Molecular characterization of clinical multiresistant isolates of Acinetobacter sp . from hospitals in Porto Alegre , State of Rio Grande do Sul , Brazil Caracterização molecular de isolados clínicos Acinetobacter sp . multirresistentes em hospitais de Porto Alegre , Estado do Rio Grande do Sul , Br

Introduction: Hospitals around the world have presented multiresistant Acinetobacter sp. outbreaks. The spread of these isolates that harbor an increasing variety of resistance genes makes the treatment of these infections and their control within the hospital environment more difficult. This study aimed to evaluate the occurrence and dissemination of Acinetobacter sp. multiresistant isolates and to identify acquired resistance genes. Methods: We analyzed 274 clinical isolates of Acinetobacter sp. from five hospitals in Porto Alegre, RS, Brazil. We evaluated the susceptibility to antimicrobial, acquired resistance genes from Ambler’s classes B and D, and performed molecular typing of the isolates using enterobacterial repetitive intergenic consensus-polymerase chain reaction (ERIC-PCR) technique. Results: A high (68%) percentage of multiresistant isolates of Acinetobacter sp. was observed, and 69% were resistant to carbapenems. We identified 84% of isolates belonging to species A. baumannii because they presented the gene blaOXA-51. The gene blaOXA-23 was detected in 62% of the isolates, and among these, 98% were resistant to carbapenems. Using the ERIC-PCR technique, we identified clones of Acinetobacter sp. spread among the four hospitals analyzed during the sampling period. Conclusions: The data indicate the dissemination of Acinetobacter sp. isolates among hospitals and their permanence in the hospital after one year.

The genus Acinetobacter is widely distributed in nature; they are endogenous organisms in various types of soil and water and are occasionally found in food [1][2] .In recent years, members of this group, particularly A. baumannii, have been associated with healthcare-associated infections (HAIs), especially in debilitated patients.The main sites of infection are the respiratory tract, urinary tract, bloodstream, wounds, and burns 3 .Patients with burns and those in intensive care units and/or using mechanical ventilation are at greater risk.According to Jawad et al. 4 , Acinetobacter isolates can survive desiccation for extended periods.This characteristic increases the incidence of infections at hospitals because it makes these bacteria capable of staying in the skin of patients, doctors, and nurses, as well as remaining for long periods on hospital equipment.
The treatment of multidrug-resistant strains can be difficult.Since 1970, HAIs caused by Acinetobacter sp. were successfully treated with aminopenicillins, ureidopenicillins, and cephalosporins of first and second generation.Since then, a steady increase in the prevalence of resistant strains has been seen, which has compromised treatment with penicillins, aminoglycosides, extended spectrum cephalosporins, and most recently, fluoroquinolones 5 .The carbapenems have been the drug of choice against Acinetobacter sp., but the number of isolates resistant to these antimicrobial agents has increased considerably.Carbapenem resistance in A. baumannii is associated with a variety of combined mechanisms, including the acquiring of β-lactamases, stable desrepression of AmpC, decreased in-membrane permeability, alteration of penicillin binding proteins, and overexpression of efflux pumps.7] .The aim of this study was to evaluate the occurrence and dissemination of multiresistant Acinetobacter sp.strains of clinical origin with acquired resistance genes in five hospitals of Porto Alegre, RS, Brazil.

