Acinetobacter species, non-fermenting gram-negative bacilli that are ubiquitous in the environment, have emerged as important nosocomial pathogens . Resistance to drying and to many commonly used antimicrobial agents are the key factors that enable these organisms to survive and spread in nosocomial environments .
The genus Acinetobacter currently comprises 30 named species and nine genomic species [1, 3]. Its most important representative, Acinetobacter baumannii, has emerged as one of the most problematic pathogens for healthcare institutions world-wide, largely because of its ability to develop resistance to several antimicrobial agents [1, 2]. Although species other than A. baumannii are also likely responsible for nosocomial disease, their role is not yet clear because insufficient simple phenotypic tests are used in diagnostic laboratories and these are able only to identify members of the Acinetobacter calcoaceticus–A. baumannii complex [4-7]. Although phenotypic identification of Acinetobacter species is unsatisfactory, several molecular methods have been shown to be adequate for identification. Among these methods, rpoB gene-sequencing analysis appears to be highly useful for identification of Acinetobacter species .
Carbapenems play an important role in the treatment of Acinetobacter infections. However, carbapenem-resistant A. baumannii has rapidly spread worldwide in the past two decades .
Many reports have described the molecular epidemiology of Acinetobacter species outside Japan. However, few such reports especially with regard to the molecular epidemiology of Acinetobacter species other than A. baumannii, have come from Japan. The aim of this study was to investigate the antimicrobial susceptibility of a Japanese blood culture collection of Acinetobacter species and to characterize the isolates on a molecular level.
In this study 123 consecutive blood culture isolates of Acinetobacter species collected between April 2003 and March 2011 from 123 patients by two university hospitals in the same geographical region of Japan were assessed. Contaminants were excluded from this study according to review of charts by infectious disease physicians. Isolates were routinely identified as Acinetobacter species using phenotypic methods, the Microscan Walkaway System (Siemens Healthcare Diagnostics Japan, Tokyo, Japan) and the Vitek 2 System (Sysmex–bioMérieux, Kobe, Japan). Only one isolate from each patient was assessed. Species identification of the isolates was performed by partial rpoB gene sequencing analysis using the primers Ac696F and Ac1093R, as previously described . All isolates were considered correctly identified when the rpoB sequence yielded ≥98% identity with the closest species sequence match in the GenBank database . Identification of A. baumannii was confirmed by PCR amplification of the blaOXA-51-like gene .
Antimicrobial susceptibility was determined using a broth microdilution method with dry plates (Eiken Chemical, Tokyo, Japan) according to 2010 Clinical Laboratory Standards Institute criteria . The breakpoint MICs (in mg/L) of each antibiotic against susceptible strains are as follows: ampicillin/sulbactam, ≤8/4; ceftazidime, ≤8; imipenem, ≤4; meropenem, ≤4; amikacin, ≤16; gentamicin, ≤4; ciprofloxacin, ≤1; levofloxacin, ≤2; and colistin, ≤2 . All isolates were further characterized by PCR. PCR amplification was performed for detection of OXA-51-like, OXA-23-like, OXA-24-like, OXA-58-like, OXA-143-like, OXA-235-like, ISAba1/blaOXA-51-like complex, ISAba1/blaOXA-23-like complex and MBL genes (including IMP, VIM, GIM, SIM, SPM and NDM) as previously reported [13-15].
Sequencing of OXA-type β-lactamase and MBL genes was performed as previously reported [11, 14, 16]. Sequencing of the QRDRs of the gyrA and parC genes was performed for isolates with MICs of >4 mg/L for ciprofloxacin or >8 mg/L for levofloxacin, as previously reported [17, 18].
