Staphylococcus aureus biofilm formation and tolerance to antibiotics in response to oscillatory shear stresses of physiological levels


  • Victoria Kostenko,

    1. Biofilm Engineering Research Group, Calgary Center for Innovative Technology, University of Calgary, Calgary, AB, Canada
    Search for more papers by this author
  • Mohammad Mehdi Salek,

    1. Biofilm Engineering Research Group, Calgary Center for Innovative Technology, University of Calgary, Calgary, AB, Canada
    2. Department of Mechanical and Manufacturing Engineering, University of Calgary, Calgary, AB, Canada
    Search for more papers by this author
  • Pooria Sattari,

    1. Department of Mechanical and Manufacturing Engineering, University of Calgary, Calgary, AB, Canada
    Search for more papers by this author
  • Robert John Martinuzzi

    1. Biofilm Engineering Research Group, Calgary Center for Innovative Technology, University of Calgary, Calgary, AB, Canada
    2. Department of Mechanical and Manufacturing Engineering, University of Calgary, Calgary, AB, Canada
    Search for more papers by this author

  • Editor: Mark Shirtliff

Correspondence: Robert John Martinuzzi, Department of Mechanical and Manufacturing Engineering, University of Calgary, MEB 212, 2500 University Drive NW, Calgary, AB, Canada T2N 1N4. Tel.: +1 403 220 6627; fax: +1 403 282 8406; e-mail:


Bacterial infections in the blood system are usually associated with blood flow oscillation generated by some cardiovascular pathologies and insertion of indwelling devices. The influence of hydrodynamically induced shear stress fluctuations on the Staphylococcus aureus biofilm morphology and tolerance to antibiotics was investigated. Fluctuating shear stresses of physiologically relevant levels were generated in wells of a six-well microdish agitated by an orbital shaker. Numerical simulations were performed to determine the spatial distribution and local fluctuation levels of the shear stress field on the well bottom. It is found that the local biofilm deposition and morphology correlate strongly with shear stress fluctuations and maximum magnitude levels. Tolerance to killing by antibiotics correlates with morphotype and is generally higher in high shear regions.


Staphylococcus aureus are among the most important pathogens causing bloodstream infections associated with indwelling medical devices, including intravascular catheters, prosthetic heart valves, and aortic prostheses (Beathard & Urbanes, 2008; Dolmans et al., 2009). Device-related bacteremia leads to secondary infections including infective endocarditis, osteomyelitis, and septic arthritis (Lesens et al., 2004; Monk et al., 2008), and is the major factor of mortality and morbidity in hospitalized patients (Salzman & Rubin, 1995). Development of the infection requires device replacement and intensive antimicrobial treatment (Kassar et al., 2009), which increases patient risk and health care costs.

The pathogenesis of device-associated S. aureus infections results from the pathogen's ability to colonize the medical device surface or compromised tissue and form biofilms, well-organized cooperative communities of bacterial cells (Jain & Agarwal, 2009). Biofilm morphology and physiology confer onto the bacteria resistance to host defense mechanisms and antimicrobial agents (Roberts & Stewart, 2004; Kostenko et al., 2007). Consequently, the treatment of biofilm-associated infections poses severe challenges. A perhaps subtle, but nevertheless important complicating factor, as observed by several researchers, is that the development, morphology, and susceptibility of biofilms are affected by the surrounding flow environment and, in particular, the local hydrodynamic shear stresses acting on the fluid–biofilm interface (Simoes et al., 2003; Tsai, 2005; Paris et al., 2006; Salek et al., 2009a).

