Characterization of Transferrable Mechanisms of Quinolone Resistance (TMQR) among Quinolone-resistant Escherichia coli and Klebsiella pneumoniae causing Urinary Tract Infection in Nepalese Children

Background Transferrable mechanisms of quinolone resistance (TMQR) can lead to fluoroquinolone non-susceptibility in addition to chromosomal mechanisms. Some evidence suggests that fluoroquinolone resistance is increasing among the pediatric population. We sought to determine the occurrence of TMQR genes among quinolone-resistant E. coli and K. pneumoniae causing urinary tract infections among Nepalese outpatient children (< 18 years) and identify molecular characteristics of TMQR-harboring isolates. Methods We performed antimicrobial susceptibility testing, phenotypic extended-spectrum β-lactamase (ESBL) and modified carbapenem inactivation method tests, and investigated the presence of six TMQR genes (qnrA, qnrB, qnrS, aac(6’)-Ib-cr, oqxAB, qepA), three ESBL genes (blaCTX−M, blaTEM, blaSHV), and five carbapenemase genes (blaNDM, blaOXA−48, blaKPC, blaIMP, blaVIM). The quinolone resistance-determining region (QRDR) of gyrA and parC were sequenced for 35 TMQR-positive isolates. Results A total of 74/147 (50.3%) isolates were TMQR positive by multiplex PCR [aac(6’)-Ib-cr in 48 (32.7%), qnrB in 23 (15.7%), qnrS in 18 (12.3%), qnrA in 1 (0.7%), and oqxAB in 1 (0.7%) isolate]. The median ciprofloxacin minimum inhibitory concentration of TMQR-positive isolates (64 µg/mL) was two-fold higher than those without TMQR (32 µg/mL) (p = 0.004). Ser-83→Leu and Asp-87→Asn in GyrA and Ser-80→Ile in ParC were the most common QRDR mutations (23 of 35). In addition, there was a statistically significant association between TMQR and two β-lactamase genes; blaCTX−M (p = 0.037) and blaTEM (p = 0.000). Conclusion This study suggests a high prevalence of TMQR among the quinolone-resistant E. coli and K. pneumoniae isolates causing urinary tract infection in children in this area of Nepal and an association with the carriage of ESBL gene. This is a challenge for the management of urinary infections in children. Comprehensive prospective surveillance of antimicrobial resistance in these common pathogens will be necessary to devise strategies to mitigate the emergence of further resistance. Supplementary Information The online version contains supplementary material available at 10.1186/s12887-023-04279-5.


Background
Fluoroquinolone (FQ) antimicrobials are important in the treatment of a range of infections, including urinary tract infections (UTIs).They have a broad spectrum of activity, high bioavailability, convenient dosing regimens, and high potency [1,2].FQs are listed as an essential medicine by World Health Organization (WHO) [3].Their extensive use has led to a marked increase in FQ resistance globally [4][5][6][7].
Quinolone/fluoroquinolone (Q/FQ) resistance in Enterobacterales is commonly attributed to chromosomal mutations in the quinolone resistance-determining region (QRDR) of the genes encoding subunits of DNA gyrase (GyrA and GyrB) and topoisomerase IV (ParC and ParE) [8].The reduction of Q/FQ concentration in the cytoplasm by chromosomal efflux pumps or permeability alterations also contributes to resistance [9].Transferrable mechanisms of quinolone resistance (TMQR) can additionally confer low-level Q/FQ resistance and promote the development of full resistance [10].TMQR determinants include seven Qnr proteins, AAC(6')-Ibcr (an aminoglycoside acetyltransferase), and two efflux pumps, QepA and OqxAB.Qnr proteins (QnrA, QnrB, QnrC, QnrD, QnrE, QnrS, and QnrVC) are dimeric proteins belonging to the pentapeptide repeat protein (PRP) family and protect DNA gyrase and topoisomerase IV from the action of quinolones [11].The AAC(6')-Ib-cr is a bifunctional variant of AAC(6')-Ib that imparts resistance to aminoglycosides and fluoroquinolones having a piperazinyl substituent, such as ciprofloxacin and norfloxacin, via acetylation of amino nitrogen in the piperazinyl ring [11].These diverse mechanisms can act in concert to confer non-susceptibility to Q/FQ.
As the use of FQ is restricted in children [12], the increase in FQ resistance among the pediatric population is important [13,14].In a recent study from our institution, 66 (41.8%) of 158 E. coli and 7 (23.3%) of 30 K. pneumoniae isolates causing UTI among children were resistant to ofloxacin [15].Studies from tertiary care centers of Nepal focusing on pediatric UTI have reported ciprofloxacin resistance in 576 (78%) of 739 and 44 (63%) of 69 isolates, and ofloxacin resistance in 104 (62%) of 168 E. coli isolates [16][17][18].We have investigated the occurrence of TMQR among quinolone-resistant E. coli and K. pneumoniae isolates causing UTI among children at our institution and sought to identify molecular characteristics of TMQR-harboring isolates.

