Acinetobacter baumannii is a non-fermentative, gram-negative bacterium reported as an pathogen, responsible for serious nosocomial infections and outbreaks.1,2 Among the risk factors, widespread usage of broad-spectrum antibiotics3,4 has accelerated the evolution and spread of antibiotic and multidrug-resistant (MDR) bacteria. Most reports supported the notion that antibiotic resistance gene is the major molecular basis for drug resistance and the genetic mobile elements in the rapid acquisition of multiple antibiotic-resistance genes, like transposon and integron,5–7 even if it was suggested the close relationship between resistance gene acquisition and resistance phenotype was unclear.8 If the antibiotics are continued to be misused, it would speed the evolution of multidrug-resistant bacteria and dissemination. Therefore, it is helpful to understand the molecular basis, origin and evolution of multi-drug resistance gene and to understand its potential to spread between strains by searching the relation between different drug resistance genes and antimicrobial spectrum.
Genotyping has been introduced into epidemiology as an important tool for nosocomial pathogen grouping. A number of genomic fingerprinting methods have been proposed, such as amplified fragment length polymorphism (AFLP),9 pulsed-field gel electrophoresis (PFGE).10 But the deficiency of both methods in specific drug resistant gene identification made them at a dispute whether are worth applied as laboratory tools in epidemiological investigation, even if PFGE has been considered as the reference method for genotypic characterization of the majority of nosocomial pathogens.10 Meantime, in certain setting or in the hospital, there is a puzzle whether there will be a similar result when using these methods to type different drug resistance strains in endemicity or sporadic case, especially that integron with gene cassette occurred to chromosome (named superintegron) or plasmid under selective pressure.
Therefore, a modified typing method of drug resistance gene clustering, drug resistance gene typing (DRGT) was proposed with specific grouping advantage according to MDR gene presence, whose typeability was comparable to PFGE and better than AFLP in a relative closed area, like hospitals. Especially, during an endemicity period, DRGT could also distinguish epidemiological strain from non-epidemiological strain after the isolation of Acinetobacter baumannii from hospitalized patients.
Acinetobacter baumannii isolates were recovered from serial surveillance cultures obtained from patients admitted to the Intensive Care Units (ICUs) of the Second Affiliated Hospital in Ningbo, from January 2005 to October 2005. Twenty-seven clinical Acinetobacter isolates were collected from samples of 27 patients by standard protocols and isolated in pure cultures on MacConkey agar plates (Oxoid Ltd., Basingstoke, UK). Bacterial cultures for the sputum, blood, urine, pus, bronchoscopic and catheter fluid were performed at regular intervals during the ICU monitoring. The experiment numbers of these isolates were coded randomly for avoiding subjective bias. Epidemiological related isolates were defined as the sequential isolates collected from different patients who had overlapping hospital admission with the same hospital floors or ICU rooms, or else these isolates were named epidemiologically unrelated isolates or sporadic isolates. P3, P4, P6, P14, P19, P20, P23, P24, P27 collected from January to March and P9, P11, P12, P13, P18, P21, P22, P25, P26 from August to October were epidemiologically linked respectively.
Bacterial isolates identification and susceptibility testing
Upon laboratory receipt, presumptive identification of the 27 isolates was performed using an automated system (ATB; bioMerieux, France). The isolates were stored at -20°C containing glycerol 50% (v/v) before start of the study.
Antimicrobial susceptibility testing was performed for all bacterial strains using Clinical and Laboratory Standards Institute (CLSI; formerly National Committee for Clinical Laboratory Standards (NCCLS, 2003)) reference methods published in M7-A6 to determine minimal inhibitory concentrations (MICs) on validated dry-form panels (TREK Diagnostics, Cleveland, OH). The antimicrobial agents tested included piperacillin, piperacillin/clavulanic acid, cepoperazon/sulbactam, ceftazidime, cefotaxim, cefepime, aztreonam, imipenem, meropenem, gentamicin, amikacin, ciprofloxacin, sulfamethoxazole. CLSI categorical interpretive criteria published in M100-S15 (CLSI, 2005) were applied for susceptibility and resistance. Strains Escherichia coli ATC 25922, Pseudomonas aeruginosa ATCC 27853, and Staphylococcus aureus ATCC 29213 were used as quality control of the susceptibility test reagents and methods.
