Biofilm may not be Necessary for the Epidemic Spread of Acinetobacter baumannii

Biofilm is recognized as a contributing factor to the capacity of Acinetobacter baumannii to persist and prosper in medical settings, but it is still unknown whether biofilms contribute to the spread of A. baumannii. In this study, the biofilm formation of 114 clinical A. baumannii isolates and 32 non-baumannii Acinetobacter isolates was investigated using a microtiter plate assay. The clonal relationships among A. baumannii isolates were assessed using pulsed-field gel electrophoresis and multilocus sequence typing, and one major outbreak clone and 5 other epidemic clones were identified. Compared with the epidemic or outbreak A. baumannii isolates, the sporadic isolates had significantly higher biofilm formation, but no significant difference was observed between the sporadic A. baumannii isolates and the non-baumannii Acinetobacter isolates, suggesting that biofilm is not important for the epidemic spread of A. baumannii. Of the multidrug-resistant (MDR) A. baumannii isolates in this study, 95.7% were assigned to international clone 2 (IC2) and showed significantly lower biofilm formations than the other isolates, suggesting that biofilm did not contribute to the high success of IC2. These findings have increased our understanding of the potential relationship between biofilm formation and the epidemic capacity of A. baumannii.

biofilm-positive isolates, 19.5% (8/41) were strong biofilm producers. In contrast, 34.6% (9/26) of the non-AB biofilm-positive isolates were strong biofilm producers, as shown in Table 2. The 32 clinical non-AB isolates showed higher biofilm formation than the 114 clinical A. baumannii isolates (Fisher's exact test, P < 0.0001). Of the non-AB isolates, 75% were non-MDR (Table 1), so we compared the biofilm formation capacities of non-AB isolates to the non-MDR A. baumannii isolates, and no significant difference was observed between them ( Table 2). The individual biofilm formation capacities are outlined in supplementary Table S1.
Pulsed-field gel electrophoresis (PFGE) analysis of the A. baumannii isolates. The clonal relationships between A. baumannii isolates were assessed using pulsed-field gel electrophoresis (PFGE). The 114 A. baumannii isolates tested herein for biofilm formation represented 42 unique PFGE types (P1~P42), as shown in Fig. 1 and Table 3. All isolates sharing the same PFGE type were isolated from the same hospital (Table 3). Compared with the MDR isolates, a higher genetic diversity was revealed in the non-MDR isolates (Fig. 1). We define an isolate as being epidemic if at least two other isolates isolated from the same hospital during the study period exhibited the same PFGE profile (with ≥ 95% similarity in their banding patterns). Isolates clustering according to these features were regarded as epidemic clones, while all other isolates were considered sporadic. A total of 6 epidemic clones were revealed (P4, P10, P7, P12, P14, P16), which covered 62.3% of the tested A. baumannii isolates, as shown in Fig. 1 and Table 3. However, one of the epidemic clones (P10) was responsible for a major outbreak involving 29 patients and this clone was termed outbreak clone. All the epidemic isolates (including the outbreak isolates) were MDR.   Table 2. Comparison of the biofilm formation capacities of clinical A. baumannii isolates and nonbaumannii Acinetobacter isolates (non-AB). Chi-square test: P < 0.0001. The Bonferroni method was used to conduct multiple comparisons. Significant differences were found between outbreak and sporadic (S) (P < 0.0001), outbreak and non-AB (P < 0.0001), epidemic and sporadic (S) (P < 0.0001), and epidemic and non-AB (P < 0.0001). a N: non-biofilm producer, W: weak biofilm producer, M: moderate biofilm producer, S: strong biofilm producer. b MDR sporadic isolates. c Non-MDR sporadic isolates. * The positive rate of biofilm formation for each group.  Table 4. All the isolates sharing the same PFGE type were also assigned to the same ST (Table 3).   A total of 89 (78%) A. baumannii isolates, representing PFGE types P1 to P19 isolates shown in Fig. 1, were assigned to ST2 of the IC2 (Table 4), which covered 95.7% of the MDR isolates, including all the epidemic isolates.
Of the IC2 isolates, 93.3% showed weak biofilm forming capacities, of which 75.3% (67/89) were non-biofilm producers and 18% (16/89) were weak biofilm producers (Table 4). Only one IC2 isolate (HN006) was a strong biofilm producer (mean OD/ODc = 13.24, Table S1), which showed a similar but unique PFGE profile within the P14 clone, which differed by an additional band (Fig. 1). Thus, this stronger IC2 biofilm producer was not widely spread during our study period. The other 5 IC2 moderate biofilm producers originated from 3 hospitals and were assigned to 5 PFGE types (Table 3). Only two of these IC2 moderate biofilm producers belonged to epidemic clones. Compared with the IC2 isolates, the other isolates (25 isolates representing 16 STs) showed significantly higher biofilm formation (biofilm-positive rate of 24.7% vs. 76%, Table 4, Fisher's exact test, P < 0.0001).

