Virulence profiles and innate immune responses against highly lethal, multidrug-resistant nosocomial isolates of Acinetobacter baumannii from a tertiary care hospital in Mexico

Virulence profiles and innate immune responses were studied in Acinetobacter baumannii from nosocomial infections collected over one year in a tertiary care hospital in Mexico. A. baumannii were identified by VITEK 2 System followed by susceptibility tests. Carbapenemase genes, active efflux mechanism to imipenem and meropenem and outer membrane proteins profile were analyzed to evaluate their role on the activity of carbapenem resistance. All isolates were genotyped by pulsed field gel electrophoresis. The ability to form biofilm was determined on a polystyrene surface. The resistance to complement was determined with a pooled human normal serum and TNFα release by infected macrophages was determined by ELISA. The 112 isolates from this study were associated with a 52% of mortality. All were resistance to β-lactams, fluoroquinolones, and trimethroprim-sulfamethoxal, 96 and 90% were resistant to meropenem and imipenem, respectively, but with high susceptibility to polymyxin B, colistin and tigecyclin. Isolates were classified in 11 different clones. Most isolates, 88% (99/112), were metallo-β-lactamases and carbapenemases producers, associated in 95% with the presence of blaOXA-72 gene. Only 4/99 and 1/99 of the carbapenem-resistant isolates were related to efflux mechanism to meropenem or imipenem resistance, respectively. The loss of expression of 22, 29, and/or 33-36-kDa proteins was detected in 8/11 of the clinical isolates with resistance to carbapenem. More than 96% (108/112) of the isolates were high producers of biofilms on biotic surfaces. Finally, all isolates showed variable resistance to normal human serum activity and were high inductors of TNFα release by macrophages. In summary, these results suggest that multidrug-resistant A. baumannii can persist in the hospital environment through its ability to form biofilms. The high mortality observed was due to their ability to survive normal human serum activity and capability to induce potent inflammatory immune response making this nosocomial pathogen a serious threat to hospitalized patients.


Introduction
Over the last few years, infections involving species of Gram-negative non-enteric and multidrug-resistant bacteria have been increasing worldwide [1]. Among these bacteria, Acinetobacter baumannii has emerged as the most important opportunistic pathogen involved in serious hospital infection outbreaks [2,3]. These bacteria have been isolated from different environmental sources, including soil, water, food products, and medical devices as well from the skin of hospital staff. [4][5][6][7][8][9][10][11][12]. A. baumannii is a bacterium that has acquired diverse mechanisms of resistance to several antibiotic families, which has led to the emergence of important multidrug or pandrug resistant phenotypes [1]. Until recently, the carbapenems were the first antibiotics of choice in the treatment of nosocomial infections, including A. baumannii [13]. The emergence of resistance by A baumannii to carbapenems has been reported from different parts of the world [14]. The carbapenem resistance by A. baumannii is due to: a) the production of class B carbapenemases also known as metallo-β-lactamases (MBLs) and class D enzymes (also known as oxacillinases), b) genetic alterations of penicillin binding proteins, c) overexpression of efflux pumps belonging to the resistance-nodulation-cell division (RND) family and d) the loss of outer membrane proteins associated with the formation of pores [15]. The increased multidrug resistance and the persistence for long periods of time in hospitals make A. baumannii a serious threat to hospitalized patients. The capability of A. baumannii to form biofilm on biotic or abiotic surfaces partially explains the persistence of these bacteria in the hospital environment [16,17].
To date, few virulence mechanisms have been described in A. baumannii [18]. The lipopolysaccharide (LPS) of Acinetobacter is a potent inducer of the inflammatory immune response through the stimulation of tumor necrosis factor alpha (TNFα) release by phagocytic cells via Toll-like receptor 4 (TLR4) signaling [19]. In addition, A. baumannii has the ability to resist the killing action of normal human serum (NHS) [20]. This serum resistance is attributed in part to LPS expression [20].
In the present study, we analyzed clinical nosocomial isolates of A. baumannii from a tertiary care hospital in Mexico City by their antimicrobial susceptibility profile, clonal relationships, ability to form biofilms on abiotic surfaces, by their capability to resist to human normal serum activity as well as their ability to induce TNFα release by phagocytic cells. Our results attempt to clarify the participation of antimicrobial resistance, virulence profiles and the innate immune response on the prevalence, endemicity and lethality of A. baumannii in critically ill patient in the hospital environment.

