Abstract
The aim of this study was to determine whether there is an association between serum resistance, O serotypes, and the production of extended-spectrum β-lactamases (ESBLs) in Klebsiella pneumoniae. Ninety ESBL-producing and 178 non-ESBL-producing K. pneumoniae isolates gathered in five European countries were O serotyped and tested for sensitivity to the serum's bactericidal effect. The frequency of serum-resistant isolates was higher among ESBL-producing strains (30%; 27/90 isolates) than among non-ESBL-producing strains (17.9%; 32/178 isolates) (P = 0.037; odds ratio [OR] = 1.96; 95% confidence interval [95% CI] = 1.08 to 3.53). Although O1 was the most common O serotype in both Klebsiella groups, its frequency among ESBL-producing strains was significantly higher (59%; 53/90 isolates) than among non-ESBL producers (36%; 64/178 isolates) (P = 0.0006; OR = 2.5; 95% CI = 1.52 to 4.29). Furthermore, the prevalence of the O1 serotype was higher among serum-resistant strains of both ESBL-producing (74%; 20/27isolates) and non-ESBL producers (75%; 24/32 isolates) than among serum-sensitive ESBL producers (52.4%; 33/63 isolates) and non-ESBL producers (27.4%; 40/146 isolates). Serum resistance among ESBL-producing strains (36%; 17/47 isolates) versus non-ESBL-producing strains (16%; 27/166 isolates) was also significantly higher after the exclusion of clonal strains (P = 0.0056; OR = 2.9; 95% CI = 1.41 to 6.01). Sixteen ESBL types were detected, among which the frequency of serum resistance was significantly lower among the SHV-producing strains (9/48 isolates) than among the TEM producers (16/35 isolates) (P = 0.016; OR = 3.65; CI = 1.3 to 9.7). Curing ESBL-coding plasmids did not influence the serum resistance of the bacteria; all six plasmid-cured derivatives maintained serum resistance. The present findings suggest that ESBL-producing strains have a greater pathogenic potential than non-ESBL-producing strains, but the linkage between O serotypes, serum resistance, and ESBL production remains unclear at this stage.
Klebsiella pneumoniae is an opportunistic pathogen that can cause severe infections in hospitalized, immunocompromised hosts with severe underlying diseases (23). Depending on the study and type of infection, the prevalence of K. pneumoniae ranges from 3 to 17% of all nosocomial bacterial infections (10).
Extensive use of antibiotics has contributed greatly to the emergence of multidrug-resistant strains that cause hospital infections. In Klebsiella species, various types of plasmid-encoded, extended-spectrum β-lactamases (ESBLs), especially TEM and SHV enzymes, have been described worldwide (13, 14). Recent surveys (17, 34) have shown that approximately 20 years after the first outbreaks in Germany and France, the prevalence of ESBL-producing K. pneumoniae isolates from intensive-care units (ICUs) varies from 9% in southern Germany to 40% in France, 49% in Portugal, and 59% in Turkey. Of particular interest are reports on the association between ESBL production and the expression of pathogenicity factors in Klebsiella. Several studies have shown that ESBL-producing Klebsiella strains have an increased ability to adhere to human epithelial cells, probably due to plasmid-coded production of fimbrial or nonfimbrial adhesins (6, 8, 9, 18). It is anticipated that Klebsiella strains will become more virulent and resistant to antibiotics in the future, an eventuality that could contribute to the spread of certain Klebsiella clones by virtue of their greater resistance to antibiotics and better adherence to host tissues.
After the onset of inflammation, invading Klebsiella strains meet the cellular and humoral bactericidal components of the innate immune system. The host's first line of defense against invading microorganisms includes the bactericidal effect of serum, which is mediated primarily by complement proteins. Lipopolysaccharides (LPS) have been implicated as a major factor in the ability of bacteria to resist serum bactericidal activity by the host (4, 19, 24, 30). Nine different LPS serotypes (O antigens) in K. pneumoniae have been described, the O1 serotype being the most common O antigen found in clinical isolates (11). In the present study, we examined whether an association exists between ESBL production and the ability of K. pneumoniae to resist the bactericidal potency of human serum. We also evaluated the distribution of the O serotypes among the different strains.
MATERIALS AND METHODS
Bacterial strains and typing procedures.
Ninety ESBL-producing and 178 non-ESBL-producing K. pneumoniae strains were consecutively isolated from clinical specimens from hospitalized ICU and non-ICU patients suffering from nosocomial Klebsiella infections. The strains were collected in five European countries (Denmark, 50 specimens; England, 32 specimens; France, 76 specimens; Germany, 48 specimens; and Spain, 62 specimens) between 1994 and 1997 and fulfilled Centers for Disease Control and Prevention criteria for nosocomial pneumonia, urinary tract infection, or primary bloodstream infection.
The origin of each strain was noted, and only one strain per patient was included in this collection. The strains were identified biochemically with the API 20E system (API bioMérieux, Nürtingen, Germany) or by conventional methods as described previously (29).