Hospitals studied and bacterial isolates
Acinetobacter sp. was isolated from clinical samples collected in five large hospitals in Porto Alegre, Brazil, and they were obtained after the project was approved by the ethics committee of each hospital involved.The hospitals included in the study were named with letters A to E. Hospital A has an area of 128,338m 2 , 897 beds, and 4,499 employees and receives 12,000 people per day.Hospital B, as hospital A, belongs to the Brazilian Health System (SUS), has 603 beds, and receives 18,000 people per day in an area of 49,000m 2 .Hospitals C, D, and E belong to the same hospital complex but are located in different regions of the city, and these hospitals receive only SUS patients.Hospital C is a general hospital, and it has the largest emergency unit in the state, with 801 beds and 4,631 employees in an area of 43,030m 2 .Hospital D primarily receives victims of traffic and work accidents and violence, as well as burn patients; it has 304 beds, 1,291 employees, and an area of 18,835m 2 .Hospital E receives only obstetric and gynecologic emergency patients, has 189 beds, 737 employees, and an area of 12,273m 2 .In Hospital A, the sampling period was from February to April 2007 and between July and October 2007.At Hospital B, samples were collected in July 2006 and between July and October 2007.The isolates from hospitals C, D, and E were obtained from October to December 2006 and from May to August 2007.Information on the biological source of the sample from the isolates was also obtained from the laboratories of each hospital.
Identification of the isolates of Acinetobacter sp.The isolates were primarily identified as Acinetobacter sp. by the laboratories of the respective hospital.A further confirmation of the genus was performed by amplifying a fragment of 16S rDNA using the primers: Acin16SF (5'-CCT TGA TGC AGA GYT AAT GC-3') and Acin16SR (5'-GTA GCA ACC CTT TGT ACC GA-3'), specific for the genus Acinetobacter.These primers were designed by aligning the sequences 16S rRNA gene from 29 species of Acinetobacter (GENBANK), using the CLC Program FreeWorkbench 2.2.Amplification reactions were performed in mixtures containing 3.5mM MgCl2, 0.2mM dNTPs, 1µM of each primer, 1 unit of Taq DNA polymerase, 1x reaction buffer of Taq DNA polymerase, and 100ng of bacterial DNA in a final volume of 30µL.The amplification conditions were as follows: an initial cycle of denaturation at 95 o C for 5min followed by 25 cycles of denaturation at 95 o C for 1min, annealing at 52°C for 1min and extension at 72 o C for 1min, and a final extension cycle of 72 o C for 8min.

Antimicrobial susceptibility test
The profile of antimicrobial susceptibility of the isolates was provided by each hospital and was determined using the disk diffusion technique according to the standards of the Clinical Laboratory Standards Institute, 2007 8 .We analyzed only the antimicrobial agents tested in common for all hospitals: amikacin (30µg), ceftazidime (30µg), piperacillintazobactam (100µg/10µg), imipenem (10µg), meropenem (10µg), cefepime (30µg), and ampicillin/sulbactam (10µg/10µg).The strains resistant to four classes of antibiotics were considered multiresistant.
Phenotypic MBL detection test Phenotypic triage of MBL production was performed by the disk approximation test using imipenem and ceftazidime as substrates and ethylenediaminetetraacetic acid (EDTA) and 2-mercaptopropionic acid as inhibitors [9][10][11] .

DNA extraction and detection of bla genes
The DNA of the isolates was obtained by the boiling method according to Misbah et al. 12 .As A. baumannii was identified by the presence of the gene bla OXA-51, its amplification was performed with all isolates used in this study.This gene is found exclusively in isolates of this species and can therefore be used to identify it 13 .Isolates that were resistant to imipenem, meropenem, and/or with a positive triage for MBL production were subjected to polymerase chain reaction (PCR) reaction to assess the presence of the bla OXA-23 , bla SPM-1 , bla VIM , and bla IMP genes using the primers listed in Table 1.The reactions were performed in mixtures containing 1x Taq Go Master Mix (Promega), 1µM of each primer, and 100ng of bacterial DNA in a final volume of 15µL.Amplification conditions were as follows: an initial cycle of denaturation at 94 o C for 5min followed by 30 cycles of 94 o C for 1min, annealing temperature as in Table 1, extension at 72 o C for 1min, and a final extension at 72 o C for 6min.PCR products were visualized on 1% agarose gel stained with ethidium bromide.

