Methicillin-Resistant Staphylococcus aureus Grown on Vancomycin-Supplemented Screening Agar Displays Enhanced Biofilm Formation

Abstract

Brain heart infusion agar containing 3 mg/liter vancomycin (BHI-V3) was used to screen for heterogeneous vancomycin-intermediate Staphylococcus aureus (hVISA). There was markedly greater biofilm formation by isolates that grew on BHI-V3 than by strains that did not grow on BHI-V3. Increased biofilm formation by hVISA may be mediated by FnbA- and polysaccharide intercellular adhesin-dependent pathways, and upregulation of atlA and sarA may also contribute to enhanced biofilm formation by hVISA upon prolonged exposure to vancomycin.

TEXT

There are growing numbers of reports from many countries regarding strains of heterogeneous vancomycin-intermediate Staphylococcus aureus (hVISA) (1, 2). Many studies have indicated that hVISA is commonly associated with vancomycin treatment failure and persistent infection; however, the mechanism of this treatment failure has yet to be determined (3, 4). Biofilm formation is one of the most common mechanisms of antibiotic resistance in bacteria. A biofilm can assist bacteria in evading the host immune system and represents a diffusion barrier for antibiotics (5, 6). Some studies have examined the biofilm-forming ability of hVISA, but the results are controversial (7, 8).

In this study, brain heart infusion agar containing 3 mg/liter vancomycin (BHI-V3) was used to screen for hVISA (9, 10). A total of 176 clinical isolates of S. aureus were screened on BHI-V3 medium. Sixteen strains that grew on BHI-V3 (GV3) were selected as the test group, and a control group of 16 methicillin-resistant S. aureus (MRSA) strains was randomly selected from among those that did not grow on BHI-V3 (NGV3). One randomly selected GV3 isolate (0534) was induced to obtain a series of vancomycin-resistant strains. Isolates that grew on BHI supplemented with 8, 16, and 32 mg/liter vancomycin were named 0534-V8, 0534-V16, and 0534-V32, respectively. All of these strains were adjusted to a McFarland standard of 0.5 and then diluted 1:100 in BHI supplemented with 2% (wt/vol) glucose. This suspension was transferred to a flat-bottom 96-well polystyrene microtiter plate and a 15-mm glass bottom cell culture dish (Nest, Hong Kong, China) and incubated at 35°C. The medium was removed according to growth curve experiments to ensure that all of the strains had the same cell density, which was determined by measuring the optical density at 600 nm (OD₆₀₀). Biofilm formation was quantified by the addition of 100 μl of 0.1% (wt/vol) crystal violet and measurement of absorbance at 570 nm (11). Confocal laser scanning microscopy (CLSM with a Zeiss LSM710; Carl Zeiss, Jena, Germany) was also used to measure biofilm structure and thickness. Biofilms were fixed with 2.5% glutaraldehyde and stained with 15 μmol/ml propidium iodide (Sigma-Aldrich, St. Louis, MO, USA), which stained nonviable cells red, and 50 μg/ml fluorescein isothiocyanate (FITC)-labeled concanavalin A type IV (Sigma-Aldrich), which stained extracellular polysaccharide green. The excitation and emission wavelengths of these dyes were 488 and 500 to 548 nm for FITC-labeled concanavalin A and 543 and 569 to 719 nm for propidium iodide. Three-dimensional biofilm images and biofilm thickness were analyzed with Imaris 7.2.3 (Bitplane, Zurich, Switzerland). Molecular characteristics, including agr, SCCmec, and spa types, were investigated as previously described (12–14). MICs of vancomycin were determined by Etest according to the Clinical and Laboratory Standards Institute (CLSI) guidelines. Expression of biofilm formation-related genes was investigated by real-time quantitative reverse transcriptase PCR (15) (for the primers used, see Table S1 in the supplemental material), and polysaccharide intercellular adhesin (PIA) production was detected by immunochemiluminescence assay to explore the mechanisms involved in enhanced biofilm formation (16). Student's t test and analysis of variance were used to compare mean values. P < 0.05 was considered statistically significant.

