Inhibition of Staphylococcus epidermidis Biofilm by Trimethylsilane Plasma Coating

Abstract

Biofilm formation on implantable medical devices is a major impediment to the treatment of nosocomial infections and promotes local progressive tissue destruction. Staphylococcus epidermidis infections are the leading cause of biofilm formation on indwelling devices. Bacteria in biofilms are highly resistant to antibiotic treatment, which in combination with the increasing prevalence of antibiotic resistance among human pathogens further complicates treatment of biofilm-related device infections. We have developed a novel plasma coating technology. Trimethylsilane (TMS) was used as a monomer to coat the surfaces of 316L stainless steel and grade 5 titanium alloy, which are widely used in implantable medical devices. The results of biofilm assays demonstrated that this TMS coating markedly decreased S. epidermidis biofilm formation by inhibiting the attachment of bacterial cells to the TMS-coated surfaces during the early phase of biofilm development. We also discovered that bacterial cells on the TMS-coated surfaces were more susceptible to antibiotic treatment than their counterparts in biofilms on uncoated surfaces. These findings suggested that TMS coating could result in a surface that is resistant to biofilm development and also in a bacterial community that is more sensitive to antibiotic therapy than typical biofilms.

INTRODUCTION

Biofilm formation on or within indwelling medical devices, such as catheters, artificial cardiovascular implants, prosthetic joints, and contact lenses, poses a critical problem for medical care (20, 24, 26). Approximately 1 million nosocomial infections are associated with indwelling devices annually in the United States, incurring enormous health care cost (23).

Biofilms consist of organized communities of multispecies microbial cells (86). Bacteria in biofilms behave differently from planktonic bacteria, demonstrating high levels of resistance to antibiotics and host immunity (24, 26). The inherent resistance of bacteria in biofilms to antibiotics and host immune responses makes the eradication of biofilms from medical devices difficult. Biofilm infections are treated by removal or replacement of the infected medical devices, combined with systemic antibiotic therapy (23). However, the resistance of biofilms to antibiotics makes antibiotic treatments less effective. In addition, antibiotic therapy could also exacerbate the antibiotic resistance problems among human pathogens (55). As a result, much effort has been made to prevent biofilm formation on medical devices.

The development of biofilms can be roughly divided into the following steps. Bacterial cells attach to and colonize the implanted medical devices in the initial attachment step, followed by proliferation, accumulation, and maturation on the device surfaces. Once a biofilm is established, it forms a persistent source for bacterial dissemination and infection because bacterial cells or intact sections of biofilm can detach from the biofilm and spread to other parts of the host (20, 29, 62). Much effort has also been put into modifying the biomaterial surface properties to prevent bacterial attachment and reduce the number of persistent pathogens (66). The most commonly used substances for surface coatings are antibiotics (3, 55, 74). However, long-term usage of antibiotics risks selecting for resistance among pathogens (41, 52). Other materials that also demonstrated antibiofilm function when used to coat biomaterial surfaces include heavy metal ions (48, 85), furanone (6, 43), hydrophilic polyethylene glycol (PEG) derivatives (38), zwitterionic polymers (15), superhydrophobic coatings synthesized from a mixture of nanostructured silica colloids and a low-surface-energy fluorinated silane xerogel (71), and antimicrobial quaternary ammonium groups fixed to solid surfaces by silane coupling agents (103). Covalently coupled quaternary ammonium silane coatings on silicone rubber yielded a positively charged, highly adhesive surface with antimicrobial effects on adherent Gram-positive and Gram-negative bacteria (33). Quaternary ammonium-functionalized silica nanoparticles also exhibited enhanced inhibition of the growth of Gram-negative and Gram-positive bacteria compared to pristine silica nanoparticles (81). A lipopeptide, polymyxin B, was covalently bound to the solid surface via a silane coating with epoxy groups, and the resulting surface exhibited high bactericidal activity for Escherichia coli without leaching into the environment (60). In another study, surfaces with terminal methyl (CH₃), amino (NH₂), and hydroxyl (OH) groups were prepared by self-assembly of alkyl silane monolayers on glass. Bacterial adhesion was found to be dependent on the monolayer's terminal functionality. It was highest on the glass with the terminal CH₃ group, next highest on the glass with NH₂, and minimal on the glass with OH for two strains of Staphylococcus epidermidis (47). A study of silane-based coatings deposited using an atmospheric pressure plasma jet system showed that the superhydrophobic surfaces exhibit antimicrobial properties and significantly reduce protein adsorption (83). Plasma-deposited coatings using silane (SiH₄) and nitrous oxide as precursor gases also showed some effectiveness in reducing bacterial adhesion to the material surface (57). Bacterial adhesion and biofilm formation were also significantly decreased by texturing polyurethane surfaces with ordered arrays of pillars with submicron geometries (100).