Clinical isolates
The clinical isolates of A. baumannii used in this study were from patients with nosocomial infections during the study period, from January to December of 2014, at Hospital General de México Dr. Eduardo Liceaga, Mexico City, Mexico. The isolates were identified by the VITEK 2 System (bioMerieux, Marcy l'Etoile, France), confirmed by API20NE (bioMeriux1SA) [21] and by detecting the intrinsic carbapenemase bla OXA-51-like gene. Only one isolate per patient episode was analyzed. All samples were collected as part of routine care, identified and handled anonymously. Nosocomial infections were defined according to the Centers for Disease Control and Prevention (CDC) criteria, and by Infectious Diseases Unit physicians [22].

Antibiotic susceptibility testing
The isolate strains were tested for their susceptibilities to 17 antibiotics: gentamicin, tobramycin, cefepime, ceftriaxone, ampicillin/sulbactam, imipenem, meropenem, ciprofloxacin, levofloxacin, minocycline, tetracycline, trimethoprim/sulfamethoxazole, colistin, polymyxin B, and tigecycline. The minimum inhibitory concentrations (MICs) was determined using the VITEK 2 System and were confirmed using an agar dilution method, according to Clinical Laboratory Standards Institute (CLSI) guidelines [23,24]. Pseudomonas aeruginosa ATCC 27853 and E. coli ATCC 25922 were used as positive and negative controls, respectively. In carbapenem-resistant isolates, MBLs were determined by EDTA disc synergy tests using meropenem and meropenem plus EDTA and the Hodge modified method [25].
Detection of bla OXA-51-like , bla IMP , bla VIM , bla OXA carbapenemase and ISAba-1 genes Chromosomal DNA was extracted from the isolates; 2-μL volume of this extract was used for PCR reactions. Amplification of bla OXA-51-like , bla IMP , bla VIM and bla OXA genes was performed using primers and conditions of amplification as previously reported [15,26,27]. The presence of the ISAba1 promoter sequence and its association with carbapenemase genes was investigated by PCR using ISAba-1 [28], ISAba1 + bla OXA-51-like and ISAba1 + bla OXA primers [15]. Amplified products were subjected to nucleotide sequencing at the Instituto de Biotecnología, Universidad Nacional Autónoma de México.

Pulsed-field gel electrophoresis (PFGE)
To determine the spread of A. baumannii within the hospital setting, all isolates were genotyped using Pulsed-Field Gel Electrophoresis (PFGE). Bacterial genomic DNAs were prepared as previously described [15], digested with ApaI (New England Biolabs, Beverly, MA) and subjected to PFGE analysis using a Gene Path system (BioRad1). Tenover criteria and Dice coefficient were used to determine similar profiles between isolates, a correlation > 85% were considered to have the same pulsotype (clones) [29].

Detection of the efflux pumps phenotype
In order to determine the role of the efflux mechanisms, all carbapenem-intermediate or-resistant isolates were selected to identify the inhibitory effect of the efflux inhibitor carbonil cyanide 3-chlorophenylhydrazone (CCCP) (Sigma, St. Louis, MO) on the change of meropenem or imipenem susceptibility had on the isolates. Briefly, Müeller-Hinton agar plates with meropenem or imipenem double serial dilution in the presence or absence of 25, 50, and 100 mg/L CCCP inhibitor were used [30]. Positive contribution of an efflux pump was defined as a fourfold decrease in the MIC of antibiotics when efflux pump inhibitor was added to the agar plates [30].

Outer membrane proteins (OMPs) profile
The A. baumannii OMPs expression from one representative isolate of each carbapenem-resistant clone identified in this study were used to support the contribution of the porin profile modification in the carbapenem-resistance. As a control we used a carbapenem-susceptible A. baumannii-176 clinical isolate [15]. The OMPs were prepared as previously described [31].
Twenty μg of extracted proteins of each sample were resolved in one-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Silver staining (Silver Stain Plus BioRad1) was used to visualize the protein bands.

Biofilm production
The biofilm production on an abiotic surface was quantified as previously described [32]. Briefly, 5 ml overnight cultures grown at 37°C were diluted to OD 600 0.003 in Luria Bertani (LB) media and triplicate 500 μl aliquots were dispensed into polystyrene tubes. Following 24 h of static incubation at 37°C, the medium was removed and the tubes were washed gently once with deionized water. Adherent bacteria were stained with 1% (w/v) crystal violet and washed three times with deionized water. The bound crystal violet was dissolved in 1 ml of 100% methanol and quantified by measuring OD 540 nm.