Partial rpoB gene sequencing analysis assigned isolates to 13 species. The 123 isolates were identified as follows: Acinetobacter pittii (n = 42); A. baumannii (n = 22); A. nosocomialis (n = 20); A. ursingii (n = 15); “A. grimontii” (A. grimontii Carr et al. 2003 is a later heterotypic synonym of A. junii Bouvet and Grimont 1986; ), (n = 7); A. oleivorans (n = 4); A. bereziniae (n = 2); A. soli (n = 2); A. johnsonii (n = 1); A. junii (n = 1); A. baylyi (n = 1); A. radioresistens (n = 1); and Acinetobacter gen. sp. 14BJ (n = 1). Four strains did not show ≥98% sequence similarity with the rpoB gene sequence of any species in the GenBank database and therefore could not be reliably identified; of these four strains, three showed 96% sequence similarity and one showed 97% sequence similarity with the rpoB gene sequence of A. baumannii. PCR amplification of the blaOXA-51-like gene was positive for 22 of the A. baumannii strains identified by partial rpoB gene sequencing analysis. Amplification of the other 101 strains did not produce PCR products. Thus, the species were surprisingly varied. Unexpectedly, the incidence of A. pittii was high (42/123; 34.1%), exceeding that of A. baumannii (22/123; 17.9%). A recent study reported that bacteremia caused by A. baumannii and non-A. baumannii species has different clinical outcomes according to the species involved [4, 7, 20-23]. Our findings emphasize the importance of accurate epidemiological investigation of non-A. baumannii species.
Antimicrobial susceptibility data for the four Acinetobacter species (n = 98) with more than 10 isolates are shown in Table 1. More than 50% of A. nosocomialis and A. ursingii were not susceptible to gentamicin and ceftazidime, respectively. Antimicrobial susceptibility data for the rest of the nine Acinetobacter species (n = 21) were as follows: not susceptible to ampicillin/sulbactam, 9.5%; ceftazidime, 19.0%; imipenem, 4.8%; meropenem, 4.8%; amikacin, 4.8%; gentamicin, 0%; ciprofloxacin, 4.8%; levofloxacin, 0%; and colistin, 4.8%. The following five of the 123 isolates (4.1%) showed reduced susceptibility to three or more of the antimicrobial agents tested: A. pittii (n = 2), A. baumannii (n = 1), “A. grimontii” (n = 1) and A. soli (n = 1). At 4.1%, the fraction of isolates that showed resistance to three or more antimicrobial agents was lower than previously reported for Japanese samples not derived from blood cultures, that is, 8.9% . Two A. pittii isolates and one A. baumannii isolate were resistant to ceftazidime, carbapenems, gentamicin and quinolones. A. baumannii was also resistant to ampicillin/sulbactam. One “A. grimontii” isolate was resistant to ceftazidime, carbapenems and amikacin. One A. soli isolate was resistant to ampicillin/sulbactam, ceftazidime and colistin.
Table 1. Antimicrobial susceptibility profiles of blood culture isolates of four Acinetobacter species
|Acinetobacter species (no. of isolates) and antibiotic||MIC (mg/L)||Non-susceptible rate (%)|
|A. pittii (42)|
|A. baumannii (22)|
|A. nosocomialis (19)|
|A. ursingii (15)|
Of the 123 isolates, the following four (3.3%) were resistant to carbapenems (MIC of ≥8 mg/L for imipenem and meropenem): A. pittii (n = 2), A. baumannii (n = 1) and ″A. grimontii″ (n = 1) (Table 2). To our knowledge, this is the first report of a clinical isolate of carbapenem-resistant “A. grimontii” in Japan. Three of these isolates were also resistant to quinolones. The remaining isolate, “A. grimontii”, was also resistant to amikacin. PCR amplification revealed that one A. baumannii isolate with a MIC of 16 mg/L for imipenem contained class D carbapenemases (OXA-51-like and OXA-23-like) and ISAba1 was detected only upstream of the blaOXA23 gene. Sequencing of the blaOXA-51-like gene showed that the A. baumannii isolate contained blaOXA-66, which is dominant among A. baumannii isolates in Japan [24, 25]. Twenty-one A. baumannii and one A. radioresistens isolates with MIC of ≤4 mg/L for imipenem contained OXA-51-like and OXA-23-like, respectively, but ISAba1 was not detected upstream of the blaOXA-51 and blaOXA-23 genes . Three non-baumannii Acinetobacter isolates with MIC of ≥16 mg/L for imipenem did not contain class D carbapenemases but did contain IMP MBL. Sequencing of the IMP genes revealed only IMP-1, which is dominant among non-baumannii Acinetobacter isolates in Japan . These findings agree with those of recent studies that detected the blaIMP-1 gene in several carbapenem-resistant non-A. baumannii isolates [24, 27]. IMP-19, which is also the predominant MBL among Acinetobacter isolates in Japan, was not found in this study . In addition, because loss of outer membrane proteins is one of the carbapenem resistance mechanisms, CarO outer membrane proteins of four carbapenem resistant isolates were evaluated . Thus, carO gene was detected in all of them; no disruption of carO gene was observed (data not shown). Despite the heavy use of carbapenems in Japan, carbapenem-resistant blood culture isolates of Acinetobacter species were relatively uncommon, given their global prevalence .
