was E. coli (66.4%; 302/455). In AST, colistin, polymyxin B and tigecycline were the most effective antibiotics. The overall prevalence of multidrug-resistance (MDR) among both of the isolates was 58.0% (199/343). In the ESBL testing, 41.1% (n=141) isolates were conrmed as ESBL-producers. The prevalence

enzyme (discovered in late 2015), which modi es the outer membrane lipopolysachahrides by adding pEtN to the phosphate groups in Lipid A thereby decreasing the net negative charges [2]. The resulting modi cation reduces the a nity of polymyxins to the target of the bacterial cell [16,17]. Unlike chromosomal mutation, acquisition of mcr is a matter of serious concern because of its potential transferability, as the gene is spread rapidly through the horizontal transfer at a higher rate than occurring through spontaneous mutation [18]. In addition, plasmids can transfer resistance to multiple classes of antibiotics [19,20]. Elevated endemicity of the mcr genes all over the world in a short span of time is attributable to their ability to proliferate at a higher pace [21].
Since mcr gene was rst isolated in an E. coli from animal sources in China, this nding indicates that the plasmid-mediated colistin resistance may have transmitted from animals (colistin was extensively used as growth promoters for long times) to humans [22]. E. coli is the most prevalent species harbouring the mcr gene, accounting for approximately 91% of the entire load of mcr-positive bacteria, which is followed by Salmonella enterica (~ 7%) and K. pneumoniae (~ 2%) [23]. Higher burden of mcr among S. enterica than K. pneumoniae also supports the fact that the former is the food-borne pathogen and is very likely to be transmitted via food chain [24]. Moreover, these drug-resistant bacteria are isolated from humans, animals, and environments so that the perspective of 'one health' has been jeopardized [25].
Implementation of effective surveillance programs and infection controls are considered as the two pillars to check the growth and spread of AMR [26]. However, in the developing countries like Nepal, circulation and co-circulation of resistant genes may go undetected, underreported and poorly characterized due to poor diagnostic facilities [27,28]. In addition, irrational use of antibiotics among humans and animals (often as growth promoters) is putting pressure of potential outbreaks in the future [10]. Moreover, there are a limited number of studies on colistin resistance and the prevalence of resistance can vary and change over the time within and between the countries. Therefore, this study was conducted in a tertiary care center with an attempt to establish the prevalence of beta-lactamases including ESBL, MBL, KPC and colistin resistance among Gram-negative MDR pathogens. At the same time, we also aimed to explore the possible role of mcr genes in conferring the resistance to colistin. Furthermore, antibiogram of the resistant strains to a variety of antibiotics was carried out to recognize the possible therapeutic option for combating superbugs.

Study design and study sample
This cross-sectional study was conducted from March to August, 2019 at Kathmandu Model Hospital, Kathmandu, Nepal. A total of 3216 clinical samples consisting of urine (n=1776), blood (n=875), pus (n=156), sputum (n=187), body uids (n=88), wound swab (n=51), tissue (n=18), catheter and other tips (n=12), and other samples including stool, urethral and vaginal swabs, bone, and bone marrow aspirate (n=53) was collected and processed during the study period. Patients of all age-groups and genders who were admitted in or visiting the study hospital (for treatment) were included in this study. All the samples with completely lled demographic information having no visible signs of contamination were included in the study. However, others were rejected and requested for repetition, if possible.