While some of the technologies demonstrated in vivo efficacy (9, 31, 36, 43, 98), many technologies have yet to translate in vitro efficacy into in vivo success. The possibilities of discovering antimicrobial technologies with improved potency against bacteria in biofilms and of potentiating the antibiofilm activity of existing antimicrobial agents also hold long-term promise. However, application of these antibiofilm technologies is limited by a lack of clear criteria for designing and performing clinical trials to evaluate the efficacy of the antimicrobial agents (or antibiotic potentiators) (55).

Staphylococcus epidermidis is one of the most prevalent pathogens involved in biofilm formation on medical devices (29, 34, 62, 63, 93). S. epidermidis is part of the normal human epithelial bacterial flora but can cause infections when the skin or mucous membrane is injured. S. epidermidis can develop into biofilms and become a persistent source of device-associated infections. Antibiotic resistance is widespread in S. epidermidis and further limits treatment options (63).

In the present work, a nanoscale thin layer of coating was applied to 316L stainless steel (SS) and grade 5 titanium alloy (Ti), which are widely used in implantable medical devices, through the low-temperature plasma deposition process using trimethylsilane (TMS) as a monomer. A variety of biofilm assays were conducted on the coated substrates and uncoated controls. The resultant TMS plasma coating shows a very promising effect on antibiofilm formation by S. epidermidis, as evidenced by the in vitro tests, which warrants further evaluation of biofilm inhibition by in vivo animal studies with a view to future clinical applications.

MATERIALS AND METHODS

SS and Ti substrates.

316L SS and high-strength Ti (grade 5, also known as Ti₆Al₄V because of the addition of aluminum and vanadium alloying elements) wafers, approximately 10 mm by 10 mm by 1 mm, were cleaned with a 3% (vol/vol) detergent 8 (Alconox, Inc., White Plains, NY) solution for 3 h at 50°C in an ultrasonic bath. During the cleaning time, samples were removed from the solution every 30 min, rinsed with distilled water, and placed in fresh detergent solution. After cleaning, the metallic wafers were rinsed with acetone and blotted dry with Kimwipes paper.

TMS plasma coating on SS and Ti wafers.

The wafers were then mounted to a sample-holding frame made of an aluminum panel with three rows of metal clips (Fig. 1). The sample holder was hung vertically inside plasma reactor. Each square wafer was gripped by a metal clip at one corner to maximize surface area for plasma treatment and coating. Additionally, three silicon wafers were mounted to the sample holding frame at the top, middle and bottom locations for coating thickness assessment in each plasma coating batch. The sample holder was then placed inside an 80-liter bell jar-type reactor and situated between two titanium plates. In this configuration, the central aluminum sample holder served as the cathode, whereas the two outer titanium panels served as the electrically grounded anodes. This arrangement is typical of a substrate-as-electrode setup in which the sample to be modified serves as a working electrode (cathode). In this scheme, the electrodes were connected to the output of a MDX-1K magnetron drive (Advanced Energy Industries, Inc., Fort Collins, CO), which served as a DC power source. For a better understanding of the low-temperature plasma process system, the entire reactor setup is indicated in Fig. 2, which shows the electrode assembly in relation to placement inside the vacuum reactor. Oxygen plasma was used to remove organic contaminants on the surfaces of the substrates, i.e., SS and Ti. The reactor was sealed and evacuated to a base pressure (<2 millitorrs) using a mechanical pump and booster pump in series. Pure oxygen (Praxair Inc., Danbury, CT) was then introduced into the reactor at a flow rate of 1 standard cubic centimeter per minute (SCCM) using an MKS mass flow controller (MKS Instruments Inc., Andover, MA) and an MKS 247C readout to set the flow rate. The pressure inside the plasma reactor was allowed to stabilize at 50 millitorrs using an MKS pressure controller. The oxygen was then excited with the DC power supply at 20 W in order to form the plasma. The treatment time was 2 min. Following surface cleaning with oxygen plasma, the reactor was evacuated to the base pressure, and TMS (Gelest Inc., Morrisville, PA) was introduced into the reactor at 1 SCCM. The reactor pressure was allowed to reach 50 millitorrs, and the TMS was excited by the DC power supply at 5 W for 15 s.

