Histotripsy Treatment of S. Aureus Biofilms on Surgical Mesh Samples under Varying Scan Parameters

A. Biofilm Preparation

Staphylococcus aureus subsp. Aureus (ATCC®25923™) biofilms were grown on Polypropylene mesh samples (PPKM301, Surgical Mesh Division, Textile Development Associates Inc., Brookfield, CT) as described in [18]. Briefly, the 10 mm × 10 mm mesh samples were individually placed in 20 ml of tryptic soy broth (TSB) that had been inoculated with 20 μl of stationary phase bacteria. The samples and TSB were then incubated for three days at 37°C. However, every 24 hours, the mesh samples were placed in a fresh solution of TSB so that the bacteria could continue to have an ample supply of nutrients. As was stated in our earlier paper, PPKM301 was selected as it is a small-diameter monofilament mesh with relatively large pores which seems to correlate with better clinical outcomes when conducting hernia repair [9, 22].

B. Ultrasound Alignment and Exposure of Infected Mesh

After growing the biofilms on the mesh samples, the biofilms were inserted into Aquaflex Ultrasound Gel Pad Standoffs (Parker Laboratories Inc., Fairfield, NJ) as shown in Figure 1.

Prior to insertion, the mesh samples were rinsed with 10 ml of sterile phosphate buffered saline (PBS) to remove any planktonic bacteria. To insert the mesh into the gel pads, the gel pads were first cut in half. Then, a slit was cut in the center of the gel pad allowing the mesh to be placed inside the pad. After insertion, 100 μl of sterile PBS was also injected into the slit with the mesh to minimize the chance of trapped air in the slit. The gel pads were used as an easily reproducible sterile tissue mimic as they have approximately the same elastic modulus as abdominal muscle [23–25]. The elastic modulus of the medium has a significant impact on the level of damage observed following ultrasound histotripsy experiments, and therefore must be matched when utilizing a material as a tissue mimic [26–28]. The gel pad with the inserted mesh was then placed inside an aluminum holder, and the position of the mesh relative to the edges of the holder was measured. The holder with the gel pad and mesh was then placed in a water bath as shown in Figure 2.

After being placed in the water bath, the focus of the ultrasound transducer was first aligned with the left and bottom internal edges of the holder as viewed when looking along the beam-axis in the direction of ultrasound propagation (left and bottom sides of the holder in Figure 1) as illustrated in Figure 2a. The alignment was done by using a low-power signal from a pulser-receiver (Panametrics 5900, Olympus Corporation, Tokyo, Japan) and noting when the signal from the edges began to drop off significantly. The transducer was then moved until the focus was approximately in the middle of the mesh based on the measurements taken of the mesh position prior to insertion into the water tank. The transducer was then moved along the beam axis in steps of 1 mm and transverse to the beam axis in steps of 350 μm until a maximum echo was received from one of the mesh fibers. This was done to ensure that the mesh was precisely positioned in the focal plane. Once aligned, the edges of the mesh were found by moving the transducer parallel to the plane of the mesh until the echo from the mesh could no longer be observed. For the alignment, the Panametrics 5900 was set to excite the transducer with a voltage spike at a pulse repetition frequency of 200 Hz. A lower voltage spike corresponding to a focal pressure of 0.076 MPa was used when finding the edges of the holder while a higher voltage spike corresponding to a focal pressure of 0.15 MPa was used when aligning on the mesh. These focal pressures were determined using our capsule hydrophone (ONDA HGL-0200, Onda Corporation, Sunnyvale, CA).

After alignment on the mesh, the mesh was randomly exposed to either a sham exposure or one of the histotripsy exposures as illustrated in Figure 2b. All the exposures were done with a single-element spherically focused transducer with an operating frequency of 1.1 MHz, a focal length of 12.5 cm, and a diameter of 12.5 cm (H-161C, Sonic Concepts Inc., Bothell, WA). The transducer was excited to transmit tone bursts at a pulse repetition frequency (PRF) of 1 kHz with a peak-compressional pressure of approximately 150 MPa and a peak rarefactional pressure of approximately 17 MPa. The therapy exposures were calibrated via a combination of hydrophone measurements and computer modelling as was explained in detail in our earlier publications [16, 26, 29, 30]. The 3 dB beam width for the transducer as measured by reflection from a wire target was 1.2 mm. For the exposures, we continuously scanned the transducer in the horizontal direction and stepped the transducer in the vertical direction once we reached the edge of the mesh treatment zone as illustrated in Figure 3. Also, rather than performing the treatment in a single pass, the treatment scan was performed 5 times for each mesh sample. Therefore, once the focal spot reached the end of the mesh, denoted Final Focal Location in Figure 3, it was returned to the Initial Focal Location, and the treatment was repeated 4 more times. To ensure that the entire mesh was exposed, the treatment dimensions varied from 12 to 16 mm on a side with 12 mm × 12 mm being the most common. The treatment dimensions for each mesh were based on the mesh edges identified from the alignment exposures.

