Molecular association of pathogenicity and resistance to multiple antimicrobials in Acinetobacter baumannii strains recovered from patients with diverse infectious diseases

1. Universidade Federal de Minas Gerais (UFMG), Minas Gerais, Brazil
2. Centro Universitário do Maranhão, Maranhão, Brazil

Endereço para correspondência

Simone Gonçalves dos Santos
Laboratório de Microbiologia Oral e de Anaeróbios; Departamento de Microbiologia; Instituto de Ciências Biológicas; Universidade Federal de Minas Gerais
Avenida Presidente Antônio Carlos, 6.627, C.P. 486; Pampulha-Campus UFMG
CEP: 31270-901; Belo Horizonte-MG, Brasil
Phone: +55 (31) 3409-2743
e-mail: simonegsantos@icb.ufmg.br

INTRODUCTION

The success of Acinetobacter baumannii infections can be attributed to several virulence factors, such as its ability to form biofilms and resist desiccation on abiotic surfaces; ability to adhere, colonize and invade human epithelial cells; its repertoire of antimicrobial resistance mechanisms that are able to self modulate promptly when required; and its ability to acquire genetic material from other microorganisms by lateral gene transfer, promoting their survival under selective pressure of antimicrobials⁽¹⁾. The outer membrane protein A (ompA), a trimeric porin, plays an adhesion role in A. baumannii on plastic and also in interaction between the microorganism and human epithelial cells and Candida albicans filaments⁽²⁾. According to Tomaras et al. (2003)⁽³⁾, the inactivation of csuE, which encodes a terminal adhesin, results in abolition of pili production and biofilm formation. This observation suggests that the operon csuE BABCDE, which mediates the formation of pili, plays a crucial role in the onset of biofilm formation by allowing the adhesion of bacterial cells and starting microcolony training prior to the full development of biofilm structures.

Acinetobacter species still exhibit a variety of enzymatic and resistance mechanisms based on changes in the plasma membrane, which may be upregulated under selective pressure. According to Young et al. (2012)⁽⁴⁾ and Vala et al. (2014)⁽⁵⁾, nine subclasses of acquired metallo-betalactamase (MβL) are known: imipenemase (IMP), Verona imipenemase (VIM), São Paulo MβL (SPM), German IMP (GIM), New Delhi MβL (NDM), Kyorine Hospital MβL (KHM-1), Australian IMP (AIM-1), Dutch IMP (DIM-1) and Seoul IMP (SIM-1). According to Bush and Jacoby (2010)⁽⁶⁾, oxacillinases are the most common resistance mediators in A. baumannii. Oxacillinases are a β-lactam group that preferentially hydrolyzes oxacillin over benzylpenicillin. Furthermore, they can effectively inactivate aminopenicillins, methicillins and in a lesser extent, some first-generation cephalosporins. The insertion element ISAba1 contains sequences that are associated with increased expression of bla_amp C and resistance to carbapenems when associated with bla_OXA51^{(7, 8)}.

In this study we evaluated, by phenotypic and genotypic parameters, the presence of virulence and resistance factors in clinical Acinetobacter baumannii strains and verified the correlation between the variables.

METHODS

Study design and population

This was a prospective cohort observational study, conducted in a with 506-bed university hospital (56 beds for adult intensive care) in Belo Horizonte, Brazil. The microbiological, phenotypic and molecular analyses were conducted in the Laboratory of Oral Microbiology and Anaerobes of the Institute of Biological Sciences/Universidade Federal de Minas Gerais (UFMG). The study was approved by the UFMG research ethics committee (ETIC 614/08).

Microbiological examination

To confirm the species of Acinetobacter baumannii-calcoaceticus complex, the detection of the bla_OXA51 gene, known marker of the specie, was performed through the polymerase chain reaction (PCR) multiplex, which has detected the presence of other oxacillinases (bla_OXA23, ₂₄, ₅₈), as described by Martins et al. (2014)⁽⁹⁾.

The bacterial deoxyribonucleic acid (DNA) and Master Mix preparation were performed as described by Martins et al. Master Mix kit (Promega^®, Madison – WI, USA) was used. Each reaction had a final volume of 25 µl with 12.5 µl of Master Mix + 1.2 µl of primers F and R + 100 ng of bacterial DNA extracted + nuclease-free water to make up 25 µl.

The strains were submitted to antimicrobial susceptibility testing and verification of expression of MβL by Epsilometer test (E-test^®) method, and were interpreted as suggest by the manufacturer.