Extraction of genomic DNA for polymerase chain reaction experiments
Isolates were grown overnight on brain heart infusion (BHI) agar plates at 37°C and five or six discrete colonies of each isolate were removed with a sterile loop and emulsified in 50 μl of sterile distilled water in a 0.2 ml Eppendorf tube. Lysis was achieved by heating the tubes for 10 minutes at 95°C with a 2400 PerkinElmer thermal cycler (Perkin-Elmer Cetus, Norwalk, Connecticut). Lysed cells were always used on the day of preparation.
The method was performed according to Koelman et al.9 Briefly, purified DNA was digested by the EcoRI and MseI, and the EcoRI and MseI adapters were added for ligation thereafter. DNA fragments were amplified with a Cy5-labelled EcoRI + A primer (Cy5-GACTGCGTACC-AATTCa-3′; a=selective A base), and an MseI + C primer (5′-GATGAGTCCTGAGTAAc-3′; c=selective C base). The ALF expressTM DNA analysis system (Amersham Biosciences, USA) was used for fragment separation that were saved as Tiff formats that were subjected to quantitative analysis with Image/J 1.37 V (NIH, USA) to find out the peak patterns. Following conversion, normalization, and background subtraction with mathematical algorithms, the peak bands were scored as “0” or “1” according to the identification standard of bands. Levels of similarity between fingerprints were calculated with the Pearson product-moment correlation coefficient (r). Cluster analysis was performed based on the unweighted pair-group method using average linkages (UPGMA) by SAS software 9.0 (SAS Institute Inc., Cary, NC, USA). AFLP genotypes were defined using an 80% clustering level, which in our hands had been proven appropriately to define outbreaks. Interpretation of AFLP patterns was blinded to the PFGE results.
PFGE was performed as previously described.9 Briefly, genomic DNA was digested with ApaI for Acinetobacter baumannii. DNA fragments were separated in a 1% agarose gel and run in 0.5 mol/L Tris-EDTA at 6 V/cm on a contour-clamped homogeneous electric field apparatus (CHEF Mapper XA System; BioRad, Richmond, CA). Pulse times ranged from 5.0 to 8.0 seconds for 24 hours for Acinetobacter species. The gels were stained with ethidium bromide and photographed under ultraviolet light. Isolates were assigned to the same strain type (e.g., A) without different bands, to a subtype (e.g., A1) if the PFGE pattern differed by fewer than three bands and to another one (e.g., A2) if it differed by three to six bands. Isolates with a PFGE pattern differing by more than six bands were assigned to a separate strain type.10
DRGT were carried out as following indicated. The selected drug resistance gene markers were based on recent epidemiological investigations1,2,4,5,7,11–20 and the susceptibility test result. The sequences of these alleles encoding antibiotic resistance were aligned and group-specific regions were identified using Bioedit software (http://www.mbio.ncsu.edu/BioEdit/bioedit.html). Some primers were used according to previous reports, OXA-23 and OXA-24,11 SHV, VEB-1, CTX-M-1, GES, TEM and PER-1,19 aac(3)-I, aac(3)-II, aac(6′)-I, ant(2″)-I21, and other primers were self-designed (Table 1). When amplification products were more than 500 bp, the amplification conditions were: initial denaturation at 93°C for 2 minutes, 35 cycles of 93°C for 60 seconds, 55°C for 60 seconds and 72°C for 60 seconds, and a final elongation at 72°C for 5 minutes. The others were 93°C pre-degeneration for 2 minutes (35 cycles), 93°C for 30 seconds, 55°C for 30 seconds, 72°C for 60 seconds, and a elongation at 72°C for 5 minutes. The accuracy of fragments was confirmed by genomic sequence and sequence alignment.