For the exposures, the speed for the continuous portion of the scan was varied as 0.4 mm/sec, 0.8 mm/sec, 1.2 mm/sec, and 1.6 mm/sec corresponding to 0.33 beam widths/sec, 0.67 beam widths/sec, 1 beam width/sec, and 1.33 beam widths/sec respectively. In addition, the step size for the vertical portion was varied as 350 μm (0.3 beam widths) and 700 μm (0.6 beam widths). Lastly, the number of cycles in the tone burst was either 3 cycles or 10 cycles. Since the speed and step size varied for the different exposures, the total treatment time also varied from about 20 minutes for the fastest exposures to about 90 minutes for the longest exposures. For the sham exposures, ultrasound was off, the scan speed was 0.4 mm/sec, and the step size was 350 μm corresponding to the longest scan times. All of the different exposure conditions are summarized in Table I. Each exposure was repeated 7 times except for the sham exposure which was repeated 9 times. The extra repetitions for the sham exposure were due to requiring at least one sham exposure every week to help ensure the bacteria biofilms were growing as expected.

C. Quantification of Colony Forming Units (CFUs)

After each exposure was completed, the gel pad with the inserted mesh was removed from the holder. The mesh was then removed from the gel pad by trimming the gel pad with a sterile blade so that it could be separated by the slit that was originally cut in the gel pad when the mesh was inserted. The mesh was then lifted off the gel pad using sterile forceps and dipped into a tube of sterile PBS to remove any loose gel fragments. The mesh was then placed in a 1.5 ml microcentrifuge tube containing 0.5 ml of sterile PBS. The tube was then vortexed for 60 seconds and placed in a Symphony Ultrasonic Cleaner (1.9 L, VWR, Radnor, PA) and sonicated for two minutes. It was then vortexed again for 60 seconds, sonicated again for two minutes, and vortexed again for 60 seconds. The vortexing/sonication cycle would release and disperse the bacteria from the mesh without killing them. After the vortexing/sonication, 0.5 ml was removed and serially diluted in sterile PBS at dilutions of 10⁻¹, 10⁻², 10⁻³, 10⁻⁴, and 10⁻⁵. 100 μl of each dilution was then plated on tryptic soy agar and incubated overnight at 37°C.

In addition to processing the mesh, the gel in the vicinity of the mesh was also scraped by a sterile blade. The scrapped gel and fragments were then placed in the same container in which the mesh had been dipped (tube with 4.5 ml of sterile PBS). The tube of gel fragments was then vortexed for 3 minutes. After vortexing, 0.5 ml was removed and serially diluted in sterile PBS at the same dilutions of 10⁻¹, 10⁻², 10⁻³, 10⁻⁴, and 10⁻⁵. Following dilution, 100 μl was once again plated on tryptic soy agar and incubated overnight at 37°C.

The number of CFUs on each plate for both the mesh and gel dilutions were counted following the overnight incubation. The dilution that had counts between 30 and 300 was then used to back calculate the total number of CFUs remaining on the mesh and gel following the exposure. If two different dilutions had counts in the 30–300 range, the numbers were averaged after accounting for the dilution process. If all the counts were outside the 30 to 300 range, then dilution that resulted in a count that was closest to 300 was used. Lastly, if no CFUs were observed on the lowest dilution plate, then the count was set to zero for that exposure. When counting CFUs, numbers in the range of 30 to 300 are the most reliable as plates with fewer than 30 colonies may not give statistically reliable results and plates with more than 300 colonies can be too crowded to allow the bacteria to form distinct colonies. Therefore, focusing on counts in this range has become standard practice in microbiology.