ERIC-PCR
The genetic similarity of isolates of Acinetobacter sp. was determined by the enterobacterial repetitive intergenic consensus-PCR (ERIC-PCR).The PCR was performed in a final volume of 30µL containing 8L of genomic DNA, 5.5mM of MgCl2, 1.25mM of each dNTP, 400ng of each primer ERIC1 and ERIC2 14 , and 2U of Taq DNA polymerase (Invitrogen).The amplification conditions were as follows: an initial denaturation cycle at 95°C for 7min, 30 cycles of 94ºC for 1min, 51°C for 1min and 72°C for 1min and 30s, and a final extension at 72°C for 15min.The amplified product was visualized on agarose gel 2% plus the polymer Synergel (BioAmerica) in Trisborate, stained with ethidium bromide 0.5µg/mL.The banding patterns obtained from ERIC-PCR were converted into a binary matrix, where score 1 means presence of a band and score 0 means absence of a band.In this analysis, each isolate was considered to be Rev Soc Bras Med Trop 44(6):725-730, nov-dez, 2011

RESULTS
one operational taxonomic unit.The results for the fingerprinting were analyzed via SPSS software (version 13 for Windows).The similarity matrix was calculated using the Dice coefficient and clustering of the similarity matrix, via the unweighted pair-group method using arithmetic averages (UPGMAs) clustering.Only the visible bands in the ERIC-PCR fingerprinting were used to construct the similarity matrix and the dendrogram.Isolates with more than 80% similarity were considered closely related and, therefore, were treated as clones; these isolates differed in one to two bands.
We analyzed 274 isolates of Acinetobacter sp.obtained from five hospitals in Porto Alegre, Brazil, in 2006/2007.Hospital A presented the largest number of isolates, 100 isolates, followed by Hospital C with 93 isolates.Hospital D had 45 isolates, Hospital B had 32 isolates, and Hospital E had only four isolates.Only one isolate from each patient was included in this study.
Among the isolates, 69% were resistant to carbapenems.The hospitals with higher levels of carbapenem resistance were Hospital C with 87% and Hospital D with 85%.Hospital A had 53% of isolates resistant to carbapenems, and Hospital B had 41%.
In analyzing the data from all hospitals, antimicrobial resistance to cefepime and piperacillin-tazobactam, 88% and 83%, respectively, was found to be higher.The antibiotics ampicillin-sulbactam showed low percentages of resistance in Hospitals A and B, with 29% and 55%, respectively, whereas Hospitals C and D showed the highest rates of antimicrobial resistance, with 100% and 97%, respectively.The four isolates from Hospital E were resistant to all antibiotics tested.The isolates of Acinetobacter sp. were obtained from various biological samples, with tracheal aspirate being the most frequent, at 26% of isolates, followed by sputum at 20%, when all isolates were compared.They were also isolated from blood samples (13%), urine (9%), catheter (8%), secretions (6%), and burns (4%) (Figure 1).Other samples containing Acinetobacter sp.such as pleural fluid, abdominal fluid, lumbar puncture, scabs, and wound secretion were not detailed in Figure 1 because of their low percentages.These samples correspond to a total of 14% of the isolates analyzed in this study.As Hospital D has a burn center, it presented seven isolates from burn patients.
Analyzing the data obtained by the antibiogram, the isolates were grouped into 18 distinct profiles based on their susceptibility to seven antimicrobials (Table 2).The greatest diversity of profiles was found in Hospital A, with 16 different profiles, eight of them with just one isolate.In analyzing the five hospitals, the profile with the largest number of isolates showed resistance to seven (61%) antimicrobials, followed by the susceptibility to all drugs, with 12% of isolates.We found 68% of multiresistant isolates, which corresponded to the profiles 1, 2, 3, 4, 5, and 7 of Table 2  The phenotypic triage for MBL production was carried out with all the isolates.Using the antimicrobial imipenem and the inhibitor EDTA, 10% of the isolates were positive.In the test using the antimicrobial ceftazidime and the inhibitor 2-MPA, the percentage of positive isolates increased to 21%.