ESBLs were detected according to NCCLS recommendations (20). All strains were phenotypically screened for ESBL production by the broth microdilution method. ESBL production in isolates for which cefpodoxime, cefotaxime, ceftriaxone and ceftazidime MICs were ≥2 mg/liter (by initial screen test) was confirmed by testing for synergy between ceftazidime or cefotaxime and clavulanic acid by the broth microdilution method, by following NCCLS criteria (20) and by the double-disk synergy test with disks (BD Biosciences) containing ceftazidime or cefotaxime (concentration of each, 30 μg) and either ceftazidime and clavulanic acid or cefotaxime and clavulanic acid (concentration of ceftazidime or cefotaxime, 30 μg; concentration of clavulanic acid, 10 μg), as described by Jarlier et al. (15). Either a ≥3-fold concentration decrease in the MIC of ceftazidime or cefotaxime tested in combination with clavulanic acid, compared to the MIC recorded when ceftazidime or cefotaxime was tested alone, or a difference in zone diameter of ≥5 mm indicated phenotypic confirmation of ESBL production.
To determine the ESBL types, DNA from all strains was prepared for PCR as previously described (25). The blaSHV gene was amplified with primers designed to cover the more-conserved terminal point of the gene. The primers shv-F (CGC CTG TGT ATT ATC TCC CTG TTA GCC) and shv-B (TTG CCA GTG CTC GAT CAG CG) were derived from GenBank accession no. AF178850, which resulted in a PCR product of 842 bp. Likewise, for the blaTEM gene, primers were made to amplify most of the single-point mutation variations of blaTEM. Primers used were tem-F (GTA TCC GCT CAT GAG ACA ATA ACC CTG) and tem-B (CCA ATG CTT AAT CAG TGA GGC ACC), obtained from GenBank accession no. AF309824. This strain yielded an amplicon of 918 bp, which was separated on 1.5% agarose gels.
PCR conditions were as follows: initial denaturation at 95°C for 120 s, followed by denaturation at 95°C for 45 s, annealing at 62°C for 45 s, and elongation at 72°C for 60 s; except for the initial denaturation, all steps were repeated for a total of 30 cycles. Control strains for positive amplifications, K. pneumoniae KP15 (blaSHV2) and K. pneumoniae KP6T (blaTEM-4), were kindly obtained from Teresa Coque, Madrid, Spain (5). The PCR products were purified with a spin column (QIAGEN PCR purification kit), and sequencing on both strands was carried out as described previously (25) with 10 pmol of one of the primers in combination with the Big Dye terminator cycle sequencing kit (PE Biosystems, Foster City, Calif.) according to the manufacturer's instructions. Sequencing reaction mixtures were analyzed with an ABI 377 acrylamid gel DNA genetic analyzer (Applied Biosystems, Foster City, Calif.). Raw sequences were compared to the reversed counterpart of the DNA, and unresolved nucleotides in one sequencing direction were ignored if the complementary position was unambiguously determined. If the position was ambiguously determined in both sequencing directions, the sequencing reactions were repeated. The sequences were processed with a BioNumerics GeneBuilder (Applied Maths BVBA, Sint-Martens-Latem, Belgium), and the consensus sequence was compared to the Entrez Nucleotides database by the use of the National Center for Biotechnology Information BLAST program (http://www.ncbi.nlm.nih.gov/BLAST/).
Expression of the different O serotypes by the strains was tested by enzyme-linked immunosorbent assay (ELISA) with O-type-specific rabbit antibodies as described by Hansen et al. (11). LPS from all strains whose O serotypes were not detectable by ELISA were visualized by sodium dodecyl sulfate-polyacrylamide gel electrophoresis as described previously (11). Strains with smooth LPS (S-LPS; non-O-typeable but with the O antigen expressed [NTO+]) were differentiated from rough strains (R-LPS; non-O-typeable but with the O antigen unexpressed [NTO−]) by the presence of high-molecular-weight LPS. K-serotype isolates were determined by the capsular swelling method with K-specific antisera (22) or by countercurrent immunoelectrophoresis (29).
All strains whose O serotypes were not detectable by ELISA were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and the gels were stained with silver. According to the resulting LPS ladder, strains were divided into groups of nontypeable strains with S-LPS (NTO+) or nontypeable strains with R-LPS (NTO−).
The clonality of the individual strains was analyzed with regard to their ribotype, K serotype, and O serotype by means of automated ribotyping (RiboPrinting) according to the manufacturer's instructions with the RiboPrinter and the restriction enzyme EcoRI as previously described by Bruce (3). Strains were regarded as clonal if they expressed an identical ribotype, K serotype, and O serotype. In order to ensure that prevalent outbreak clones could not skew the results and to ensure clonal diversity, only one isolate of each clone (strains with identical K and O serotypes and ribotypes) was included in sequential analysis. NTO+ antigens and untypeable K antigens were considered different from strains expressing typeable O and K antigens (12). Strains which expressed no O antigen (NTO−) were assigned to the clones of those strains with otherwise identical typing characteristics.
Serum bactericidal assay.
