Comparison of the Virulence-Associated Phenotypes of Five Species of Acinetobacter baumannii Complex

Abstract

In this study, we compared the virulence-associated factors of Acinetobacter baumannii complex species. Sixty-three isolates of five A. baumannii complex species, including 19 A. baumannii, 15 A. nosocomialis, 13 A. seifertii, 13 A. pittii, and 3 A. calcoaceticus isolates, were included in this study. For all isolates, biofilm formation, A549 cell adherence, resistance to normal human serum, and motility were evaluated. A. baumannii complex isolates showed diversity in biofilm formation, A549 cell adherence, and serum resistance, and no strong positive relationships among these virulence characteristics. However, A. seifertii showed relatively consistent virulence-associated phenotypes. In addition, A. baumannii clone ST110 exhibited consistently high virulence-associated phenotypes. Motility was observed in seven isolates, and all four A. baumannii ST110 isolates showed twitching motility. Although some inconsistencies in virulence-associated phenotypes were seen, high virulence characteristics were observed in A. seifertii, which has been mainly reported in Korea and shows high rates of colistin resistance.

Introduction

Species of the genus Acinetobacter are aerobic, non-motile, gram-negative coccobacillary rods. Acinetobacter spp. are non-fermentative and can survive under a wide range of environmental conditions for extended periods of time. The genus currently includes 39 validly named species (http://www.bacterio.net/). The A. calcoaceticus-A. baumannii complex or A. baumannii complex consists of several species, including A. baumannii, A. calcoaceticus, A. nosocomialis, A. pittii, and A. seifertii (formerly Acinetobacter genomic species “close to 13TU”) [6,28]. The species in this complex cannot be easily distinguished by conventional phenotypic identification methods; however, recent studies using molecular identification techniques have clearly differentiated them as distinct species [6]. All these species, except A. calcoaceticus, are clinically important nosocomial pathogens causing human disease, and even A. calcoaceticus has recently been isolated in nosocomial settings [29]. In particular, A. seifertii, which has recently been validly named [19], has been frequently identified in Korea [22]. In particular, A. seifertii may be intrinsically resistant to colistin, which is considered as a last-line therapy for Acinetobacter infections [8]. Limited information is available on the virulence-associated factors of Acinetobacter. Whereas the virulence features of A. baumannii have been reported in some studies, only a few papers have investigated the virulence of non-A. baumannii species [5,23].

Biofilm formation is thought to be associated with increased bacterial survival; thus, it may play a significant role in the pathogenicity of A. baumannii [17]. Bacterial biofilms are highly organized communities that are formed through a series of coordinated steps. Bacteria in biofilms are more resistant to antimicrobial agents than planktonic forms, and the biofilm allows them to persist on biotic and abiotic surfaces and helps them to evade the host response [11]. The ability of bacteria to adhere to eukaryotic cells is critical for pathogenesis. Although biofilm formation and adherence to cells are common features in clinical isolates of A. baumannii [14,24], few studies have explored the biofilm-forming activity and eukaryotic cell adherence of other species in the A. baumannii complex.

The complement system is central to the innate defense against bacterial infection [10]. Pathogenic bacteria are capable of evading the host immune system, and they can avoid the bactericidal activity of serum. It has been reported that the serum resistance of A. baumannii is significantly related to patient mortality and could be used to predict the outcome of patients infected with multidrug-resistant A. baumannii isolates [14]. Therefore, survival in normal human serum (NHS) could be a marker of Acinetobacter virulence.

Motility is another potential factor that contributes to virulence. Although Acinetobacter species have been described as non-motile, more than a few studies have demonstrated motility in A. baumannii isolates [7]. It is interesting that A. baumannii shows differences in motility in response to illumination and iron limitation [7,18]. It is unclear whether motility plays a significant role in the virulence of Acinetobacter because it has been rarely reported.

In this study, we examined the virulence characteristics, such as biofilm formation, binding to eukaryotic cells, resistance to NHS, and motility, of 63 Acinetobacter isolates from Korea, to investigate the virulence traits of five species belonging to the A. baumannii complex.

