Antibiotic resistance and detection of plasmid mediated colistin resistance 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 Klebsiella 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 modified Kirby-Bauer disc diffusion method. Drug-resistant isolates were further screened for extended-spectrum beta-lactamase (ESBL), metallo-beta-lactamase (MBL), carbapenemase and K. pneumoniae carbapenemase (KPC) production tests. All the suspected enzyme producers were processed for phenotypic confirmatory 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) and K. pneumoniae isolates were 9% (41/455). 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 confirmed as ESBL-producers. The prevalence of ESBL-producing E. coli was 43% (130/302) whereas that of K. pneumoniae was 26.8% (11/41). Similarly, 12.5% (43/343) of the total isolates, 10.9% (33/302) of E. coli and 24.3% of (10/41) K. pneumoniae were resistant to carbapenem. Among 43 carbapenem resistant isolates, 30.2% (13/43) and 60.5% (26/43) were KPC and MBL-producers respectively. KPC-producers isolates of E. coli and K. pneumoniae were 33.3% (11/33) and 20% (2/10) respectively. Similarly, 63.6% (21/33) of the E. coli and 50% (5/10) of the K. pneumoniae were MBL-producers. In MIC assay, 2.2% (4/179) of E. coli and 10% (2/20) of K. pneumoniae isolates were confirmed as colistin resistant (MIC ≥ 4 µg/ml). Overall, the prevalence of colistin resistance was 3.1% (6/199) and acquisition of mcr-1 was 16.6% (3/18) among the E. coli isolates. Conclusion High prevalence of drug-resistance in our study is indicative of a deteriorating situation of AMR. Moreover, significant prevalence of resistant enzymes in our study reinforces their roles in the emergence of drug resistance. Resistance to last resort drug (colistin) and the isolation of mcr-1 indicate further urgency in infection management. Therefore, extensive surveillance, formulation and implementation of effective policies, augmentation of diagnostic facilities and incorporation of antibiotic stewardship programs can be some remedies to cope with this global crisis. Supplementary Information The online version contains supplementary material available at 10.1186/s13099-021-00441-5.