For purposes of standardization of minimum inhibitory concentration (MIC), and considering the peculiarities of the institution involved in the study, all samples were again subjected to antimicrobial susceptibility testing and interpretation by E-Test^® method, as recommended by the producer. The following drugs were tested: ampicillin-sulbactam (AB), ceftazidime (TZ), gentamicin (GM), meropenem (MP), polymyxin B (PO) and tigecycline (TGC). The cut-off points used were those defined by the Clinical and Laboratory Standards Institute (CLSI)⁽¹⁰⁾ and the British Society for Antimicrobial Chemotherapy (BSAC)⁽¹¹⁾.

The PCR reaction was performed for detection of the following genes: bla_OXA23, bla_OXA24, bla_OXA51, bla_OXA58, bla_OXA143, bla_VIM-1, csuE, ompA and insertion element ISAba1, as shown in Table 1. The bacterial DNA extraction and preparation of the Master Mix were performed as recommended by Martins et al. (2014)⁽⁹⁾. Genetic diversity among the strains was evaluated by enterobacterial repetitive intergenic consensus (ERIC)-PCR, as proposed by Martins et al., and phenotyping assays of biofilm formation on polystyrene microplate, as described by Ferreira (2009)⁽¹²⁾.

For microtiter, plates of inert polystyrene were used, containing 96 wells (Nunclon, Nunc InterMed). Bacterial strains were analyzed for depletion seeded on trypticase soy agar and incubated for 24 h at 37°C. As a positive control, we used the reference strain A. baumannii ATCC 19606. As a negative control, tryptic soy broth (TSB) without bacterial rate was used. After this period, a heavy 24-h growth was picked into tubes containing 3 ml of TSB. Then 180 µl TSB were applied to each well in triplicate, plus 20 µl aliquots of bacterial strain in 0.5 McFarland scale. Subsequently, the plates were incubated for 4 h at 37°C to evaluate quick biofilm production in 24 h/37°C, 24 h/25°C and 48 h/25°C. After incubation, plate contents were aspirated; the wells were washed three times with 200 µl phosphate-buffered saline (PBS) (0.01 M, pH 7.2) and placed at room temperature to dry. Then, 30 µl 0.5% crystal violet solution was added to each well for 10 minutes. After staining the biofilm, the dye was aspirated and the wells were washed three times with 200 µl PBS (0.01M, pH 7.2). The plate was kept at room temperature to dry and the dye was eluted with 200 of absolute ethanol for 10 minutes; 150 µl of the solution contained in each of the plate wells were transferred to another clean, dry board. The optical density (OD) of the biofilm was measured by an enzyme-linked immunosorbent assay (Elisa) reader (BioRad, model 550) at a wavelength of 570 nm and strains were classified accordingly.

Stepanović et al. (2000)⁽¹⁸⁾ into four categories according to the average OD relates to the results obtained for the optical density of the negative control (Doc-). The categories were based on the following criteria: nonadherent (NA) ≤ when the DOC; weakly adherent (+) when the doc < DO ≤ 2× DoC; moderately adherent (++) when 2× DoC < DO ≤ 4× DoC or strongly adherent (+++) when 4× DoC < DO.

Data analysis

For a description of the variables contained in the results of molecular biology, phenotypic testing and resistance profile of the absolute and relative frequencies were used. To verify the correlation between the variables, polychoric and tetrachoric correlations were calculated^{(19, 20)}. The polychoric correlation is used to estimate the correlation from two ordinal variables, while the tetrachoric correlation is a special case of the former wherein the two dichotomous variables are ordinal. To check which correlations were significant, the linear association test⁽²¹⁾ was performed. In order to graphically display how the variables were related, a perceptual map was made via principal component analysis⁽²²⁾. The software used in the analysis was the R version 3.0.2.

RESULTS

Regarding the profile of the susceptibility E-test^® method, the majority of strains (98.4%) were resistant to MP, 71% to TZ and 61.3% to AB. In contrast, 98.4% of strains were sensitive to PO, 67.7% to GM and 48.4% to TGC. The search of MβL was positive in 95.2% of evaluated strains. All tested strains were positive for bla_OXA51 gene and 98.4% for bla_OXA23, 83.9% were positive for bla_VIM-1 and 90.3% for ISAba1. All strains were negative for bla_OXA24, bla_OXA58 and bla_OXA143 genes. Regarding the association of the studied genes with virulence factors, 43.5% of the strains showed the csuE and 53.2% the ompA gene. The frequency of the combination of genes studied in A. baumannii strains showed that bla_OXA23/bla_OXA51/ bla_VIM-1/ISAba1/csuE/ompA and bla_OXA23/bla_OXA51/bla_VIM-1/ISAba1 were the most frequent. Among the strains, 88.7% showed ISAba1 upstream of bla_OXA23 and bla_OXA51, concomitantly, and all of these strains were resistant to meropenem.