Biofilms formed on blood indwelling devices are exposed to shear stresses caused by blood flow and device flush. Normally, cardiovascular flow is predominantly laminar, with wall shear stress magnitude ranging from 0.5 to 10 Pa (Dyverfeldt et al., 2009). Under these (physiological) conditions, bacteria were observed to colonize endothelial and blood cells (Isberg & Barnes, 2002). However, adhesion is weak and bacterial cells are easily removed by blood flow (Ymele-Leki & Ross, 2007). Device insertion causes flow disturbances, which results in local turbulent flow conditions (Sotiropoulos & Borazjani, 2009; Tuka et al., 2009), increased wall shear stresses (Mareels et al., 2007), nonuniform distribution of the shear stress magnitudes over the wall (Ge et al., 2008), and oscillation of wall shear stress magnitudes (August et al., 2007). These flow irregularities play a role in the pathogenesis of atherosclerosis and thrombus formation, and increase the risk of hemolysis (Malek et al., 1999). Variations and fluctuations in hemodynamically induced surface shear stress distributions have also been shown to play a role in mammalian cell infection. For example, even the relatively small physiological fluctuations in shear stresses that occur in cerebral capillary blood flow have been shown to accelerate the colonization of the endothelial cells by Neisseria meningitidis. The attached bacterial cells were resistant to the external forces generated by the bloodstream (Mairey et al., 2006), similar to biofilms developed under a continuous turbulent flow (Simoes et al., 2009). However, the role of blood flow disturbances in the pathogenesis of device-associated infections remains poorly understood.

The present study investigated the impact of oscillatory surface shear stresses of physiological level on methicillin-resistant S. aureus (MRSA) biofilm formation, morphology, and tolerance to killing by antibiotics. For this purpose, biofilm properties were analyzed under different shear levels and gradients of shear magnitude. Oscillating surface shear stress fields were generated along the bottom face of a well of six-well microdishes through the fluid motion induced by the translation of an orbital shaker operating at different speeds and by varying the fluid volume contained in the well. Additionally, spatial distributions of biofilm deposition, morphology, and tolerance to antibiotics were analyzed with respect to variations in shear distribution. The shear fields were characterized based on numerical simulations validated by experimental measurements.

Materials and methods


MRSA UC18 (ST5-MRSA-II), isolated from a catheter-associated infection site, was kindly provided by the Biofilm Research Group (University of Calgary). MRSA UC18 was determined to carry genes fnbA, fnbB, clfA, ebpS, and map/eap encoding surface proteins, the ica gene, and type II of the agr locus. Gene profiling was performed by JoAnn McClure and Kunyan Zhang (Center for Antimicrobial Resistance, University of Calgary). Previously, MRSA UC18 was shown to form biofilms of different morphology and extracellular polymeric substances (EPS) production according to environmental conditions such as substratum surface topography and the presence of antimicrobials (V. Kostenko, unpublished data).

Biofilm formation

MRSA biofilms were developed on circular glass coupons in six-well tissue culture dishes filled with either 2 or 4 mL of the bacterial suspension of 107 CFU mL−1 in tryptic soy broth (TSB). The coupons (nominal diameter of 35 mm) covered the entire bottom of wells and were not free to move relative to the well wall. Plates were agitated with an orbital shaker (orbital radius of R=9.5 mm) at either 100 or 200 r.p.m. for 24 h at 37 °C. Then, biofilm-covered coupons were removed from the plates and washed with PBS to remove loosely attached cells. To determine the overall biofilm deposition, six independent coupons for each condition were transferred into fresh TSB (1 mL), sonicated with an Aquasonic sonicator 170HT (VWR International) at 40 kHz for 10 min to resuspend bacterial cells from biofilms as described by Ceri et al. (1999). The numbers of resuspended cells were determined with plate counts. For determining the spatial distribution of the biofilms, pieces of 4 mm2 were cut around the center of the coupon, and at distances of 6 and 12 mm from the center. Three cut-outs from six independently cultured coupons were analyzed for each condition. Viable cell numbers for these samples were determined as described above.

Tolerance to antibiotics

Biofilms were grown on coupons in six-well plates with agitation of either 100 or 200 r.p.m. (six independent coupons for each condition) for 24 h at 37 °C as described above. Then, biofilm-covered coupons were washed and transferred into challenge plates with 2 mL of Mueller–Hinton broth supplemented with 10, 100, and 500 × minimal inhibitory concentrations (MICs) of ciprofloxacin or vancomycin. MIC were determined according to Clinical and Laboratory Standards Institute (2005) guidelines. The MICs of ciprofloxacin and vancomycin were 0.25 and 2 μg mL−1, respectively. The plates were incubated for 24 h under static conditions at 37 °C. Then, viable cell numbers within biofilms deposited around the center of the well and 6 and 12 mm away from the center (three cut-outs from six independent coupons) were determined as above. Tolerance to killing by antibiotics was determined as the percentage of viable cell numbers within the biofilms having survived antibiotic treatment relative to the number of cells within the biofilms before treatment.

