Coexistence of blaOXA-48, blaVIM, and blaSHV genes in Klebsiella pneumoniae and Escherichia coli isolated from urinary tract infections: Microbiological and epidemiological analysis

Objective: To investigate antimicrobial resistance mechanisms of isolated bacterial strains, and their correlation with virulence profile. Method: The cross-sectional study was conducted in January 2020 at outpatient health centres in Kafrelsheikh Governorate of Egypt, and comprised urine samples from patients regardless of age and gender. Midstream samples were collected into sterile swaps which were kept in ice-cooled boxes until transported to the laboratory within 5h. Antimicrobial resistance profile of the isolated Enterobacteriaceae was done using Kirby–Bauer disk diffusion method and was confirmed with Vitek compact 2. The phenotypic of carbapenemases and extended-spectrum beta lactamase was determined, and polymerase chain reaction was used, as appropriate. Data was analysed using SPSS 20. Results: Of the 199 patients, 101(50.7%) were females and 98(49.3%) were males. The majority 73(36.6%) were aged 30-50 years. Urinary tract infection was found in 68(34.2%) patients. In 28(41.2%) of these patients, there were 32 isolates of Enterobacterales ; 21(65.62%) Klebsiella pneumoniae , 7(21.87%) Escherichia coli and 4(12.5%) Enterobacter cloacae . Of the 28(41.2%) patients, 24(85.7%) were infected with a single strain; 17(70.8%) Klebsiella pneumoniae, 4(16.7%) Escherichiacoli and 3(12.5%) Enterobacter cloacae. In 3(10.7%) cases, there was co-infection with Escherichia coli and Klebsiella pneumoniae, and 1(3.6%) sample had mixed infection with Klebsiella pneumoniae and Enterobacter cloacae . The other 40(58.8%) patients had other causative agents. Housewives, agricultural workers and those aged >50 years had a higher risk of urinary tract infections ( p <0.05) Among Klebsiellapneumonia isolates, 6(28.5%) possessed carbapenemase-related genes and 4(19.1%) extended-spectrum beta lactamase-related genes. The carbapenemase-related genes were bla-Verona integron-encoded metallo beta lactamase 6(100%) bla-New Delhi metallo beta lactamase-1 4(66.6%) and bla-oxacillinase-48 2(33.3%). The 4(19.1%) cases of extended-spectrum beta lactamase-related genes had bla-temoneira gene 3(75%) and bla-sulfhydryl variable gene 4(100%). In Escherichia coli isolates, bla-oxacillinase-48 and bla-Cefotaximase genes were observed in 2(28.5%) cases. Virulence genes uridine diphosphate-glucose 4-epimerase, fimbrial adhesion and mannose-resistance adhesin of Klebsiella spp genes in Klebsiella pneumoniae isolates were positive in in 16(76.2%), 14(66.7%) and 10(47.6%) cases, respectively. All 21(100%) isolates of Klebsiella pneumoniae were negative for mucoviscosity-associated gene A. Conclusion: There was evidence of the coexistence of bla-oxacillinase-48, bla-Verona integron-encoded metallo beta lactamase and bla-sulfhydryl variable genes in Klebsiella pneumoniae and Escherichia coli isolates from mixed urinary tract infection samples


Introduction
organisms. [15][16][17] Carbapenemases are beta (β) lactamase enzymes, which were shown to be slow to emerge in Enterobacteriaceae, but are now increasingly causing serious infections leading to mortalities in hospitalised patients. [18][19][20] The most common carbapenemases are K. pneumoniae carbapenemases (KPC), New Delhi metallo-βlactamase (NDM), imipenem-resistant Pseudomonas (P.)-type carbapenemases (IMP), Verona integron-encoded metallo-β-lactamase (VIM), and oxacillinase (OXA-48-like) types. They are coded by blaKPC, blaNDM, blaIMP, blaVIM, and blaOXA-48 genes, respectively. 21 The detection of carbapenemases is difficult; the isolate may be reported as sensitive by phenotypic tests while still harbouring carbapenemases but released in low levels 22,23 , particularly for meropenem (MEM) and imipenem IMP) groups. 