Study design and setting
This is a retrospective study conducted at Siddhi Memorial Hospital (SMH), Bhaktapur, Nepal.SMH is a 50-bedded secondary care maternal and pediatric hospital with 10 pediatric ICU beds, serving about 16,000 pediatric OPD visits annually.E. coli and K. pneumoniae isolates obtained from UTI patients less than 18 years old attending the outpatient department (OPD) of the hospital were included in the study.An anonymized dataset, with personal identifiers removed, containing the patient's age, sex, name of the pathogen, and the susceptibility result to nalidixic acid from June 2018 to February 2021 was retrieved from the microbiology laboratory.

Microbiological methods at the time of isolation of the isolates
Clean catch mid-stream urines were collected from children suspected of UTI as per the pediatrician's discretion as a part of routine patient diagnosis.Urine cultures were performed by semi-quantitative method on a cysteinelactose-electrolyte deficient agar (CLED) plates which were then incubated at 37 o C for 18-24 h aerobically.Urine cultures with a growth of ≥ 10 5 CFU/mL were considered for further processing.
The presumptive identification of the pathogens was performed by Gram stain, colony morphology, and a panel of in-house biochemical tests.Susceptibility to nalidixic acid (NA) was performed by the Kirby Bauer disk diffusion method [19].Significant isolates were stored at -40 o C at the time of isolation.
E. coli and K. pneumoniae isolate resistant to NA were sub-cultured from the frozen stocks on a MacConkey agar and sheep blood agar till uniform well-isolated colonies were obtained.The investigations carried out in this study include antimicrobial susceptibility testing and molecular investigations for the detection of β-lactamases, TMQR genes, and mutations in gyrA and parC.

Nucleic acid extraction
The genomic DNA was extracted using Qiagen DNA mini kit (Qiagen, Hilden, Germany) following the procedures described by the manufacturer with the only exception that the final elution was made with 150 µl of nuclease-free water (NFW).The DNA extracts were quantified using Qubit 4 Fluorometer (Invitrogen, Thermo Fisher Scientific) following the manufacturer's recommendations.

Characterization of β-lactamases
The phenotypic determination of extended-spectrum β-lactamase (ESBL) production was first performed by a combination disc diffusion method with cefotaxime, cefotaxime-clavulanic acid, ceftazidime, and ceftazidimeclavulanic acid (D62C and D64C, Mast group Ltd, Liverpool, UK).The results were interpreted as described in the CLSI guideline [19].DNA samples of the ESBL-positive isolates were analyzed by PCR to detect bla CTX−M [22], bla SHV (for E. coli only), and bla TEM by the assays described elsewhere [23].
The modified carbapenem inactivation method (mCIM) was used to confirm carbapenemase production among imipenem non-susceptible isolates as described in the CLSI guideline [19].DNA samples of these isolates were analyzed by PCR to detect bla NDM , bla OXA−48 , bla KPC , bla IMP , and bla VIM using the primers published previously [24].
Escherichia coli ATCC 25922 and clinical strains confirmed to harbor bla CTX−M , bla TEM , and bla SHV β-lactamase genes were used for quality control for ESBL phenotyping and genotyping.Previously characterized strains confirmed to harbor bla NDM and bla OXA−48 were used as positive controls for mCIM.DNA extracted from the control strains was used as a positive control, and Escherichia coli ATCC 25922 DNA was used as a negative control in PCR assays.