For each strain, a panel of 25 drug resistance genes together with integron and Tn21/Tn501 transposon, was amplified with PCR and checked by 1% agarose gel whose image result was transformed in a binary way (namely, positive specific band was “1”, and negative band was “0”), used as polygene cluster analysis by Q-type cluster. An original data matrix was composed of all the data collected and then was clustered for judging the relationships between these 27 isolates, which was used for the polygene cluster analysis by Q-type cluster. Similar the criterion of PFGE, isolates were assigned to the same strain type (e.g., A) if the DRGT pattern had the same genes and to a subtype if it differed by a gene (e.g., A1). Isolates with a DRGT pattern differing by more than two genes were assigned to a separate strain type.
Identifying drug resistance gene from the clinical isolates
The antibiotic susceptibility test was performed to find out the drug resistant property of the clinical isolates (Table 2). The incidence of resistance to sulfamethoxazole was the highest, followed by cefotaxime and aztreonam, which were 96.3%, 92.6% and 92.6% respectively. Moreover, the minimal incidence of drug resistance was also as high as 55.6% about imipenem or meropenem. Interestingly, 25 isolates were with the multi-drug resistant and 15 strains resisted to all antibiotics tested.
In concordant with most of antibiotic susceptibility test, multi-drug resistant genes were simultaneously detected in a single isolate, or the isolates except P1 and P2, mainly contained the β-lactamase genes and aminoglycosides modification genes (Table 3). Twenty-three (85.2%) of 27 isolates was detected β-lactamase genes, including TEM 22 (81.5%), OXA-23 12 (44.4%), ADC 23 (85.2%). Twelve (52.2%) isolates with ADC gene was also found to carry TEM and OXA23 genes, while other 10 (43.5%) with ADC gene took TEM gene. Furthermore, the proportion of aminoglycosides modification genes were 85.2% totally, among which, 23 (85.2%) isolates were positive for aac(3)-I, 18 (66.7%) for aac(6′)-I and 22 (81.54%) for ant(3″)-I. Twenty-three (85.2%) were carried qacE ⊲ 1 and sul1 gene. Noticeably, only intI1 (Figure) was obtained in these genetic markers of integron and Tn/Tn501 transposon, amounting for 85.2%.
Although most drug resistance strains supported the exact causal relationship between genotype and phenotype, there was no any drug resistance gene detected in P15 (the number of strain), which was resistant to gentamicin and amikacin. The similar cases were found in P8.
Comparison between DRGT, PFGE and AFLP typeability
Following the clustering analysis of the genotyping results, typeability of DRGT and AFLP were compared with PFGE (Table 3). Both PFGE and AFLP had three types, and the AFLP clustering was affect by the cutoff value of homology. Furthermore, using PFGE as reference, the concordance rate of typing was 88.9% (24/27) for DRGT, and 77.8% (21/27) for AFLP that has the maximal approximation. However, some strains could not be typed by DRGT, including P1, P2, P8 and P15, even if they were strains resistant to cefotaxime and sulfamethoxazole.
Epidemic strain identification
Now, PFGE is one of the standard genotyping methods to discriminate pathogen outbreak or epidemic. For all A-type of PFGE, 22 strains except p10 were overlapped with DRGT A-type, accounted as high as 95.7% (22/23) for PFGE A-type, but AFLP A-type had 82.6%. However, there were a few differences between A-subtype. Considering the epidemiological data, there was two epidemic events of the 19 isolates of A-type in PFGE contained nearly two epidemic events based on the event interval definition that is larger than 3 months, which was never identified by PFGE typing. DRGT had two A subtypes, A and A1 subtypes, which might reflect the real epidemic situation, though a slight difference was present in P9, P12 and P13 without TEM gene (Table 3). In contrast, though AFLP also showed the same epidemic trend as PFGE, It included more other types (e.g. p15 belonged to B in PFGE) and obscured the bounds between subtypes and types.