There was no significant difference between the genetic backgrounds of the GV3 and NGV3 groups. Most of the GV3 (15/16, 93.6%) and NGV3 (15/16, 93.6%) isolates were agr group II, and SCCmec type III was most common among both GV3 and NGV3 isolates (Table 1). All of the strains were susceptible to vancomycin according to current CLSI guidelines, while the percentage of GV3 group isolates with an MIC of ≥1 mg/liter was significantly higher than that of the NGV3 group (75% versus 18.75%; P < 0.05).

TABLE 1.

Clinical MRSA strains used in this study

Characteristic	No. (%) of clinical isolates in:
	GV3 group (n = 16)	NGV3 group (n = 16)
Vancomycin MIC (mg/liter) of:
≤0.5	4 (25.00)	13 (81.25)
1	7 (43.75)	3 (18.75)
1.5	4 (25.00)	0
2	1 (6.25)	0
agr group
I	1 (6.25)	1 (6.25)
II	15 (93.75)	15 (93.75)
SCCmec type
I	3 (18.75)	2 (12.50)
II	4 (25.00)	5 (31.25)
III	9 (56.25)	8 (50.00)
IV	0	1 (6.25)
spa type
t030	15 (93.75)	15 (93.75)
t289	1 (6.25)	0
t189	0	1 (6.25)

We next compared the biofilm formation of the GV3 and NGV3 group strains. Quantification of crystal violet staining by measurement of OD₅₇₀ showed markedly greater biofilm formation by GV3 isolates than by NGV3 isolates. The mean absorbance values of the GV3 and NGV3 isolates were 0.336 ± 0.088 and 0.109 ± 0.036, respectively (P < 0.05) (Fig. 1A). We further tested the capacity of the series of vancomycin-resistant induced strains to form biofilms. The mean absorbance values of isolates 0534-V8 (1.64 ± 0.03), 0534-V16 (2.13 ± 0.10), and 0534-V32 (2.49 ± 0.06) were higher than that of isolate 0534 (0.43 ± 0.03) (Fig. 2A), and the biofilm formation abilities of resistant strains increased significantly with increasing vancomycin MICs (P < 0.05). As shown in Fig. 2B, a thin layer of isolate 0534 was visible and all of the strains with induced vancomycin resistance formed a compact, thick biofilm on the surface. The average biofilm thickness of isolate 0534-V32 was greater than that of isolate 0534 (12 ± 0.7 versus 2 ± 0.4 μm; n = 3). These results are consistent with the findings of Sakoulas et al. (17). However, Howden et al. presented a conflicting report, in which clinical hVISA/VISA strains were significantly less able than the parental VSSA strain to form a biofilm (8). There are some factors that may be responsible for these divergent results. The first is incubation time, as hVISA and VISA isolates are characterized by slow growth (18). We removed the medium according to the growth curve experiments to ensure that all of the strains had adequate growth. The second factor is medium composition. All of the growth media in the present study were supplemented with 2% glucose, which may promote the expression of a biofilm-positive phenotype (19). We also noticed that most of the strains in this study were agr group II, whereas all of the isolates in study of Howden et al. were agr type I. Genetic background differences might also account for the discrepancies between our study and those of other groups.

FIG 1.

Biofilm formation and levels of biofilm formation-associated gene mRNA expression of all of the clinical strains used in this study. (A) Quantification of biofilm formation was done by measurement of crystal violet staining with a plate reader (OD₅₇₀). (B) icaA, icaR, and fnbA transcript levels were measured by quantitative reverse transcriptase PCR. P < 0.05.

FIG 2.

