A Swordless Knight: Epidemiology and Molecular Characteristics of the bla_KPC-Negative Sequence Type 258 Klebsiella pneumoniae Clone

Abstract

In June 2010, a bla_KPC-negative, ertapenem-resistant ST-258 Klebsiella pneumoniae strain was isolated from a patient in the Laniado Medical Center (LMC). Our aims were (i) to describe its molecular characteristics and resistance mechanisms and (ii) to assess whether the bla_KPC-negative ST-258 K. pneumoniae clone spreads as efficiently as its KPC-producing isogenic strain. In a prospective study, surveillance of all ertapenem-resistant, carbapenemase-negative K. pneumoniae (ERCNKP) isolates was conducted from June 2010 to May 2011 at LMC (314 beds) and from July 2008 to December 2010 at the Tel Aviv Sourasky Medical Center (TASMC) (1,200 beds). Molecular typing was done by arbitrarily primed PCR, pulsed-field gel electrophoresis (PFGE), and multilocus sequence typing (MLST). A total of 8 of 42 (19%) ERCNKP isolates in LMC and 1 of 32 (3.1%) in TASMC belonged to the ST-258 clone. These strains carried the bla_CTX-M-2 or the bla_CTX-M-25 extended-spectrum β-lactamase (ESBL) gene. Sequencing of the ompK genes showed a frameshift mutation in the ompK35 gene. The fate of the bla_KPC-carrying plasmid, pKpQIL, was determined by S1 analysis and by PCR of the Tn4401 transposon, repA, and the truncated bla_OXA-9. Plasmid analysis of the ERCNKP ST-258 isolates showed variability in plasmid composition and absence of the Tn4401 transposon and the pKpQIL plasmid. In addition, the ST-258 clone was identified in 35/35 (100%) of KPC-producing K. pneumoniae isolates but in none of 62 ertapenem-susceptible K. pneumoniae isolates collected in the two centers. Our results suggest that ERCNKP ST-258 evolved by loss of the bla_KPC-carrying plasmid pKpQIL. ERCNKP ST-258 appears to have low epidemic potential.

INTRODUCTION

The first cases of Klebsiella pneumoniae carbapenemase (KPC)-2- and KPC-3-producing K. pneumoniae were reported in the mid-Atlantic coastal region of the United States between 1997 and 2000 (6, 48, 49). In 2006, a strain of KPC-3-producing K. pneumoniae, exhibiting resistance to nearly all antimicrobial agents, spread in all major Israeli hospitals (28, 35). This strain, designated pulsed-field gel electrophoresis (PFGE) clone Q, ultimately identified as sequence type (ST)-258, was identical to the epidemic strain that has spread across the United States (26), Europe, and South America (3, 4, 13). Unlike the situation for other carbapenemases, such as OXA-48, where the spread of resistance has been mediated mainly by transmission of mobile genetic elements (2, 9), the success of the ST-258 clone has been a key factor in the dissemination of KPC, making it by far the most important class A carbapenemase (36, 37). This predominance of a single clone in the dissemination of resistance has several precedents, such as the spread of extended-spectrum β-lactamase (ESBL)-producing Escherichia coli via the ST-131 clone (41) and the spread of methicillin-resistant Staphylococcus aureus (MRSA) via the USA-300 clone (16). The molecular mechanisms behind the success of the KPC-producing ST-258 clone have remained obscure. An intriguing finding in this regard is that to date, carbapenem-susceptible ST-258 has never been reported in the literature as associated with any particular factor involved with either virulence or epidemiological success (7, 15, 39, 46). In June 2010, the National Center for Infection Control (NCIC) in Israel received a report concerning a high incidence of ertapenem-resistant, carbapenemase-negative K. pneumoniae (ERCNKP) infections at the Laniado Medical Center (LMC). Ten patient-unique isolates were typed by PFGE, and surprisingly, two isolates were identified as ST-258 (see below). Following this finding, the NCIC, LMC, and the Tel Aviv Sourasky Medical Center (TASMC) initiated an epidemiological investigation, designed to explore the epidemiology of ERCNKP in these hospitals. In particular, we aimed to assess whether the carbapenemase-negative ST-258 K. pneumoniae strain had achieved a degree of epidemiological success comparable to that observed in its KPC-producing isogenic strain and to describe its molecular characteristics and resistance mechanisms.

MATERIALS AND METHODS

Settings and infection control practices.