Scanning electron microscopy and energy-dispersive X-ray analysis (EDAX) of the biofilms

Biofilms were grown as described above. Then, the biofilms on the coupons were fixed with 5% glutaraldehyde, dehydrated in serial ethanol dilutions, and were viewed with a FEI ESME XL30 SEM (Microscopy and Imaging Facility, University of Calgary) to observe biofilm morphology and distribution. Five view fields around the center of the coupon and 6 and 12 mm away from the center were analyzed for six independent coupons for each experimental condition. The size of the colonies and areal porosity (the void area within a colony over the total area covered by the colony) were analyzed using imagej image analysis software. The content of carbon (C) and phosphorus (P) in biofilm colonies was determined using EDAX as atomic percentages. The P/C ratio was used to characterize biofilm composition with respect to the presence of bacterial cells and EPS. As determined by Fagerbakke et al. (1996), the P/C ratio for bacterial cells varied from 0.04 to 0.09. We assume that a decrease in the P/C ratio below this level indicates an increase in the amount of C-containing material such as EPS.

Statistical analysis

At least six independent experiments were performed. A one-way anova was used to analyze the data. Coefficients of P<0.05 were considered statistically different. Spearman's rank correlation coefficient test was used to characterize a relationship between hydrodynamic parameters and biofilm properties.

Wall shear stress analysis in agitated six-well plates

The flow field and wall shear stress distribution as a function of time within a single well of a six-well plate undergoing orbital motion was simulated using the commercial CFD software package fluent (Version 6.3.26, Fluent Inc., Lebanon, NH). The well geometry was modeled and the numerical grids were generated in gambit. The well diameter and depth are 35 and 18 mm, respectively. The unsteady, incompressible form of the governing equations for conservation of mass (continuity) and momentum were solved in a stationary frame of reference. A no-slip condition was imposed at the walls of the container and a free-surface (constant pressure, zero net shear) condition at the liquid– air interface. The container is rigid and the motion of the container can be calculated directly from the motion of the orbital shaker. Because each well undergoes the same orbital motion, modeling was performed for a single well. Forces resulting from the movement of the container are transmitted through the boundary conditions for the velocity field at the solid surfaces. The orbital shaker imposes an in-plane, two-dimensional (2-D) sinusoidal translation on the table and thus the plate walls. The acceleration of the walls appears as a force or an impulse source term, which is imposed at the fluid domain boundaries. Essentially, the velocity of the fluid at the solid boundary takes on that of the moving wall given as: vx=−RΩsinΩt, vy=−RΩcosΩt, vz=0, where x, y indicate horizontal and z vertical components; R is the radius; and Ω the rate of gyration of the orbital shaker. R was 9.5 mm. A dynamic mesh technique was used to implement the orbital motion. The transient boundary conditions for orbiting wells were defined using an external user-defined function linked to fluent.

A critical aspect of the simulation is the prescription for the dynamic constraints at the free surface, which impact the flow field and rate of mixing. The volume of fluid (VOF) method was chosen to simulate the free surface, which essentially requires that two-phase flow dynamics (liquid and air) be resolved. The volume fraction of each phase is determined in each computational cell and then the mean properties are calculated based on the volume fraction. Hence, just one set of momentum equations is solved, which includes both phases through the fluid properties of each computational cell. This method thus allows for the free-surface deformation and local fluid height fluctuations to be calculated at each step. Further simulation details can be found in Salek et al. (2009b). Simulations were conducted for Ω=100 and 200 r.p.m. for liquid volumes of 2 and 4 mL, the latter corresponding to static liquid levels of h=2.1 and 4.1 mm, respectively.