In those strains (9.7%) wherein ISAbal was not verified upstream of any of the two genes (bla_OXA23 or bla_OXA51), there was maintenance of resistance to carbapenems. However, the strains in which bla_OXA23 gene was not present, even in the presence of ISAbal and bla_OXA51, sensitivity profile to MP was observed. When evaluating biofilm formation, it was verified that 96.8% of strains incubated at 37°C showed biofilm formation after 4 h of incubation. Among these, 38.7% were poorly adherent, 45.2% moderately adherent, 12.9% strongly adherent and 3.2% were nonadherent under these conditions. The majority of strains (95.2%) grew at 37°C for 24 h, with 29%, 38.7% and 27.4% of strains considered weakly, moderately and strongly adherent, respectively, and 4.8% being nonadherent. In relation to incubated strains at room temperature, 96.8% formed biofilm when incubated for 24 h.

Out of these, 20.9% were weakly adherent, 38.7% moderately adherent, and 37.1% strongly adherent. A total of 3.2% of the tested strains were characterized as non-adherent. When incubated for 48 hours at room temperature, 1.62% were considered nonadherent, 98.4% of the strains showed biofilm formation (20.9% of those were weakly and moderately adherent, and 56.4% were characterized as strongly adherent). All strains negative to csuE gene (56.5% of total) produced biofilm. When csuE and ompA were absent (40.3%), concomitantly, biofilm production was also observed, as well as when csuE was absent and ompA was present.

Table 2 presents the polychoric and tetrachoric correlations for the study variables. The p-value of the linear association test is shown in Table 3.

In order to better visualize the relationship between the variables, a perceptual map was made via the principal component analysis (Figure 1). The amount of variance explained by the two main components was 69.4%, which indicates that it is possible to make a good interpretation of the relationship between variables by the two-dimensional graph of the principal component analysis.

Figure 1 − Perceptual map via the main components for the variables Dim: dimension; PR: profile of resistance; ampic: ampicillin; tigec: tigecycline; EF: environmental factors; ceftaz: ceftazidime; GN: gene normalization; gentam: gentamicine; OMP: outer membrane protein; metal: metallo-betalactamase; VIM1: Verona imipinemase.

A tendency was observed of the strains resistant to gentamicin not to carry the csuE gene associated with biofilm formation. On the other hand, gentamicin resistance was positively correlated to biofilm formation in 24 h/37°C, 24 h/48 h, 25°C/25°C. The presence of the ompA gene, also associated with biofilm formation, was negatively correlated to biofilm formation in 24 h/37°C, 24 h/48 h, 25°C/25°C. So, the ompA gene was loosely associated to biofilm formation during the incubation periods from 24 to 48 hours. The ompA gene was positively correlated with the biofilm formation variable in 4 h/37°C, but without statistical significance (p > 0.05). It was also noted that the csuE gene was positively correlated with ISAba1. Therefore, positive ISAba1 strains tended to have the csuE gene. Finally, TGC resistance was positively correlated to ampicillin resistance and, the strains which showed resistance to ampicillin also presented to TGC.

Figure 2 shows the result of sample clustering of the bands obtained by PCR profiles. For band share analysis, the bands were seen in a range between approximately 200 and 1600 base pairs (bp). The Dice⁽²³⁾ matrix of similarity coefficient generated by the Unweighted Pair Group Method with Arithmetic mean (UPGMA) dendrogram showed that they belonged to three distinct clades: A, B and C. Of the 62 amplified strains, 33 (53.2%) were combined in clade A, 21 (33.9%) in clade B and eight (12.9%) in clade C. In profile A, the fragments ranged from 250 to 1600 bp; in profile B, the variation of the molecular weight of the fragments was 200-1600 bp; and in profile C, 250 to 1500 bp.

