Characterization of biofilm production in different strains of Acinetobacter baumannii and the effects of chemical compounds on biofilm formation

Bacterial strains

All of the bacterial strains used in this study were isolates from our previous studies (Lin et al., 2014; Lin, Lin & Lan, 2015; Lin et al., 2015; Lin, Lin & Lan, 2017a; Lin et al., 2017b) or American Type Culture Collection (ATCC), as shown in Table 1. The studied strains contained quality control strain ATCC 19606, ATCC 17978 and 21 clinical isolates. Five antibiotic-induced resistant strains from ATCC 17978 or tigecycline resistant clinical isolate ABhl1 were also included. Since biofilm formation is associated with antibiotic resistance and baeR as well as emrA/emrB have been shown related with tigecycline and colistin resistance respectively in our previous studies (Lin et al., 2014; Lin, Lin & Lan, 2017a), so we also enrolled the mutant strains for these genes.

Biofilm quantification

Biofilm formation was performed in 24-well polystyrene plates as previously described with some modifications (Marti et al., 2011; Orsinger-Jacobsen et al., 2013; Kim et al., 2015). Biofilm formation was determined in Mueller Hinton Broth (MHB) using an initial OD₆₀₀ of 0.01, and the plates were incubated at 37 °C for 24 h without shaking (24-well, one mL). Two wells were left uninoculated and used as negative controls. The culture media was removed by inversion, and the wells were washed twice with phosphate buffered saline (PBS).

Both crystal violet and XTT assays show an excellent application for quantification of biofilms. The Crystal violet assay is cheap, easy and is usually used for the quantification of biofilms formed by microorganisms but XTT is more reliable and repeatable (Hendiani, Abdi-Ali & Mohammadi, 2014). In this study, XTT assay was used to measure the metabolic activity in order to estimate the burden of viable cells. The electron transport system in the cellular membrane of live bacteria reduces the XTT tetrazolium salt to XTT formazan, which results in a colorimetric change measured at 492 nm (Orsinger-Jacobsen et al., 2013). 250 µL XTT solution (0.0625 mg/mL XTT, 0.5 mM menadione in PBS) was added to each well of 24-well microtiter plate for 30 min at 37 °C. After incubation, 200 µL of the XTT supernatant was transferred to a fresh 96-well plate, and the calorimetric absorbance was measured (OD_{490nm(XTT)biofilm}) at 490 nm using iMarkTM Microplate Absorbance Reader (Bio-Rad laboratories, Inc. USA). The negative control (OD_{490nm(XTT)control}) was used to reduce the background absorbance optic density values in the XTT assay. Since different bacterial strains had different growth rates, the total cell numbers would be different at the end of incubation, thus influencing the total biofilm mass or metabolic activities. To correct the bias due to cell numbers, the OD_595nm (OD_595nm(cells)) was determined to estimate the total cells in each corresponding well of 24-well microtiter plate after transferred to a new 96-well plate for optic density measurement. The ability to form a biofilm was expressed using a biofilm formation index: [BFI = (OD_{490nm(XTT)biofilm}–OD_{490nm(XTT)control})/OD_595nm(cells)]. The ODc was defined as three standard deviations above the mean OD of the negative control. Each isolate was classified as follows: non-biofilm producer: OD ≤ ODc; weak biofilm producer: ODc <OD ≤ 2 × ODc; moderate biofilm producer: 2 × ODc < OD ≤4 × ODc; or strong biofilm producer: OD > 4 × ODc (Hu et al., 2016).

Biofilm imaging

Biofilms were examined by SEM and CLSM (Djeribi et al., 2012). For SEM examination, biofilms were developed at 37 °C for 24 h in polystyrene coverslips placed into a 24-well plate with MHB containing bacterial cells (OD_600nm 0.01) added. After fixed in 3.7% formaldehyde for 40 min at room temperature and then rinsed twice with PBS, further fixation was performed with 1% osmium tetroxide for 15 min at room temperature and then rinsed twice with PBS. The samples were then dehydrated by passing them through different concentrations of ethanol: 35%, 50%, 70%, 80%, and 95%, each for 5 min, followed by 100% ethanol twice for 10 min each time. The samples were then dried at 60 °C for 24 h. After being coated with gold-palladium (via sputter coating), the samples were examined under a bioscanning electron microscope (Hitachi, S-4700, Type II).

For CLSM examination, biofilms formation in polystyrene coverslips and fixation in 3.7% formaldehyde were the same as mentioned above. Then the samples were stained with a prepared 100 nM solution of Syto-9 stain (Invitrogen, Carlsbad, CA) for 30 min (Luo et al., 2015b). Syto-9 stained biofilms were excited with a 488 nm solid-state laser, and fluorescence was captured between 500–550 nm. The images of biofilms were rendered and assembled using appropriate computational software (ZEISS LSM780).