The study was conducted as part of a joint epidemiological investigation by the Israeli NCIC (45), LMC, and TASMC. LMC, a 300-bed hospital, is the only acute-care medical center in the city of Netanya, covering a population of approximately 200,000 people. TASMC is a 1,200-bed tertiary care hospital in Tel Aviv, Israel. TASMC serves as a referral center for the greater Tel Aviv area, including Netanya. Thus, there is a constant flow of patients between the institutions.

In both centers, patients infected or colonized by carbapenem-resistant Enterobacteriaceae (CRE) were put under contact isolation; patients colonized by a carbapenemase-producing isolate of the Enterobacteriaceae were kept as cohorts in separate rooms, according to the national Israeli guidelines (45).

Active surveillance for CRE was performed in the following cases: (i) in patients located in the proximity of a newly identified CRE-colonized/infected patient (contact investigation) at both centers and (ii) in all newly admitted patients with a history of previous admission (within 1 year) to a health care facility at TASMC only.

The total consumption of carbapenem antibiotics (in defined daily doses [DDD] per 1,000 hospital days) at TASMC in 2009 and 2010, respectively, was as follows: meropenem, 12.66 and 8.88; imipenem, 10.05 and 9.72; ertapenem, 7.71 and 7.72.

The total consumption of carbapenem antibiotics (in DDD per 1,000 hospital days) in LMC in 2009 and 2010, respectively, was as follows: meropenem, 4.12 and 6.76; imipenem, 3.73 and 1.32; ertapenem, 4.71 and 3.19.

Microbiological data were collected using the computerized data systems of TASMC, LMC, and the NCIC. Strains from LMC were shipped to the laboratory at TASMC for molecular studies.

Bacterial strains.

ERCNKP was defined based on the following criteria: (i) species (K. pneumoniae), (ii) resistance to ertapenem, with or without nonsusceptibility to imipenem or meropenem, and (iii) lack of carbapenemase (see details below). All ERCNKP isolates found at LMC from June 2010 to May 2011 were prospectively collected and studied. At TASMC, 34 of 81 ERCNKP isolates identified from July 2008 to December 2010 were available for study. Patient-unique, KPC-producing or ertapenem-susceptible K. pneumoniae isolates were randomly collected from June 2010 to December 2011 in the two centers. KPC-producing ST-258 control strains were available from our collection (28).

Microbiological methods.

Rectal surveillance cultures were performed at both centers as previously described (1). Identification and antimicrobial susceptibility testing of bacterial strains were performed by the Vitek-2 system using GN-ID and GN09 cards (bioMérieux, Marcy l'Etoile, France). Ertapenem, imipenem, meropenem, and colistin MICs were verified by Etest (AB Biodisk, Solna, Sweden). Susceptibility was determined using MIC breakpoints of the Clinical and Laboratory Standards Institute (CLSI) 2011 criteria (12).

Molecular epidemiology studies.

Arbitrarily primed PCR (AP-PCR) (5) was used to screen for clone Q (see Fig. S1 in the supplemental material). The results were confirmed by PFGE (28). MLST was performed on representative PFGE clone Q isolates (17) and analyzed using the K. pneumoniae MLST website (http://www.pasteur.fr/recherche/genopole/PF8/mlst/Kpneumoniae.html).

Determination of resistance mechanisms.

A lack of carbapenemase was determined based on a negative result on a modified Hodge test (MHT) (12), negative EDTA synergy testing (22), and a negative bla_KPC (43) and bla_OXA-48 (40) PCR in all ERCNKP isolates. The ERCNKP clone Q isolates were additionally tested for synergy between ertapenem and boronic acid and ertapenem and dipicolinic acid (22) and for carbapenem-hydrolyzing activity by the imipenem hydrolysis assay (30). All isolates were tested by bla_KPC-PCR, excluding LMC MHT-positive isolates, among which only representative isolates (n = 17) were available for testing. In addition, ERCNKP clone Q isolates were tested by PCR for bla_NDM-1, bla_VIM, bla_IMP, bla_SIM-1, bla_GIM-1, bla_SPM-1 (18), and mobile ampC (14). Amplification of genes belonging to the bla_CTX-M group was done using multiplex PCR (47) followed by sequencing (34). Characterization of bla_SHV and bla_TEM was done by PCR and sequencing (44). The outer membrane protein (OMP) genes ompK35, ompK36, and ompK37 were studied by sequencing of their PCR products (25).