Surface shear rate simulation predictions were validated experimentally using an optical shear rate sensor (MicroS Sensor, Scientific Measurement Enterprise) flush-mounted at the bottom of a modified well. This nonintrusive sensor operates based on a modified Laser Doppler Anemometry technique. A linearly diverging interference fringe pattern is produced by a pair of closely spaced parallel slits illuminated with a laser source. With a known fluid viscosity, the frequency of the scattered light by each seed particle is directly related to the instantaneous wall shear stress component. The principles of sensor operation and design have been described in detail by Sattari (2008). Measurements were obtained at three radial locations from the center of the well (1, 6, and 12 mm) for all test conditions. The predicted average shear rates agreed with the experimental results.


Biofilm formation under oscillatory shear stresses

Biofilm deposition in the wells of six-well microdishes was analyzed under different cultivation conditions: (1) 4 mL of bacterial suspension agitated at 100 r.p.m., (2) 2 mL at 100 r.p.m., (3) 4 mL at 200 r.p.m., and (4) 2 mL at 200 r.p.m. The overall biofilm depositions on the well bottom are given in Table 1. The number of cells contained within biofilms grown at 200 r.p.m. were higher (P=0.02) when compared with those formed at 100 r.p.m., but the differences between biofilms in plates filled with either 4 or 2 mL at the same agitation speed were not statistically significant (P=0.35).

Table 1.   Average wall shear stress and biofilm density and ‘percent survival’ in a well under different experimental conditions (n=6)
Experimental conditions100 r.p.m. (4 mL)100 r.p.m. (2 mL)200 r.p.m. (4 mL)200 r.p.m. (2 mL)
  • *

    Number of cells before treatment with antibiotics.

Average shear stress on the bottom of the well (Pa)
Number of cells (log CFU per well)*6.1 ± 0.46.4 ± 0.27.1 ± 0.36.8 ± 0.4
% Survival after antibiotic exposure
Ciprofloxacin 10 × MIC55.9 ± 2.690.3 ± 1.495.3 ± 2.192.9 ± 0.8
Ciprofloxacin 100 × MIC29.5 ± 0.661.2 ± 1.168.6 ± 2.466.2 ± 1.2
Ciprofloxacin 500 × MIC16.4 ± 1.144.2 ± 1.655.2 ± 0.457.2 ± 0.8
Vancomycin 10 × MIC54.9 ± 1.476.3 ± 2.684.8 ± 1.883.5 ± 1.2
Vancomycin 100 × MIC6.9 ± 0.228.5 ± 0.838.1 ± 1.235.3 ± 0.4
Vancomycin 500 × MIC6.3 ± 0.127.1 ± 0.631.8 ± 0.828.9 ± 1.1

Biofilm deposition varied radially across the wells. The numbers of cells deposited within 4 mm2 segments around points at the center, 6 and 12 mm away from the center of the well are shown in Fig. 1. Biofilms formed in the wells filled with 4 mL and agitated at 100 r.p.m. contained 4.0 to 4.3 log CFU per segment within the region extending from the center to approximately a radius of 6 mm. The biofilm cell density increased significantly (P<0.01) outside this region. Biofilms developed in the wells filled with 2 mL and agitated at 100 r.p.m. contained a low number of cells (3.7 log CFU per segment) around the center of the well. However, the biofilm cell density increased significantly (P<0.01) and proportionally to the radial distance from the center. MRSA biofilms formed at 200 r.p.m. were similar for 2 and 4 mL media volumes and showed a relatively low cell density around the center. Outside a radius of approximately 4 mm from the center, the cell density was higher (P<0.01) and uniform (the difference between density at 6 and 12 mm was not statistically significant with P>0.05).

Figure 1.

 Number of MRSA cells accumulated within biofilms developed on the segments of 4 mm2 around either the center of the well (inline image) or the points 6 mm (inline image) and 12 mm (inline image) away from the center under different experimental conditions: (1) in a well filled with 4 mL of bacterial suspension and agitated with 100 r.p.m.; (2) 2 mL agitated with 100 r.p.m.; (3) 4 mL agitated with 200 r.p.m.; and (4) 2 mL agitated with 200 r.p.m. *Significant difference in the population density (P<0.01) within the block of data obtained for a separate experimental condition (n=18).