Influence of antibiotics, LL-37 and tannic acid on biofilm formation

The activity of the various antibiotics, including amikacin, imipenem, colistin and tigecycline, LL-37 and tannic acid, on bacterial growth was tested in 24-well microtiter plates containing the above compounds at different concentrations. The concentrations chosen for different compounds were based on individual MIC with two-fold serial dilution. However, if biofilm formation was not observed at the higher concentrations of compounds, then the lower concentration was adopted. Biofilm production was checked by the same method as described above. Finally, A. baumannii ATCC 17978 without any compound added was used as a control for biofilm formation to calculate the relative biofilm formation capacity (relative biofilm formation index [RBFI] = BFI/BFI_{A.baumannii ATCC 17978}).

Planktonic susceptibility test

A planktonic susceptibility test was performed as previously described to determine the minimum inhibitory concentration (MIC) of A. baumannii (Lin et al., 2014). In brief, bacteria were inoculated into one mL cation-adjusted Mueller-Hinton Broth (CAMHB) (Sigma-Aldrich, St. Louis, MO) containing different concentrations of antibiotics, including amikacin, imipenem, colistin and tigecycline, to reach ≈ 5 × 10⁵ CFU/mL, and the cultures were incubated at 37 °C for 24 h. The lowest antibiotic concentration that completely inhibited bacterial growth was defined as the minimum inhibitory concentration, and growth was determined by unaided eyes and by measuring optical densities using a spectrophotometer.

Biofilm susceptibility test

The biofilm susceptibility test was performed as previously described with some modifications (Kim et al., 2015). In brief, preparation of the antibiotic stock solution and preparation of the inoculum were the same as the method used in checking the MIC. Biofilm susceptibility tests were performed in flat-bottom, polystyrene cell-culture microtiter plates containing 5 × 10⁵ CFU/mL in MHB with final well volumes of 100 µL. The plates were incubated at 37 °C without shaking. After 24 h, all the suspension fluid was removed from the plates and then washed with PBS. Equal amounts of appropriate antibiotic dilutions were added. Twofold dilutions of the tested antibiotics in CAMHB were diluted from 256 to 0.125 µg/mL. The tested antibiotic concentrations would be increased if necessary. Then, the plates were incubated at 37 °C for 24 h without shaking. After 24 h, inhibition of bacteria growth was determined by no turbidity observed by unaided eyes. Regrowth of bacteria was further determined by spot assay. The minimum biofilm eradication concentration (MBEC) represents the lowest drug concentration at which bacteria failed to regrow (Olson et al., 2002).

Statistical analysis

The comparison of biofilm formation, including BFI and RBFI, was analyzed using Student’s t-test. All of the data were from three independent experiments and analyzed using Statistical Package for the Social Sciences version 16 (SPSS Inc., Chicago, IL, USA). p < 0.05 was considered statistically significant.

Biofilm quantification and its relationship with the antimicrobial susceptibility

Biofilm quantification of the 33 different strains of A. baumannii (Table 1), including 21 clinical isolates, by the biofilm formation index is presented in Fig. 1. There are obvious differences in biofilm formation among the clinical isolates of A. baumannii. Of the 22 clinical isolates including A. baumannii ATCC 17978, twenty isolates were strong biofilm producers (4 × ODc < OD_490nm) except VGH1 and CT11 (2 × ODc < OD_490nm ≤ 4 × ODc) were moderate biofilm producers (data shown in the raw data set).

The antimicrobial susceptibility of the clinical isolates in this study has been published previously (Lin et al., 2017b) as shown in Table S1. Susceptibility breakpoints to antimicrobial agents were determined by the MIC interpretive standards for Acinetobacter spp. of Clinical and Laboratory Standards Institute (CLSI, 2012). The MIC interpretive criteria for each individual antimicrobial agents except cefazolin, cefmetazole and tigecycline (MIC µg/mL) (S = susceptible, I = intermediate and R = resistant) are listed respectively as the following: Ampicillin/sulbactam (S ≤ 8/4, I 16/8 and R ≥ 32/16); piperacillin/tazobactam (S ≤ 16/4, I 32∕4 − 64∕4 and R ≥ 128/4) ; cefotaxime (S ≤ 8, I 16-32 and R ≥ 64); ceftazidime (S ≤ 8, I 16 and R ≥ 32); cefepime (S ≤ 8, I 16 and R ≥ 32); imipenem (S ≤ 4, I 8 and R ≥ 16); meropenem (S ≤ 4, I 8 and R ≥ 16); amikacin (S ≤ 16, I 32 and R ≥ 64); gentamicin (S ≤ 4, I 8 and R ≥ 16); ciprofloxacin (S ≤ 1, I 2 and R ≥ 4); levofloxacin (S ≤ 2, I 4 and R ≥ 8); trimethoprim/sulfamethoxazole (S ≤ 2/38 and R ≥ 4/76). A. baumannii is naturally resistant to cefazolin, whereas the isolates of A. baumannii with cefmetazole MICs of ≥ 32µg/mL are considered resistant (Jones et al., 1986). The provisional MIC breakpoints for tigecycline are ≤2, 4 and ≥8 µg/mL to designate susceptible, intermediate and resistant strains, respectively (Pachon-Ibanez et al., 2004). All of the 21 A. baumannii clinical isolates are not susceptible to many of the tested antimicrobial agents. Although ATCC 17978 is the least resistant strain among the isolates studied, its BFI ranked second. Two isolates, VGH2 and VGH7, are pan-drug resistant. However, their biofilm formation ability BFI differed with VGH2 ranking first and VGH7 ranking sixteenth. It seems the capacity of biofilm production does not relate to antimicrobial susceptibility as shown in Table S1. To validate the inference, the correlation between biofilm formation ability and resistance to the 15 antimicrobial agents in A. baumannii was analyzed using the Wilcoxon rank-sum test. Figure 2 depicts the box plots with the comparison of the biofilm formation indices of the susceptible and resistant strains to the individual antibiotics. It reveals no relationship with statistical significance (p > 0.05) between the ability to form biofilm and the antimicrobial susceptibility of the A. baumannii clinical isolates.

