Correlation of Multi-drug Resistance, Integron and blaESBL Gene Carriage With Genetic Fingerprints of Extended-Spectrum β-Lactamase Producing Klebsiella pneumoniae

Background: Some genetic and phenotypic variables are associated among distinct microbial populations. Objectives: The associations between multi-drug resistance (MDR) phenotypes, prevalence of antibiotic resistance integrons (ARIs), blaSHV, blaTEM and blaCTX-M gene carriage and genetic fingerprints of random amplified polymorphic DNA (RAPD), confirmed by pulsed field gel electrophoresis (PFGE), were investigated among extended-spectrum β-lactamases (ESBL)-producing nosocomial isolates of Klebsiella pneumoniae. Materials and Methods: Susceptibility of 35 ESBL-producing K. pneumoniae nosocomial isolates to 22 antimicrobial agents was determined. Integron carriage was detected using specific primers for intI1, intI2 and intI3 genes by PCR. Results: All isolates were resistant to piperacillin and susceptible to imipenem. MDR phenotype was observed in 91.4% of the isolates. Class 1 integrons were detected in 21 (60%) and class 2 integrons in 3 (8.57%) of the isolates. Two of the isolates carried both classes and none harbored class 3 integrons. Significant correlations were observed between resistance to aminoglycosides, fluoroquinolones and sulfonamides, and between genotype groups with carriage of ARIs, MDR phenotype and blaSHV gene carriage. ARI carriage was also significantly associated with MDR phenotype. Conclusions: Our findings suggest the possible co-carriage of some blaSHV genes and ARIs on the same plasmids harboring the MDR genes. Possible role of integrons in dissemination of ESBL-encoding blaSHV genes among ESBL-producing K. pneumoniae nosocomial isolates may be inferred.


Background
Klebsiella pneumoniae is responsible for up to 10% of all nosocomial infections (1,2). The importance of the organism in hospital settings has been increasing due to the emergence and progressive spread of multidrug resistance; specifically the extended-spectrum β-lactamase (ESBL)-producing strains (3). More than 600 ESBL variants have been described and the majority of them belong to the SHV, TEM and CTX-M families (http://www.lahey.org/ studies/webt.htm) (3). Horizontal gene transfer due to mobile genetic elements such as insertion sequences, transposons and conjugative plasmids, mediates intra and interspecies dissemination of not only the genes encoding ESBLs but also other antibiotic resistance determinants which are likely to form part of an antibiotic resistance integron (ARI) (3)(4)(5).
Three classes of ARIs (classes 1, 2, and 3) have been histor-ically involved in multi-drug resistant (MDR) phenotypes and are identified based on their respective integrase genes (5). Various typing methods have been applied to understand transmission patterns of resistance genes and management of nosocomial infections (6). We have previously developed an optimized RAPD-PCR protocol for genotyping K. pneumoniae, comparable to PFGE (7). To understand the associations between phenotypic and genetic characteristics of multi-drug resistant pathogens can be useful for reliable detection of these bacteria in epidemiological studies. Some reports have suggested associations between ESBL production and resistance to several classes of antibiotics, as well as bla ESBL with ARI genes carriage in K. pneumoniae (4,8).

Objectives
In this study, the association between MDR phenotypes, prevalence of ARIs, bla ESBL genes and RAPD profiles were investigated among ESBL-producing K. pneumoniae nosocomial isolates.

Genetic Fingerprinting and Characterization of blaESBL Genes
Genetic profiles of the isolates by RAPD, confirmed by PFGE have been reported in our previous article (7). Presence of bla ESBL genes (bla SHV , bla TEM and bla CTX-M ) and the sequencing result for the isolates were also previously reported (9).

Statistical Analyses
To assess the strength and statistical significance of correlations between the studied variables including patient gender, type of specimen, antimicrobial susceptibility, MDR phenotypes (resistance to 6 or more antibiotics), carriage of ARIs, bla SHV , bla TEM and bla CTX-M genes and genotype grouping, and also measure the association between resistance to each of the aminoglycoside, quinolone and sulfonamide antibiotics, separate bivariate analyses were performed by use of the non-parametric Spearman's rank correlation test. To confirm the association between each pair of significantly correlated variables after factoring out the effect of other effective variables, partial correlation analyses were used. To interpret the results of correlation analyses, we considered correlation coefficients (r values) as well as the levels of significance (P values).