Background
Extensive and irrational use of antibiotics has led to the emergence and spread of antimicrobial resistance (AMR)-a condition in which pathogenic strains of bacteria develop resistance to the therapeutic antibiotics prescribed against it [1]. Realising the AMR as a global crisis, World Health Organization (WHO) and US Centres of Disease Control and Prevention (CDC) have already warned of an imminent global disaster and possibility of returning to the post-antibiotic era [2]. Gramnegative bacteria, mainly Enterobacterales, Acinetobacter baumannii and Pseudomonas aeruginosa are able to produce enzymes such as extended-spectrum beta-lactamases (ESBLs), AmpC beta-lactamases, carbapenemase and metallo-beta-lactamases (MBL), which enable the host bacteria to develop resistance to most classes of antibiotics in use [3]. All of these enzymes possess a similar mechanism to hydrolyse β-lactam ring of the antibiotics. ESBLs are class A β-lactamases that are responsible for resistant against oxy-imino cephalosporins (cefotaxime, ceftazidime, ceftriaxone, cefuroxime, and cefepime) and monobactams (aztreonam) [4]. Carbapenemase are the members of class A, B and D β-lactamases. Class A and D carbapenemase bring out the serine based hydrolytic mechanism while the class B carbapenemase are metallo-β-lactamases (MBL) that contain zinc based hydrolytic mechanism [5]. A novel kind of MBL, known as New Delhi metallo-β-Lactamase (NDM) possesses the ability to resist virtually all β-lactam antibiotics (except aztreonam) and carbapenems [6]. Similarly, K. pneumoniae carbapenemase (KPC), a derivative of carbapenemase (class A β-lactamases) also has become prominent because of their ability to inactivate carbapenems [7]. Although KPC is prevalent among K. pneumoniae, the enzyme has been frequently isolated from other Gramnegative bacilli [8].
Multidrug-resistant (MDR) bacteria-better known as superbugs-seriously limit the treatment options and thus are associated with increased mortalities, morbidities and economic burden [9]. On the other hand, narrowed treatment option is forcing clinicians to rely upon the "last line" drugs, primarily colistin (a polymyxin E antibiotics), which is reintroduced to counter the rapidly surging carbapenemase-producing Gram-negative bacteria [10]. Polymyxins (polymyxin B and polymyxin E) are cyclic lipopeptide discovered in the late 1940s [11] that were introduced in the treatment of infections caused by Gram-negative bacteria. However, they were no longer used due to their neuro-and nephro-toxicity and also due to the availability of comparatively 'safer' drugs such as beta-lactams [12]. Although their toxic effects were standstill, polymyxins were re-introduced in 1990s to counter the uncontrolled emanation of carbapenem resistant bacteria [13]. Polymyxins are now extensively used in modern clinics due to paucity of novel, effective and safer antibiotics [14]. Like with all other antibiotics, bacteria have managed to develop resistant to colistin, as a large number of studies suggest the emergence and globalization of colistin-resistance [10].
Until the first report of a variant of noble plasmidmediated mobilized colistin resistance gene (mcr-1) in late 2015 in China, polymyxin (particularly, colistin) resistance was solely attributed to the regulatory changes mediated by the chromosomal genes (phoPQ, pmrAB, and mgrB) [2]. Since the first identification of mcr-1, several variants (from mcr-1 to mcr-9) have been reported from more than 40 countries across five different continents [15]. The mcr-1 encodes for phosphoethanolamine (pEtN) transferase enzyme (discovered in late 2015), which modifies the outer membrane lipopolysaccharides by adding pEtN to the phosphate groups in Lipid A thereby decreasing the net negative charges [2]. The resulting modification reduces the binding affinity of polymyxins to the bacterial cell wall [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 resistant to multiple classes of antibiotics can be transferred to other bacteria [19,20]. Elevated endemicity of 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 first isolated in an E. coli from animal sources in China, 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 drugresistant 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 cocirculation 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 determine 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 resistance to colistin. Furthermore, antibiogram of the resistant strains to a variety of antibiotics was carried out to recognize the possible therapeutic options for combating superbugs.

Study design and study samples
This cross-sectional study was conducted from March to August, 2019 at Kathmandu Model Hospital, Nepal. A total of 3216 clinical samples consisting of urine (n = 1776), blood (n = 875), pus (n = 156), sputum (n = 187), body fluids (n = 88), wound swab (n = 51), tissue including femur, tibia, intestine region, and appendicular sites (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 gender who were admitted in or visiting the hospital for treatment were included in this study. All the samples with completely filled demographic information and 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 to 6 °C.

Culture, isolation and identification of bacteria
Each sample was processed by following the standard microbiological guidelines [29,30].

Urine sample
Urine samples were inoculated into Cysteine Lactose Electrolyte Deficient (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 into 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 first 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 h [31].

Sputum and throat swab
Sputum samples were inoculated into BA, chocolate agar (CA) and MA plates. For sputum, in CA plate a 5 gµ optochin disc and a 10U bacitracin disc were added to screen S. pneumoniae and H. influenzae respectively whereas for throat swab, 0.05U bacitracin disc was added to the plate to screen Streptococcus pyogenes. CA 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 24 h. 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 fluids: Body fluid 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 5-10% CO 2 enriched atmosphere and MA plates were incubated aerobically at 37 °C overnight [31].

Identification of the bacterial isolates
Following incubation, culture plates were observed for possible microbial growth. Isolates were presumably identified on the basis of Gram's staining and colony characteristics. Further confirmation of the isolates were based on biochemical tests such as IMViC (Indole production, Methyl red test, Voges-Proskauer test and Citrate utilization), H 2 S production, catalase test, coagulase test, and oxidase test [30,32].

Screening of the ESBL production
ESBL-producers were screened by using Ceftazidime (30 µg) and Cefotaxime (30 µg) in the AST. Isolates showing reduced susceptibility to one or both of these drugs with diameter of the zone of inhibition for ceftazidime ≤ 22 mm and cefotaxime ≤ 27 mm were considered as potential ESBL-producers [33]. Suspected strains were further processed using confirmatory assay.