The ability of the isolates to resist killing by serum was tested as described previously (22). Bacteria were grown in nutrient broth, harvested during the early logarithmic phase of growth, and adjusted to a concentration of 2 × 106 bacteria/ml of physiological saline. Twenty-five microliters of the bacterial suspension was mixed with 75 μl of pooled normal human serum in microtiter plates. Viable counts (VCs) of bacteria were determined for a period of 3 h. Responses were graded as follows. For grade 1, the VC after 1 and 2 h was <10% of the inoculum; after 3 h, it was <0.1%. For grade 2, the VC after 1 h was 10 to 100%; after 3 h, it was <10%. For grade 3, the VC after 1 h was >100% of the inoculum; after 2 and 3 h, it was <100%. For grade 4, the VC after 1 and 2 h was >100%; after 3 h, the VC was <100%. For grade 5, the VC after 1, 2, and 3 h was >100%; but the VC fell sometime during the 3-h period. For grade 6, the VC after 1, 2, and 3 h was >100% of the inoculum and rose throughout the 3-h period. Each strain was tested at least three times. A strain was considered serum resistant or serum sensitive if the grading was the same in all experiments.
Preparation of plasmid DNA and plasmid curing.
To analyze the relation between harboring ESBL-coding plasmids and serum resistance in Klebsiella, six strains producing ESBL type TEM3, TEM 8, TEM16, SHV4, SHV5, or SHV12, all of which were shown to be resistant to serum (grade 5 or 6), were subjected to plasmid curing as described previously (9). The preparation of the plasmids was performed with a QIAGEN preparation kit for high-molecular-weight plasmids, according to the manufacturer's instructions. The loss of the plasmids was visualized by agarose gel electrophoresis as described previously (9). All plasmid-cured derivatives were screened for ESBL production and for their ability to resist killing by serum, as described above.
Statistical analysis.
The significance of differences between groups of isolates was evaluated by Yates' correction of the chi-square test. For tables showing results where less than five positive isolates were found within a cell, Fisher's exact test was used. The odds ratio (OR) and 95% confidence intervals (CI) were determined, and values were taken as gauges of association.
RESULTS
Clonality of the strains and ESBL types.
A total of 197 clones were detected, from which 18 clones comprised 2 to 22 strains (Table 1). After exclusion of clonal strains, a total of 166 non-ESBL-producing and 47 ESBL-producing strains were defined and subjected to the same analysis as the entire bacterial collection (see below).
TABLE 1.
Characterization of clonal groups among ESBL-producing and non-ESBL-producing K. pneumoniae strains, defined by identical ribotype O and K serotypesa
| Clone or ribotype | n | Country(ies) of isolate origin | No. of ESBL-producing strains | Serotype(s)
|
SR | |
|---|---|---|---|---|---|---|
| O | K | |||||
| 1 | 23 | D, ES, DK, F | 17/22 | 1 | 2 | 11/22 |
| 3 | 2 | F | 0/2 | 2a | 18 | 0/2 |
| 4 | 4 | D, ES, DK, UK | 0/4 | 2a | 18 | 0/3 |
| 6 | 2 | DK | 0/2 | 1 | 10, 61 | 1/2 |
| 8 | 2 | DK, F | 2/2 | 2a | 20 | 0/2 |
| 9 | 2 | ES | 2/2 | 3 | 14 | 1/2 |
| 10 | 4 | ES | 4/4 | NTO+ | 52 | 1/4 |
| 11 | 2 | DK, D | 0/2 | 1 | 1 | 0/2 |
| 14 | 2 | D | 0/2 | 2a | 24 | 0/2 |
| 16 | 2 | F | 0/2 | 1 | 35 | 1/2 |
| 17 | 2 | F | 0/2 | 2a | 27 | 0/2 |
| 18 | 21 | F, ES | 21/21 | 1 | 25 | 2/21 |
| 21 | 2 | D | 0/2 | 2a | 17 | 1/2 |
| 23 | 2 | F | 0/2 | 4 | 15 | 0/2 |
| 24 | 2 | F | 2/2 | 1 | 68 | 2/2 |
| 26 | 2 | D | 0/2 | 1 | 43 | 0/2 |
| 26 | 10 | F | 7/10 | 1 | 24 | 3/10 |
| 28 | 2 | ES | 2/2 | 3 | 46 | 0/2 |
D, Germany; DK, Denmark; ES, Spain; F, France; n, number of isolates; SR, serum resistance. Boldface type indicates that the frequency of the clone or ribotype was significantly higher than the frequencies of those in roman type.
Sixteen ESBL types were identified, among which SHV5 was the most common ESBL type (23.3%; 21/90 isolates tested), followed by TEM3 (15.6%; 14/90 isolates), SHV4 and SHV12 (each 10% and 9/90 isolates), and TEM24 (7.8%; 7/90 isolates). All other ESBL types were detected at a frequency of ≤4% (Table 2). For seven strains, ESBLs were not determined, as the sequences of the PCR products of the blaTEM or blaSHV genes were not homologous to published sequences.
TABLE 2.