Results
The antibiotic susceptibility results are shown in Figure 1. As observed, all isolates were resistant to piperacillin followed by 97.1% resistance to co-amoxiclav, 94.3% to aztreonam, 88.6% to kanamycin and cefotaxime, 85.7% to cefpodoxime, 82.9% to tobramycin and ceftazidime, 74.3% to ceftriaxone and ampicillin/sulbactam, 71.4% to spectinomycin and cotrimoxazole, 68.6% to cefepime, 60% to norfloxacin, 48.6% to gentamicin, 45.7% to ciprofloxacin, 28.6% to amikacin, 25.7% to piperacillin/tazobactam and 11.4% to nitrofurantoin and colistin. All isolates were susceptible to imipenem. Streptomycin resistance was not observed but 31.4% of the isolates showed intermediate resistance. The most active antibiotic was imipenem followed by streptomycin, colistin and nitrofurantoin. Significant associations were observed between resistance to kanamycin, tobramycin, gentamicin, amikacin, norfloxacin, ciprofloxacin and cotrimoxazole (Table 1). Class 1 integrons were detected in 21 isolates (60%) and class 2 integrons in 3 isolates (8.57%). Two of the isolates carried both classes and none harbored class 3 integrons.

Discussion
Infections due to ESBL-producing strains, have been most commonly reported regarding K. pneumoniae (3). ESBL encoding genes are usually located on plasmids which may also carry other antibiotic resistance determinants. Reports have suggested a close association between ESBL production and ciprofloxacin resistance in K. pneumoniae (8). Co-resistance with other classes of antibiotics such as fluoroquinolones, aminoglycosides, tetracyclines, chloramphenicol and sulfonamides are also widespread among ESBL producing strains (12). This may explain the significant associations found between resistance to aminoglycosides (kanamycin, tobramycin, gentamicin and amikacin) in this study. The same trend was observed for the association of resistance between norfloxacin with kanamycin, tobramycin, ciprofloxacin and cotrimoxazole. Similarly, resistance to ciprofloxacin and gentamicin were associated showing a relationship as a sign of co-carriage.
Bivariate correlation analyses followed by partial correlation analyses in order to distinguish between direct and indirect interactions, confirmed the results. Despite high heterogeneity observed among the isolates of this study, genotyping results were strongly correlated with carriage of ARIs and bla SHV genes. Although almost all K. pneumoniae isolates carry chromosomal non-ESBL bla SHV-1 , nearly all ESBL encoding bla SHV genes found in K. pneumoniae are plasmid borne (13,14). In this study, RAPD profiles were strongly correlated with the presence of bla SHV genes suggesting that plasmid mediated bla SHV-5 and bla SHV-12 (the two prevalent ESBL encoding bla SHV genes among our isolates) had some influence on RAPD patterns. Possible contribution of plasmid DNA to RAPD patterns was suggested in K. pneumoniae (15). However, Elaichouni et al. found no influence of plasmid DNA on the RAPD profiles in Escherichia coli and claimed that the amount of chromosomal DNA per cell in natural conditions inhibits observable plasmid amplification (16). The association of bla ESBL genes with ARIs occurs when both form parts of complex integrons or are located on the same plasmid (4,17). We found a positive association between class 1 integrons and bla SHV-11 , bla SHV-5 and bla SHV-12 at the confidence level of 90% (P < 0.1). Since genotyping results were highly correlated with the carriage of both ARIs and bla SHV , it could be concluded that ARIs and bla SHV genes are carried on the same plasmids, or bla SHV genes are located within ARIs at least among some of our isolates. Association between ARIs and bla SHV-5 as well as co-location of bla SHV-12 and a class 1 integron on the same plasmid have been reported (17,18). However, other investigators have found a low rate of association between integrons and ESBL genes with the exception of bla CTX-M-9 (19).
Presence of plasmids that carry ESBL encoding genes as well as integron mediated antibiotic resistance has been reported among nosocomial isolates of K. pneumoniae (17,19,20). In most of these studies, ESBL encoding genes were located on plasmids but not within the integrons. Although most of the findings so far suggest contribution of integrons in the acquisition and transmission of resistance genes among bacteria, further investigations are needed to evaluate the involvement of other factors in transmission of linked resistance genes.