Biofilm formation by A. baumannii ATCC 17978 after baeR, emrA, emrB and ompA gene knockout

Many proteins, including two-component regulatory system BfmRS (Luo et al., 2015a), CsuA/BABCDE usher-chaperone assembly system (Tomaras et al., 2008), biofilm-associated proteins (Loehfelm, Luke & Campagnari, 2008), OmpA protein (Gaddy, Tomaras & Actis, 2009), AdeFGH (He et al., 2015) and blaPER (Lee et al., 2008), have been demonstrated to be related with biofilm formation in A. baumannii. To understand the scope of involvement further, biofilm formation caused by A. baumannii ATCC 17978 after gene knockout of another two-component regulatory system gene baeR, efflux pump genes emrA/emrB and ompA is compared as shown in Fig. 3A. The ΔbaeR, ΔemrA, ΔemrB and ΔompA mutant strains all had less biofilm formation than the wild type strain. This result is further confirmed by the SEM and CLSM images shown in Figs. 3B and 3C, respectively. Of these three categories of genes, ompA knockout led to the most marked decrease biofilm formation in A. baumannii. The SEM figure of ΔompA strain shows that only few cells are clustered and its images of CLSM exhibit the thinnest biofilm. Although the influence on biofilm formation reduction after baeR, emrA, or emrB gene knockout was not so obvious as the ΔompA strain, the degree of decrease was still statistically significant while being compared with the ATCC 17978 strain (p < 0.05 for ΔbaeR, p < 0.01 for ΔemrA, and p < 0.001 for ΔemrB).

Influence of different compounds on biofilm formation

Table 2 presents minimum inhibitory concentrations (MICs) of A. baumannii ATCC 17978 and VGH2 to amikacin, imipenem, colistin, tigecycline, LL-37 and tannic acid. The influence of different compounds, including amikacin, imipenem, colistin, tigecycline, LL-37 and tannic acid, on biofilm formation in A. baumannii ATCC 17978 and VGH2 is shown in Fig. 4 and Fig. 5. Amikacin, imipenem, colistin and LL-37 were tested at 1/4, 1/2 and 1 x MIC whereas tigecycline was tested at 1/16, 1/8, 1/4 x MIC. Since tannic acid had very high MIC (>300 µg/mL), the concentrations of 200, 100 and 50 µg/mL were chosen arbitrarily for testing. The addition of amikacin (>0.5 µg/mL), colistin (>0.5 µg/mL), LL-37 (>4 µg/mL), or tannic acid (>50 µg/mL) decreased the biofilm formation ability of A. baumannii ATCC 17978, whereas the addition of amikacin (>256 µg/mL), imipenem (>4 µg/mL), colistin (>0.125 µg/mL), LL-37 (>2 µg/mL), or tannic acid (>50 µg/mL) decreased the biofilm formation ability of VGH2. In contrast, the addition of tigecycline increased the biofilm formation ability of A. baumannii ATCC 17978 (>62.5 ng/mL) and VGH2 (>0.5 µg/mL). Figure 6 presents the SEM and CLSM images of biofilm formation by A. baumannii ATCC 17978 in the presence of 1 µg/mL colistin or 125 ng/mL tigecycline. With 1 µg/mL colistin added to the ATCC 17978 strain, the SEM image shows the cells become scattered and the CLSM images exhibit decreased thickness of biofilm. In contrast, 125 ng/mL tigecycline exposure has the opposite impact on biofilm formation of A. baumannii ATCC 17978. Both of the SEM and CLSM images validated the findings of biofilm quantification by RBFI analysis (Figs. 4 and 5).

MICs and MBECs of A. baumannii ATCC 17978 and VGH2

Besides MICs, MBECs of A. baumannii ATCC 17978 and VGH2 to amikacin, imipenem, colistin and tigecycline is also shown in Table 2. In this study, the MIC values for antimicrobials were determined using a broth microdilution assay. The ATCC 17978 strain had low MICs to all of the four tested antimicrobial agents, whereas the VGH2 strain had higher MICS for amikacin (1,024 µg/mL) and imipenem (16 µg/mL). When biofilms cells of the same strains were tested, MBECs rose astonishingly with at least fourfold higher relative to the planktonic MICs. Of all the four antibiotics tested, colistin had the most obvious MBEC increase from 0.5 to >8,192 µg/mL for VGH2. This result suggested that biofilm cells of bacteria would become more resistant to tested antibiotics according to MBEC.