Detection of TMQR genes
Previously validated multiplex PCR assay for the detection of TMQR genes was used for the detection of qnrA, qnrB, qnrS, oqxAB (reported only for E. coli), qepA, and aac(6')-Ib-cr [25].Briefly, the PCR reaction mixture of 50 µl was prepared with 25 µl of multiplex PCR master mix (2X) (Qiagen, Hilden, Germany), 5 µl of the pool of primers containing 2 µM of each primer, 5 µl of template of concentration of 20 ng/µl, and 15 µl of NFW (Ambion™ Nuclease-Free water, Invitrogen, Thermo Fisher Scientific).The PCR amplification was carried out in Veriti 96 Well Thermal Cycler (appliedbiosystems, Thermo Fisher Scientific) with 15 min of initial denaturation at 95 o C followed by 30 cycles of denaturation at 94 o C for 30 s, annealing at 63 o C for 90 s, and extension for 10 min at 72 o C. The amplification products were first resolved by gel electrophoresis (1.5%, w/v) at 100 V for 40 min and visualized in a gel documentation system (Major Science, California, USA).All PCR amplicons of qnr genes were sequenced and confirmed by BLAST (Basic Local Alignment Search Tool).

Detection of mutations in gyrA and parC
A convenience sample of thirty-five TMQR-positive isolates with representative ciprofloxacin's interpretive categories (susceptible, intermediate, and resistant) for E. coli and K. pneumoniae were selected for the amplification of the gene fragment covering the QRDR of the gyrA [26] and parC [27].None of the K. pneumoniae isolates with a TMQR gene were susceptible to ciprofloxacin.The gyrA and parC amplicons were purified and subjected to bidirectional DNA sequencing by capillary electrophoresis (Macrogen, South Korea).
The chromatograms were visualized and processed in BioEdit software, and the sequences were then imported into MEGA11 software.In MEGA11 alignment explorer, the sequences were aligned by the ClustalW algorithm followed by codon-based nucleotide alignment.The substitutions in the QRDR of GyrA and ParC were determined by comparing the amino acid sequences of the isolates to the amino acid sequences of E. coli ATCC 25922 (GenBank Accession number NZ_CP032085 for gyrA, NZ_CP009072 for parC) and Klebsiella pneumoniae ATCC 13883 (GenBank Accession number DQ673325 for gyrA, KFJ75438 for parC).

Data analysis
The data were collected in a Microsoft Excel spreadsheet and imported to IBM SPSS Statistics for Windows v.20 (IBM Corp, Armonk, NY).A chi-squared test of independence or Fisher exact test was performed to determine whether there was a significant relationship between TMQR and other categorical variables.The difference in ciprofloxacin MIC among TMQR positive and negative isolates was investigated by the Mann-Whitney U test.A cutoff value of ≤ 0.05 for the P-value was considered for statistical significance.

Bacterial isolates
There were 522 unique uropathogens isolated from children with a UTI within the study period.Of the 522 isolates, there were 362 E. coli isolates and 74 K. pneumoniae isolates.Nalidixic acid resistance was present in 130/362 (35.9%) of E. coli and 24/74 (32.4%) of K. pneumoniae.Five E. coli and two K. pneumoniae isolates were not recovered in the sub-culture.The final sample size of this study was 147 isolates.The isolates included in this study were obtained from 100 female (68%) and 47 (32%) male children.The median (inter-quartile range (IQR)) age of the children was 6 (2-9) years.