Frequency of serum resistance in relation to ESBL type
| ESBL type | n | Country(ies) of isolate origin | Serotype(s)
|
% SR (no. of isolates positive/no. tested) | |
|---|---|---|---|---|---|
| O | K | ||||
| SHV2 | 2 | F | NTO+, 3 | 1, 23 | 0 |
| SHV2A | 1 | ES | NTO+ | 33 | 0 |
| SHV4 | 9 | F | NTO−, 1 | 2, 18, 25, 42 | 33.3 (3/9) |
| SHV5 | 21 | DK, ES, F, UK | NTO+, NTO− 1, 2a | NT, 8, 25, 44, 45, 47, 68 | 9.5 (2/21) |
| SHV12 | 9 | UK, ES | NTO+, NTO−, 1, 3, 5, 2a | 8, 21, 41, 46, 47, 51, 52, 62 | 33.3 (3/9) |
| SHV36 | 2 | ES | 3 | 14 | 50 (1/2) |
| SHV40 | 4 | ES | NTO+, 4 | 42 | 0 |
| TEM3 | 14 | F | NTO−, 1, 4 | NT, 2, 15, 24, 68 | 50 (7/14) |
| TEM5 | 3 | F, ES | 1 | NT, 2 | 100 (3/3) |
| TEM8 | 3 | F | NTO+, 1 | 2, 38 | 33.3 (1/3) |
| TEM12 | 1 | F | 1 | 24 | 0 |
| TEM13 | 4 | F | 1 | 24, 68 | 25 (1/4) |
| TEM15 | 1 | F | 4 | 15, | 0 |
| TEM16 | 1 | F | NTO− | 2, | 100 (1/1) |
| TEM18 | 1 | F | 1 | 24 | 0 |
| TEM24 | 7 | F | 1, 2a | NT, 2, 25 | 42.9 (3/7) |
| NT | 7 | DK, F, ES | NTO+, 1, 4, 2a | 25, 33, 41, 51, 52, 62 | 28.6 (2/7) |
D, Germany; DK, Denmark; ES, Spain; F, France; UK, United Kingdom; SR, serum resistance; NT, not typeable; n, number of isolates with ESBL type.
Serum resistance in relation to ESBL production.
Bacterial survival in normal human serum was recorded over a 3-h period and arranged into six grades. The strains were classified as highly sensitive (grades 1 and 2), intermediately sensitive (grades 3 and 4), or resistant (grades 5 and 6). In the vast majority of the strains (98%; 263/268 strains), the results were reproducible in all three experiments. A few strains (2%; 5/268 strains), which were revealed to have different serum resistance values in two of the three consecutive experiments, were freshly thawed from the strain collection and retested for their serum resistance property at least three times. Old noncongruent results were not considered.
The percentage of serum-resistant isolates (grades 5 to 6) was significantly higher among the ESBL-producing strains than among the non-ESBL-producing strains (P = 0.037; OR = 1.96; 95% CI = 1.08 to 3.53). All told, 30% (27/90 strains) of the ESBL-producing strains but only 17.9% (32/178 strains) of the non-ESBL-producing strains showed serum resistance (Table 3).
TABLE 3.
Distribution of serum-resistant strains among ESBL-producing and non-ESBL-producing Klebsiella isolates with and without clonal strains, collected in five European countries
| Type | No. of strains | % of positive isolates (total isolates tested)
|
|
|---|---|---|---|
| Serum sensitive (grades 1-4) | Serum resistant (grades 5-6)a | ||
| Non-ESBL | |||
| Total | 178 | 82.1 (146) | 17.9 (32) |
| After clonal revision | 166 | 83.7 (139) | 16.3 (27) |
| ESBL | |||
| Total | 90 | 70 (63) | 30 (27)* |
| After clonal revision | 47 | 63.8 (30) | 36.2 (17)** |
*, significantly higher incidence of serum-resistant strains among ESBL-producing K. pneumoniae isolates (P = 0.037; OR = 1.96; 95% CI = 1.08 to 3.53). **, significantly higher incidence of serum-resistant strains among ESBL-producing K. pneumoniae isolates (P = 0.006; OR = 2.9; 95% CI = 1.41 to 6.01).
Although no association between certain ESBL types and increased serum resistance was shown, the frequency of serum-resistant strains among TEM producers (45.7%; 16/35 strains) was significantly higher than among SHV producers (18.75%; 9/48 strains) (P = 0.016; OR = 3.65; 95% CI = 1.3 to 9.7).
Distribution of the O serotypes related to ESBL production and serum resistance.
The O1 serotype had the highest incidence among both ESBL-producing (58.8%; 53/90 strains) and non-ESBL-producing (35.9%; 64/178 strains) Klebsiella isolates (P = 0.0006; OR = 2.5; 95% CI = 1.52 to 4.29) (Table 4). The O serotypes O2a (19.1%; 34/178 strains), O3 (18%; 32/178 strains), and O5 (10.7%; 19/178 strains) had significantly higher frequencies among the non-ESBL-producing isolates than among ESBL producers (for O2a, 6.7%; 6/90 strains; for O3, 6.7%; 6/90 strains; and for O5, 1.1%; 1/90 strains) (P < 0.02) (Table 4). Details of the frequencies of all other serotypes are listed in Table 4.
TABLE 4.