Sample Collection and transport
All the samples were aseptically collected following the standard microbiological procedure. Individual collection procedures varied in accordance to the type of samples. Generally, samples were collected in a dry, wide-mouthed, and leak-proof container and were sent to microbiology department without delay. In case of unwarranted delay, clinical specimens were refrigerated at 4°C to 6°C.
Culture, isolation and identi cation of bacteria Each sample was processed by following the standard microbiological guidelines [29,30].
Urine sample: Urine samples were inoculated into Cysteine Lactose Electrolyte De cient (CLED) agar using sterile and standard calibrated loop. The plates were incubated at 37°C overnight.
Blood and (endotracheal and catheter) tips: Blood specimen was inoculated aseptically onto BHI broth at the ratio of 1: 10 and was incubated at 37°C for 7 days and routinely inspected twice a day for at least rst three days for microbial growth. Then broth from the culture bottle showing visible signs of microbial growth was sub-cultured in Blood agar (BA), MacConkey agar (MA) and Salmonella-Shigella agar (SS) and plates were incubated at 37°C for 24 hrs [31].
Sputum and throat swab: Sputum samples were inoculated into BA, CA and MA plates. For sputum, in CA plate a 5 μg optochin disc and a 10U bacitracin disc were added to screen out S. pneumoniae and H. in uenzae respectively whereas for throat swab, 0.05U bacitracin disc was added to the plate to screen Streptococcus pyogenes. Chocolate agar and BA were incubated at 37°C overnight in 5-10% CO 2 environment whereas the MA plate was incubated at 37°C in an aerobic condition [31].
Pus, pus swab, and wound swab: These samples were inoculated into BA and MA plates, and inoculated at 37°C for 24hrs. In case of swab, an initial inoculum was made by rubbing the swab over the media plate in order to transfer maximum number of organisms. Then the streaking was performed [31].
Body uids: Body uid samples were centrifuged before culture. The sediment after centrifugation was inoculated into BA, MA and CA plates. The BA and CA plates were incubated in a 5-10% CO 2 enriched atmosphere and MA plates were incubated aerobically at 37°C overnight [31].
Identi cation of the isolates Following incubation, culture plates were observed for possible microbial growth. Isolates were presumably identi ed on the basis of Gram's staining and colony characteristics. Further con rmation of the isolates were based on biochemical tests such as IMViC (Indole production, Methyl red test, Voges-Proskauer test and Citrate utilization), H2S production, catalase test, coagulase test, and oxidase test [30,32].
Screening of the ESBL production ESBL-producers were screened by using Ceftazidime (30µg) and Cefotaxime (30µg) in the AST. Isolates showing reduced susceptibility to one or both of these drugs with diameter of the zone of inhibition (ZOI) for ceftazidime ≤22 mm and cefotaxime ≤27 mm were considered as potential ESBL-producers [33].
Suspected strains were further processed for con rmatory assay.

Con rmation of the ESBL-producers by phenotypic method
Combination disc test (CDT) as prescribed by the CLSI was the choice of protocol for con rmation of the ESBL-producing strains. In this method, cefotaxime (30µg) and ceftazidime (30µg) discs alone and in combination with clavulanic acid (10µg) ( ceftazidime plus clavulanic acid, 30/10 µg and cefotaxime plus clavulanic acid, 30/10 µg) were used. The ZOI of cephalosporin discs alone was compared with their respective cephalosporin/clavulanic acid (combined) discs. An increase in ZOI by ≥5mm in the presence of clavulanic acid was considered as con rmed ESBL production test [33].
Phenotypic con rmatory test for carbapenemase and/or KPC producers Inhibitor-based method was followed for the con rmation of carbapenemase and KPC production. In this method, combined disc test of carbapenem with and without PBA was employed. An increase in the diameter of ZOI by ≥ 5mm in combined disc (carbapenem disc supplemented with PBA) than single disc (only carbapenem disc) was considered as con rmed test for carbapenem or KPC production [35,36].
Phenotypic con rmatory test for MBL production Similar to carbapenemase production test, con rmation of MBL production was made by inhibition method in which Ethylene Diamine Tetra Acetic Acid (EDTA) was used as an inhibitor. Two imipenem (10µg) discs were placed on MHA and 10µl of 0.5 M EDTA solution was added to one of the discs to obtain the desired concentration. After overnight incubation, an increase in the diameter of ZOI by 7 mm in combined disc (imipenem disc supplemented with EDTA) than single one (only imipenem disc) was considered as con rmed test for MBL production [37].

Determination of MIC of colistin
Agar dilution was used to determine the minimum inhibitory concentration of colistin. In this method, different concentrations of colistin ranging from 2 µg/ml to 32 µg/ml were prepared in the agar medium. Bacterial inoculum was then applied readily onto the agar surface and the plates were incubated at 37°C up to 18 hours. The MIC end point was determined as the lowest concentration of antibiotics that completely inhibits the visible growth. Isolates having a MIC of ≤4 μg/mL is considered colistin susceptible while MIC of >4 μg/mL is considered colistin resistant [22,38].