Drug resistance evaluation
Based on the resistance examined by antibiotics susceptibility test, relationship between genotyping and drug resistance was further analyzed. DRGT provided more resistance information than PFGE did in nosocomial infections. Among A subtype of PFGE, the subgroup (P3, P4, P6, P14, P19, P20, P23, P24, P27) suggested the same clone dissemination as the other subgroup (P11, P18, P21, P22, P25, P26). But in DRGT, A-subtype, A1 and A2 subtypes shared aac(3)-I, aac(6′)-I, ant(3″)-I, qacE ⊲ 1-sul1, and intI1 genes as stem, which was diversified by TEM and OXA-23 like gene presence. For example, A subgroup had both genes, while A1 possesses only TEM and A2 lacks both genes. However, P7 and P17 in A-subtype of DRGT were divided into the A1 subtype of PFGE, though they had all detected drug resistant genes. And there were three isolates with different types between PFGE and DRGT, which were P8, P10 and P15. In AFLP, it was uncertain about the relationship between genotyping and drug resistance.
Recently, Acinetobacter baumannii becomes a frequent nosocomial infection species, and it is important to investigate the relationship between genotyping and drug resistance from epidemic or outbreak strains in the hospitals.1,4,5,11–19,21–23 Moverover, Acinetobacter baumannii could acquire different drug resistance with positive findings of TEM, SHV, OXA, PER, VEB genes, which were considered as one of the major drug resistance mechanisms. The study also supported that there were a spectrum of Acinetobacter baumannii drug resistance genes that contributed to the outbreaks and sporadic cases of infection in the hospital. Twenty-seven isolates not only possessed the higher incidence of β-lactamase genes, like TEM, OXA-23, ADC than others,11,12,19 but also had more kinds of drug resistant genes than other studies,1,4,5 including aac(3)-I, aac(6′)-I, ant(3″)-I, qacE ⊲ 1-sul1, et al. Hujer et al14 suggested the similar results that MDR Acinetobacter baumannii isolates with at least eight identified resistance determinants were recovered from 49 of the 75 patients. In this study, DRGT is a good way for the epidemic strain genotyping with additional information on antibiotic genes. AFLP genotyping varies depending on the cutoff value and has no relation with antibiotics. Although PFGE was used as a reference when analyzing typeability, its typing result could not be supported by antibiotic susceptibility test in some cases.
The fundamental feature of a molecular typing method is its ability to differentiate between epidemiologically related and unrelated strains. The results of this study suggest that DRGT is comparable to PFGE and better than AFLP in distinguishing between different strains of Acinetobacter baumannii in an endemic setting though some differences are observed. The main possibility of inability of AFLP to differentiate between strains may reflect the amplification of only a subset of chromosomal DNA fragments, allowing the examination of only a portion of potentially varying sites between different strains. In studying the relation between drug resistance gene and PFGE in typing, Noppe-Leclercq et al21 ever suggested the use of fingerprinting based on PCR amplification of aminoglycoside resistance genes as the rapid screening of Acinetobacter baumannii isolates. However, some discrepancies between resistance gene profiles of strains from the same pulsotypes were also observed. The similar phenomenon is showed in the study. Interestingly, the strains among the A-type by PFGE were divided two groups by DRGT, the latter of which showed more drug resistance gene blaOXA-23 than former. In fact PFGE typing method is often aimed for all the chromosome of bacteria or plasmid and could not provide all the perspective of drug resistance strains and maybe mislead the decision of treatment. DRGT had a better method to combine discriminatory ability with drug resistance in nosocomial infections. In addition to a high discriminatory power, DRGT has other requirements for a molecular typing technique including typeability and reproducibility. DRGT had the same three different strain types as PFGE and less than six ones of AFLP. With regards to the reproducibility of DRGT, the use of primers that match their target site perfectly and stringent annealing temperatures ensure that similar DNA profiles are generated upon repeat testing. The results of this study indicate that DRGT and PFGE are comparable molecular typing methods for Acinetobacter baumannii isolates.
Additional epidemiological information was provided by DRGT that type 1 integrons were detected in the two kinds of Acinetobacter baumannii epidemic strains, but not in the sporadic strains C, in accordance with several studies presenting integrons as a marker for outbreak strains.20,24 Integron only is analyzed by carrying gene cassette and ignoring of other drug resistance genes, however DRGT method can be used to make up for their methods and provided more information about drug resistance.