The detection of the bla OXA-51like gene was performed on all isolates analyzed in this study, and 230 (84%) isolates were positive for the presence of this gene.Hospitals A and C had the greatest number of positive isolates, both with 84 isolates.The four isolates from Hospital E were positive for this gene, whereas Hospital B had 27 positive isolates, and Hospital D had 31.It should be mentioned that among the isolates positive for the bla OXA-51 gene, a high percentage of multidrug resistance was also observed.In Hospitals A and B, 16 and 3 isolates, respectively, were not identified as A. baumannii because they were negative for the gene bla OXA-51 .Of these, 14 were sensitive to all antimicrobial agents analyzed.In Hospitals C and D, 22 isolates were negative for the bla OXA-51 gene, but they showed an opposite behavior when compared with the isolates from the other hospitals because only two isolates were sensitive to all antimicrobials tested.
The presence of the bla SPM-1 , bla VIM , and bla IMP genes were analyzed using PCR in the 214 isolates that were resistant to carbapenems and/or positive in the phenotypic test for detection of MBL.None of the isolates was positive for these genes, although there was a large percentage of positive isolates in phenotypic tests (28%).The same 214 isolates were tested for the presence of the bla OXA-23 gene, and it was found in 61.2% of the isolates.The four isolates from Hospital E had the bla OXA-23 gene.The hospital with the greatest number of positive isolates was Hospital C, with 56 (64%) of the 87 positive isolates examined.Hospital B showed the lowest number (12) of tested isolates, and from these, eight (62%) were positive for the gene.Hospital A had 59% (66 tested) of positive isolates, and Hospital D had 46.6% (45 tested).Twenty-one isolates that were susceptible to carbapenems and positive for the phenotypic screening for MBLs were evaluated for the presence of the bla OXA-23 and bla OXA-51 genes.Eight of these showed the two genes, three isolates from Hospital A, three from Hospital D, and two from Hospital C.
The ERIC-PCR technique was selected for the molecular typing of the isolates that were bla OXA-23 positive.The banding pattern found for these isolates showed high diversity, with patterns of 4 to 17 fragments ranging from 211 to 1,600pb.Twenty-five isolates had a single ERIC-PCR profile, and 124 were grouped into 30 distinct groups (80% similarity as a cutoff) in Table 3.The largest group showed nine isolates of Acinetobacter sp.belonging to four different hospitals (Group AV), and the isolates were from different sampling periods (five and four isolates from 2006 to 2007).Other groups showed similar results; the isolates were from various hospitals and remained in the hospital for a long period (Table 3).Some isolates from the same hospital showed 100% similarity, but this was also observed among isolates from different hospitals, as in Group A. In other groups, isolates with 100% similarity belonged to different hospitals: Hospitals A and B (Group B), Hospitals C and D (Group H), and Hospitals A and E (Group T).This distribution of the isolates in 55 ERIC-PCR profiles indicated that they were grouped neither by the hospital nor by the year of collection.It was not possible to determine any relationship with the susceptibility profiles found.The five isolates that were negative for the bla OXA-51 gene were classified into distinct groups.