Incidence of O serotypes among ESBL-producing and non-ESBL-producing strains before and after exclusion of clonal strains
| O serotype | % of tested isolates with serotypea
|
|||
|---|---|---|---|---|
| Before exclusion
|
After exclusion
|
|||
| Non-ESBL | ESBL | Non-ESBL | ESBL | |
| 1 | 35.96 (64/178)* | 58.9 (53/90)*,♦ | 34.3 (57/166)* | 34 (16/47)* |
| 2a | 18.5 (33/178)♦ | 6.7 (6/90) | 18.7 (31/166) | 10.6 (5/47) |
| 3 | 17.9 (32/178)♦ | 6.7 (6/90) | 19.3 (32/166) | 8.5 (4/47) |
| 4 | 2.8 (5/178) | 4.4 (4/90) | 3.0 (5/166) | 8.5 (4/47) |
| 5 | 10.7 (19/178)♦ | 1.1 (1/90) | 11.4 (19/166) | 2.1 (1/47) |
| 12 | 0.56 (1/178) | 0 | 0.6 (1/166) | 0 |
| NTO− | 6.74 (12/178) | 11.1 (10/90) | 6.6 (11/166) | 21.3 (10/47)♦ |
| NTO+ | 6.18 (11/178) | 11.1 (10/90) | 6.0 (10/166) | 14.9 (7/47) |
Ratios in parentheses indicate number of isolates testing positive for a given O serotype per total number of isolates tested. *, significantly higher incidence of the O1 serotype (P < 0.02) among ESBL-producing and non-ESBL-producing strains than that of the other O serotypes. ⧫, significantly higher incidence (P < 0.02) of the indicated serotype in the ESBL-producing or non-ESBL-producing strains.
The frequency of the O1 serotype was higher than other serotypes among serum-resistant strains of both ESBL-producing (74%; 20/27 strains) and non-ESBL-producing (75%; 24/32 strains) Klebsiella groups than among serum-sensitive ESBL-producing (52.4%; 33/63 strains) and non-ESBL-producing (27.4%; 40/146 strains) strains (Table 5). These differences were highly significant for non-ESBL-producing strains (OR = 7.95; 95% CI = 3.3 to 19.15) (P < 0.0001), but not for ESBL producers (OR = 2.6; 95% CI = 0.96 to 7.01) (P = 0.09). Details of the frequencies of the other serotypes in both groups and the significance of these frequencies are listed in Table 5.
TABLE 5.
Incidence of O serotypes among serum-resistant and serum-sensitive isolates in ESBL-producing and non-ESBL-producing strains before and after exclusion of clonal strains
| O serotype | % of tested isolates with serotypea
|
|||
|---|---|---|---|---|
| Non-ESBL
|
ESBL
|
|||
| SS | SR | SS | SR | |
| Entire collection | ||||
| 1 | 27.3 (40/146) | 75 (24/32)* | 5.2 (33/63)*♦ | 74 (20/27)* |
| 2a | 22.6 (33/146)♦ | 3.1 (1/32) | 9.6 (6/63) | 0 |
| 3 | 20.5 (30/146)♦ | 6.25 (2/32) | 6.4 (4/63) | 7.4 (2/27) |
| 4 | 3.4 (5/146) | 0 | 6.4 (4/63) | 0 |
| 5 | 10.3 (15/146) | 12.5 (4/32) | 1.6 (1/63) | 0 |
| 12 | 0.7 (1/146) | 0 | 0 | 0 |
| NTO− | 8.2 (12/146) | 0 | 11.1 (7/63) | 11.1 (3/27) |
| NTO+ | 6.8 (10/146) | 3.1 (1/32) | 12.6 (8/63) | 7.4 (2/27) |
| After clonal revision | ||||
| 1 | 26.6 (37/139) | 74 (20/27)* | 16.7 (5/30)* | 64.7 (11/17)* |
| 2a | 22.3 (31/139) | 0 | 20 (6/30) | 0 |
| 3 | 21.6 (30/139)♦ | 7.4 (2/27) | 6.7 (2/30) | 11.8 (2/17) |
| 4 | 3.6 (5/139) | 0 | 13.3 (4/30) | 0 |
| 5 | 10.8 (15/139) | 14.8 (4/27) | 3.3 (1/30) | 0 |
| 12 | 0.7 (1/139) | 0 | 0 | 0 |
| NTO− | 7.9 (11/139) | 0 | 23.3 (7/30)♦ | 17.6 (3/17) |
| NTO+ | 6.5 (9/139) | 3.7 (1/27) | 20 (6/30)♦ | 5.8 (1/17) |
Ratios in parentheses are the numbers of isolates testing positive for a given O serotype per total numbers of isolates tested. *, significantly higher incidence of the O1 serotype (P < 0.02) among ESBL-producing and non-ESBL-producing strains than that of the other O serotypes. ♦, significantly higher incidence (P < 0.02) of the indicated serotype in the ESBL-producing or non-ESBL-producing strains. SS, strains sensitive to serum; SR, strains resistant to serum.
Serum resistance in relation to clonality.
After exclusion of clonal strains, a total of 166 non-ESBL-producing and 47 ESBL-producing strains were subjected to the same analysis as the entire bacterial collection.
As with the entire bacterial collection, the clonally revised collection was found to have a significantly higher frequency of serum-resistant strains among ESBL-producing strains (36%; 17/47 strains) than among non-ESBL-producing strains (16%; 27/166 strains) (P = 0.0056) (OR = 2.9; 95% CI = 1.41 to 6.01) (Table 3). Among serum-resistant strains, the O1 frequency in both ESBL-producing strains (65%; 11/17 strains) and non-ESBL-producing strains (74%; 20/27 strains) was significantly higher among serum-resistant strains than among ESBL-producing (16.7%; 5/30 strains) and non-ESBL-producing (25.4%; 35/138 strains) serum-sensitive strains (P < 0.0025) (Table 4 and 5). The data on the frequencies of all other serotypes are detailed in Tables 4 and 5.