Confirmation of the ESBL-producers by phenotypic method
Combination disc test (CDT) as prescribed by the CLSI was used for the confirmation of ESBL-producing strains. In this method, cefotaxime (30 µg) and ceftazidime (30 µg) discs alone and in combination with clavulanic acid (10 µg) (ceftazidime plus clavulanic acid, 30/10 µg and cefotaxime plus clavulanic acid, 30/10 µg) were used. The zone of inhibition of cephalosporin disc alone was compared with their respective cephalosporin/clavulanic acid (combined) disc. An increase in zone of inhibition by ≥ 5 mm in the presence of clavulanic acid was considered as confirmed ESBL production [33].

Phenotypic confirmatory test for carbapenemase and/ or KPC producers
Inhibitor-based method was followed for the confirmation of carbapenemase and KPC production. In this method, combined disc test of carbapenem with and without phenyl boronic acid (PBA) was employed. An increase in the diameter of zone of inhibition by ≥ 5 mm in combined disc (carbapenem disc supplemented with PBA) than single disc (only carbapenem disc) was considered as confirmed test for carbapenem or KPC production [35,36].

Phenotypic confirmatory test for MBL production
Confirmation of MBL production was made by inhibition method in which Ethylene Diamine Tetra Acetic Acid (EDTA) was used as an inhibitor. Two imipenem (10 µg) discs were placed on MHA and 10 µl 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 zone of inhibition by 7 mm in combined disc (imipenem disc supplemented with EDTA) than single one (only imipenem disc) was considered as confirmed test for MBL production [37].

Determination of MIC of colistin
Agar dilution method was used to determine the minimum inhibitory concentration of colistin. Different concentrations of colistin ranging from 2 µg/ml to 32 µg/ml were prepared in the agar medium. Bacterial inoculum was applied readily onto the agar surface and the plates were incubated at 37 °C upto 18 h. 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
Isolated colonies of bacteria were inoculated in Luria-Bertani (LB) broth and incubated overnight at 37 °C. After incubation, alkaline-lysis method was adopted to extract the DNA. Plasmid DNA of E. coli and K. pneumoniae was extracted by using phenol-chloroform method [39]. Extracted plasmid DNA and genomic DNA was then suspended in TE buffer and preserved at − 20 °C until further processing [39].

PCR amplification of mcr-1 gene
Amplification of mcr-1 gene was carried out by conventional PCR using primers: 5′-CGG TCA GTC CGT TTG TTC -3′ as forward primer and 5′-CTT GGT CGG TCT GTA GGG -3′ as reverse primer [2]. A PCR mixture having the final volume of 25 µl (3 µl of template DNA, 0.5 µl each forward and reverse primers and 21 µl of PCR master mix) was used for the reaction mixture.
The thermal condition for amplification was initial heat activation of 95ºC for 15 min followed by 35 cycles of denaturation at 94 ℃ for 30 s; annealing at 57ºC for 90 s; extension at 72 ℃ for 90 s; and final extension at 72 ℃ for 10 min. The amplified products were subjected to gel electrophoresis (2.0% agarose gel stained with ethidium bromide) at 100 V for 60 min 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 Sciences (SPSS) software (Version 25). The results have been presented in the form of tables and figures. Chi-square (χ2) test was applied to test the association between the variables. A p value of < 0.05 was considered as statistically significant.

Antibiotic susceptibility pattern of E. coli and K. pneumoniae
All of the tested isolates of E. coli and K. pneumoniae were susceptible to colistin, polymixin B and tigecycline. E. coli isolates were also susceptible to antibiotics 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, K. pneumoniae isolates were resistant to amoxycillin and azithromycin. Resistance of K. pneumoniae towards cephalosporin antibiotics include: cefepime (57.1%), ceftazidime (55.6%) and cefotaxime (51.2%). None of the isolates was resistant to ciprofloxacin, colistin, polymyxin-B and tigecycline (Table 3).
In contrast, none of the phenotypically colistin resistant K. pneumoniae harboured mcr-1 gene.