Plasmid analysis.

Plasmid DNA from the nine ERCNKP clone Q isolates was extracted and studied by S1 analysis (23). All representative plasmids (according to size) were gel extracted as previously described (42) for further individual analysis. To determine whether pKpQIL, the bla_KPC-harboring plasmid (29, 31), was modified or absent, three genetic elements of this plasmid were screened by PCR: the repA gene (8), the truncated bla_OXA-9 gene (F′ primer, GCTGCATATGTTGGTGTTCG; R′ primer, TTGCTCCTTGGGAGATATGG), and the Tn4401 transposon, tested at an internal region (F′ primer, TCTCCAGCACCCACTGTCTG; R′ primer, TGCCGTGATGAAGCGTGT) and at its flanking regions as previously described (10).

RESULTS

Epidemiology of carbapenem-resistant K. pneumoniae strains.

The monthly incidences of ERCNKP, carbapenemase-positive K. pneumoniae, and carbapenem-susceptible K. pneumoniae in clinical cultures at LMC are presented in Fig. 1A. The average monthly incidences of ERCNKP and carbapenemase-positive K. pneumoniae were 3.04 and 3.64 per 10,000 patient days, respectively. The proportions of ERCNKP and carbapenemase-positive isolates among all K. pneumoniae isolates were 7.7% and 9.3%, respectively.

Fig 1.

Molecular epidemiology of ertapenem-resistant, carbapenemase-negative K. pneumoniae (ERCNKP) strains. (A) Monthly incidence of ERCNKP and carbapenemase-producing and carbapenem-susceptible K. pneumoniae infections at the Laniado Medical Center, June 2010 to May 2011. *, clone Q ERCNKP isolate. (B) Quarterly incidence of ERCNKP and KPC-producing and carbapenem-susceptible K. pneumoniae infections at the Tel Aviv Sourasky Medical Center, July 2008 to December 2010. *, clone Q ERCNKP isolate.

The quarterly incidences of ERCNKP, KPC-producing K. pneumoniae, and carbapenem-susceptible K. pneumoniae in clinical cultures at TASMC are presented in Fig. 1B. The average monthly incidences of ERCNKP at TASMC were 0.17, 0.60, and 0.5 per 10,000 patient days in 2008, 2009, and 2010, respectively. The average monthly incidences of KPC-producing K. pneumoniae at TASMC were 1, 0.82, and 0.98 per 10,000 patient days in 2008, 2009, and 2010, respectively. The proportions of ERCNKP among all K. pneumoniae isolates were 0.6%, 2.3%, and 2.0% in 2008, 2009, and 2010, respectively. The proportions of KPC-producing isolates among all K. pneumoniae isolates were 3.3%, 3.2%, and 3.9% in 2008, 2009, and 2010, respectively.

Frequency of clone Q in different K. pneumoniae populations.

From June 2010 to May 2011, a total of 42/43 ERCNKP isolates from LMC (clinical isolates, 35; surveillance culture isolates, 7) were typed by AP-PCR and PFGE as described (see Fig. S1 in the supplemental material). Eight isolates (19%) were identified as belonging to clone Q (Table 1, Fig. 1A). All clone Q isolates were from noninvasive sites (Table 1), mainly from urine samples collected from indwelling catheters (n = 6), with four of them isolated from a single internal medicine ward during a 2-month period (February-March, 2011) (Fig. 1A). In addition, a retrospective analysis of 32 (clinical isolates, 24; surveillance culture isolates, 8) of 81 ERCNKP isolates collected at TASMC from July 2008 to December 2010 was performed. A single blood culture isolate (3.1%) from September 2009 was identified as belonging to Q (Table 1 and Fig. 1B). None of the patients infected with the ERCNKP Q strain had a documented history of infection by KPC-producing K. pneumoniae.

Table 1.