Number of isolates per hospital ERIC-PCR groups*
*ERIC-PCR: enterobacterial repetitive intergenic consensus-polymerase chain reaction; UPGMA: unweighted pair-group method using arithmetic averages.

DISCUSSION
In Brazil, as in many parts of the world, outbreaks of infections caused by multidrug-resistant Acinetobacter have been reported.In this study, we found a high percentage of multidrug-resistant (68%) and carbapenem-resistant (69%) Acinetobacter sp.isolates, which are still the best treatment option in cases of infection by this microorganism.The antimicrobial resistance in this genus is associated with several mechanisms, but the production of β-lactamases has been considered the main one 18 .
Although Acinetobacter sp. is implicated in a variety of HAIs, including meningitis, urinary tract infections, and surgical infections, it is the predominant etiologic agent of nosocomial pneumonia, especially in patients with ventilator-associated pneumonia, who have been admitted to treatment in intensive care units 19 .This observation concurs with observations obtained in this study regarding isolation of Acinetobacter sp., mainly from sputum and tracheal aspirate samples.Other studies also found similar results to ours, such as Papa et al. 20 , who found 20% of isolates obtained from tracheal aspirate, 17% from blood culture, and 15% from catheters.Another study found 32% of isolates from the respiratory tract, 19% from wounds, 19% from urine, 16% from blood, and 13% from catheter 21 .
Our study found 28% of positive isolates for the phenotypic screening test for MBL, and none of the isolates exhibited the bla IMP , bla VIM , and bla SPM genes.The disk approximation techniques proposed by Arakawa et al. 10 and Lee et al. 12 for Gram-negative bacilli were initially effective for the detection of metallo-enzymes β-lactamases.As the technique was used to detect Acinetobacter sp., studies have shown low specificity for this genus [22][23] .Therefore, it was suggested that positive results in the phenotypic test for detection of MBL should be interpreted carefully in samples of Acinetobacter sp.
The bla OXA-51 gene has been found in all isolates of A. baumannii tested and is being considered as a natural component of the species chromosome 8,24 .Thus, the presence of this gene has been used to identify A. baumannii.This gene may be associated with resistance to carbapenems when ISAbaI-type insertion sequences, which carry strong promoters, are found upstream to bla OXA genes, resulting in increased expression and concomitant resistance to carbapenems 25 .In our study, we found a large number of isolates positive for the bla OXA-51 gene (84%), and results similar to that were found by Wisplinghoff et al. 26 , who observed 86% of isolates positive for the bla OXA-51 gene.In many studies, the species A. baumannii is the most prevalent and associated with HAIs, and it is also the species with the highest rate of resistance to antimicrobials [27][28] .
Carbapenem resistance in isolates of Acinetobacter sp. is most often mediated by oxacillinase (OXA)-type enzymes and less frequently by MBLs 29 .As in ours, other studies have searched for enzymes like MBLs, but the OXA-type enzymes, such as bla OXA-23 , bla OXA-24 , bla OXA-51 , and bla OXA-58 , are the most frequent in this genus 7,[30][31] .In another study also held in Porto Alegre, Brazil, with Acinetobacter sp., 53 isolates were analyzed, and all of them showed the genes bla OXA-23 and bla OXA-51

27
. In another study with A. baumannii in Curitiba, Brazil, the bla OXA-23 gene was detected in eight strains resistant to carbapenems.They also found that isolates susceptible to carbapenems did not have the bla OXA 23 gene 32 .In our study, we observed eight strains that were sensitive to carbapenems and had the bla OXA-23 and bla OXA-51 genes.OXA-type enzymes exhibit a weak hydrolysis of carbapenems and may not always show resistance profile, but when they are associated with insertion elements, they may have an increase in its expression and show resistance to carbapenems 29 .
The ERIC-PCR technique is a low-cost, alternative and rapid method that showed good results for the genetic characterization of Acinetobacter sp.isolates.By ERIC-PCR, we identified clones of Acinetobacter sp. in different hospitals, as well as the permanence of isolates from one year to another in the same hospital.These data can be related to the dissemination characteristics of this microorganism, its permanence for long periods in hospitals, its main form of dissemination, human contact, and the mobility of patients and staff within and between hospitals.In a work carried out by Jeong et al. 33 using the ERIC-PCR in isolates of A. baumannii, identical profiles were observed for seven isolates producing OXA-23 in a hospital in Korea.In another study with seven isolates of A. baumannii producing OXA-23 from the same hospital, two distinct groups were identified using ERIC-PCR, one group consisting of five isolates with identical profiles and the other consisting of two isolates 6 .
With the spread of multiresistant isolates and the diversity of resistance genes, the choice of appropriate treatment becomes increasingly difficult.It is of great importance to study the local epidemiology of Acinetobacter sp.isolates to establish the best treatment and use the correct epidemiological control, to avoid the spread of these microorganisms, because studies like this have shown the dissemination of clones in different hospitals.Efficient control measures such as washing hands, keeping patients in isolation, cleaning equipment properly, and ensuring awareness of the technical staff are needed to reduce the outbreaks of Acinetobacter sp.In our study, we can observe the behavior of the isolates from different hospitals analyzed through the antimicrobial resistance and molecular typing profiles and resistance genes found.The five hospitals exhibited dissemination of multiresistant and carbapenemresistant Acinetobacter sp.isolates harboring the bla OXA-23 and bla OXA-51 genes, probably acting as the main causing agent of the high rate of carbapenem resistance observed in this study.

TABLE 1 -Polymerase chain reaction primers and annealing conditions used in this study for detection of the resistance genes.
.
FIGURE 1 -Percentage of isolates for each biological material used in the isolation of Acinetobacter sp. per hospital (A, B, C, and D).