Background

The rates of extended-spectrum beta-lactamase (ESBL)-production in Klebsiella pneumoniae are very high in hospitals worldwide, and highly resistant strains such as K. pneumoniae carbapenemase (KPC) or New Delhi metallo-beta-lactamase (NDM)-producing K. pneumoniae have been rapidly spreading between countries. Treatment of infections caused by multidrug-resistant K. pneumoniae has been a challenge, and there is increasing concern about the economic impact of these bacteria.

Serotype K1 K. pneumoniae has been reported as the predominant serotype among those isolates causing liver abscesses [1, 2]. Most K1 strains isolated from liver abscess cases belong to sequence type (ST) 23 in Asian countries [3, 4]. Although ESBL K. pneumoniae and KPC-producing K. pneumoniae have become globally widespread, a distinctive characteristic of K1 K. pneumoniae ST23 has been good susceptibility to most antibiotic classes (except ampicillin and piperacillin) [5], however, ESBL-producing K1 K. pneumoniae strains have begun to be reported recently [6, 7]. In this study, we report serotype K1 K. pneumoniae strains co-producing the plasmid-mediated AmpC beta-lactamase (PAB) DHA-1 and ESBL.

Materials and methods

Three ESBL-producing serotype K1 K. pneumoniae ST23 strains were identified from strains in the Asian Bacterial Bank (Asia Pacific Foundation for Infectious Diseases, Seoul, Korea), which had been collected during bacteremia studies in Korea between 2006 and 2008.

In vitro antimicrobial susceptibility testing was performed using the broth microdilution method according to the Clinical and Laboratory Standards Institute (CLSI) guidelines [8]. Ten antimicrobial agents were tested, including ampicillin (AMP), amikacin (AMI), ceftazidime (CAZ), tetracycline (TET), cefepime (CPM), cefotaxime (CTX), ciprofloxacin (CIP), rifampicin (RIF), and piperacillin/tazobactam (P/T). For ESBL-positive candidates, which presented a ceftazidime or cefotaxime minimum inhibitory concentration (MIC) ≥2 mg/L, production of ESBL was confirmed by the double-disk synergy test using BBL™ Sensi-Disc™ kits (Becton Dickinson, Franklin Lakes, NJ, USA) according to CLSI guidelines [8]. Quality control for ESBL production was performed using E. coli ATCC 25922 and K. pneumoniae ATCC 700603.

Double-disk synergy test-positive isolates were further tested by polymerase chain reaction (PCR) and sequence analyses to determine the gene responsible for the ESBL phenotype in the ESBL producers. In the ESBL-producing K. pneumoniae isolates, bla TEM, bla SHV, bla CTX-M, bla GES, bla PER, and bla VEB genes were tested by PCR. In addition, bla DHA-1, bla CMY, bla FOX, bla MOX, bla IMP, bla VIM, bla OXA, and bla NDM genes were tested. The sequences were compared with those in the GenBank nucleotide database for subtyping.

To determine serotype K1, PCR was conducted using a primer pair specific for magA, which is a gene specific for the K1 antigen. The primers were chosen as previously described: forward, 5′GGTGCTCTTTACATCATTGC-3′, and reverse, 5′-GCAATGGCCATTTGCGT TAG-3′ [9]. Multilocus sequence typing (MLST) was conducted using the nucleotide sequences of seven housekeeping genes (gagA, infB, mdh, pgi, phoE, rpoB, and tonB) as previously described [10]. For pulsed field gel electrophoresis (PFGE), agarose-embedded bacterial genomic DNA was digested with 20 U of XbaI. The restriction fragments were separated by gel electrophoresis in 0.5 × Tris-borate-EDTA buffer. Electrophoresis was performed using a CHEF Mapper XA (Bio-Rad Laboratories, Hercules, CA, USA). The PFGE patterns were analyzed using GelCompar II version 6.1 (Applied Maths, Belgium).

Results

Of the 120 K. pneumoniae isolates, 20 (16.7%) were ESBL-producers. Among the ESBL-producers, only three were determined to be serotype K1 (KPN1, KPN2, and KPN3). Antimicrobial susceptibilities of the three isolates are shown in Table 1. All three strains were resistant to ampicillin, piperacillin-tazobactam, cefotaxime, ceftazidime, and rifampin. Two strains (KPN2 and KPN3) were also resistant to amikacin and ciprofloxacin. One strain (KPN2) was resistant to imipenem. In all three strains, no ESBL genes, including bla TEM, bla SHV, bla CTX-M, bla GES, bla PER, and bla VEB, were detected. Among the PAB genes (bla DHA-1, bla CMY, bla FOX, and bla MOX), only bla DHA-1 was detected in two strains (KPN2 and KPN3). Carbapenemase genes were also not detected in all three isolates (Table 2). Clonal relatedness was investigated by PFGE for the two strains. Analysis of the PFGE patterns showed that the two isolates carrying bla DHA-1 were closely related (84% similarity).

Table 1 Antimicrobial susceptibilities in K1 K. pneumoniae ST23 strains
Full size table
Table 2 Detection of ESBL genes and plasmid-mediated AmpC β-lactamase genes in K1 K. pneumoniae ST23 strains
Full size table

Discussion

Serotype K1 K. pneumoniae ST23 is, a highly virulent pathogenic strain causing invasive community-acquired infections, that is widespread in its geographical distribution in Asia [5]. Fortunately, ST23 strains have shown good susceptibility to most antibiotics. However, the advent of multidrug resistance with high transmission potential in K. pneumoniae serotype K1 causes serious concerns. A few strains of ESBL-producing serotype K1 K. pneumoniae have already been reported in Asian countries [6, 7, 11]. In the present study, we report serotype K1 K. pneumoniae strains co-producing the PAB DHA-1 and ESBL. We found three isolates that were phenotypically ESBL producers, but we could not detect any bla genes responsible for the ESBL phenotype. Therefore, other ESBL genotypes may be involved. In addition, two of the three isolates were identified as PAB. Recently, gram-negative organisms that produce both ESBLs and PAB enzymes have increasingly been described worldwide [12–15]. To our knowledge, this is the first report of serotype K1 K. pneumoniae producing both ESBL and PAB. Both ESBLs and PAB enzymes are associated with broad, multidrug resistance because multiple antibiotic-resistance genes exist on the same plasmid [16]. One isolate (KPN3) in our study showed carbapenem resistance in the broth microdilution test. Subsequently, we used PCR analysis to determine the presence of carbapenemase genes, such as bla IMP, bla VIM, bla OXA, and bla NDM; however, none of these genes were detected. The frequency of PAB transmission may be higher than initially thought, especially if the spread of resistance mimics the trend that we have seen occurring over the past few years for ESBLs. Moreover, porin alteration, combined with the production of ESBL or PAB, has been demonstrated to confer carbapenem resistance [17, 18]. We also characterized antimicrobial susceptibility between non-serotype K1 and serotype K1 ESBL-producing K. pneumoniae. Although the number of total isolates was too small to draw definite conclusions, the resistant rates of piperacillin-tazobactam and imipenem in K1 isolates tended to be high. It is uncertain whether the increase of antimicrobial resistance in serotype K1 K. pneumoniae is connected with any one particular resistance gene.

The emergence of multidrug-resistant strains with high transmission potential in serotype K1 K. pneumoniae is of great concern due to limited alternative treatment options and the possibility of global dissemination. Careful surveillance of resistant strains and adequate infection prevention and control measures are necessary.