Biofilm morphology in response to oscillatory shear stresses

Figure 2 shows the different types of biofilms formed by MRSA exposed to varying shear stress fields. MRSA biofilms consisting of single cells or cell clumps <100 μm in diameter (Fig 2a) were observed inside a radius of 6 mm from the center of the well when incubated in 4 mL of a bacterial suspension at 100 r.p.m. (Fig. 3a) and inside a radius of 2 mm when incubated in 2 mL at 100 r.p.m. (Fig. 3b). P/C ratios ranged from 0.050 to 0.052, indicating low amounts of EPS within the colonies. As the radial distance increased, MRSA produced a biofilm consisting of large monolayer cell aggregates (0.1–0.4 × 0.8–2.5 mm) with <20% areal porosity (Fig. 2b), and elongated monolayer cell clumps up to 20 μm wide, organized in a web with 62–85% areal porosity (Fig. 2c). Both colony types contained low amounts of EPS as indicated by the P/C ratio, 0.052–0.056. Similarly, biofilms consisting of large aggregates (0.3–0.8 × 1.2–1.8 mm) within a web formed by elongated cell clumps (Fig. 2d and e) were observed over the coupon surface when incubated at 200 r.p.m. in either 2 or 4 mL (Fig. 3c and d). However, cell aggregates formed at 200 r.p.m. contained proportionally more EPS (as indicated by reduced P/C ratios ranging from 0.028 to 0.032) with thick 3-D structures, which are significantly thicker at the exterior of the edge than within the colony.

Figure 2.

 Biofilm colony types observed under different shear stress magnitude and shear gradient conditions. (a) Cell clumps without a preferred direction of growth; (b) large monolayer cell aggregates; (c) elongated monolayer cell clumps organized in a web; (d) large 3-D cell aggregates; (e) multilayer cell clumps. Scale bar=10 μm.

Figure 3.

 Biofilms developed around the center of the well or the points 6 and 12 mm away from the center under different conditions: (a) in 4 mL of suspension agitated at 100 r.p.m.; (b) 2 mL at 100 r.p.m.; (c) 4 mL at 200 r.p.m.; and (d) 2 mL at 200 r.p.m. The scale on the left shows the distance from the center to the wall of the well in mm. Images represent the average colonies observed for five view fields at six independent coupons. Scale bar=100 μm.

Tolerance of the biofilms developed under oscillatory shear stresses to killing by antibiotics

Table 1 summarizes the tolerance to ciprofloxacin and vancomycin for MRSA cells within biofilms grown under the test conditions described above. Note that the antibiotic challenges were conducted under static conditions. The survival of the biofilm cells grown at 100 r.p.m. was significantly lower than for those grown at 200 r.p.m. (P=0.001 for biofilms grown in 4 mL of suspension and P=0.01 for biofilms grown in 2 mL). The relative number of bacterial biofilm cells surviving the antibiotic challenge was higher for biofilms formed with 100 r.p.m. in 2 mL of the bacterial suspension than for those grown in 4 mL (P=0.001), whereas no significant difference was observed between survival of the MRSA within biofilms grown in 2 or 4 mL with agitation of 200 r.p.m. (P=0.73). The number of bacterial cells within biofilms grown in 4 mL of decreased at 100 r.p.m. decreased as the concentrations of either ciprofloxacin or vancomycin increased. In contrast, antibiotics reduced the bacterial populations at concentrations less than 100 × MICs under other conditions (2 mL at 100 and 200 r.p.m.; 4 mL at 200 r.p.m.), but further increases in antibiotic concentration did not affect biofilm viability (P>0.05).