Figure 2 − Dendrogram of genetic similarity derived from Dice similarity coefficient showing the relationship between the evaluated A. baumannii strains

DISCUSSION

In the present study were successfully employed PCR assays and ERIC-PCR with specific primers for the identification and characterization of the genetic profile of the 62 A. baumannii strains. Martins et al. (2014)⁽⁹⁾ reviewed the clonal spread of 64 multidrug-resistant (MDR) A. baumannii strains in Belo Horizonte hospitals also by ERIC-PCR with the same primers used in this study, suggesting that the same clone was present between the hospitals participating in the study. Furthermore, it was indicated that these strains can be spread by different routes, such as health professionals and the transfer of patients between hospitals and between sectors of the same institution. The agreement and the level of discrimination provided by ERIC-PCR, when compared to multilocus sequence typing (MLST) is relatively high, and validates the identification of clonal strains by ERIC-PCR¹⁴.

The prevalence of Acinetobacter resistant strains is high in countries such as Turkey and Greece, where carbapenems are widely used, with the lowest prevalence in countries such as the Netherlands and the Scandinavian countries, which adopt restrictive policies on the use of antimicrobials⁽²⁵⁾. Our data corroborate the literature, since 98.4% of our strains were resistant to MP. In the local epidemiological context, Neves et al. (2012)⁽²⁵⁾ identified 58.1% of imipenem resistance in the same hospital where our study was conducted. In the study by Fonseca et al. (2013)⁽²⁶⁾, as in the present study, they obtained high percentage of presence of bla_OXA23 and bla_OXA51 and ISAba1 genes, which are involved in resistance to carbapenems.

Overall, our results corroborate the data presented in the literature, given the global spread of bla_OXA23, and warn about the prevalence of this gene in Brazil. In this study, 100% of the analyzed strains were positive for the gene bla_OXA23, followed by bla_VIM-1 (83.9%) and ISAba1 (90.32%). No tested strain was positive for the bla_OXA24, bla_OXA58 and bla_OXA143 genes.

Martins et al. (2014)⁽⁹⁾ found similar results in Belo Horizonte, with positivity of 92.1% for bla_VIM-1 gene and 93.7% for bla_OXA23.

The elements also shown are implicated in resistance to antimicrobials. In the present study we found 90.3% of positivity for the presence of ISAba1, and this was upstream to bla_OXA51 and bla_OXA23 in 88.7% of strains. Interestingly, tin the only strain in which bla_OXA23 was absent, sensitivity to meropenem by E-test^® method was observed. Pagano et al. (2013)⁽²⁷⁾ investigated the potential contributing factor to antimicrobial resistance in 58 clinical strains of A. baumannii in Porto Alegre, Brazil. ISAba1 was found upstream of bla_OXA23-like gene only in strains resistant to carbapenems and was present upstream of bla_OXA51-like both in susceptible and resistant strains. These authors also suggest that _ISAbat/bla_OXA51 alone is insufficient to confer resistance to carbapenems. Our data corroborate this statement, since it also identified the carbapenem susceptibility profile in the strains that showed _ISAbat/bla_OXA51.

Azizi et al. (2015)⁽²⁸⁾ investigated the association between the presence of oxacillinases and biofilm formation in polystyrene microtiter plate by A. baumannii strains and found it to be positive for bla_OXA51 and bla_OXA23, concomitantly, but negative for bl_OXA24/4D, formed biofilms loosely or non-adherent. In our study, the presence of the csuE gene was positively correlated with the presence of ISAba1 and negatively to the phenotypic expression of gentamicin resistance (p < 0.05). However, resistance to gentamicin was positively correlated to biofilm formation variables for 24 h/37°C, 24 h/48 h 25°C/25°C (p < 0.05). Thus, it is clear, from these statistical analyses, that resistance to gentamicin strains is associated with the formation of biofilms 24 to 48 hours, both at 25°C and at 37°C, suggesting the expression of other genes associated with formation of biofilm, different from csuE. The csuE gene was also found to be positively correlated with ISAba1. Therefore, positive strains ISAba1 tend to have the csuE gene (p < 0.05).

However, resistance to gentamicin was positively correlated to biofilm formation variables for 24 h/37°C, 24 h/48 h 25°C/25°C (p < 0.05). Despite the many variables that affect the expression of csuE and ISAba1, we can infer that carriers in the strains ISAba1 can be more virulent and resistant to certain antibiotics such as ampicillin, for associating positively with the presence of the csuE gene. All negative strains for the presence of gene csuE in our research showed biofilm formation with some degree of adhesion.

