Multidrug resistance genes , including bla KPC and bla CTX-M-2 , among Klebsiella pneumoniae isolated in Recife , Brazil

Introduction: The prevalence of cephalosporins and carbapenem-resistant Klebsiella pneumoniae strains is rising in Brazil, with potential serious consequences in terms of patients’ outcomes and general care. Methods: This study characterized 24 clinical isolates of K. pneumoniae from two hospitals in Recife, Brazil, through the antimicrobial susceptibility profile, analyses of β-lactamase genes (blaTEM, blaSHV, blaCTX-M, blaKPC, blaVIM, blaIMP, and blaSPM), plasmidial profile and ERIC-PCR (Enterobacterial repetitive intergenic consensus-polymerase chain reaction). Results: ERIC-PCR and plasmidial analysis grouped the isolates in 17 and 19 patterns, respectively. Six isolates from one hospital presented the same pattern by ERICPCR, indicating clonal dissemination. All isolates presented blaSHV, 62.5% presented blaCTX-M-2, 29% blaTEM, and 41.7% blaKPC. Metallo-β-lactamase genes blaVIM, blaIMP, and blaSPM were not detected. Eleven isolates were identified carrying at least 3 β-lactamase studied genes, and 2 isolates carried blaSHV, blaTEM, blaCTX-M-2, and blaKPC simultaneously. Conclusions: The accumulation of resistance genes in some strains, observed in this study, imposes limitations in the therapeutic options available for the treatment of infections caused by K. pneumoniae in Recife, Brazil. These results should alert the Brazilian medical authorities to establish rigorous methods for more efficiently control the dissemination of antimicrobial resistance genes in the hospital environment.

The carbapenems are often used for the treatment of infections caused by K. pneumoniae ESBLs producers.However, this species has presented an efficient mechanism of resistance to carbapenems, known as Klebsiella pneumoniae carbapenemase (KPC), which was initially detected in the United States and Israel and later in other countries 6 .In Brazil, KPC enzymes have been described since 2009 7 when originally reported in isolates of K. pneumoniae from Recife 8 .Considering that Monteiro et al. 8 , reported the one-time identification of KPC enzymes in Recife and analyzed only 4 K. pneumoniae isolates and 3 of them were clones, is of utmost importance the study of the

Plasmid analysis and ERIC-PCR
Plasmid DNA was extracted by the UltraClean Endotoxin-Free Mini Plasmid Prep Kit (Mo Bio Lab, USA).The primers described by Duan et al. 18 were used for ERIC-PCR.The amplification reactions were prepared in a total volume of 25µl, containing 100ng of genomic DNA, 1.0U of Taq DNA polymerase (Promega, USA), 200µM of each dNTP (Promega, USA), 1.52mM of MgCl 2 , 0,4µM of each primer and 1X reaction buffer (final concentration).The PCR amplifications were performed in a thermocycler (Biosystems) as follows: 95ºC for 3min

Antimicrobial susceptibility profile
The antimicrobial susceptibility profiles of the 10 K. pneumoniae isolates from the public hospital (Hospital A) showed that the carbapenems presented the highest levels of activity; imipenem inhibited all isolates and meropenem inhibited 9 (90%) isolates.The aminoglycosides, amikacin and streptomycin inhibited 80% of the isolates, and gentamicin inhibited 50% of the isolates.The other antimicrobials inhibited less than 50% of the isolates.The aminoglycosides were the most effective antimicrobials against the isolates from the private hospital (Hospital P), followed by the carbapenems.Interestingly, the isolates K2A and K7A, and K8P and K12P showed pair wise identical antimicrobial resistance profiles and were not clones according to the ERIC-PCR results.

ERIC-PCR
Six K. pneumoniae isolates from Hospital P (K10P, K12P, K13P, K15P, K16P, and K22P) presented the same band pattern by ERIC-PCR (profile E11) indicating clonality (Figure 1).All other isolates from both hospitals showed distinct genetic profiles (Table 3), with a maximum of 60% similarity (Figure 1).These other isolates were considered genetically unrelated by differing in at least seven bands, which indicates changes that are consistent with the occurrence of three or more independent genetic events.and 40 cycles of 1min at 92ºC, 1min at 36°C and 8min at 72ºC.A final extension step of 16min at 72ºC was performed.The products of plasmid extraction and ERIC-PCR were analyzed by electrophoresis on agarose gel 0.7% and 1.5%, respectively.The band patterns generated by ERIC-PCR were analyzed according to Tenover et al. 19 and the software DARwin 5.0 was used to generate dendrogram.