Tolerance to antibiotics for the MRSA biofilms grown at an agitation of 100 r.p.m. depended on the location within a well (Fig. 4). The bacterial populations grown within 6 mm from the center in 4 mL of bacterial suspension demonstrated low survival capacity. Biofilms were completely eradicated by 100 × MICs of vancomycin and 500 × MICs of ciprofloxacin. In contrast, a significant number of biofilm bacteria survived 500 × MICs of antibiotics in the peripheral areas of the well. In plates filled with 2 mL of bacterial suspension, populations around the center of the well also exhibited lower survival capacity than biofilms developed in distal areas. However, these populations exhibited greater tolerance to antibiotics than those developed in 4 mL at the same location (P<0.01). Tolerance to antibiotics for the biofilm cells grown at 200 r.p.m. was independent of the location in the well and was comparable with the tolerance levels of the peripheral biofilm populations incubated in 2 mL at 100 r.p.m. (P>0.01).

Figure 4.

 Tolerance to vancomycin (left column) and ciprofloxacin (right column) for MRSA within biofilms deposited around the center of the well (inline image) or 6 mm (inline image) and 12 mm (inline image) away from the center under different experimental conditions: (1) in a well filled with 4 mL of bacterial suspension and agitated with 100 r.p.m.; (2) 2 mL agitated with 100 r.p.m.; (3) 4 mL agitated with 200 r.p.m.; and (4) 2 mL agitated with 200 r.p.m. Tolerance is represented as the percentage of viable cell number within the biofilms that survived antibiotic treatment compared with the number of cells within the biofilms before treatment. Antibiotics were applied at 500 × MIC (a); 100 × MIC (b) and 10 × MIC (c). *Significant difference in the population density (P<0.01) within the block of data obtained for a separate experimental condition (n=18).

Shear stress distribution on the bottom wall of a well

The wall shear stress distribution in a well of a six-well plate undergoing orbital motion was simulated for the four test conditions described above. The magnitude of the surface shear stress on the bottom of the well averaged over the well bottom area is constant and is given in Table 1. This average shear stress depends strongly on the rate of rotation. At 100 r.p.m., the liquid volume in the well significantly affects the average shear stress, whereas at 200 r.p.m., its effect is second order.

The average shear stress on the bottom of the well is a poor guide for interpreting the biological results. For any given point on the bottom surface of the well, the shear stress is periodic in time and varies with the location. Figure 5 shows the shear magnitude field for two instantaneous snapshots for each of the four test cases studied separated in time by half an orbital period. The stress field undergoes a rigid body rotation about the well center, even though the well does not rotate (the orbital motion is a pure translation). The acceleration of the peripheral well walls imparts an impulse to the fluid, which induces a traveling wave within the well that completes a circuit during each orbital period, and thus, the instantaneous flow field is adequately represented by a single snapshot. The resulting wave form (time series) of the shear stress magnitude at a fixed point on the well bottom is periodic. Around the center of the plate, the wave form is sinusoidal, while in the peripheral areas, the wave form resembles more a saw tooth (Fig. 6).

Figure 5.

 Bottom surface shear stress magnitude (Pa) generated under different conditions: (a) 4 mL at 100 r.p.m.; (b) 2 mL at 100 r.p.m.; (c) 4 mL at 200 r.p.m.; and (d) 2 mL at 200 r.p.m. Contours in the left column represent t=0; in the right column, t=0.5T. T is the period.

Figure 6.

y, wall shear stress (Pa) as a function of t/T over two cycles at 100 r.p.m., 2 mL (a) at x=1 mm from the center and (b) x=12 mm from the center. T is the period.

The rather complex shear stress distribution is a result of the free-surface dynamic behavior and shape. The shape of the free surface depends on both the rotation rate and the liquid volume in the well. Typically, the maximum shear occurs where the liquid level is the lowest, and variations in the local liquid levels result in fluctuations in shear levels. Table 2 provides a summary of the mean, peak-to-peak amplitude, maximum shear stress magnitudes, and the average temporal gradient (rate of change) of the shear stress at radial locations corresponding to those for biofilm analysis. The mean shear stress levels at 100 r.p.m. are nearly uniform over the entire well, but it is noted that these are higher for the test with 2 mL of liquid than for those with 4 mL. In contrast, at 200 r.p.m., the mean shear stress levels increase radially, while differences due to the liquid volume are second order. Fluctuations in the free-surface height above the well bottom increase radially from the center to the wall of the well. Shear levels close to the center of the well mainly undergo low-amplitude fluctuations over a cycle, but distal areas are exposed to a higher range (amplitude) of shear stresses. More significantly, the increase in the fluctuation amplitude implies higher local maxima and shear magnitude temporal gradients. Overall, changes in the dynamic parameters, rather than the mean shear stress levels, can be directly related to the differences observed in the biofilm morphology and susceptibility as shown in Table 3 and discussed below.