Relationship among ESBL-coding plasmids and serum resistance.
To better assess the contribution of the ESBL-coding plasmid in conferring serum resistance to bacteria, six serum-resistant strains that harbor the ESBL-coding plasmid were subjected to a plasmid-curing procedure. The plasmid-cured derivatives were retested for ESBL production and for their ability to resist the bactericidal activity of serum, as described above. The loss of the ESBL-coding plasmids in the strains was demonstrated by agarose gel electrophoresis (data not shown). To confirm the consecutive lack of ESBL production, MICs of cefpodoxime, ceftazidime, and cefotaxime for all plasmid-cured derivatives were determined, by following NCCLS guidelines. In all strains, the MICs of the tested cephalosporins were <2 mg/liter. The absence of the blaTEM and blaSHV ESBL genes was confirmed by PCR. All six plasmid-cured derivatives showed the same serum resistance grade (5 or 6), indicating that plasmid curing did not influence the ability of the strains to resist the bactericidal effect of serum (data not shown).
DISCUSSION
The pathogenicity of microorganisms that invade the bloodstream is partly a function of their ability to evade the bactericidal effect of serum, which is mediated by the complement cascade. Commensal microorganisms are generally vulnerable to the bactericidal effect of serum, while nosocomial bacteria tend to be much more serum resistant (21, 26, 27).
In the present study, we investigated whether there is a relationship between ESBL production and the serum resistance of K. pneumoniae isolates. ESBL-producing Klebsiella strains proved to be significantly more resistant to the bactericidal effect of human serum than their non-ESBL-producing counterparts. Moreover, our results demonstrate that the O1 serotype, the most frequent O serotype among clinical K. pneumoniae isolates (11, 32), predominates among both ESBL-producing and non-ESBL-producing Klebsiella strains. Its frequency among ESBL-producing strains, however, was significantly higher (59%) than among non-ESBL producers (36%).
Because both serum resistance and expression of the O1 serotype are thought to be factors in the pathogenicity of Klebsiella (22), it can be hypothesized that the higher frequency of the O1 serotype among ESBL-producing strains and their superior resistance to the bactericidal activity of serum are indicators of the greater pathogenic potential of ESBL-producing strains than that of non-ESBL-producing strains.
The molecular basis of serum resistance is not entirely clear. O side chains of the LPS of pathogenic bacteria are thought to play a major role in conferring serum resistance. Gram-negative bacteria which express intact O side chains (S-LPS) are generally more serum resistant than those which do not express O side chains (R-LPS) (28). For K. pneumoniae, a number of investigations have shown that serum sensitivity is determined by its LPS molecule (1, 2, 4, 16, 19, 33). Cell-free LPS from serum-resistant K. pneumoniae O1 strains inhibit serum's bactericidal effect on serum-sensitive strains but does not render serum-resistant strains susceptible to serum (4).
Interestingly, our analysis of the distribution of O serotypes in serum-resistant isolates of both ESBL-producing and non-ESBL-producing strains showed they had equally high incidences of the O1 serotype. If it is assumed that S-LPS confer serum resistance on bacteria, further studies are needed to clarify why all O serotypes do not protect the bacteria against the bacteriolytic activity of complement. Further studies are also needed to show how the concept of the O1 serotype as a factor that mediates serum resistance harmonizes with the fact that the O1 serotype is expressed by 52% of ESBL-producing and by 27% of non-ESBL-producing serum-sensitive isolates.
One aspect of this study was to examine whether the high incidence of particular serotypes and serum-resistant isolates among ESBL-producing strains can be assigned to clonal groups. Among all isolates tested, we were able to define 18 clones, most from among ESBL-producing strains.
The higher incidence of the O1 serotype in ESBL-producing strains of the entire bacterial collection was apparently due to the three major clones 1, 18, and 26. A sequential analysis revealed that serum-resistant strains were significantly more frequent among ESBL-producing isolates than among non-ESBL-producing isolates, even after the exclusion of clonal strains. Among ESBL-producing isolates, the TEM producers showed a significantly higher frequency of serum resistance than the SHV producers, although no correlation between the expression of certain ESBL types and serum resistance could be detected.
The linkage between the O serotypes, especially O1, serum resistance, and the presence of certain ESBL-encoding plasmids is not clear at this stage.
Because the incidence of the O1 serotype among ESBL producers after the exclusion of clonal strains was not higher than that among non-ESBL producers, it is conceivable that the higher frequency of serum-resistant strains among ESBL producers was due to factors other than the O1 antigen. Moreover, it appears to be unlikely that the ESBL plasmids also code for specific factors that confer serum resistance, because curing the R plasmid did not impair the ability of the bacteria to resist killing by serum and because no association could be shown in the tested strains between serum resistance and the harboring of particular ESBL types.