β-lactamase phenotypic and genotypic detection
Out of the 24 K. pneumoniae clinical isolates studied, 9 (37.5%) were ESBL producers by the DDST (double-disc synergy test).The non-ESBL producers by DDST presented at least one of the three studied genes, bla SHV , bla TEM, or bla CTX-M .All isolates that were positive by PCR for bla CTX-MA were positive for bla CTX-M-2 .The gene bla CTX-M-1 was not detected.All 24 K. pneumoniae isolates presented the bla SHV , 15 (62.5%) presented the bla CTX-M-2 , and 7 (29%) had the bla TEM (Table 3).Based on two characteristics, the presence of the bla CTX-M-2 gene (15 isolates) and a positive DDST (K3P), 67% (16/24) of the isolates were identified as ESBL producers.The most frequently observed combination of genes was bla SHV and bla CTX-M-2 (42%), followed by bla SHV and bla TEM (8%); the bla CTX-M-2 and bla TEM combination was not observed.The three β-lactamase genes studied were identified simultaneously in five isolates (21%): K1A, K5A, K10A, K1P, and K14P (Table 3).The ERIC-PCR results showed that these isolates were not clones.

KPC phenotypic and genotypic detection
The gene bla KPC was identified in 10 (71%) isolates from the Hospital P, on the other hand, this gene was not detected in the isolates from the hospital A. The isolate K14P did not show resistance to carbapenems, was negative in the MHT, however, it carried the bla KPC gene (Table 3).The K8P, K12P, K16P, and K20P isolates showed negative MHT results but presented resistance to carbapenems and contained the bla KPC gene.The isolate K5P was resistant to carbapenems, however, it was MHT negative and did not carry the bla KPC gene.The bla VIM , bla IMP , and bla SPM genes were not detected in any of the isolates analyzed in this study.

576
DISCUSSION Multidrug resistance among ESBL-producer isolates is predictable because the genes that encode for ESBLs are usually located in self-transferable plasmids, which can also carry other antimicrobial resistance genes 20 .All isolates were tested against 9 classes of antimicrobials, however, for the majority, the therapeutic options were restricted to a maximum of 4 classes, evidencing the problematic aspects of treatment of multidrug resistant K. pneumoniae isolates.It was possible