Antibiotic resistance profiles of colistin resistant and sensitive isolates
Antibiotic resistance profiles of colistin resistance E. coli and K. pneumoniae were determined. Colistin resistant 100% E. coli isolates were resistant to cefotaxime, cefixime, amoxycillin, ofloxacin, levofloxacin, amoxycillin/ clavulanate, and doxycycline where as colistin resistant 100% K. pneumoniae isolates were resistant to nitrofurantoin, cefotaxime, cefixime, amoxycillin, ofloxacin, and levofloxacin (Fig. 3, Table 6).  acquisition of mcr-1 in colistin resistant isolates. In this study, most of the isolates were resistant to commonly prescribed broad-spectrum antibiotics. In addition, prevalence of colistin resistance and acquisition of mcr-1 among the drug-resistant isolates was also observed in this study. In this study, 16.4% (529/3216) specimens showed bacterial growth in which prevalence of Gram-negative  bacteria from different clinical specimens was much higher than the Gram-positives. This finding 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 flora 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 factors encoded by pathogenic strains of E. coli enable 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 findings [46,47]. These classes of antibiotics can be effective drugs in the management of 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 reasons 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 findings have been reported by some previous studies [47,49,50]. Higher rate of resistance to cephalosporins can be attributable to their irrational prescription and uses [51].
Resistance rate of E. coli to fluoroquinolones in this study ranged from 47 to 55% which are in agreement with earlier studies from Nepal [52,53] and India [54]. Resistance to fluroquinolones among MDR Gram-negative bacteria is common and is expected to sustain and perhaps accelerate even if other antibiotics are used [55]. The prevalence of fluroquinolone resistance is related to the intensity of antibiotics used, which may reduce the efficacy of drug in a 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% to 21.0% in some of the previous studies from Nepal [50,52,53] whereas 100% sensitivity towards imipenem was reported in some other studies [47,57]. Production of beta-lactamase enzymes and the upregulation of efflux pump are suggested as the reasons for reduced susceptibility [58]. However, comparatively low resistance to carbapenem antibiotics reported in this study could be due Table 6 Antibiotic resistance profiles of colistin resistant and sensitive isolates  to the lower use of these antibiotics in the treatment of infections [59]. In AST assay of K. pneumoniae, two-third of the isolates were susceptible to fluoroquinolones and gentamicin. This finding is similar to another study reported from Nepal [60]. However, some other studies have reported lower sensitivity 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 findings were reported in other studies [52,60]. In addition, reduced sensitivity 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 beta-lactamases, or efflux 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 findings 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 findings [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 kinds of β-lactamases primarily ESBL, AmpC and MBL, and carriage of resistance trait for quinolones and aminoglycosides 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% to 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% isolates were KPC and MBL-producers respectively. Our findings are comparable to a previous finding [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 carbapenem resistance genes [59].
This study reported comparatively low prevalence (3.0%) of colistin resistant E. coli and K. pneumoniae isolates (3.0%). Similar result was reported in previous findings of 0.3% in Switzerland [81], 0.7% in Spain [82] and in a previous report of global antimicrobial surveillance programs [83,84]. Low prevalence of the resistant isolates in our study may be attributable to the presence of low number of mcr-1 positive bacteria and lesser use of colistin use for the treatment of community acquired infections [81]. However, other studies from India [85] and Thailand [86] reported the prevalence rate of colistin resistant as high as 32.0% and 71.3% respectively. In this study, rates of colistin resistance were different among the bacterial species. Higher prevalence of colistin resistance was reported in K. pneumoniae (10.0%) when compared to E. coli (2.2%). Different studies also showed that higher prevalence of colistin resistant K. pneumoniae than E. coli [87,88]. The main risk factor for the development of colistin resistance is related to the extensive and irrational use of colistin in antimicrobial therapy [89].
MIC range of colistin for E. coli ranged from ≤ 2 µg/ml to 8 µg/ml which was lesser than that of K. pneumoniae (≤ 2 µg/ml to 16 µg/ml). This finding 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/l 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 was high in this study. 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 first 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 findings 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 some 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 reflect 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 drug-resistance in the country.