24,25 Extended-spectrum beta lactamase (ESBL)-producing bacteria represent a public health concern because of the variable prevalence and emergence particularly in K. pneumoniae. These enzymes hydrolyze b-lactam antimicrobials, third-and fourth-generation cephalosporins and monobactams, and lead to failure of treatment. 26 ESBLs are grouped into four classes A, B, C and D enzymes. Cefotaximase (CTX-M), temoneira (TEM) and sulfhydryl variable (SHV) are class A ESBLs. 27 Phenotypic detection of ESBLs between Enterobacteriaceae species is important for epidemiological purposes and to limit the spread of resistance mechanisms. 28,29 Multiple bacterial determinants have been implicated in the pathogenesis of K. pneumoniae. They include capsular serotypes, especially K1 or K2. 30 Examples include mucoviscosity-associated gene A (magA) 31 , mannoseresistance adhesin of Klebsiella spp (mrkA) gene 32 , and uridine diphosphate (UDP)-glucose 4-epimerase (uge). 33 Fimbrial adhesion (fimH) is a major virulence factor implicated in many urinary and systemic pathological consequences, particularly during K. pneumoniae colonisation, biofilm formation, and persistence in UTIs. 34,35 Although K. pneumoniae, E. coli and E. cloacae are very important causes of UTI, no information reflects these bacteria in the Kafrelsheikh Governorate of northern Egypt. The current study was planned to determine the epidemiological data obtained from UTI patients and to investigate the antimicrobial resistance mechanisms of isolated Enterobacteriaceae strains. Also genotypic characterisation of carbapenemase, metallo-β-lactamase (MBL) and ESBL-producing isolates were planned, and the correlation of profile of K. pneumoniae with its virulence profile.

Patients and Methods
The cross-sectional study was conducted in January 2020 at outpatient health centres in Kafrelsheikh Governorate of Egypt ( Figure 1). After approval from the institutional ethics review committee, the sample size was calculated using 7% expected prevalence 36 and acceptable margin of error 5%. 37 The sample was inflated by 97%. Urine samples were obtained from patients regardless of age and gender who had clinical evidence suggestive of UTI. Informed consent was obtained from all patients or from parents of patients aged <18 years. Those not willing to participate were excluded. Inadequate urine samples and results showing >3 pathogens, reflecting a contamination more than a real UTI 38 , were also excluded.
Demographic data, including age, gender and occupation, was recorded. Urine was collected from patients using the midstream clean catch method. The patients were instructed by the medical staff to void the first portion of the urine stream into the toilet and then collect the midstream into the sterile container.
The samples were kept in ice-cooled boxes until transported to the laboratory within 5h.
All the samples were diluted using 10-fold serial dilution method with the help of peptone water, and the dilution was spread on cysteine lactose electrolyte-deficient (CLED) (Oxoid, UK) agar plates before being incubated at 37°C for 24-48h. A significant colony count was considered on the basis of a pure growth of ≥10 5 CFU/mL. 3,[39][40][41] All urine samples were streaked on CLED agar. Each suspected sample was streaked on eosin methylene blue (EMB) agar and MacConkey agar (Oxoid, UK). All plates were incubated at 37°C for 18-48h. 42 The isolates were identified using classical biochemical methods by standard laboratory techniques. 43 Multidrug resistence (MDR) was characterised if a strain during in vitro antimicrobial susceptibility test (AST) showed resistance to three or more antimicrobial classes. 41 The AST test was carried out using the Kirby-Bauer disk diffusion method, as recommended by the Clinical & Laboratory Standards Institute (CLSI) [44][45][46] , which was further confirmed by an automated system (Vitek Compact 2; ASTcard, Biomerieux®, France) following the manufacturer's instructions.