Plasmid and genomic DNA extraction
Plasmid DNA of E. coli and K. pneumoniae was extracted by using phenol-chloroform method [39]. Firstly, isolated colonies of bacteria were inoculated in Luria-Bertani (LB) broth and incubated overnight at 37 o C in the presence of air. After required incubation, alkaline-lysis method was adopted to extract the DNA. Extracted plasmid DNA was then suspended in TE buffer and preserved at −20°C until further processing [39].

PCR ampli cation of mcr-1 gene
Ampli cation of mcr-1 gene was carried out by conventional PCR using primers: 5′-CGGTCAGTCCGTTTGTTC-3′ as forward primer and 5′-CTTGGTCGGTCTGTAGGG-3′ as reverse primer [2]. A PCR mixture having the nal volume of 25µl (3µl of template DNA, 0.5µl each forward and reverse primes and 21µl of PCR master mix) was used for the reaction mixture.
The thermal condition for ampli cation was initial heat activation of 95ºC for 15 minutes followed by 35 cycles of denaturation at 94ºC for 30 seconds; annealing at 57ºC for 90 seconds; extension at 72ºC for 90 seconds; and nal extension at 72ºC for 10 minutes. The ampli ed products were subjected to gel electrophoresis (2.0% agarose gel stained with µl ethidium bromide) at 100V for 60 minutes and visualized under UV transilluminator [22].

Quality control during antimicrobial susceptibility and MIC assays
Each batch of media, reagents and antibiotic discs were checked for their lot number, expiry date, and proper storage. Similarly, purity plates were used to ensure the pure culture of test organisms. Control strain of E. coli ATCC 25922 was used during AST.

Data analysis
All the data were entered in the worksheet of statistical package for social science (SPSS) software (Version 25) and presented in the form of tables and gures. Chi-square (χ2) test was applied to predict the relationship between the variables. A p value of <0.05 was considered as statistically signi cant. Antibiotic susceptibility pattern of E. coli and K. pneumoniae

Distribution of bacterial isolates
In susceptibility testing, colistin, polymyxin B and tigecycline were most effective antibiotics as all of the tested isolates of E. coli and K. pneumoniae were susceptible towards these drugs. E. coli isolates were also susceptible to antibiotics like nitrofurantoin (93.5%; only for uropathogens), gentamycin (87.8%) and amikacin (89.8%). However, all of the tested E. coli isolates were resistant to azithromycin. Similarly, prevalence of drug-resistance was documented in K. pneumoniae isolates. All K. pneumoniae isolates were resistant to amoxycillin and azithromycin. This was followed by cephalosporin antibiotics: cefepime (57.1%), ceftazidime (55.6%) and cefotaxime (51.2%). None of the isolates was resistant to cipro oxacin, colistin, polymyxin-B and tigecycline (  Figure 3). MIC of colistinresistant isolates of E. coli and K. pneumoniae ranged from ≤2µg/ml to 8 µg/ml and from ≤2µg/ml to 16µg/ml respectively.
This study also screened for harbouring of mcr-1 by colistin susceptible isolates, however, only those isolates having MIC break point 2 µg/ml were included for ampli cation. In this way 18 isolates were taken for further processing, all of which belonged to the isolates of E. coli as none of the K. pneumoniae isolates had the desired MIC. Among the tested isolates, 16.6% (3/18) of them harboured mcr-1 gene.