It could be speculated that virulent “successful” Klebsiella strains possessing certain serotype genes have higher competence for acquiring R plasmids. Such an association between particular genetic background and plasmid compatibility has been shown. For instance, although β-lactam antibiotics have been used extensively for the last 4 decades to treat infections caused by Streptococcus pyogenes, the bacteria fail to acquire the β-lactamase-encoding plasmid, and all S. pyogenes isolates identified to date are de facto sensitive to β-lactam antibiotics (7, 31). Alternatively, both biological selection by the innate immune system (e.g., complement) and the increased use of β-lactam antibiotics are responsible for the emergence of strains harboring ESBL plasmids and exhibiting serum resistance.
The fact that different ESBL types are harbored in strains belonging to one clone supports the notion that the general plasmid compatibility phenomenon also applies for Klebsiella. However, in order to clarify whether or not certain Klebsiella serotypes are better fit for acquiring R plasmids, additional conjugation experiments which consider all Klebsiella serotypes are needed.
Footnotes
This study was submitted in memory of Vicente Javier Benedí, who died prior to submission of the manuscript.
REFERENCES
- 1.Albertí, S., G. Marqués, S. Camprubí, S. Merino, J. M. Tomás, F. Vivanco, and V. J. Benedí. 1993. C1q binding and activation of the complement classical pathway by Klebsiella pneumoniae outer membrane proteins. Infect. Immun. 61:852-860. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Albertí, S., D. Alvarez, S. Merino, M. T. Casado, F. Vivanco, J. M. Tomás, and V. J. Benedí. 1996. Analysis of complement C3 deposition and degradation on Klebsiella pneumoniae. Infect. Immun. 64:4726-4732. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bruce, J. 1996. Automated system rapidly identifies and characterizes microorganisms in food. Food Technol. 50:77-81. [Google Scholar]
- 4.Ciurana, B., and J. M. Tomás. 1987. Role of lipopolysaccharide and complement in susceptibility of Klebsiella pneumoniae to nonimmune serum. Infect. Immun. 55:2741-2746. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Coque, T. M., A. Oliver, J. C. Pérez-Diaz, F. Baquero, and R. Cantón. 2002. Genes encoding TEM-4, SHV-2, and CTX-M-10 extended-spectrum β-lactamases are carried by multiple Klebsiella pneumoniae clones in a single hospital (Madrid, 1989 to 2000). Antimicrob. Agents Chemother. 46:500-510. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Darfeuille-Michaud, A., C. Jallat, D. Aubel, D. Sirot, C. Rich, J. Sirot, and B. Joly. 1992. R-plasmid-encoded adhesive factor in Klebsiella pneumoniae strains responsible for human nosocomial infections. Infect. Immun. 60:44-45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.De Melo, M. C., A. M. Sa Figueiredo, and B. T. Ferreira-Carvalho. 2003. Antimicrobial susceptibility patterns and genomic diversity in strains of Streptococcus pyogenes isolated in 1978-1997 in different Brazilian cities. J. Med. Microbiol. 52:251-258. [DOI] [PubMed] [Google Scholar]
- 8.Di Martino, P., V. Livrelli, D. Sirot, B. Joly, and A. Darfeuille-Michaud. 1996. A new fimbrial antigen harbored by CAZ-5/SHV-4-producing Klebsiella pneumoniae strains involved in nosocomial infections. Infect. Immun. 64:2266-2273. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Di Martino, P., D. Sirot, B. Joly, C. Rich, and A. Darfeuille-Michaud. 1997. Relationship between adhesion to intestinal Caco-2 cells and multidrug resistance in Klebsiella pneumoniae clinical isolates. J. Clin. Microbiol. 35:1499-1503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Gikas, A., G. Samonis, A. Christidou, J. Papadakis, D. Kofteridis, Y. Tselentis, and N. Tsaparas. 1998. Gram-negative bacteremia in non-neutropenic patients: a 3-year review. Infection 26:155-159. [DOI] [PubMed] [Google Scholar]
- 11.Hansen, D. S., F. Mestre, S. Albertí, S. Hernández-Allés, D. Álvarez, A. Doménech-Sánchez, J. Gil, S. Merino, J. M. Tomás, and V. J. Benedí. 1999. Klebsiella pneumoniae lipopolysaccharide O typing: revision of prototype strains and O-group distribution among clinical isolates from different sources and countries. J. Clin. Microbiol. 37:56-62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hansen, D. S., R. Skov, V. J. Benedí, V. Sperling, and H. J. Kolmos. 2002. Klebsiella typing: pulsed-field gel electrophoresis (PFGE) in comparison with O:K-serotyping. Clin. Microbiol. Infect. 8:397-404. [DOI] [PubMed] [Google Scholar]
- 13.Jacoby, G. 1996. Antimicrobial-resistant pathogens in the 1990s. Annu. Rev. Med. 47:169-179. [DOI] [PubMed] [Google Scholar]
- 14.Jacoby, G. 1998. Epidemiology of extended-spectrum beta-lactamases. Clin. Infect. Dis. 27:81-83. [DOI] [PubMed] [Google Scholar]
- 15.Jarlier, V., M. H. Nicolas, G. Fournier, and A. Philippon. 1988. Extended broad-spectrum β-lactamases conferring transferable resistance to newer β-lactam agents in Enterobacteriaceae: hospital prevalence and susceptibility patterns. Rev. Infect. Dis. 10:867-878. [DOI] [PubMed] [Google Scholar]
- 16.Kelly, R. F., W. B. Severn, J. C. Richards, M. B. Perry, L. L. MacLean, J. M. Tomas, S. Merino, and C. Whitfield. 1993. Structural variation in the O-specific polysaccharides of Klebsiella pneumoniae serotype O1 and O8 lipopolysaccharide: evidence for clonal diversity in rfb genes. Mol. Microbiol. 10:615-625. [DOI] [PubMed] [Google Scholar]
- 17.Livermore, D., and M. Yuan. 1996. Antibiotic resistance and production of extended-spectrum β-lactamases amongst Klebsiella spp. from intensive care units in Europe. J. Antimicrob. Chemother. 38:409-424. [DOI] [PubMed] [Google Scholar]
- 18.Livrelli, V., C. De Champ, P. Di Martino, A. Darfeuille-Michaud, C. Forestier, and B. Joly. 1996. Adhesive properties and antibiotic resistance of Klebsiella, Enterobacter, and Serratia clinical isolates involved in nosocomial infections. J. Clin. Microbiol. 34:1963-1969. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Merino, S., S. Camprubí, S. Albertí, V. J. Benedí, and J. M. Tomás. 1992. Mechanisms of Klebsiella pneumoniae resistance to complement-mediated killing. Infect. Immun. 60:2529-2535. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.National Committee for Clinical Laboratory Standards. 2002. Performance standards for antimicrobial susceptibility testing. Supplemental tables. M100-S12. National Committee for Clinical Laboratory Standards, Wayne, Pa.