GyrA and ParC substitutions
The amino acid substitution profiles observed in the QRDR of GyrA and ParC of E. coli and K. pneumoniae along with ciprofloxacin MIC values and the presence of TMQR genes are presented in Table 3. Twenty of 30 E. coli and three of five K. pneumoniae isolates had double residue substitutions (Ser-83→Leu and Asp-87→Asn) in GyrA and single substitution in ParC (Ser-80→Ile).Three E. coli isolates had double mutations in ParC (Ser-80→Ile and Glu-84→Val) in addition to GyrA double mutations (Ser-83→Leu and Asp-87→Asn).One E. coli isolate also had double-double mutations, but the alteration at the 84th position in ParC was from glutamic acid to glycine (Glu-84→Gly).One K. pneumoniae isolate had an alteration from aspartic acid to glycine at the 87th position in addition to Ser-83→Tyr in GyrA and Ser-80→Ile in ParC.

Discussion
This study demonstrates alarmingly high levels of FQ resistance among E. coli and K. pneumoniae isolates causing UTI in children attending the outpatient department of Siddhi Memorial Hospital, Bhaktapur, Nepal.Half of the isolates were TMQR positive which suggests that TMQR genes may have an important role in the emergence of quinolone resistance in E. coli and K. pneumoniae isolates within our study population.TMQR genes were found to have a statistically significant association with two β-lactamases, bla CTX−M and bla TEM .
We found a high prevalence of TMQR in diverse gene combinations among study isolates.Similar high proportions of the TMQR genes have been reported in previous studies, while few studies have comparatively lower proportions.The proportion of TMQR positivity among FQ-resistant isolates we report, 68/132 (51.5%) of ciprofloxacin-resistant isolates, is similar to a study from the Netherlands (29/ 56, 51.8%) [28], higher than in Korea (13/122, 10.7%) [29], Taiwan (37/248, 14.9%) [30], and China (137/302, 45.4%) [31], and lower than in Iran (54/60, 90%) [32], South Africa (47/48, 98%) [33], and Egypt (90/90, 100%) [34].Studies from China, Korea, and Taiwan investigated solely the E. coli isolates and the Iran study included E. coli and K. pneumoniae.The rest of the three studies had various Enterobacterales isolates.The proportion and distribution of TMQR genes vary among different studies possibly due to the heterogeneity in the isolate selection criteria, the specific TMQR genes investigated, and the study population.Also, the actual proportion of TMQR could be slightly higher than reported in this study among uropathogens at our institution because they can be present even among nalidixic acidsusceptible Enterobacterales [11].Since we only included nalidixic-resistant isolates, we might have missed isolates with such phenotype.
A recent study demonstrated that possession of aac(6')-Ib-cr gives a selective advantage to E. coli ST131 in the presence of ciprofloxacin [35].In addition, alone or in  combination with chromosomal mutations, QnrS1 has been shown to increase bacterial fitness while QnrA1 and QepA2 decrease fitness [11].These observations could explain the predominance of aac(6')-Ib-cr and qnrS, and the low prevalence of qnrA and qepA.
Our data show that TMQR-positive isolates have higher FQ MIC than those that lack them similar to studies from Iran and Korea [36,37].The ciprofloxacin MIC values were significantly higher in TMQR positive isolates (Median = 64 µg/mL, n = 74) compared to TMQR negative isolates (Median = 32 µg/mL, n = 73) (Mann-Whitney U = 1969, Z=-2.871, p = 0.004, but with a small effect size of r = 0.24).Results from the analysis of 35 representative isolates suggest that the concomitant presence of GyrA and ParC substitutions accompanies TMQR genes to result in high levels of FQ resistance (Table 3).Similar findings of multiple substitutions in GyrA and ParC along with TMQR genes leading to high FQ resistance have been shown in other studies [34,38].
The statistically significant association of TMQR with ESBL observed in this study mirrors previous findings from several other studies [36,37,39,40].The β-lactamase genotypes, bla TEM and bla CTX−M , showed an independent association with TMQR, but the difference in the proportion of bla TEM was remarkably high between TMQR positive and TMQR negative group (47.3% vs. 1.4%) (Additional file 3: Table 1).Notably, of 57 ESBL-producing TMQR positive isolates, 29 (50.9%)co-harbored both bla TEM and bla CTX−M while no TMQR negative isolate had more than one β-lactamase gene (Table 2).These observations suggest that the association of β-lactamase and TMQR is driven by the coexistence of multiple β-lactamase genes rather than a single genotype in the