However, as the supplement of PFGE, DRGT has its limitations. It is noteworthy that integration of insertion sequence within a resistance gene may lead to false-negative results. In such an eventuality, the PCR should yield either a product of larger size or negative results. Sometimes, there were no drug resistance gene detected, which requires a full list of the potential drug resistance genes in case of some unexpected epidemiological events, such as P15 and P8. Although P10 with B-type in DRGT had four different drug-resistant genes, it belonged to A-type in PFGE. It suggested the existence of unknown effects. Nevertheless, the caution about the cost-efficiency and the difficulty of clinical practice also maybe limit the application of DRGT method. Certainly, small sample sizes maybe weaken the effect of discrimination power.
For nosocomial infection, it is a key to administrate antibiotics in a proper way and reduce the nosocomical drug resistance occurrence. In contrast to PFGE and AFLP, DRGT directly reflects the presence of the drug resistant gene whose typing results could be correlated with antibiotics administration. Supposing the successive isolates were collected in a surveillance hospital, the origin of the new isolates could be identified by discriminatory analysis and then studied further the original, dissemination and evolution of certain drug resistance gene. In sum, DRGT is important for bacteria dissemination control and patient treatment in the case of nosocomial epidemic and outbreak, and is also useful in prediction and prevention of the epidemic events.
1. Mammeri H, Poirel L, Mangeney N, Nordmann P. Chromosomal integration of a cephalosporinase gene from Acinetobacter baumannii
into Oligella urethralis
as a source of acquired resistance to ß-lactams. Antimicrob Agents Chemother 2003; 47: 1536-1542.
2. Wisplinghoff H, Edmond MB, Pfaller MA, Jones RN, Wenzel RP, Seifert H. Nosocomial bloodstream infections caused by Acinetobacter
species in United States hospitals: clinical features, molecular epidemiology and antimicrobial susceptibility. Clin Infect Dis 2000; 31: 690-697.
3. Koeleman JG, Parlevliet GA, Dijkshoorn L, Savelkoul PH, Vandenbroucke-Grauls CM. Nosocomial outbreak of multi-resistant Acinetobacter baumannii
on a surgical ward: epidemiology and risk factors for acquisition. J Hosp Infect 1997; 37: 113-123.
4. Vahaboglu H, Ozturk R, Aygun G, Coskunkan F, Yaman A, Kaygusuz A, et al. Widespread detection of PER-1-type extended-spectrum beta-lactamases among nosocomial Acinetobacter
and Pseudomonas aeruginosa
isolates in Turkey: a nationwide multicenter study. Antimicro Agents Chemother 1997; 41: 2265-2269.
5. Marilyn CR. Multidrug-resistant genes are associated with an 86-kb island in Acinetobacter baumannii.
Trends Microbiol 2006; 14: 375-378.
6. Naiemi NA, Duim B, Savelkoul PH, Spanjaard L, de Jonge E, Bart A, et al. Widespread transfer of resistance genes between bacterial species in an intensive care unit: implications for hospital epidemiology. J Clin Microbiol 2005; 43: 4862-4864.
7. Wu TL, Chia JH, Su LH, Kuo AJ, Chu C, Chiu CH. Dissemination of extended-spectrum beta-lactamase-producing Enterobacteriaceae
in pediatric intensive care units. J Clin Microbiol 2003; 41: 4836-4838.
8. Cousin SJ, Whittington WL, Roberts MC. Acquired macrolide resistance genes in pathogenic Neisseria spp.
isolated between 1940 and 1987. Antimicrob Agents Chemother 2003; 47: 3877-3880.
9. Erika MC, Monique MG, Tang YW. Comparsion of pulsed-field gel electrophoresis
and amplified fragment-length polymorphism for epidemiological investigations of common nosocomial pathogens. Infect Control Hosp Epidemiol 2001; 22: 550-554.
10. Tenover FC, Arbeit RD, Goering RV. The Molecular Typing Working Group of the Society for Healthcare Epidemiology of America. How to select and interpret molecular strain typing methods for epidemiological studies of bacterial infections: a review for healthcare epidemiologists. Infect Control Hosp Epidemiol 1997; 18: 426-439.
11. Afzal-Shah M, Woodford N, Livermore DM. Characterization of OXA-25, OXA-26, and OXA-27, molecular class D β-lactamases associated with carbapenem resistance in clinical isolates of Acinetobacter baumannii.