Fig 1.

Aluminum frame with three rows of metal miniclips for holding wafers to be coated with TMS plasma inside the plasma reactor.

Fig 2.

Plasma reactor system used for TMS coating deposition.

TMS coating thickness assessment.

The thicknesses of TMS plasma coatings on SS and Ti wafers were determined based on the measurements of coatings deposited on Si wafers for each coating batch. The thicknesses of TMS coatings on silicon wafers were assessed using an AutoEL-II automatic ellipsometer (Rudolph Research Corporation, Flanders, NJ), which is a null-seeking type device utilizing a helium-neon laser with a wavelength of 632.8 nm and directing the laser beam coming to the silicon surface at an incident angle of 70°. The AutoEL-II ellipsometer can be used to find the refractive index and thickness of a single thin film on a reflective substrate such as a silicon wafer. The ellipsometer measures the change in state of a polarized laser beam upon reflection from a substrate surface. The state of polarization is determined by two intermediate parameters: the amplitude ratio of the parallel and perpendicular components of the light and the phase shift difference between the two components. The two parameters are used by a computer program embedded in the instrument to calculate the thickness and refractive index of a single-layer coating on substrates.

Surface characterization.

Surface characterization was performed using a noncontact optical profilometer, X-ray photoelectron spectroscopy (XPS), and contact angle to evaluate the surface topography, chemical composition, and surface wettability, respectively.

Surface chemistry analysis.

In order to better understand how plasma coating affects the surface chemistry of SS and Ti, all the plasma-coated wafers and uncoated controls were analyzed using XPS at the Material Research Center, Missouri University of Science & Technology, Rolla, MO. The XPS analysis of a surface provides qualitative and quantitative information on all the elements present (except hydrogen and helium) from the binding energies of the main lines and the peak area, respectively. A Kratos AXIS 165 X-ray photoelectron spectrometer (Kratos Analytical Inc., Chestnut Ridge, NY) equipped with a monochromatic Al Kα X-ray (1,486.6 eV) source operating at 150 W was used to characterize the elemental composition and chemical bonding states of the elements present at the substrate surfaces. The takeoff angle of the X-ray source was fixed at 90° to the substrate surface, for an area of 200 μm by 200 μm to be analyzed at a depth of 1 to 10 nm from the top surfaces of the substrates. Survey spectra, from a 0- to 1,200-eV binding energy, were recorded at 160 eV pass energy, a dwell time of 500 ms, and 1 scan, whereas the high-resolution spectra were taken at 20 eV pass energy, 0.1 eV/step, a dwell time of 500 ms, and a total of 12 scans averaged. The relative atomic concentration of elements detected by XPS was quantified on the basis of the peak area in the survey spectra with sensitivity factors for the Kratos instrument used. High-resolution spectra were charge compensated by setting the binding energy of the C1s peak to 284.5 eV. Peaks (C1s) were fitted (Gaussian/Lorentzian curves) after background subtraction (Shirley type with CASA XPS [Casa Software Ltd.] version 2.3.15), taking into consideration Scofield sensitivity factors, so as to determine the peak components or chemical states and their elemental concentrations.

Surface contact angle.

The static water contact angle was determined at room temperature using deionized water. The contact angle formed between a sessile drop and its supporting surface is directly related to the forces at the liquid-solid interface, indicating the hydrophilic or hydrophobic characteristics of the surface. The water droplet size used in the contact measurements was 1 μl, and the measurements were performed and recorded using a computer-aided VCA-2500XE video contact angle system (AST Products Inc., Billerica, MA).

Surface profilometry.

Surface morphology was measured using a Wyko NT9100 optical profilometer (Veeco Instruments, Inc., New York, NY) and Vision (version 4.10) software. This optical profiler performs noncontact, three-dimensional (3-D) surface measurements using a vertical scanning interferometry (VSI) mode for a wide variety of topologies. The samples were mounted in a horizontal position for measurements. The measurements were made over an approximately 125-μm by 94-μm area. Each sample was scanned in 5 locations. The scan resolution was 500 nm laterally and 0.5 nm vertically. Before the topography parameters were calculated, raw data were processed with a tilt correction. From the corrected and smoothed data, the surface roughness parameters R_a (arithmetic average of the absolute values of vertical deviations from a mean plane), and R_q (the root mean square roughness) were derived. R_a and R_q were used to characterize the surface roughness before and after TMS plasma coating. The values were expressed in units of height (nm).