Table 2.   Oscillating shear stress field characteristics in a well of six-well microdishes at different rotation rates (100 and 200 r.p.m.) and media volume (4 and 2 mL)
 100 r.p.m. (4 mL)100 r.p.m. (2 mL)200 r.p.m. (4 mL)200 r.p.m. (2 mL)
Center6 mm12 mmCenter6 mm12 mmCenter6 mm12 mmCenter6 mm12 mm
Mean shear magnitude, Pa0.
Peak to Peak amplitude in shear magnitude in a cycle, Pa0.
Maximum shear magnitude, Pa0.160.200.300.280.440.471.
Average temporal gradient of shear magnitude, Pa s−
Table 3.   Correlation coefficients for the relationships between oscillatory shear stress field characteristics and biofilm properties
Peak to
of shear
Number of bacterial cells0.470.750.640.77
Biofilm morphotype0.830.860.890.88
Susceptibility to antibiotics0.770.620.770.64


In the present paper, the influence of fluctuating shear stress of physiological levels on MRSA biofilm formation and tolerance to antibiotics was studied in six-well microdishes agitated by an orbital shaker. These conditions can represent blood flow oscillations as can be generated by some cardiovascular pathologies or the insertion of indwelling devices (Malek et al., 1999; August et al., 2007), which are known to induce blood stream infections (Mairey et al., 2006; Dolmans et al., 2009). Previously, Azevedo et al. (2006) modeled the flow inside the well to study the effect of shear stress on the adhesion of Helicobacter pylori to stainless steel and polypropylene coupons. They demonstrated that the average wall shear stresses at the bottom of the wells increased with the rotation speed. The present study further shows that the dynamic behavior of the wall shear stress field is very important. For any given point on the bottom surface of the well, the shear stress is periodic in time and is a function of the location. Shear stress fluctuations result from the 2-D translation imposed by the orbital motion. The acceleration of the walls imparts a cyclical motion to the fluid, which induces a traveling wave within the well, which completes a circuit during each orbital period, generating an oscillatory stress field on the well walls. The occurrence of a surface wave introduces an additional, perhaps subtle, parameter: the liquid media level. The variations in the local liquid levels (free-surface dynamic) result in fluctuations in shear levels over the well. Thus, the resulting spatial distribution and temporal evolution of the flow and shear stress fields is complex and depends on the location within the well, radius and rate of rotation, and liquid level.

Depending on the rate of rotation and liquid level, the shear stress field under which the MRSA biofilms develop changes drastically (see Fig. 5), resulting in significantly different local mean and dynamic shear field characteristics (Table 2). The MRSA biofilms displayed different population densities, morphotypes, and tolerances to killing by antibiotics according to the shear regimes (Table 3). The cell density in the biofilm increased as both the shear stress magnitude and the temporal shear stress gradient increased (correlation coefficients of ρ=0.64 and ρ=0.77, respectively), but poorly correlated with mean shear magnitude (ρ=0.47). The biofilm morphotype correlated similarly well with the local shear level (ρ=0.89), shear temporal gradient (ρ=0.88), but also with the mean shear values (ρ=0.83). In the central area of the plate agitated at 100 r.p.m., where shear magnitudes and temporal gradients are low, MRSA formed cell clumps without a preferential direction of growth, which are typical for stagnant and laminar flow conditions (Francolini et al., 2004). Higher shear stress and fluctuation values in peripheral areas are associated with the formation of long monolayer colonies surrounded by elongated cell aggregates organized in an interconnecting porous web. Further increase in shear stress magnitudes and gradients at a higher rotation rate (200 r.p.m.) results in the formation of dense and thick cell aggregates embedded in EPS, a type of biofilm typically observed under continuous turbulent flow conditions (Paris et al., 2006).