Antimicro Agents Chemother 2001; 45: 583-588.
12. Chu YW, Afzal-Shah M, Houang ET, Palepou MI, Lyon DJ, Woodford N, et al. IMP-4, a novel metallo-β-lactamase from nosocomial Acinetobacter spp.
collected in Hong Kong between 1994 and 1998. Antimicro Agents Chemother 2001; 45: 710-714.
13. Heritier C, Poirel L, Lambert T, Nordmann P. Contribution of acquired carbapenem-hydrolyzing oxacillinases to carbapenem resistance in acinetobacter baumannii.
Antimicrob Agents Chemother 2005; 49: 3198-3202.
14. Hujer KM, Hujer AM, Hulten EA, Bajaksouzian S, Adams JM, Donskey CJ, et al. Analysis of antibiotic resistance genes in multidrug-resistant Acinetobacter spp.
isolates from military and civilian patients treated at the Walter Reed Army Medical Center. Antimicrob Agents Chemother 2006; 50: 4114-4123.
15. Jeon BC, Jeong SH, Bae IK, Kwon SB, Lee K, Young D, et al. Investigation of a nosocomial outbreak of imipenem-resistant Acinetobacter baumannii
producing the OXA-23 beta-lactamase in korea. J Clin Microbiol 2005; 43: 2241-2245.
16. Landman D, Quale JM, Mayorga D, Adedeji A, Vangala K, Ravishankar J, et al. Citywide clonal outbreak of multiresistant Acinetobacter baumannii
and Pseudomonas aeruginosa
in Brooklyn, NY. Arch Intern Med 2002; 162: 1515-1520.
17. Lee K, Yong D, Yum JH, Lim YS, Bolmstrom A, Qwarnstrom A, et al. Evaluation of Etest MBL for detection of blaIMP-1 and blaVIM-2 allele-positive clinical isolates of Pseudomonas spp.
and Acinetobacter spp.
J Clin Microbiol 2005; 43: 942-944.
18. Mammeri H, Poirel L, Mangeney N, Nordmann P. Chromosomal integration of a cephalosporinase gene from Acinetobacter baumannii
into Oligella urethralis
as a source of acquired resistance to ß-lactams. Antimicrob Agents Chemother 2003; 47: 1536-1542.
19. Nass T, Coignard B, Carbonne A, Blanckaert K, Bajolet O, Bernet C, et al. French nosocomial infection early warning investigation and surveillance network. VEB-1 extended-spectrum ß-lactamase-producing Acinetobacter baumannii,
France. Emerg Infect Dis 2006; 12: 1214-1222.
20. Turton JF, Kaufmann ME, Glover J, Coelho JM, Warner M, Pike R, et al. Detection and typing of integrons in epidemic strains of Acinetobacter baumannii
found in the United Kingdom. J Clin Microbiol 2005; 43: 3074-3082.
21. Noppe-Leclercq I, Wallet F, Haentjens S, Courcol R, Simonet M. PCR detection of aminoglycoside resistance genes: a rapid molecular typing method for Acinetobacter baumannii.
Res Microbiol 1999; 150: 317-322.
22. Poirel L, Cabanne L, Vahaboglu H, Nordmann P. Genetic environment and expression of the extended-spectrum beta-lactamase blaPER-1 gene in gram-negative bacteria. Antimicrob Agents Chemother 2005; 49: 1708-1713.
23. Poirel L, Menuteau O, Agoli N, Cattoen C, Nordmann P. Outbreak of extended-spectrum beta-lactamase VEB-1-producing isolates of Acinetobacter baumannii
in a French hospital. J Clin Microbiol 2003; 41: 3542-3547.
24. Agodi A, Zarrilli R, Barchitta M, Anzaldi A, Di Popolo A, Mattaliano A, et al. Alert surveillance of intensive care unit-acquired Acinetobacter
infections in a Sicilian hospital. Clin Microbiol Infect 2006; 12: 241-247.
Keywords:© 2009 Chinese Medical Association
Acinetobacter baumannii; drug-resistance gene typing; pulsed-field gel electrophoresis; amplified fragment length polymorphism fingerprinting