Microbiological and molecular characteristics of ERCNKP clone Q strains

Strain no.	Medical center	Isolation date (mo/yr)	Isolation site	MICa				Gentamicin resistanceb	ESBL genec	ompK35 sequence	Estimated plasmid sizes (kb)
				Ertapenem	Meropenem	Imipenem	Polymyxin B
4186	TASMC	9/09	Blood	12	1.5	0.19	<2	R	blaCTX-M-2	121 insd	10, 40, 160, 170
863	LMC	7/10	Urine	6	1	0.19	<2	R	blaCTX-M-2	121 ins	10, 40, 75, 170
872	LMC	8/10	Urine	4	0.75	0.19	<2	R	blaCTX-M-2	121 ins	10, 40, 150, 170
849	LMC	12/1	Urine	6	0.75	0.19	16	R	blaCTX-M-2	121 ins	10, 40, 160
1036	LMC	2/11	Urine	3	0.5	0.19	<2	S	blaCTX-M-25	121 ins	10, 40, 150
1041	LMC	2/11	Urine	4	0.75	0.19	<2	R	blaCTX-M-25	121 ins	10, 40, 150, 170
1037	LMC	3/11	Rectal	16	2	0.75	<2	S	blaCTX-M-25	121 ins	10, 40, 150
1039	LMC	3/11	Urine	2	0.38	0.25	<2	R	blaCTX-M-2	121 ins	10, 40, 60, 150
1050	LMC	3/11	Wound	12	1.5	0.38	<2	R	blaCTX-M-2	121 ins	10, 150, 160

^aMICs (mg/liter) were determined by the Etest method.

^bR, resistant; S, susceptible.

^cOther non-ESBL β-lactamase genes found in all isolates included bla_SHV-11 and bla_TEM-1.

^d121 ins, 1-bp insertion at position 121.

The frequency of clone Q was also studied in 35 KPC-producing isolates (TASMC, 18; LMC, 17) and 62 carbapenem-susceptible isolates (TASMC, 50; LMC, 12). This clone comprised 100% of the KPC-producing isolates and 0% of the carbapenem-susceptible isolates.

Resistance mechanisms and plasmid analysis of the nine ERCNKP clone Q isolates.

All nine ERCNKP clone Q isolates were resistant to ertapenem, and three were intermediate to meropenem (Table 1). Isolates were resistant to all classes of antimicrobials except for tigecycline and colistin; only two isolates were susceptible to gentamicin. No synergy was detected between ertapenem and either EDTA, dipicolinic acid, or boronic acid. Hydrolysis assay using protein extracts showed a lack of carbapenemase activity. All isolates were negative by PCR to all carbapenemase and ampC genes tested. All isolates carried bla_SHV-11 and bla_TEM-1; six carried bla_CTX-M-2, and two carried bla_CTX-M-25. OMP sequence analysis of ompK35, ompK36, and ompK37 revealed a 1-bp insertion (G) at position 121 of the ompK35 gene in all clone Q isolates, including the KPC-producing strains, leading to frameshift and early termination of the protein at amino acid 89, as previously described (10, 21, 27). In addition, the ompK36 gene contained the Asp135 and Gly136 insertions in the L3 loop region as described previously (20, 21).

The plasmid profile of the nine ERCNKP clone Q isolates was compared to that of three KPC-producing clone Q isolates (Fig. 2). S1 analysis showed plasmids of variable size; all isolates harbored an ∼10-kb and an ∼40-kb plasmid. None of the ERCNKP isolates harbored the 113-kb bla_KPC-carrying plasmid, pKpQIL (29), present in the KPC-producing strains (Fig. 2). PCRs for the Tn4401 transposon, repA, and the truncated bla_OXA-9 gene were negative in all ERCNKP plasmids, confirming the absence of pKpQIL.

Fig 2.

S1 analysis of plasmids in type Q strains. Group A, KPC-producing isolates; group B, carbapenemase-negative isolates. Isolate 1041 was analyzed separately. *, pKpQIL, the bla_KPC-carrying plasmid.

DISCUSSION

In the present study, we examined the incidence of the ERCNKP clone Q in 2 medical centers in Israel, for a period of up to 2.5 years. For comparison, we examined the prevalence of the same clone in two different populations: the KPC-producing and the carbapenem-susceptible K. pneumoniae isolates. In LMC, clone Q comprised 19% of the ERCNKP population, with half isolated during a small-scale outbreak in a single ward. As these isolates differed in their ESBL genes and plasmid profiles, it is likely that this outbreak originated from more than one source. These isolates were not associated with invasive infection and were not detected in the last 2 months of the study. In the retrospective survey conducted at TASMC, a single clone Q ERCNKP isolate was identified in 2009, but none were identified in 2010. Clone Q was completely absent from the carbapenem-susceptible population. This is in sharp contrast with the overwhelming predominance of this clone among the KPC-producing population. Also, only a few cases of infection caused by KPC-negative ST-258 strains have been reported (10), despite the global predominance of its isogenic KPC-producing strain (3, 4, 13). Together, these findings indicate a relatively low epidemic potential of the KPC-negative ST-258 clone. Phylogenetically, the ST-258 clone is not related to known virulent K. pneumoniae clones associated with liver abscesses or pneumonia, such as the ST-23 and the ST-82 clonal complexes (CC) (7, 46). It is, however, closely related to a single-locus variant strain, ST-11, that has been reported as a common hospital-associated clone from several European countries (15, 24, 39). Further genomic studies are required in order to explore whether ST-258 has indeed evolved from ST-11.