- 21.Olling, S. 1977. Sensitivity of gram-negative bacilli to the serum bactericidal activity: a marker of the host parasite relationship in acute and persisting infections. Scand. J. Infect. Dis. 10:1-40. [DOI] [PubMed] [Google Scholar]
- 22.Podschun, R., D. Sievers, A. Fischer, and U. Ullmann. 1993. Serotypes, hemagglutinins, siderophore synthesis, and serum resistance of Klebsiella isolates causing human urinary tract infections. J. Infect. Dis. 168:1415-1421. [DOI] [PubMed] [Google Scholar]
- 23.Podschun, R., and U. Ullmann. 1998. Klebsiella spp. as nosocomial pathogens: epidemiology, taxonomy, typing methods, and pathogenicity factors. Clin. Microbiol. Rev. 11:589-603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Porat, R., M. A. Johns, and W. R. McCabe. 1987. Selective pressures and lipopolysaccharide subunits as determinants of resistance of clinical isolates of gram-negative bacilli to human serum. Infect. Immun. 55:320-328. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Sandvang, D. 1999. Novel streptomycin and spectinomycin resistance gene as a gene cassette within a class 1 integron isolated from Escherichia coli. Antimicrob. Agents Chemother. 43:3036-3038. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Schoolnik, G. K., T. M. Buchanan, and K. K. Holmes. 1976. Gonococci causing disseminated gonococcal infection are resistant to the bactericidal action of normal human sera. J. Clin. Investig. 58:1163-1173. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Simberkoff, M. S., I. Ricupero, and J. J. J. Rahal. 1976. Host resistance to Serratia marcescens infection: serum bactericidal activity and phagocytosis by normal blood leucocytes. J. Lab. Clin. Med. 87:206-217. [PubMed] [Google Scholar]
- 28.Taylor, P. W. 1983. Bactericidal and bacteriolytic activity of serum against gram-negative bacteria. Microbiol. Rev. 47:46-83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Toivanen, P., D. S. Hansen, F. Mestre, L. Lehtonen, J. Vaahtovuo, M. Vehma, T. Mottonen, R. Saario, R. Luukkainen, and M. Nissila. 1999. Somatic serogroups, capsular types, and species of fecal Klebsiella in patients with ankylosing spondylitis. J. Clin. Microbiol. 37:2808-2812. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Tomás, J. M., V. J. Benedí, B. Ciurana, and J. Jofre. 1986. Role of capsule and O antigen in resistance of Klebsiella pneumoniae to serum bactericidal activity. Infect. Immun. 54:85-89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Tomasz, A., and R. Munoz. 1995. Beta-lactam antibiotic resistance in gram-positive bacterial pathogens of the upper respiratory tract: a brief overview of mechanisms. Microb. Drug Resist. 1:103-109. [DOI] [PubMed] [Google Scholar]
- 32.Trautmann, M., M. Ruhnke, T. Rukavina, T. K. Held, A. S. Cross, R. Marre, and C. Whitfield. 1997. O-antigen seroepidemiology of Klebsiella clinical isolates and implications for immunoprophylaxis of Klebsiella infections. Clin. Diagn. Lab. Immunol. 4:550-555. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Williams, P., P. A. Lambert, M. R. W. Brown, and R. J. Jones. 1983. The role of the O and K antigens in determining the resistance of Klebsiella aerogenes to serum killing and phagocytosis. J. Gen. Microbiol. 129:2181-2191. [DOI] [PubMed] [Google Scholar]
- 34.Winokur, P. L., R. Canton, J. M. Casellas, and N. Legakis. 2001. Variations in the prevalence of strains expressing an extended-spectrum beta-lactamase phenotype and characterization of isolates from Europe, the Americas, and the Western Pacific region. Clin. Infect. Dis. 32:94-103. [DOI] [PubMed] [Google Scholar]