Discussion
Drug-resistance, especially pan-drug resistance and MDR is emerging as a major challenge in treatment of the infections caused by Gram-negative bacteria. As surveillance of the AMR and early response to the infection control are crucial steps to address the issues, this study also aimed to determine the status of MDR among E. coli and K. pneumoniae isolates and to investigate possible acquisition of mcr-1 in colistin resistant isolates. In this study, most of the isolates were resistant to commonly prescribed broadspectrum antibiotics. In addition, prevalence of colistin resistance and acquisition of mcr-1 among the drug-resistant isolates was also observed in our study.
In this study, 16.4% (529/3216) specimens showed bacterial growth in which prevalence of Gramnegative bacteria from different clinical specimens was much higher than the Gram-positives. This nding concords with previous studies reported from Nepal [9,10,[40][41][42][43][44]. Higher prevalence of E. coli in comparison to other species could be due to being normal ora of human gut which is highly opportunistic in immunocompromised patients. When E. coli reaches out to the tissues other than its common site, it serves as an opportunistic pathogen. A number of virulence factor encoded by pathogenic strains of E. coli enables them to colonize the human body in spite of effective host defence [45].
In this study, all of the isolates of E. coli and K. pnemoniae were susceptible to colistin, polymyxin B and tigecycline, which is comparable to some previous ndings [46,47]. These classes of antibiotics can be effective drugs in management of the Gram-negatives. Conversely, all of the isolates were resistant towards azithromycin. Previous exposure of the isolates to these antibiotics as well as the state of resistance genes of corresponding antibiotics may be the reason for their susceptibility patterns [48].
In this study, increased resistance to third-generation cephalosporins was observed, as more than half of the isolates were non-susceptible to those drugs. Similar ndings have been reported by some previous studies [47,49,50]. Higher rate of resistant to cephalosporin can be attributable to their irrational prescription and uses [51].
Resistant rate of E. coli to uoroquinolones in this study ranged from 47.0-55.0% which is in agreement with earlier studies from Nepal [52,53] and India [54]. Resistance to uroquinolones among MDR Gramnegative bacteria is common and is expected to sustain and perhaps accelerate even if other antibiotics are used [55]. The prevalence of uroquinolones resistance is related to the intensity of antibiotics used which may reduce the e cacy of drug in progressive manner [56].
In this study, almost one fourth of E. coli isolates were resistant to carbapenem antibiotics. The resistance rate towards these antibiotics ranged from < 3.0-21.0% in some of the previous studies from Nepal [50,52,53] whereas 100% sensitivity towards imipenem was reported in some others [47,57]. Production of beta-lactamase enzymes and the upregulation of e ux pump are suggested as the reasons for reduced susceptibility [58]. However, comparatively low resistance to carbapenem antibiotics reported in this study could be due to lower use of these antibiotics in the treatment of the infection [59].
In AST assay of K. pneumoniae, two-third of the isolates were susceptible to uoroquinolones and gentamicin. This nding is similar to another study reported from Nepal [60]. However, some other studies have reported lower susceptibility rates (less than 50.0%) [57,61]. This variation may be due to the difference in the specimens included in the study as well as the exposure of isolates towards the antibiotics. All of the K. pneumoniae isolates were resistant to amoxicillin. Similar ndings were reported in other studies [52,60]. In addition, reduced susceptibility towards cephalosporin and carbapenem antibiotics was observed in this study. High resistant rate towards cephalosporins was also reported in previous studies [50,52,62,63]. Multiple factors such as extensive use of drugs, production of betalactamases, or e ux pumps (which actively pump out these antibiotics) are attributable to the rise in the resistance against carbapenems [26, 64].
Among the total (343) isolates of E. coli and K. pneumoniae, more than half (58.0%) were MDR. MDR strains were predominant among the isolates of E. coli in comparison to K. pneumoniae. This result was in consistent with previous ndings which also reported the rate of MDR in a range of 41.0 %-67.7 % [9,20,28,52,60] while lower than some other ndings [8,10]. In this study, the prevalence of MDR E. coli and K. pneumoniae was 59.3% and 48.8% respectively. Common risk factors associated with development of MDR are poor hygiene, misuse of antibiotics and absence of antimicrobial surveillance program [65,66]. Higher rate of antibiotic resistance among E. coli and Klebsiella spp. is associated with their ability to produce different kind of β-lactamases primarily ESBL, AmpC and MBL, and carriage of resistance trait for quinolones and aminoglycoside in the plasmid [67]. In several hospitals in Nepal, the antibiotics used for the treatment of infected patients are effective in curing only a half of the cases whereas other half of the treatment course shows no response [68]. In addition, development of partial resistance by bacteria, most antibiotics intended to cure people are becoming less effective which might also be the reason of increasing prevalence of MDR reported in this study [26].
In this study, the prevalence of ESBL producing strains was found to be 41.1% among Gram-negative isolates. This result is comparable to some previous reports from Nepal [52,69]. The prevalence of ESBL production was reported low in other studies [70][71][72]. The difference in the prevalence of ESBL production can be partially due to geographical variations, type of specimens processed, and local practices of antibiotics prescription and use [73].
In our study, among 343 isolates of E. coli and K. pneumoniae, 12.5% were resistant to carbapenem. In earlier study reported from Kathmandu Model Hospital, the prevalence of carbapenem resistant ranged from 4.5-20.0% among the members of Enterobacteriaceae [74]. However, higher rate of carbapenem resistant among Enterobacteriaceae was reported in other studies [75,76]. The difference in utilization of carbapenem antibiotics to treat infections in different study settings may be responsible for these variations [75].
Carbapenem resistant isolates were subjected to KPC and MBL production test phenotypically. In this assay, 30.2% and 60.5% of the isolates were KPC and MBL-producers respectively. Our ndings are comparable to a previous nding [77]. The prevalence of KPC production in E. coli and K. pneumoniae was 33.3% and 20.0% respectively in our study. Similarly, MBL production was reported 63.6% in E. coli and 50.0% in K. pneumoniae. In a previous study, KPC production in E. coli was reported as 14.4% and 7.1% in K. pneumoniae [8]. Accordingly, another study reported comparatively lower incidence of MBL producing-E. coli and K. pneumoniae in different clinical samples in Central Nepal [78]. Similarly, lower rate of MBL (9.0%) and KPC (6.5%) production was reported in E. coli [79]. Another study from Iran reported 80.5% of K. pneumoniae isolates as KPC-producers [80]. This study revealed higher prevalence of MBL and KPC production in E. coli and K. pneumoniae which may be due to dissemination of plasmid encoded carbapenems resistance gene [59].
This study reported comparatively low prevalence MIC range of colistin for E. coli ranged from ≤ 2µg/ml to 8µg/ml which was lesser than that of K. pneumoniae (≤ 2µg/ml to 16µg/ml). This nding is in agreement with a previous study from China [90]. MIC range of mcr-1 positive Enterobacteriaceae typically have a moderate level 4-16 mg/ml of colistin resistant strains [91]. MIC of colistin resistant isolates carrying mcr-1 was lower in this study. Exceptionally, MIC range of colistin in colistin resistant K. pneumoniae without mcr-1 in this study was high. This result suggests that colistin resistance in K. pneumoniae might be associated with chromosomal mutations in mgrB, phoP/phoQ, pmrA, pmrB, pmrC and crrABC [92]. High MIC may also be due to strong selective pressure in the isolates. These strains may carry another variant of mcr gene [93].