These results suggest that the absence of csuE and ompA are not decisive for the interruption of biofilm formation on polystyrene plates, and the mechanisms developed by these strains still need to be elucidated. Findings from de Breij et al. (2009)⁽²⁹⁾ corroborate our data, once the pili formation dependent of csuA/ BABCDE was not involved in adhesion of A. baumannii ATCC 19606 epithelial cells in the respiratory tract. In our study, we did not investigate the presence of bfmS, Bap, quorum sensing or bla_PER-1. Hence, strains whose csuE gene was absent and yet proved to be good producers of biofilm in phenotypic assays on polystyrene plate can be regulated by other genes and mechanisms not evaluated in this study.

Our study has several notable strengths. It was the first study in Brazil to associate the presence of virulence and resistance genes of A. baumannii, and a well conducted molecular analysis, tracing a strong epidemiological profile of A. baumannii in our region.

Some limitations are worth noting. Resistance in A. baumannii might be expressed for different mechanisms such as efflux pump, alterations in outer membrane proteins, or penicillin-binding protein modification. We did not conduct molecular analysis at this level to determine accurately which other mechanisms could be involved in the expression of resistance.

CONCLUSION

This study showed that the presence of ISAba1 upstream of bla_OXA51 seems to be crucial to carbapenem resistance phenotype. It was identified that the absence of csuE and ompA did not alter the biofilm formation capacity on polystyrene microplate for analyzed strains, demonstrating that other mechanisms are responsible for the regulation of biofilm formation in A. baumannii. The presence of the csuE gene was associated with the presence of ISAba1 and lack of resistance to gentamicin in tested strains (p < 0.05), and resistance to gentamicin was significant in the strains forming biofilm on polystyrene plates at 24 h/37°C, 24 h/48h 25°C/25°C (p < 0.05). Due to the clinical importance of A. baumannii and the emergence of extensively drug-resistant (XDR) clones, further studies on virulence, pathogenesis and acquisition of antimicrobial resistance by this microorganism are required. These studies will provide valuable information on the basic biology of the human pathogen, which can be used as a target for the development of new treatments and prevention strategies.

ACKNOWLEDGEMENTS

This work was supported by Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Fundação de Amparo à Pesquisa do Estado de Minas Gerais (Fapemig), Fundação de Amparo à Pesquisa e Desenvolvimento Científico do Maranhão (Fapema) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (Capes).

DISCLOSURE STATEMENT

There are no conflicts of interest.

REFERENCES

1. Cerqueira GM, Peleg AY. Insights into Acinetobacter baumannii pathogenicity. IUBMB Life. 2011; 63: 1055-60

2. Gaddy JA, Tomaras AP, Actis LA. The Acinetobacter baumannii 19606 OmpA protein plays a role in biofilm formation on abiotic surfaces and in the interaction of this pathogen with eukaryotic cells. Infect Immun. 2009; 77: 3150-60.

3. Tomaras AP, Dorsey CW, Edelmann RE, Actis LA. Attachment to and biofilm formation on abiotic surfaces by Acinetobacter baumannii: involvement of a novel chaperone-usher pili assembly system. Microbiology. 2003; 149: 3473-84.

4. Yong D, Toleman MA, Bell J, et al. Genetic and biochemical characterization of an acquired subgroup B3 metallo-β-lactamase gene, blaAIM-1, and its unique genetic context in Pseudomonas aeruginosa from Australia. Antimicrob Agents Chemother. 2012; 56: 6154-59.

5. Vala MH, Hallajzadeh M, Hashemi A, et al. Detection of Ambler class A, B and D β-lactamases among Pseudomonas aeruginosa and Acinetobacter baumannii clinical isolates from burn patients. Ann Burns Fire Disasters. 2014; 37: 8-13.

6. Bush K, Jacoby GA. Updated functional classification of β-lactamases. Antimicrob Agents Chemother. 2010; 54; 969-1076.

7. Jacobs AC, Sayood K, Olmsted SB, Blanchard CE, Hinrichs S, Russell D, Dunman PM. Characterization of the Acinetobacter baumannii growth phase-dependent and serum responsive transcriptomes. FEMS Immunol Med Microbiol. 2012; 64:403-412.

8. Pogue JM, Mann T, Barber KE, Kaye KS. Carbapenem-resistant Acinetobacter baumannii: epidemiology, surveillance and management. Expert Rev Anti Infect Ther. 2013; 11: 383-93.

9. Martins HS, Bomfim MR, França RO, et al. Resistance markers and genetic diversity in Acinetobacter baumannii strains recovered from nosocomial bloodstream infections. Int J Environ Res Public Health. 2014; 11: 1465-78.

10. Clinical and Laboratory Standards Institute (CLSI). M100-S23 — Performance standards for antimicrobial susceptibility testing; twentythird informational supplement; 2014.