Comparison of biofilm formation capacities between outbreak and epidemic A. baumannii isolates.
During our study period, no A. baumannii infection outbreak was identified except for one hospital. A total of 54 isolates isolated during this outbreak period were used in this study, which were typed into 11 PFGE types (P4, P7, P10, P1~3, P6, P13, P31, P40, P39, Table 3). Among them, the P10 clone which covered 29 isolates was identified to be responsible for this outbreak. To determine whether biofilm was one possible reason for this outbreak, we compared the biofilm formation of the P10 clone with other epidemic clones that did not cause higher isolation rates than the exception. Contrary to our expectation, although there was no significant difference, a lower biofilm-positive rate was observed for the P10 clone (10.3% vs. 31%), as shown in Table 2. Therefore, biofilm formation did not contribute to the high isolation of this outbreak clone.

Comparison of biofilm formation capacities between epidemic and sporadicA. baumannii isolates.
A total of 43 A. baumannii isolates representing 36 unique PFGE types were identified as sporadic isolates, which showed significantly higher biofilm-forming capacity than the epidemic isolates (biofilm-positive rate of 58.1% vs. 31%, Fisher's exact test, P = 0.0047, Table 2). Of the biofilm-negative sporadic isolates, 72.2% (13/18) were MDR; therefore, a sub-classification according to drug resistance was performed. For the biofilm-positive sporadic A. baumannii isolates, the OD/ODc ratios ranged from 1.03 to 24.08 for the non-MDR sporadic clones and from 1.06 to 4.9 for the MDR sporadic clones (Table S1). Although a higher biofilm-positive rate was observed in non-MDR sporadic isolates than in the MDR sporadic isolates (76.2% vs. 41%), no significant difference was observed between them (Table 2). However, a significant difference was observed between the non-MDR sporadic isolates and the epidemic isolates (Table 2).
Taking into account that we could not exclude the possibility that the MDR sporadic isolates would cause an epidemic at another time or in another hospital, we compared the biofilm formation capacities between all the MDR and non-MDR isolates. A significantly higher biofilm formation capacity was observed in the non-MDR  isolates (biofilm-positive rate of 76.2% vs. 26.9%, Fisher's exact test, P < 0.0001). However, no significant difference was noted among the three MDR isolate groups (outbreak, epidemic and MDR sporadic, Table 2).

Discussion
There have been some reports on the variations in biofilm formation capacity among clinical isolates of A. baumannii [9][10][11][12] , but the quantitative differences in biofilm formation among clinical isolates, in association with the epidemic capacity of strains, have been poorly investigated thus far. In this study, the biofilm formation capacity was evaluated in a large set of well-described clinical Acinetobacter isolates. Contrary to our expectation, the non-AB isolates showed a higher biofilm formation than did the A. baumannii isolates. Among the A. baumannii isolates, the non-MDR ones showed a higher biofilm formation capacity than the MDR isolates, including all the epidemic clones. Even when comparing the non-AB isolates to only the non-MDR A. baumannii isolates, there was still no higher biofilm formation observed for A. baumannii, suggesting that biofilm-forming capacity could not explain the clinical success of A. baumannii. For A. baumannii isolates, the strong biofilm producers were less frequently resistant to antibiotics and seemed to be less epidemic, suggesting that biofilm is not necessary for the epidemic spread of A. baumannii. A high proportion (95.7%) of the MDR A. baumannii isolates used in this study were assigned to IC2 by MLST, which agreed with previous reports that multidrug resistance is often associated with isolates that belong to international clones [13][14][15] . Distinct genetic diversity among the IC2 isolates was revealed by PFGE, with only some of those isolates demonstrating epidemicity during our study period; no significant difference was observed between the epidemic and sporadic IC2 isolates. Although the other ST lineages revealed in this study were not as successful as the IC2, which is widely spread worldwide and include strains that are usually MDR and associated with outbreaks, a significantly higher biofilm formation capacity was observed for non-IC2 than for the IC2 isolates, suggesting that biofilm does not contribute to the success of IC2.
It remains an open question whether A. baumannii were first to develop MDR and then lost their biofilm-forming capability or whether weak biofilm isolates were more prone to develop MDR, promoted by survival pressure. A recent study of isogenic mutants from a susceptible A. baumannii clinical isolate demonstrated the overproduction of resistance-nodulation-cell division (RND)-type efflux systems, AdeABC and AdeIJK, which pump out a wide range of antimicrobial compounds and are associated with multidrug resistance in A. baumannii 16 , resulting in the acquisition of antibiotic resistance and decreased biofilm formation 17 . This observation demonstrated the hypothesis that A. baumannii lost their biofilm-forming capability after developing MDR, but this model still needs further confirmation. However, the mechanism maybe more complicated than our speculation and cannot be answered with only one hypothesis. Whatever the truth is, we can speculate that compared with the MDR isolates, the non-MDR A. baumannii isolates are easily cleared after infection, so the capacity to grow as a biofilm may play a more important role in their persistence. Therefore, although high genetic diversity was revealed in the non-MDR isolates, a high proportion of them still maintained strong biofilm-forming capabilities.
In conclusion, the sporadic A. baumannii isolates have significantly greater biofilm-forming capabilities than the outbreak and epidemic A. baumannii isolates, but they showed biofilm formation capabilities that were similar to the other Acinetobacter species, suggesting that biofilm formation could not explain the clinical success of A. baumannii and is not important for the epidemic spread of A. baumannii. The IC2 isolates showed significantly lower biofilm formation capacity than other isolates, suggesting that biofilm did not contribute to the success of IC2. These findings have refreshed our understanding of the relationship between biofilm formation and A. baumannii epidemic capacity and may serve as caveats for future studies to understand the transmission of this pathogen.