Changes in the biofilm morphology strongly correlated with the tolerance of the embedded bacterial cells to antibiotics (ρ=0.89), while the direct correlation of the tolerance with shear magnitude and shear gradient was lower (ρ=0.77 and ρ=0.64, respectively). Because the tolerance was investigated under static conditions, the impact of the flow-driven mass transfer on antibiotic efficacy was negligible. Thus, the increased tolerance of the shear-exposed biofilm cells is attributed to morphological alterations such as the high density of cell aggregates and the presence of EPS induced by shear levels and their gradients.

The MRSA biofilm morphology features and tolerance to antibiotics observed in the present study under low (physiological), but oscillatory shear stresses (typically observed under some cardiovascular pathologies and indwelling device insertion) were similar to those observed in biofilms developed under high shear stresses in continuous flow systems. Continuous turbulent and high shear stress conditions were determined to induce the formation of the biofilms, which are denser and more resistant to mechanical stresses and antimicrobial treatment than biofilms grown under laminar low shear stress conditions (Simoes et al., 2003, 2009). In the present study, it is observed that oscillations of the shear stresses, even at low levels, promote biofilm communities to behave as the turbulent-flow phenotype. These alterations make bacterial communities recalcitrant to antimicrobial compounds and mechanical forces, at levels typically found in blood flows, and create a huge challenge in the control of blood stream infections, especially those associated with indwelling devices. While high shear rates are often associated with turbulent flows, it is possible to sustain turbulence at low shear rates. Turbulence is characterized by inherent unsteadiness in the velocity field and thus the instantaneous shear rates fluctuate about a mean. The root-mean-square fluctuations for near-wall turbulence typically range from 10% to 30% of the mean values (Hinze, 1975). These strong fluctuations enhance mass and momentum (i.e. shear) transport. A similar enhancement is observed in the present experiments due to oscillations (the flow remains essentially laminar). Hence, fluctuations in the velocity field, either arising due to turbulence or forced as in the present case, appear to play a similar role.


The growth, morphology, and tolerance to antibiotics for MRSA UC18 biofilms, formed at different locations on the bottom of wells of six-well microdishes agitated by an orbital shaker, were investigated. Four test conditions were considered: agitation of 100 and 200 r.p.m. (radius of gyration of 9.5 mm) for liquid volumes of 2 and 4 mL. The shear stress field was determined from numerical simulations. The translational motion of the shaker induces a rotation of the fluid about the center of the well and a traveling surface wave. As a result, the radial and tangential levels of the shear stresses on the well bottom depend on the location, time, rotational speed, and liquid fill volume.

It is found that changes in the deposition, morphology, and survival of the biofilms correlate strongly with the local maximum magnitude and the amplitude of fluctuations of the shear level. In particular, as the magnitude and fluctuation levels increase, the density and thickness of the biofilm colonies increase. The cell aggregates are generally elongated and contain a significant amount of EPS. The tolerance of the biofilms to vancomycin and ciprofloxacin is higher for biofilms grown under higher shear magnitude and fluctuation levels. Because the antibiotic challenges were conducted under static conditions, it can be concluded that tolerance results correlate with the biofilm morphology.

The present results thus show that shear stress fluctuation levels play an important role in determining the local biofilm morphology and, by consequence, tolerance to antibiotics. Thus, irregularities in the flow field introduced through indwelling devices, especially under pulsatile flow conditions, or the appearance of turbulent regions in the flow, can give rise to strong shear stress fluctuation levels that promote the formation of more tolerant biofilms and thus engender difficult-to-treat infection sites.


R.J.M. acknowledges the support of the Canadian Natural Sciences and Engineering Research Council (NSERC) through the Industrial Research Chair program. M.M.S. acknowledges funding provided by Alberta Ingenuity, now part of Alberta-Innovates Technology Future, through their Graduate Scholarship program.