The combined disk test (CDT) was used for the phenotypic detection of MBLs in carbapenem-resistant gram-negative bacteria. [51][52][53] Two 10mg IMP disks were placed on Müller-Hinton agar, and 4µl of 0.5M concentrated ethylenediaminetetraacetic acid (EDTA) solution was added to one of the disks. The inhibition zones of IMP and IMP-EDTA disks were compared after 16-18h incubation at 35°C. CDT compared the increased inhibition zone ≥7mm with the IMP-EDTA disk to the IMP disk alone, which was considered MBL-positive.
Double-disk synergy test (DDST) was performed for IMP-EDTA. 51,54 One blank filter paper disk containing 10mL EDTA was placed at 20mm from the centre of the IMP disk, and the inhibition zones of the IMP and the blank disk with EDTA were compared. Enhancement of the inhibition zone in the area between IMP and the EDTA disk compared to the inhibition zone on the far side of the drug was interpreted as a positive result.
All isolates were screened for ESBL production by CDT following CLSI guidelines 55,56 . Susceptibility to CTX (30μg), CTX/clavulanate (30/10μg), CAZ (30μg) and CAZ/clavulanate (30/10μg (Oxoid, UK) was determined on Müller-Hinton agar (Oxoid, UK). ESBL-producing strains were identified by at least 5mm increase in zone diameter around CTX/clavulanate and CAZ/clavulanate disks compared to disks without clavulanic acid. 55,56 Conventional PCR was performed using specific primers (Metabion, Germany) ( Table 1) to detect genes encoding carbapenemases, MBL (blaIMP, blaVIM, blaNDM, blaKPC, and blaOXA-48), and ESBL (blaTEM, blaSHV, and blaCTX-M) 57 . In order to obtain DNA from the isolated strains, a single colony from each plate was taken and inoculated into 5ml of tryptone soya broth (Oxoid, UK) overnight at 37°C. Then 1ml of bacterial culture broth was centrifuged in a micro-centrifuge tube at 13,000rpm for 1min. The supernatant was discarded, and the bacterial pellets were homogenised with nuclease-free water and heated at 95°C for 10min. The boiled lysates were finally centrifuged, and the supernatant was collected as DNA templates, which were kept at -80°C until use.
Primers (Table 2) were used in 25µl uniplex PCR mix, comprising 12.5µl of EmeraldAmp Max PCR Master Mix (Takara, Japan), 1µl of each primer (20pmol), 5.5µl of water, and 5µl of DNA template 55 . The reaction was performed in a thermal cycler (Applied Biosystems 2720). The cycling condition started with primary denaturation at 94°C for 5min followed by 35 cycles and a final extension at 72°C for 10min. Positive controls were represented by field samples that were previously confirmed to be positive by PCR for the related genes in the reference laboratory for veterinary quality control on poultry production, animal health research institute. 57-60 Sterile water was added to the PCR mix with each primer pair as control negative.
The products were separated by electrophoresis in 1% agarose gel with 1X Tris/borate/EDTA, stained with a safe stain load dye, and visualised under ultraviolet illumination. PCR for virulence-associated genes led to extraction of bacterial DNA using EmeraldAmp GT PCR Mastermix (Takara, Japan; Code No. RR310A) as was the case with resistance genes.
Data was analysed using SPSS 20. Univariate logistic regression model was built to examine the associations between dependent and independent variables. Variables with p<0.2 were subjected to multivariate logistic regression model, in the light of a 2019 study by Abdelwahab et al. 61 . In K. pneumoniae and E. coli isolates, the relative risk (RR) for resistance to both ERT and IMP was worked out using the formula (D1/N1) * (N2/D2), where D1 was the number of isolates that were phenotypically or genotypically positive for resistance to a particular antibiotic/group of antibiotics, D2 was the number of isolates phenotypically or genotypically positive to both ERT and IMP resistance (baseline group), N1 was the number of isolates resistant to a particular antibiotic/group of antibiotics, and N2 was the number of isolates resistant to both ERT and IMP. Similar forumla was used for K. pneumoniae and E. coli for CTX and CAZ resistance. The relative association between the resistance type and the virulence gene of K. pneumoniae isolates was calculated using univariable ordinal logistic regression analysis. The resistance type was established as the response variable and the predictor variables were virulent for mrkA (fimbriae), fimH (fimbriae), uge (capsule) and magA (capsule). Odds ratios (ORs) were calculated. Level of significance was set at p<0.05.