Materials and Methods
Bacterial strains. A collection of 114 well-characterized A. baumannii isolates and 32 non-AB isolates were used (4 A. bereziniae isolates, 8 A. nosocomialis isolates, 13 A. pittii isolates, and 7 A. junii isolates, Table 1). The A. baumannii isolates included in the present study were from a collection of clinical isolates recovered during epidemiological surveys (from 4 Chinese cities, one hospital per city, not more than 2 months). All isolates were identified by matrix-assisted laser desorption/ionization time-of-flight (MALDI TOF) mass spectrometry 18 and were verified using sequence analysis of the 16S-23S ribosomal DNA intergenic spacer 19 .
The antimicrobial susceptibilities of the tested Acinetobacter isolates to 11 antimicrobials were performed using an Etest on Mueller-Hinton agar. If a strain was resistant to at least three classes of antimicrobial agents, including all penicillins and cephalosporins (including inhibitor combinations), fluoroquinolones, and aminoglycosides, then that strain was called MDR. An MDR strain also resistant to carbapenems was called extensively drug-resistant (XDR) 20 .

PFGE and MLST.
The clonal relationships between A. baumannii isolates were assessed using PFGE, as previously described 21 . The PFGE patterns were analysed with BioNumerics software (Applied Maths) using the Dice coefficient and the unweighted-pair group method with average linkages (UPGMA), a 1.5% tolerance limit and 1.5% optimization. MLST was performed according to the published Pasteur protocols 22 . Biofilm formation. Biofilm formation was examined by the semi-quantitative determination of biofilm formation in a 96-well microtiter plate assay, as previously described 12 . Cultures were inoculated in Luria-Bertani broth (LB) and adjusted to an optical density at 600 nm of ~0.1. Each well of sterile 96-well polystyrene microtiter plates was filled with 200 μ L of bacterial suspension. Wells containing only the medium were used as negative controls. After static incubation at 37 °C for 24 h, the plates were washed gently three times with phosphate-buffered saline to remove unattached bacteria, air-dried and stained with 0.1% crystal violet solution for 20 min, then scanned at 570 nm to determine the OD of the stained biofilms. The same protocol was followed to quantify the biofilm after prolonged incubation for 48 and 72 hours, and the maximum values obtained under the three incubation times were selected to represent the biofilm-forming capacity to avoid variations due to differences in biofilm formation rate. All assays were performed in triplicate at three independent time-points using fresh samples each time. The ODc was defined as three standard deviations above the mean OD of the negative control 23 . Each isolate was classified as follows 23 : non-biofilm producer (N): OD ≤ ODc; weak biofilm producer (W): ODc < OD ≤ 2 × ODc; moderate biofilm producer (M): 2 × ODc < OD ≤ 4 × ODc; or strong biofilm producer (S): OD > 4 × ODc.
Statistical analysis. All statistical analyses were conducted in SAS9.2 software (SAS Institute Inc., Cary, NC, USA). All statistical tests were two-sided, and P < 0.05 was considered statistically significant. The chi-square test and Fisher's exact test were selected to analyse the biofilm formation differences among groups.The Bonferroni method was used to conduct multiple comparisons.