Bacterial strain.

The S. epidermidis strain ATCC 35984 (RP62A) was isolated from a patient with device-associated sepsis (18). This strain was demonstrated to be a high biofilm producer (17). RP62A was kindly provided by the Network on Antimicrobial Resistance in Staphylococcus aureus program (NARSA).

Assessment of biofilm formation by crystal violet staining.

Biofilm formation was first measured by crystal violet staining (13). Four groups of 1- by 1-cm wafers (uncoated SS, TMS-coated SS, uncoated Ti, and TMS-coated Ti) were used in the experiments. Wafers were sterilized with UV lamps at a wavelength of 253.7 nm for 20 min on each side and then coated with 20% (vol/vol) human plasma in 50 mM carbonate buffer (pH 9.5) overnight (12). The contents of one capsule of carbonate-bicarbonate buffer (Sigma-Aldrich, St. Louis, MO) were dissolved in 80 ml distilled and deionized water and mixed with 20 ml human plasma to generate 20% human plasma in 50 mM carbonate buffer (pH 9.5). The human plasma was purchased from Innovative Research (Innovative Research, Inc., Novi, MI), with sodium citrate as the anticoagulant.

After human plasma adsorption, wafers were placed into wells of 24-well flat-bottomed sterile microtiter plates (TPP Techno Plastic Products AG, Trasadingen, Switzerland). An overnight culture of S. epidermidis was diluted at 1:200 in Todd-Hewitt broth containing 0.2% yeast extract (THY) medium with 0.5% glucose. Aliquots (1 ml) of the diluted bacterial suspensions were inoculated into the wells containing wafers precoated with human plasma and incubated for 48 h at 37°C, and the medium was changed every 12 h. Triplicate wafers of each group (uncoated SS, TMS-coated SS, uncoated Ti, and TMS-coated Ti) were used.

The wafers were washed four times with phosphate-buffered saline (PBS) to remove nonadherent bacterial cells. Biofilms on wafers were dried at 37°C for 1 h and stained at room temperature with 2.3% (wt/vol) crystal violet (CV) (Sigma-Aldrich, St. Louis, MO) for 30 min. The wafers were rinsed four times with PBS to remove excess stain. Biofilm formation was quantified by solubilization of the CV stain in 100% ethanol. The concentration of CV was determined by measuring the optical density at 595 nm (OD₅₉₅) with a microplate reader (Molecular Devices, Sunnyvale, CA). The experiment was performed three times to obtain means and standard errors of the means.

Biofilm slime assay.

Biofilms formed on four groups of wafers (uncoated SS, TMS-coated SS, uncoated Ti, and TMS-coated Ti) as described above, and slime was measured as reported before (6, 90). To exclude the possibility that TMS coating could affect the binding of toluidine to metal wafers, wafers were also incubated with culture medium without S. epidermidis inoculum as a blank control. Triplicate wafers of each group were used in each experiment. Biofilm samples on the wafers were fixed with Carnoy's solution (glacial acetic acid, chloroform, and absolute alcohol; 1:3:6 [vol/vol]) for 30 min and stained with 0.1% toluidine solution (Sigma-Aldrich, St. Louis, MO) for 30 min. The wafers were subsequently incubated in 0.2 M NaOH solution and heated in a water bath at 85°C for 1 h, and OD₅₉₀ was measured. The mean OD₅₉₀ of control wafers in each group was used as a blank to calibrate the value of slime formation in each group. The experiment was performed three times to obtain means and standard errors of the means.

Quantification of biofilm formation.

Biofilms were formed on the four groups of wafers in triplicate as described above with or without human plasma pretreatment. The wafers were washed four times with PBS to remove nonadherent bacterial cells. The wafers were put into tubes with PBS. The biofilms on wafers were detached and disaggregated with an ultrasonic bath treatment (35, 50). The wafers were sonicated six times in an ultrasonic bath (120 V, 50/60 Hz) (Fisher Scientific, Pittsburgh, PA) for 30 s and vortexed on the highest setting for 30 s after each sonication. The number of bacterial cells in PBS was quantified using the spread plate technique. The experiment was performed three times to obtain means and standard errors of the means.