 

Materials and Methods

Bacterial Isolates

A total of 63 Acinetobacter species isolates were included in this study (Table 1). They were isolated from patients with bacteremia in two tertiary care hospitals in Korea from between August 2003 and February 2010. We selected them randomly from more than 450 blood isolates. Acinetobacter species identifications using the 16S rRNA gene and partial rpoB sequences were performed in our previous study [22] (Fig. 1). The isolates included the following species (no. of isolates): A. baumannii (19), A. nosocomialis (15), A. seifertii (13), A. pittii (13), and A. calcoaceticus (3). Multilocus sequence typing (MLST) analysis was used to determine the genotypes of the A. baumannii isolates, as previously described (http://www.pubmlst.org/abaumanii), with some modifications [21]. Growth curves were generated by diluting equal numbers of CFUs of three isolates of each species (approximately 5 × 106 CFU/ml) in LB broth and incubating at 37°C with constant shaking (180r pm). The OD600 of each culture was then measured at 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 12, and 24 h post-dilution. Each curve was performed in duplicates.

Table 1.aST, sequence type; NA, not available. bSMC, Samsung Medical Center; CNUH, Chonnam National University Hospital.

Fig. 1.A phylogenetic tree reconstructed on the basis of nucleotide sequences of the rpoB gene (GenBank accession numbers, KP938444 to KP938506).

Biofilm Assay

Biofilm formation was assayed as described previously [25], with minor modifications. A single colony was inoculated into Luria-Bertani (LB) broth containing 0.25% glucose and incubated overnight at 3 7℃. A liquots (200μl) were subsequently diluted 1:100 with fresh LB broth in 96-well U-bottom polystyrene plates and incubated for 18 h at 37℃. A well containing sterile LB without bacteria served as the negative control. Adherent cells were washed once with phosphate-buffered saline (PBS), dried for 30mi n, and stained with 0.1% crystal violet for 15 min. They were solubilized by adding 200 μl of 95% ethanol. The absorbance at 600 nm (A600) of the extracted crystal violet was then measured with a microplate reader to determine biofilm formation relative to the control. We defined the biofilm-forming strain as a strain showing higher OD value than that of A. baumannii ATCC 19606. Experiments were repeated with three independent cultures, each tested in duplicate.

Eukaryotic Cell Adherence Assay

The adherence of Acinetobacter isolates to A549 cells was determined as described previously [7]. Here, A549 human type 2 pneumocytes were utilized because they have been previously used to mimic adherence to the human lung and as a model for pneumonia caused by A. baumannii [16]. Cells were grown in Dulbecco’s modified Eagle’s medium (Hyclone, UT, USA) supplemented with 10% fetal bovine serum. Washed A549 monolayers grown in 24-well tissue culture plates were subsequently infected with bacterial inoculums containing 1 × 107 CFU. The average number of A549 cells at the time of infection was 1.6 × 105, and the multiplicity of infection in our experiment was 93.75. After incubation at 37℃ and in 5% CO2 (v/v) for 1 h, the culture medium was removed and the monolayers were washed three times with 1 ml of PBS. Cell monolayers were detached from the plate by treatment with 1 ml of sterile DW. Eukaryotic cells were plated on LB agar to determine the number (CFU) of adherent bacteria per well. The adherence assays were repeated three times independently.

Serum Bactericidal Assay

Susceptibility to NHS (Innovative Research, MI, USA) was determined using the method previously described [12], with modifications. Bacterial cultures were grown to mid-log phase (OD600 = 0.4~0.6), and then the cells were suspended in PBS to a final density of 1 × 107 CFU/ml. PBS (200 μl) containing 20% NHS was prepared, and then 10 μl of the cell suspension (1 × 105 cells) was added. We used 20% NHS because most bacterial cells were killed under 40% NHS and too many cells survived under <10% NHS in preliminary tests using several strains. The mixtures were incubated with shaking at 37℃ for 3 h. The number of surviving bacteria in each mixture was determined by plating serial dilutions on LB agar plates that were incubated at 37℃ overnight. Heat-inactivated human serum (HIS) was used as a control for determining the bactericidal effect of NHS. Serum susceptibilities were characterized by calculating the survival rate of each strain by plotting the average survival percentage, as number of CFUs survived in media containing NHS versus number of CFUs survived in media containing HIS. The experiments were repeated three times independently.

Motility Assay

Motility was investigated using LB media containing different concentrations of agar (0.3% for swimming, 0.5% for swarming, and 1.0% for twitching) as previously described [26]. Bacterial cultures were grown to mid-log phase (OD600 = 0.4~0.6), and then a 1-μl drop of the culture was placed on the center of the plate. Plates were incubated at 37℃ for 36 h. Motility assays were repeated for isolates where no motility was observed.