The incidence of ERCNKP infection at TASMC increased in 2009 compared to 2008 and remained stable in 2010; the incidence of the KPC-producing strains was constant throughout the study. These findings are consistent with the containment of new incidence of carbapenem-resistant Enterobacteriaceae (CRE) in Israel following the successful national intervention that began in 2007 (45). At LMC, the incidence of CRE of both types was higher than at TASMC, as was the proportion of carbapenem-resistant K. pneumoniae isolates among the overall population of K. pneumoniae isolates. Interestingly, the incidence of ERCNKP compared to carbapenemase-producing isolates was higher at LMC than at TASMC. The reasons for the differences between the institutions are not clear, but they may be due to several factors. Considering that patients are regularly transferred between centers and that consumption of carbapenem is lower at LMC than at TASMC (see “Settings and infection control practices”), it appears that differences in infection control practices, specifically, the absence of CRE admission screening at LMC during the study period, might have played a role in determining these differences in the incidence of CRE. Whereas KPC-producing strains are unlikely to evolve de novo but are rather transmitted from other patients or staff members, the evolution and transmission of the ERCNKP strains are less understood. In addition to the small-scale outbreak described in our study, only a single ERCNKP outbreak case, due to the ST-37 clone, has been described (20). This suggests that in addition to transmission from other patients, de novo evolution of ertapenem resistance (25) may play a role in the spread of these organisms.

The source and the evolution of the ERCNKP clone Q in Israel are not entirely clear, but several findings revealed in this study may help in their elucidation. (i) Clone Q ERCNKP was not detected in a previous study of ERCNKP strains in Israel (data not shown) (30). (ii) The ERCNKP strains were all extremely drug resistant, susceptible only to colistin, tigecycline, and gentamicin (two cases), similar to the KPC-producing isogenic strain (32). (iii) Both the KPC-producing and their isogenic ERCNKP strains exhibited a 1-bp insertion in ompK35, identical to previous reports from Italy and the United States (21, 27) (GenBank accession number FJ577672), indicating its presence in this strain, prior to its worldwide dissemination. (iv) The two types of ESBL genes (bla_CTX-M-2 and bla_CTX-M-25) found are common in, though not exclusive to, Israel (11, 34), in contrast with bla_SHV-12, which was identified in a similar strain in New York (10). (v) The common bla_KPC-harboring plasmid in Israel, pKpQIL (29), as well as the Tn4401 transposon, was absent in all ERCNKP isolates. Based on these facts, we postulate that the ERCNKP clone Q evolved from the KPC-producing isogenic strain following its dissemination into Israel, by expulsion of the pKpQIL plasmid. The variability in the ESBL genes and plasmids suggests that such an event might have happened more than once. The lack of KPC-producing K. pneumoniae infections in these patients suggests that the ERCNKP clone Q strains were probably acquired from other patients.

There are several parallel examples of antibiotic-resistant bacterial clones that, via global dissemination, have contributed immensely to the spread of antibiotic resistance, similar to the spread of carbapenem resistance by the KPC-producing K. pneumoniae ST-258 clone. Two notable examples are the spread of methicillin resistance in Staphylococcus aureus by the USA-300 MRSA clone (16) and the spread of vancomycin resistance by Enterococcus faecium CC-17 (19). In both these cases, antibiotic-susceptible isogenic strains were identified (19, 33, 38), but unlike the ERCNKP ST-258 clone, these strains were associated with significant morbidity as well. This raises the question of whether there are additional genetic differences between the two isogenic ST-258 strains, besides the presence or absence of the bla_KPC gene, that contribute to the epidemiological differences between them. Further studies, including full genomic analyses, are required in order to explore this question.