Serum resistance assay
The assay was conducted as previously reported [33] with few modifications. Briefly, A. baumannii isolates were grown in Müller Hinton (MH) broth to the mid-log phase. We adjusted the inoculum to 1x10 7 bacteria with 40% [in phosphate buffered saline (PBS)] either pooled normal human serum (NHS) (Cedarlane Laboratories Limited, Ontario, Canada) or with heat-inactivated human serum (HIS), the samples were incubated for 3 h at 37°C without shaking. After the incubation, the number of surviving bacteria in each sample was determined by a serial dilution and plating in MH agar at 37°C for 18 h. The serum bactericidal effect was calculated with the following formula: (CFUs-NHS / CFUs-HIS) Ã 100. NHS and HIS were assessed using Salmonella typhi 9:12:Vid as a sensitive strains to determine the complement pathway activated by these bacteria [34]. All experiments were performed in triplicate and results were expressed as percent of survival.
TNFα release assay RAW 264.7 (ATCC TIB-71) macrophages were grown in 24-well plates at 2.5x10 5 cells per well and incubated during 24 h. Overnight bacterial cultures were grown in MH broth at 37˚C with shaking. Bacterial cultures were washed twice with RPMI plus 10% fetal bovine serum (FBS) and used to infect macrophages monolayers at an MOI of 100 as described [35]. Briefly, to synchronize the infection, plates were centrifuged for 1 min at 1400 rpm and incubated for 3 h at 37˚C under 5% carbon dioxide. After infection, the supernatants were removed and centrifuged at 14000 rpm to remove bacteria. The supernatants were evaluated for the TNFαrelease by ELISA (R&D systems) as described [36].

Statistical analysis
The data represents the mean of the standard deviation (SD) and were analyzed by one-way ANOVA followed by a post hoc Tukey's comparison. In some cases, data with normal distribution were analyzed using Student t test. The p values, 0.05 was considered significant. The data were analyzed using GraphPad Prism 6 software.

Antibiotic susceptibility
All A. baumannii isolates were confirmed by amplification of bla OXA-51-like gene. The isolates were 100% resistant to β-lactams, fluoroquinolones and trimethoprim-sulfamethoxazole, 96.4% with resistance to meropenem and 89.2% to imipenem ( Table 1). All isolates were susceptible to colistin and tigeciclyne and only the 2.7% of them were resistant to polymyxin B (Table 1). MBL activity was found in 88.4% (99/112) carbapenem-resistant isolates ( Table 2). By PCR and sequencing we detected in 95% (94/99) of the isolates the bla OXA-72 gene (Gen-Bank accession number: JX968505.1). None of the isolates were carrying either bla VIM or bla IMP genes ( Table 2). Insertion sequences ISAba-1 were detected in all of the isolates, however, bla OXA-51-like or bla OXA-72 genes were not adjacent to the ISAba-1.

Genotyping of A. baumannii isolates
PFGE fingerprinting analysis revealed 11 clones in the 112 A. baumannii clinical isolates studied that were classified by assigning them letters from A to K (Fig 1). Nine clones were endemic in the Pulmonary Medicine Ward (Fig 2A). Clone B was the most frequently detected in 30% (34/112) of the clinical isolates ( Fig 2B) and was responsible for an outbreak between February and March (Fig 2A and 2B). Clone B was also detected 7 times in the Pulmonary Medicine Ward and 4 times in the Intensive Care Unit (Fig 2A). Clone C was the second most frequently identified with 20.5% (23/112) of the isolates. This clone was detected between February-July and was also associated with the outbreak of February-March ( Fig 2B). This clone was identified 7 times in the Pulmonary Medicine Ward and 5 times in the infectious diseases unit (Fig 2A). The third most frequent clone was the clone F with 15 isolates 13%. Members of this clone were identified at the end of the study ( Fig 2B).

Efflux pump phenotype
Results indicated that in the minority of the carbapenem-intermediate or -resistant isolates, efflux pump expression was not related to the carbapenem resistant phenotype (Table 2). Only 4/112 isolates exhibited a four-fold or greater decrease in MICs to meropenem. These isolates belonged to three different clones, two were from clone I, one from clone D and one from clone H. Only one isolate (0.89%) from clone F exhibited efflux pump activity to imipenem ( Table 2). None of the isolates resistant to imipenem and meropenem exhibited an efflux pump to both carbapenems.

Serum resistance assay
Evidence from the literature indicates that clinical isolates of A. baumannii show a differential capacity to survive NHS activity [33,37]. In this study, we quantified the capacity of all clinical   Virulence profiles and innate immune responses against A. baumannii TNFα release. The results presented in the Fig 5 show that all A. baumannii isolates were able to induce more than 1500 pg/ml of TNFα and only 9.8% (11/112) isolates were higher inducers of TNF-α release (more than 6000 pg/ml). These results indicate that A. baumannii can be a potent inductor or inflammatory immune response.