Strength and limitations
This is the rst study from Nepal which also determined the prevalence of mcr-1 among colistin susceptible isolates. In addition, this study is one among a handful of studies that attempted to investigate the role of the beta-lactamases (ESBL, MBL, and carbapenemase) and acquisition of mcr-1 among Gram-negative isolates. The ndings of this study can be an important reference tool for policy makers and clinicians which can ameliorate the practice of antibiotic prescription and use in the country.
However, this study suffers from several limitations. Firstly, this study was conducted among small population of a tertiary care centre so that the reported rates cover a limited geographical region and may not re ect overall picture of epidemiology across the country. Secondly, this study tested the acquisition of only one variant (mcr-1) of mcr gene; isolation and characterization of all variants (mcr-1 to mcr-9) is suggested in future studies to predict the role of genes in conferring resistance. In addition, this study could not predict the origin and possible transferability of resistant strains. The insertion sequence ISApl1 plays an important role in the mobilization of this mcr-1 gene. However, in this study only mcr-1 gene was analysed. Therefore, further molecular analysis can better predict other possible mechanisms responsible for colistin resistance and role of insertion sequence in dissemination of this gene.

Conclusion
High burden of MDR strains in our study could be due to the pervasive and irrational practices of antibiotic prescription and use, as half of the Gram-negative isolates were found as drug-resistant.
Moreover, the presence of colistin resistance and acquisition of mcr-1 among clinical isolates warrants an imminent threat of no antibiotic era. Therefore, prompt action is recommended for proper infection control. Augmentation of diagnostic facilities, AMR surveillance, and antibiotic stewardship programs can be some effective measures to address the problem of uncontrolled drug-resistance in the country.        Distribution of mcr-1 gene in plasmids and chromosomes of both COL-R and COL-S isolates Note: Total bacterial isolates= 24 (6 COL-R with MIC≥4 µg/ml; 18 COL-S with MIC=2 µg/ml)