Plasmidial profile analysis
All isolates harboured 1 to 10 plasmids of different sizes (Table 4).Plasmids of >150 kb, 150kb, and ≤3.4kb were identified in 50%, 54%, and 83% of the isolates, respectively.Plasmids of 100kb, 80kb, and 12kb were identified in 57%, 78%, and 50% of the isolates from the Hospital P, suggesting a dissemination of these plasmids.It should be noted that the isolates K4A and K10A; K12P and K13P; K14P and K19P; and K15P and K16P, presented the same plasmidial profile, defined by the number and molecular weight of the bands (Table 4).
to establish an association between antimicrobial resistance profiles and β-lactamases genes because the isolates K2A and K7A, and K8P and K12P showed the same β-lactamase genes and the same resistance profile.All isolates from the Hospital A that had two or three genes (bla SHV , bla TEM, and bla CTX-M-2) were resistant to the oxyiminocephalosporins tested (ceftazidime, cefotaxime and cefepime) and to aztreonam, whereas isolates that carried only bla SHV were resistant to only one of these antimicrobials.On the other hand, this association was not observed between the isolates from the Hospital P because they were resistant to all tested cephalosporins and monobactam, regardless of the number of genes carried.
The DDST was effective in detecting the ESBL producers among all isolates from the Hospital A (PCR positive for bla CTX-M-2 ) that were KPC negative.On the other hand, this test failed to detect 7 ESBL producers from the Hospital P. The failure of DDST to detect ESBLs in the K1P, K10P, K13P, K14P, K15P K16P Hospital P isolates can be explained by the concomitant presence of the KPC carbapenemase.Tsakris et al. 21howed that the phenotypic detection of ESBLs is hampered when the isolates also produce KPC; therefore, these authors suggest the addition of boronic acid to the test to inhibit the KPC enzymes, thus facilitating the detection of the ESBL enzymes.
The negative result in the ESBL detection through the DDST for the isolate K5P, which was positive for the bla CTX-M and negative for the bla KPC , can be explained by the fact that: I) this is not a confirmatory test for the detection of ESBLs 9 because it relies on the bacterial phenotype that can be altered by the conditions of in vitro cultivation; II) additionally, the test only detects ESBLs that are inhibited by the clavulanic acid (36 of 135 TEM enzymes, and 5 of 72 SHV enzymes have currently demonstrated resistance to the clavulanic acid and related inhibitors, however, no CTX-M has demonstrated this feature so far) 2 ; and III) the molecular methods present higher sensitivity for the detection of ESBLs compared to the phenotypic methods.
The MHT did not detect carbapenemases in five isolates carrying the bla KPC gene and resistant to carbapenems.In Brazil, it is more common to encounter false positive results for carbapenemase production, for example, in isolates of K. pneumoniae cefotaximases producers that simultaneously present loss of porins 22 .The problematic detection of KPC in Enterobacteriaceae has been reported by other authors, supporting the suggestion of phenotypic detection of KPC using boronic acid 23 .
In spite of an early dissemination of KPC-2-producing K. pneumoniae strains in Brazil, bla KPC gene was not detected in the isolates from Hospital A. This fact may indicate a good clinical practice in Hospital A. The bla KPC gene occurrence was 71% in the isolates from the Hospital P, which is equal to the occurrence of the ESBL producers in this hospital.Probably, this fact can limit the therapeutic options and increase the mortality rates.Another relevant finding was the lack of expression of resistance to carbapenems by K14P isolate that was positive for bla KPC .Similar results were reported by Peleg et al. 24 , showing that only 5 out of 19 isolates carrying carbapenemases genes expressed resistance to carbapenems.The acquisition of genes that encode for carbapenemases is not always associated with high levels of resistance to carbapenems 25 .This variable susceptibility can be explained by a number of factors such as: I) co-presence of other mechanisms of resistance 26 ; II) genetic suppression leading to a silenced gene; or III) dosage of the gene that is dependent on the plasmid copy number 24 .
The presence of identical plasmid profile in different isolates from Hospital P suggests that plasmids with the same molecular weight may have disseminated among these isolates.Aktas et al. 27 , considered that K. pneumoniae is a nosocomial reservoir and the source for the transmission of resistance plasmids.According to Sharma et al. 28 , there is a correlation between the number of plasmids harbored by an isolate and the multidrug resistance.This association was not observed in this study; the isolates showing resistance to the greatest number of drugs (K8P and K12P) did not carry the greatest number of plasmids, and the isolates resistant to fewer drugs (K3A and K4A) were not the ones carrying the smallest number of plasmids.Additionally, the isolates that presented the same antimicrobial resistance profile did not present the same plasmidial profile.
No correlation was observed in a quantitative analysis between the plasmids and the bla SVH , bla TEM , and bla CTX-M genes, because some isolates were positive for only one gene, but carried one to nine plasmids.Other isolates presented two genes with varying numbers of plasmids.K1P was positive for all β-lactamases genes studied (bla SHV, bla TEM , bla CTX-M , and bla KPC ) and carried only two plasmids (100kb and 8kb).According to Wei et al. 29 , the number of antibiotic resistance genes in plasmids of multidrug-resistant K. pneumoniae can reach up to five.The presence of high molecular weight plasmids (60 to 180Kb) in clinical isolates of K. pneumoniae is quite common and has often been associated with the production of ESBLs 29,30 .
The ERIC-PCR revealed that 6 isolates from the Hospital P showed clonal relationship even when presenting different antimicrobial resistance phenotypes, contents of resistance genes, and plasmidial profiles.Molecular typing data reported in the literature corroborate these results.Ben-Hamouda et al. 31 , reported isolates with identical PFGE or ERIC-PCR profiles and different drugs susceptibility profiles.Bennett et al. 32 did not observe any correlation between PFGE profiles and resistance genes.Souza Lopes et al. 30 identified RAPD subtypes correlated with different plamidial profiles.These discrepancies in the typing techniques result from the fact that, the evidences for clonality are best considered as relative instead of absolute, because of the potential of genetic changes detectable only by DNA sequencing or other specific analyses.
The finding that 11 (46%) isolates were carriers of at least 3 of the β-lactamase studied genes is worrying.It should be noted that the isolates K1P and K14P presented the bla SHV, bla TEM , bla CTX-M , and bla KPC genes and were not clones.The simultaneous production of three β-lactamases by K. pneumoniae deserves to be highlighted 33 .
The accumulation of resistance genes in some strains, observed in this study, imposes limitations in the therapeutic options available for the treatment of infections caused by K. pneumoniae in Recife, Brazil.The presence of intra-hospital clones was another remarkable observation.These results should alert the Brazilian medical authorities to establish rigorous methods for detection and thus, more efficiently control the dissemination of antibiotic resistance genes in the hospital environment.Moreover, measures to control the excessive use of antimicrobials in hospitals must also be taken.

TABLE 2 -Primers used for polymerase chain reaction to detect β-lactamase genes. Primer name Primer sequence (5'-3') Temp* Reference Gene
*Temp: annealing temperature.All primer names were described according the respective references F