Discussion
During the last two decades A. baumannii has emerged as an important nosocomial pathogen, especially in hospitalized patients requiring intensive care [38]. One of the main factors that characterize clinical A. baumannii isolates is the abundance of plasmids, transposons and  frequent genetic exchanges that confer multidrug-resistance, limiting therapeutic options which often leads to high mortality rates in hospitalized patients [1,39]. This work was carried out in one of the major tertiary hospitals in Mexico City. The hospital has 1,245 beds, with an average of 45,000 discharges per year. During 2014, 112 clinical A. baumannii isolates from nosocomial infections were collected. The isolates exhibited high resistance to multiple antibiotics including 88% to carbapenems. These results are very similar to those we reported previously in a tertiary hospital in western Mexico [15] as well as reports from others countries [40][41][42].
Carbapenem resistance in the isolates studied was associated mainly with MBLs 88.4% and bla OXA-72 (83.9%) production. Since 2012, bla OXA-72 has been reported in Asia and currently in different parts of the world, including western Mexico [15]. Although all isolates carried blaOXA-51, and a high proportion presented bla OXA-72 , the high expression of these two genes was not colinear with the ISAba1, similar to what has been reported previously [43]. In this study, bla VIM-1 or bla IMP genes were not detected in MBLs producing A. baumannii isolates, indicating that these isolates carried other type of MBLs, not tested.
Regarding other mechanisms of carbapenem resistance, we detected a lesser proportion of efflux pump expression associated with meropenem (3.57%) or imipenem (0.89%) resistance; these results differ from previous studies which showed drug export by efflux pump reduced meropenem susceptibility among the vast majority 65.7% of nosocomial A. baumannii isolates [30]. We noted a decrease in the number of isolates expressing this mechanism when compared with the nosocomial A. baumannii isolates studied in previous work [15]. Regarding the analysis of OPMs profiles, we observed that a high number of isolates 72.1% (8/11) Virulence profiles and innate immune responses against A. baumannii representing each of the carbapenem-resistant clones showed the absence of 1, 2 or 3 porins. This suggests that this mechanism contributes significantly to the resistance to carbapenems in A. baumannii isolates causing nosocomial infections in this hospital in central Mexico.
In this study, patients with A. baumannii infection showed higher mortality rate (52.8%) than those with A. baumannii infection into the Hospital Civil de Guadalajara, Mexico (28.2%) reported previously [32]. In other countries, the mortality rate due to bloodstream imipenemresistant A. baumannii infections is ranges from 52.2 to 86.7% [39,44]. We demonstrated that the spread of clones B and C were responsible for an outbreak during February and March 2014 and these clones persisted during the first six months of that year. Epidemiological analysis showed that both clones were the most frequent and together represented half of the isolates 50.8% (57/112) identified in this study. The vast majority of these clones 94.7% produced MBLs, which explains why at least half of A. baumannii infections producing MBLs are due to outbreaks. Interestingly, when we compared eleven PFGE profiles of the clones detected in this study with the PFGE profiles of the clones of A. baumannii recently reported in a previous study in hospital from western Mexico [15] we found that clone H, the fourth most frequent clone with 10 isolates, was identical to clone 22, the most frequent clone previously described [15]. These results demonstrate that this multidrug-resistant clone has spread at least between these two distant geographically area of Mexico. To better understand how this clone emerged or where it has spread, we compared the genotype of clone H with those obtained by enzymatic restriction enzymatic with ApaI and PFGE reported for clones disseminated internationally. We used the genotypes of European clones I-III representative strains of outbreaks from United Kingdom, Spain, the Netherlands, France, and Israel previously reported [45], and found no similarities between these and clone H. Our results show that genotyping is a Virulence profiles and innate immune responses against A. baumannii methodology that allows us to monitor the geographic spread of epidemic pathogens, and understand their epidemiology.
The persistence of A. baumannii in the hospital environment is associated with their ability to form biofilms on diverse biomedically relevant surfaces [46]. A study developed by Rodríguez-Baño et al. [47] showed that the 63% of unrelated clinical isolates were able to produce biofilm on abiotic surfaces. In contrast, 92.2% of the members of a highly prevalent clone of A. baumannii in the Hospital Civil de Guadalajara, Mexico were able to produce biofilm [32]. Our results show that almost all unrelated isolates were biofilm producers. The high percentage of biofilm formation by unrelated clinical isolates could be due by the presence of one or more of the mechanism associated to biofilm formation including the presence of bla PER-1 gene [48,49] or to the presence of genes that encode for the pili assembly system (csuC, csuD and csuE) [50], the autoinducer synthase gene (abaI) [51] or the OmpA expression [52].
Based on our results, the high prevalence of nosocomial isolates of A. baumannii with high mortality rate observed in this study was due to multidrug-resistant phenotypes in conjunction of their ability to form biofilm on abiotic surfaces, their high resistance to normal human serum and potent capacity to induce macrophage TNFα release.