Statistics

Differences of each group were calculated using the t-test and one-way ANOVA (SAS Statistical Software Package ver. 9.1). Differences were considered statistically significant at a p value < 0.05 for all tests.

 

Results

Biofilm Formation

No differences in the growth rate of isolates of five A. baumannii complex species were found (not shown). However, significant variation in the biofilm-forming activity was observed among different strains (Fig. 2A). As evidenced by the absorbance at 600 nm, 12 of the 63 isolates (19.0%) showed higher biofilm-forming activity than the reference strain, A. baumannii ATCC 19606, which does not form biofilms. Five isolates that belonged to A. seifertii showed the highest biofilm-forming activity. No A. pittii isolates were positive for biofilm formation. Among 19 A. baumannii isolates, only four were biofilm-positive, andthese four isolates belonged to ST110 in the MLST analysis (Fig. 3A). However, the difference of biofilm-forming activity between species was not significant (Fig. 2A).

Fig. 2.Data of virulence-associated phenotypes in five Acinetobacter species.

Fig. 3.Comparison of virulence-associated phenotypes between ST110 and non-ST110 A. baumannii isolates.

A549 Cell Adherence

Similar to the results for biofilm formation on an abiotic surface, significant differences in adherence to A549 eukaryotic cells were observed between the A. baumannii complex isolates (Fig. 2B). Differences of the average were significant among the five species (p, 0.004). In particular, A. seifertii and A. calcoaceticus showed significantly higher adherence activity to A549 cells than the other species (Fig. 2B), although there were large variations between isolates of the two species. All A. baumannii ST110 isolates except one (H06-1088) also showed high adherence to A549 cells (Fig. 3B), as was observed for biofilm-forming activity. However, the difference of A549 cell adherence was not significant between ST110 and non-ST110 A. baumannii isolates.

Serum Resistance

In this study, isolates with >50% survival in NHS were considered as serum-resistant. Based on this criterion, 34 of the 63 isolates (53.9%) were resistant to NHS (Fig. 2C), including 11 A. seifertii isolates (84.6%), 13 A. baumannii isolates (68.4%), eight A. nosocomialis isolates (53.3%), and two A. pittii isolates (15.4%). No A. calcoaceticus isolates showed >50% survival rate in NHS. Differences of the average were significant among the five species (p, <0.001). The survival rate of A. seifertii was significantly higher than that of A. nosocomialis, A. pittii, and A. calcoaceticus (p values, 0.042, <0.000, and 0.003, respectively). A. baumannii and A. nosocomialis showed significantly higher survival rates than A. pittii and A. calcoaceticus (p values, <0.001 to 0.006). In addition, four A. baumannii ST110 isolates showed serum resistance (Fig. 3C). However, the difference of serum resistance was not significant between ST110 and non-ST110 A. baumannii isolates.

Motility

In this study, seven isolates (11.1%) were motile (Table 1); six of these were A. baumannii and one was A. nosocomialis. Five of the A. baumannii isolates exhibited twitching motility (motility in 1.0% agar), whereas one A. baumannii and one A. nosocomialis isolate showed swarming motility (motility in 0.5% agar). Notably, all four of the A. baumannii ST110 isolates showed twitching motility (Table 1).

 

Discussion

In this study, we investigated and compared the in vitro virulence-associated phenotypes of isolates from five A. baumannii complex species. A. baumannii is the most frequently isolated and currently one of the most important pathogens in human infections [6]. A. nosocomialis and A. pittii are also important clinically in several countries. However, the mortality rates for infection with these species are lower than those for infection with A. baumannii [3]. A. calcoaceticus was previously considered as an environmental species; however, now it is frequently isolated from human infections [22]. Lastly, A. seifertii had formerly been named as Acinetobacter genomic species “close to 13TU” and was recently validly named [19].

Antunes et al. [1] suggested the multifactorial nature of A. baumannii virulence, with no unique virulence factor accounting for the pathogenicity of the bacterium. The diversity of virulence characteristics among A. baumannii isolates has been previously reported [1,7]. Our results demonstrate the significant diversity in the virulence-associated phenotypes of clinical isolates of both A. baumannii isolates and non-A. baumannii isolates. Diversity was observed for all the characteristics tested in this study: biofilm formation, eukaryotic cell adherence, and serum resistance. However, the virulence-associated phenotypes may not reflect directly the clinical outcome in patients. Because the clinical characteristics of infection were not evaluated in this study, it is unknown whether infections with the isolates that exhibited high virulence features in this study showed worse clinical outcomes.