Prevalence of virulence genes among K. pneumoniae isolates, mrkA, fimH and uge genes were investigated in K. pneumoniae isolates, with 10(47.6%), 14(66.7%) and 16(76.2%), respectively, being positive. Conversely, all 21(100%) isolates of K. pneumoniae were negative for the magA gene (Table 9; Figure 4). The association between the resistance and any of the virulence genes in K. pneumoniae isolates was not significant, but it showed a tendency for an increase in resistance with the presence of virulence genes (Table 10).

Discussion
To the best of our knowledge, the current study is the first evidence of the coexistence of blaOXA-48, blaVIM and blaSHV genes in K. pneumoniae and E. coli isolated from mixed UTI in Egypt. The study indicated that K. pneumoniae was the most frequent pathogen isolated from UTIs, which was consistent with earlier findings. 62 In addition, UTI prevalence was significantly higher in patients aged >50 years than the younger groups. A similar trend was observed by Medina et al. 63 in Italy.
Rural-urban differences in using appropriate antibiotic durations may be related to patient and provider-level factors, such as distance to healthcare facility. 64 In the current study, rural patients were more likely to experience longer treatment durations, which is in agreement with literature. 65,66 This study found that most of the strains were MDR as they were resistant to most antimicrobial agents tested. 67    frequency of antibiotic use was more for AM, AMX and AMC, which are the most commonly prescribed in hospitals even before urine analysis results arrive.
A decline in CIP activity against E. cloacae, E. coli and K. pneumoniae strains would be especially problematic given the ability of gram-negative bacilli to gain resistance to all other classes of antimicrobials. These findings agree with those reported earlier. 15,68 FT had low resistance rates for K. pneumoniae, E. coli and E. cloacae and this was supported by a 2017 study. 69 K. pneumoniae showed a resistance rate of 52.7% against NOR, which was consistent with a study done in Italy. 26 There was a decline in STX activity, with resistance ranging from 76.8% and 85.7% to 100% for K. pneumoniae, E. coli and E. cloacae, respectively. This was in line with literature. 15 Compared to carbapenem sensitivity of >90 reported from Mexico 38 , the current study showed a reduction in sensitivities ranging from 33.3% to 76.2%.
In this study, aminoglycosides, including GM and AN, showed high efficacy against tested UTI pathogens E. coli and E. cloacae. The K. pneumoniae isolates also showed a low resistance rate against GM and AN. This may be due to the low and irregular treatment regimens for recorded aminoglycosides. 39 This study found that the most carbapenemase-related genes detected in K. pneumoniae tested isolates were blaVIM and blaNDM, followed by blaOXA-48. No blaNDM-1 gene was detected in E. coli strains. The blaIMP and blaKPC genes were not found in any of the bacterial isolates. This was in kine with a study in Tehran. 70 In this study, two mixed or co-infected samples containing K. pneumoniae and E. coli strains carried the blaSHV gene. In addition, the two bacterial strains carried the blaVIM and blaOXA-48 genes in the other mixed or co-infected sample. This may show gene transfer among different bacterial UTI strains. 71 One K. pneumoniae isolate simultaneously carried one ESBL gene with one carbapenemase gene. The coexistence of two ESBL-related genes and one carbapenemase-related gene was observed in one E. coli isolate of UTI.
Overall, uge was the most commonly detected putative virulence gene, followed by fimH gene. This result agrees with Asani A et al. 72 Also, this analysis found that 11 isolates harboured one virulence gene at least with one of the carbapenemase-related genes and/or ESBL-related genes. Four isolates harboured genes with at least one ESBLrelated gene. Three isolates harboured genes with at least

Conclusions
There was evidence of the coexistence of blaOXA-48, blaVIM and blaSHV genes in Klebsiella pneumoniae and Escherichia coli isolates from mixed UTI samples. Data on antibiotic resistance and virulence mechanisms could be further developed and used in a national antibiotic sales and antimicrobial resistance monitoring programme.