After determining the number of CFUs for each exposure, we calculated log₁₀(count+1) to translate the count into log number for comparison. This transformation facilitates the calculation of the log₁₀ reduction in the number of CFUs. In addition, the transformed data for the gel for each of the exposures (including the sham exposure) follows an approximately normal distribution whereas the original count data does not. It is important for the data to follow a normal distribution if traditional statistical analysis approaches are to be used [31].

The transformed data for the gel was analyzed using a one-way analysis of variance (ANOVA) and the JMP software (Pro 12, The SAS Institute). The data included 51 observations, 9 for the sham and 7 for each of the six histotripsy exposures. Dunnett’s method [32] was used for simultaneously comparing the six histotripsy exposures with the sham, with an overall type I error rate of no more than 5%. The plot of residuals versus predicted values and the normal probability plot of the residuals indicate that the model was adequate.

The transformed data for the mesh also included 51 observations, 9 for the sham and 7 for each of the six histotripsy exposures. Because most of the exposed mesh samples had no CFUs, most of the transformed values are zero, invalidating the normality and equal-variance assumptions for the standard one-way ANOVA. Thus a more robust procedure was used [33] for pairwise comparisons between the seven exposures, with an overall type I error rate of no more than 5%. The procedure also gave 95% simultaneous confidence intervals for all pairwise differences. We used the R software (version 3.1.0, The R Foundation for Statistical Computing) to analyze the transformed data for the mesh because the robust procedure was implemented in the R software [33].

A. Mesh Results

Figures 4, 5, and 6 show the numbers of CFUs remaining on the mesh for the different treatments as a function of scan speed, step size, and the number of cycles in the tone burst respectively.

As was given in Table I, for the scan speed comparison, all of the speeds used a step size of 350 μm and 10-cycle tone bursts. Also, when comparing step size, the scan speed was maintained at 0.8 mm/sec with 10-cycle tone bursts. Likewise, the tone burst comparison used a 0.8 mm/sec scan speed and a step size of 350 μm. The sham exposure, while duplicated in each plot, was the same data set corresponding to a 350 μm step size and a scan speed of 0.4 mm/sec. The 0.8 mm/sec scan speed, 350 μm step size, 10-cycle exposure (the exposure condition B in Table I) in each plot was also the same data (with no CFUs on all the seven mesh samples). The error bars in each figure correspond to one standard deviation. The numbers above each bar in each plot give the number of repetitions for which CFUs were found on the mesh. Therefore, a 0/7 means that of the 7 repetitions, none of the meshes had any surviving CFUs for that exposure.

From these results, it is clear that the vast majority of the exposures had no CFUs remaining on the mesh following the treatment and all of the exposed meshes were statistically different from the sham exposures. The 0.8 mm/sec scan speed, 700 μm step size, 10-cycle exposure had the most meshes with CFUs (3 out of 7) and the highest average CFU count at 1.1±1.4-log10. However, none of the average CFU counts for the six histotripsy exposures were statistically different from each other. Table II summarizes the average reduction in CFUs for each of the exposures relative to the sham exposure.

The reduction in CFUs ranged from 3.2-log10 for the 700 μm step size to 4.3-log10 with an average reduction of 3.8-log10.

B. Gel Results

In addition to quantifying the number of CFUs surviving on the mesh samples, the surviving bacteria released onto the gel surrounding the mesh was also quantified. As a reminder, biofilms continuously release planktonic bacteria. Therefore, CFUs on the gel would include both spontaneously released bacteria and bacteria released from the biofilm due to the ultrasound exposure.

Figures 7, 8, and 9 show the numbers of CFUs remaining on the gel for the different treatments as a function of scan speed, step size, and the number of cycles in the tone burst respectively.

For the scan speed results, 0.4 and 0.8 mm/sec scan speeds had statistically fewer average surviving CFUs than the sham exposure. The 1.2 and 1.6 mm/sec exposures were not statistically different from the sham exposure. For the step size comparison, the average numbers of CFUs on the gel are not statistically different for the 700 μm step and the sham. However, there is a statistically significant difference between the 350 μm step and the sham as was already stated. Likewise, for the cycle number comparison, the 3-cycle exposure was not statistically different from the sham; however, the 10-cycle and sham exposures were statistically different from each other. In addition, we qualitatively observed less damage to the gel with the 3-cycle exposures which was consistent with our earlier observations regarding the dependence of collateral damage on pulse duration [18].