Table 2. Characteristics of carbapenem-non-susceptible Acinetobacter species
|Isolates||Species||Years isolated||β-Lactamase||MIC (mg/L)|
|OXA-51 like||OXA-23 like||MBL||Imipenem||Meropenem|
Five of the 123 isolates (4.1%) were resistant to quinolones (MICs of ≥4 mg/L for ciprofloxacin and ≥8 mg/L for levofloxacin); they belonged to A. pittii (n = 4) and A. baumannii (n = 1) (Table 3). Sequencing of the QRDRs showed that all five isolates had a point mutation on the gyrA gene that converted the serine at position 83 (Ser-83) to leucine (Leu) in GyrA, which is consistent with a fluoroquinolone-resistant phenotype. No additional amino acid substitutions were identified for GyrA, not even at other “hot spot” amino acid positions (Gly-81, Ala-84 and Glu-87) known to contribute to fluoroquinolone resistance (Table 3). Sequencing of the parC genes indicated that leucine had replaced Ser-80 in two isolates and tryptophan had replaced Ser-80 in two isolates. No additional amino acid substitutions were identified for ParC, not even at the other “hot spot” amino acid position (Ser-84). Although one A. pittii isolate with moderate fluoroquinolone resistance had only one amino acid substitution in GyrA, the remaining isolates demonstrating strong fluoroquinolone resistance had amino acid substitutions in both GyrA and ParC. This finding is consistent with a previous report .
Table 3. Characteristics of quinolone-non-susceptible Acinetobacter species
|Isolates||Species||Years isolated||Amino acid substitutions due to point mutations in the QRDRs||MIC (mg/L)|
|Gly-81 (GGT)||Ser-83 (TCA)||Ala-84 (GCT)||Glu-87 (GAA)||Ser-80 (TCG)||Ser-84 (GAA)||Ciprofloxacin||Levofloxacin|
|t10||A. pittii||2005||–||Leu (TTA)||–||–||–||–||4||2|
|t41||A. pittii||2009||–||Leu (TTA)||–||–||Leu (CTG)||–||32||8|
|d30||A. pittii||2008||–||Leu (TTA)||–||–||Trp (TGG)||–||64||8|
|d42||A. baumannii||2009||–||Leu (TTA)||–||–||Leu (TTG)||–||64||8|
|d45||A. pittii||2010||–||Leu (TTA)||–||–||Trp (TGG)||–||64||8|
Isolates tested in this study were collected from two university hospitals, between which no differences in antimicrobial susceptibility profiles were found. However, differences were found in the number of species identified; 6 versus 13 species. These hospitals had almost the same number of beds; we were unable to identify an explanation for this difference.
In conclusion, the present study showed a predominance of various non-A. baumannii species and low prevalence of carbapenem resistance and quinolone resistance among blood culture isolates of Acinetobacter species in two Japanese university hospitals. Although our study suggests that carbapenemase genes are not yet widespread among blood culture isolates of Acinetobacter species, antibiotic selection pressure and plasmid-mediated dissemination of carbapenemase may still occur. Therefore, it is important to continue surveillance of Acinetobacter species.