Comparison of the biofilm formation and biofilm formation-associated gene mRNA expression levels of isolate 0534 and a series of induced isolates. (A) Quantification of biofilm formation by measurement of crystal violet staining with a plate reader (OD₅₇₀). (B) Three-dimensional reconstructions of CLSM images of biofilm formation. Bacterial cells were stained with propidium iodide, which stained nonviable cells red, and FITC-labeled concanavalin A type IV, which stained extracellular polysaccharide green. (C) icaA, icaR, fnbA, sar, atlA, and lrgA transcript levels were measured by quantitative reverse transcriptase PCR. P < 0.05. (D) Quantification of PIA production by clinical strains and a series of induced strains.

Biofilm formation is a complex, multifactorial process. Initially, bacteria adhere directly to a material's surface. This process is followed by proliferation, accumulation, and intercellular interactions between bacterial cells to form multilayered structures (20–22). To ascertain the molecular mechanism involved in enhancing biofilm formation, we investigated the expression of fnbA, atlA, and sarA, which are involved in the primary attachment phase, and icaA, icaR, lrgA, and sarA, which are involved in proliferation and formation. The icaA and fnbA expression levels were 1.67- and 3.44-fold higher, respectively, in the GV3 group than in the NGV3 group (P < 0.05). In contrast, the icaR transcript levels in the GV3 group were 1.88-fold lower than those in the NGV3 group (Fig. 1B). Overall analysis of the six biofilm-associated genes in the series of induced strains also showed upregulation of icaA and fnbA and downregulation of icaR compared with the parent isolate. The increase in gene expression level was associated with the degree of MIC elevation (Fig. 2C). Clinical strains and all of the induced strains were further assayed for the ability to produce PIA. Figure 2D shows that PIA production by isolate 0534-32 was 5.31 times as high as that by isolate 0534 (P < 0.05). The fnbA gene encodes FnbA, which is a large, multidomain protein that binds to fibronectin-coated surfaces. FnbA is essential for S. aureus colonization of host tissues and implanted biomedical devices (1, 23). Recent reports have revealed that FnbA can promote biofilm formation at both the primary attachment and intercellular accumulation stages (7, 24). Expression of icaA is involved in PIA production (25–27). It clusters around cells and protects them against both host immune defenses and antibiotic treatment (28). The icaR gene is located upstream of icaA and is transcribed divergently, encoding a transcriptional repressor that negatively regulates icaA expression in S. aureus (20, 29). Therefore, we speculate that the reduced expression of icaR stimulates an increase in icaA expression, which leads to increased PIA synthesis in GV3 strains.

In the series of strains with induced resistance, we found that, besides the upregulation (fnbA and icaA) or downregulation (icaR) of the above genes, the expression of atlA and sarA also increased with the vancomycin MIC. The atlA gene encodes autolysin, which is a bifunctional peptidoglycan hydrolase that plays an important role in daughter cell separation following cell division (30). Autolysin mediates primary attachment to plastic surfaces by promoting the release of extracellular DNA from bacterial cells (22). The sarA gene encodes the DNA-binding protein SarA. Several studies have demonstrated that SarA enhances the transcription of icaA and fnbA, as well as other genes that encode surface-associated binding proteins (31, 32), and this may directly influence the production of adhesins and multilayered structures that contribute to biofilm formation (33).

hVISA infections are more prevalent in patients with indwelling medical devices (34), which can lead to prolonged stay lengths, a greater risk of persistent bacteremia, and a high frequency of treatment failure (3, 35–37). Our work has provided insights into the correlation between clinical characteristics of hVISA infections and biofilm formation. This is believed to be the first comprehensive analysis of the biofilm-forming capacity of hVISA clinical isolates and strains laboratory-induced vancomycin resistance, based on semiquantitative detection of biofilms and expression of genes related to biofilm formation. The increase in biofilm formation by hVISA strains may proceed through FnbA- and PIA-dependent pathways, and the upregulation of atlA and sarA may also contribute to the enhanced biofilm formation of hVISA upon prolonged exposure to vancomycin.