study population.In a study from Iran investigating UTI caused by Enterobacterales, 72 (43.6%) isolates had the co-existence of bla CTX−M and bla TEM among 165 ESBL-producing TMQR positive E. coli and K. pneumoniae isolates [36].In contrast, among 155 ESBL-positive TMQR harboring K. pneumoniae (originating from various clinical specimens) in a study in Algeria, all 155 isolates had bla CTX−M only [41].Geographic, demographic, and differences in clinical specimens could account for this disparity.The predominance of bla CTX−M as the most common ESBL gene associated with TMQR is in an agreement with both Iranian and Algerian studies.We also demonstrate carbapenemase genes among TMQR-positive isolates, in contrast to previous findings; most studies either had no or negligible TMQR-positive isolate resistant to carbapenem [31,34,39,42].Two-thirds (14/21, 66.7%) of the carbapenemaseproducing isolates were TMQR-positive.Although, this was not a statistically significant association (Additional file 3: Table 1).WHO's GLASS report 2022 showed that more than 90% of antimicrobial use in Nepal in 2018 was attributed to oral administration reflecting their use in the community setting [43].Ciprofloxacin and cefixime were the second and third most consumed oral antimicrobials, respectively.Pathogens harboring resistance mechanisms for either or both of these two antimicrobials most likely thrived under such high selective pressure in the community.A recent study from a tertiary care center in Nepal with a large sample size (n = 2153) showed a high prevalence of isolates with overlapping resistance to extendedspectrum cephalosporin and fluoroquinolone in both inpatient and outpatient settings that is consistent with the hypothesis that these two groups of genes are cospreading in Nepal [44].TMQR and ESBL genes can be located in the same conjugative plasmid with other antimicrobial-resistant determinants, and this facilitates their simultaneous spread and contributes to the emergence of MDR [11].Such co-localization implies that the use of either quinolones or β-lactams could also promote the selection of these strains as suggested in a study from Vietnam [45].Considering that fluoroquinolone is typically avoided in children, high levels of fluoroquinolone resistance may be explained by the high prevalence of such strains, promoted by selective pressure in the community, and by the spread of strains with co-localization of TMQR and β-lactamases within the same conjugative plasmids.
TMQR genes seem to have community origin [42,46], and several studies have shown that commensal gut microbiota frequently harbors these genes [46,47], as do isolates from other body surfaces [48].With the growing appreciation of the involvement of the gut microbiome [49] and urinary microbiome in causing UTI [50], the detection and characterization of TMQR genes from the microbiome of these niches could be a future investigation at our institution.Characterization of the genetic background of TMQR (such as TMQR copy number, and expression level), conjugation experiments, phylogrouping, and MLST could be another aspect of focus for further research.The lack of data on whether the patients had consumed antimicrobials prior to hospital visits is a limitation of this study.In addition, we have not characterized other mechanisms of quinolone resistance, such as chromosomal efflux pumps, permeability alterations, and the role of biofilms.

Conclusions
Our study demonstrates alarmingly high levels of FQ resistance among E. coli and K. pneumoniae causing UTI in Nepalese children indicating the presence of high selective pressure in the community to promote their resistance.High prevalence and diversity in combinations of TMQR genes among quinolone-resistant E. coli and K. pneumoniae suggest an important role of these genes in the emergence of Q/FQ resistance.Also, the findings of this study highlight the dissemination of TMQR along with β-lactamases among the pediatric population in Nepal, amplifying multidrug resistance.The increase in FQ resistance is a challenge for the management of UTIs.Comprehensive prospective surveillance of antimicrobial resistance in these common pathogens will be necessary to understand the origin and spread of TMQR genes and to devise strategies to mitigate the emergence of further resistance.

Table 1
Proportion of TMQR positivity stratified by sex, age groups, and pathogens

Table 2
Co-occurrence of transferrable mechanisms of quinolone resistance determinants with β-lactamase gene combinations

Table 3
Distribution of GyrA and ParC substitutions in 30 E. coli and 5 K. pneumoniae isolates