Several studies on the relationship between Acinetobacter species and infection outcome have shown inconsistent results, which may result partly from the fact that mortality is dependent on the patient’s condition and the quality of medical care. Although such studies should be performed prospectively with large numbers of patients, most of them were not. While many studies demonstrated that A. baumannii is associated with worse clinical outcomes than non-A. baumannii species [3,9,25], other studies did not [2,15,23]. Such inconsistent data are thought to result from the inclusion of isolates with different virulence features. The high A549 cell adherence of A. calcoaceticus, which is an environmental species, may be because the three isolates we used were from patients with bacteremia. In addition, differences in virulence-associated phenotypes have also been observed. The lack of a strong relationship between the phenotypes investigated in this study suggests a distinct mechanism for each virulence-associated phenotype [7] and may confirm the multifactorial nature of the pathogenicity of A. baumannii complex species, including A. baumannii [1]. There have been discrepancies between in vitro and in vivo virulence phenotypes [22].

Our results showed the high variability of biofilm formation, eukaryotic cell adherence, and serum resistance between species of A. baumannii complex. In addition, the variability was found among the isolates within the same species. The diversity of biofilm-forming activity between and within Acinetobacter species has been reported by de Breij et al. [5]. In addition, de Breij et al. [5] showed that biofilm formation was not different among the species, which was also identified in this study. Peleg et al. [23] also showed the diversity of virulence-associated phenotypes in A. baumannii complex species. However, they revealed that biofilm formation and persistence on human skin was significantly lower in A. calcoaceticus than in the other species. The diverse genotype of A. baumannii complex isolates has been reported [27], which may be reflected in phenotypes. Nonetheless, it is noteworthy that A. seifertii showed consistently high in vitro virulence-associated phenotypes. In particular, most A. seifertii isolates showed resistance to NHS (>50% survival in 20% NHS). As indicated earlier, A. seifertii comprises a relatively high proportion of Acinetobacter infections, especially in Korea [20,22], which suggests that this species may be endemic in certain localities. It is characterized by its high resistance to polymyxins [22], which are considered to be a last line of defense against carbapenem-resistant Acinetobacter infections. Thus, dissemination of A. seifertii is of great clinical concern. However, its clinical characteristics have not been investigated. Despite these limited data and the absence of an animal model, our results indicate that A. seifertii may be highly virulent in vitro. Thus, further studies on the virulence and clinical characteristics of this species as well as other non-A. baumannii species are needed. In addition, the in vitro virulence-associated phenotypes tested in this study indicated that A. pittii is avirulent, although all the A. pittii isolates were patients with bacteremia. It may indicate that other factors contribute to the virulence of A. pittii.

Another notable finding was the high virulence-associated phenotype of the A. baumannii ST110 clone. Despite the overall inconsistent virulence-associated phenotypes, four isolates of A. baumannii ST110 showed high virulence-associated phenotypes, particularly in the activity to form biofilm. In addition, all the isolates showed twitching motility, which is mediated by the extension and retraction of type IV pili and is a major component of A. baumannii motility [4]. In A. baumannii, motility is believed to influence its success as a pathogen, and twitching motility was identified as a common trait of A. baumannii global clone 1 [4]. In this study, only 11.1% of isolates (16.3% of A. baumannii isolates) showed motility phenotypes, which suggests that motility may not be a major contributing factor for the pathogenesis of A. baumannii complex species, including A. baumannii. However, the high biofim-forming activity and motility of A. baumannii ST110, which belongs to CC92, global clone 2, led us to speculate that it is a highly virulent clone, although it may be a single epidemic strain. It remains to be determined whether this clone is highly virulent in vivo and shows a worse clinical outcome. In addition, the mechanisms of virulence, including twitching motility, are good subjects for future studies.

In conclusion, we showed the significant variability among five A. baumannii complex species. Although no strong relationships among the virulence-associated phenotypes were revealed, A. seifertii showed consistently high virulence-associated phenotypes, including biofilm-forming activity, adherence to eukaryotic cells, and resistance to NHS. A. baumannii ST110 also showed consistently high virulence-associated phenotypes and twitching motility. Based on our findings, continuous monitoring of Acinetobacter species and clones is warranted. In this study, we did not investigate the virulence mechanisms and we did not compare the virulence of the isolates in an in vivo model. Therefore, further investigations are required to compare the in vivo virulence and to determine the mechanism of each virulence-associated phenotype.