11. British Society for Antimicrobial Chemotherapy (BSAC). Methods for antimicrobial susceptibility testing; 2014.

12. Ferreira JAB. Diversidade genética, perfil de resistência aos antimicrobianos e produção de biofilme de amostras de Pseudomonas aeruginosa isoladas da água utilizada em unidades de terapia renal substitutiva [thesis]. Rio de Janeiro: Fundação Oswaldo Cruz; 2009.

13. Yan JJ, Hsueh PR, Ko WC, et al. Metallo-beta-lactamases in clinical Pseudomonas isolates in Taiwan and identification of VIM-3, a novel variant of the VIM-2 enzyme. Antimicrob Agents Chemother. 2001; 45: 2224-28.

14. Woodford N, Fagan EJ, Ellington MJ. Multiplex PCR for rapid detection of genes encoding CTX-M extended-spectrum (beta)-lactamases. J Antimicrob Chemother. 2006; 57: 154-55.

15. Higgins PG, Lehmann M, Seifert H. Inclusion of OXA-143 primers in a multiplex polymerase chain reaction (PCR) for genes encoding prevalent OXA carbapenemases in Acinetobacter spp. Int J Antimicrob Agents. 2010; 35(3): 305-14.

16. Turton JF, Ward ME, Woodford N, et al. The role of ISAba1 in expression of OXA carbapenemase genes in Acinetobacter baumannii. FEMS Microbiol Lett. 2006; 258: 72-7.

17. Turton JF, Gabriel SN, Valderrey C, Kaufamann ME, Pitt TL. Use of sequence-based typing and multiplex PCR to identify clonal lineages of outbreak strains of Acinetobacter baumannii. Clin Microbiol Infect. 2007; 13: 807-15.

18. Stepanovic S, Vukovic D, Dakic I, Savic B, Svabic-Vlahovic M. A modified microtiter-plate test for quantification of staphylococcal biofilm formation. J Microbiol Methods. 2000; 40: 175-9.

19. Lee SY, Poon WY, Bentler PM. A two-stage estimation of structural equation models with continuous and polytomous variables. Br J Math Stat Psychol. 1995; 48: 339-58.

20. Bonett DG, Price RM. Inferential methods for the tetrachoric correlation coefficient. J Educ Behav Stat. 2005; 30: 213-25.

21. Agresti A. Categorical data analysis. 2 ed. John Wiley & Sons; 2002. p. i-xv.

22. Kidd TJ, Grimwood K, Ramsay KA, Rainey PB, Bell SC. Comparison of three molecular techniques for typing Pseudomonas aeruginosa isolates in sputum samples from patients with cystic fibrosis. J Clin Microbiol. 2011; 19: 263-8.

23. Dice LR. Measures of the amount of ecologic association between species. Ecology. 1945; 26(3): 297-302.

24. Adriaenssens N, Coenen S, Versporten A, et al. European surveillance of antimicrobial consumption (ESAC): outpatient antibiotic use in Europe (1997-2009). J Antimicrob Chemother. 2011; 66: vi3-vi12.

25. Neves FAC. Características clínico-epidemiológicas e microbiológicas relacionadas a cepas de Acinetobacter baumannii resistentes ao imipenem isoladas de pacientes internados em Unidade de Terapia Intensiva do Hospital das Clínicas da UFMG. Universidade Federal de Minas Gerais. 2012; 12: 1-122.

26. Fonseca EL, Scheidegger E, Freitas FS, Cipriano R, Vicente AC. Carbapenem-resistant Acinetobacter baumannii from Brazil: role of carO alleles expression and blaOXA-23 gene. BMC Microbiol. 2013; 13: 1-7.

27. Pagano M, Martins AF, Machado AB, Barin J, Barth AL. Carbapenemsusceptible Acinetobacter baumannii carrying the ISAba1 upstream blaOXA-51-like gene in Porto Alegre, southern Brazil. Epidemiol Infect. 2013; 41: 330-3.

28. Azizi O, Shakibaie MR, Modarresi F, Shahcheraghi F. Molecular detection of class-D OXA carbapenemase genes in biofilm and non-biofilm forming clinical isolates of Acinetobacter baumannii. Jundishapur J Microbiol. 2015; 8: e21042.

29. de Breij A, Gaddy J, van der Meer J, et al. CsuA/BABCDEdependent pili are not involved in the adherence of Acinetobacter baumannii ATCC19606(T) to human airway epithelial cells and their inflammatory response. Res Microbiol. 2009; 160: 213-8.