In a separate experiment, biofilms were formed on the four groups of wafers precoated with human plasma in triplicate as described above. The wafers were washed eight times with PBS, and bacteria were enumerated as described above. The experiment was performed three times to obtain means and standard errors of means.

In order to assess the efficiency of sonication at disassociating bacteria from wafers, biofilms were formed on six uncoated SS wafers as described above. Three wafers were sonicated as described above. Bacteria on the wafers without sonication, bacteria in the sonication solution and bacteria on the wafers after sonication were digested with lysostaphin and proteinase K, and their genomic DNA was isolated with an EZ cell DNA isolation kit (EZ Bioresearch, St. Louis, MO). SYBR green-based quantitative PCR (qPCR) was performed with DNA samples from the three groups. The 16S rRNA gene was used to measure the amount of DNA. The primers for 16S rRNA gene were 5′CTGGTAGTCCACGCCGTAAAC3′ (forward) and 5′CAGGCGGAGTGCTTAATGC3′ (reverse). Forty cycles of PCR were carried out at 94°C for 45 s, 65°C for 30 s, and 72°C for 30 s on a CFX96 real-time PCR detection system (Bio-Rad, Richmond, CA). The relative DNA concentrations were calculated using the 2^−ΔΔCT method (54).

Primary attachment assay.

Primary attachment assay was carried out following a previously reported method (42, 61, 96) with minor modification. The four groups of wafers in triplicate were sterilized and coated with human plasma as described above. Overnight cultures of S. epidermidis were diluted in fresh THY with 0.5% glucose to an OD₆₀₀ of 0.02 and grown at 37°C to an OD₆₀₀ of 0.5. Aliquots (1 ml) were then pipetted into sterile wells containing wafers and incubated at 37°C for 1 h, 2 h, and 4 h, respectively. Culture supernatants were gently removed with a pipette. The wafers were washed four times with PBS and then put into tubes with PBS. The wafers were sonicated twice in an ultrasonic bath (120 V, 50/60 Hz) (Fisher Scientific, Pittsburgh, PA) for 30 s and vortexed twice on the highest setting for 30 s each time. The number of bacterial cells in PBS was quantified using the spread plate technique. The experiment was performed three times to obtain means and standard errors of means. In order to assess the efficiency of sonication of 1 min at disassociating bacteria from wafers, biofilms were allowed to form on six uncoated SS wafers for 4 h as described above. Three wafers were sonicated as described above. DNA from bacteria on the wafers without sonication, bacteria in the sonicated solution, and bacteria on the wafers were isolated and quantitated by qPCR as described above.

Fluorescent staining of adherent bacteria.

Biofilms were allowed to form on the four groups of wafers in triplicate as described above. The wafers were gently washed four times with PBS to remove nonadherent bacterial cells. Adherent bacterial cells were stained using a Live/Dead BacLight viability kit (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions, followed by four PBS washings to remove nonspecific stain. Fluorescence-adherent bacteria were visualized with a Zeiss LSM 510 confocal laser scanning microscope (CLSM) (Carl Zeiss MicroImaging GmbH, Jena, Germany). Images were acquired from random locations within the biofilms formed on the wafers. Three-dimensional structural reconstruction of CLSM image stacks was performed using Imaris 4.0 (Bitplane AG, Zurich, Switzerland).

Scanning electron microscopy (SEM).

Biofilms were formed on the four groups of wafers for 48 h in triplicate as described above. The wafers with biofilms were gently washed four times with PBS to remove nonadherent bacteria and fixed with 2% glutaraldehyde–2% paraformaldehyde in 0.1 M cacodylate buffer (pH 7.4) for 2 h at 4°C. The surfaces were washed three times with 0.1 M cacodylate buffer (pH 7.4) and subsequently fixed with 0.1% osmium tetroxide for 1 h. After being washed in ultrapure distilled water, the biofilm samples were dehydrated by replacing the buffer with increasing concentrations of ethanol (20%, 50%, 70%, 90%, 95%, 100%, 100%, and 100%) for 15 min each. After critical-point drying and coating by platinum sputter, samples were examined using a Quanta 600F scanning electron microscope (FEI Company, Hillsboro, OR).

Antibiotic susceptibility assay of S. epidermidis cells attached to wafers.

The susceptibility of bacterial cells in biofilms to antibiotics was studied with a protocol established previously, with modifications (1). Biofilm formation on the four groups of wafers was performed in triplicate as described above. The wafers were gently washed four times with PBS to remove nonadherent bacterial cells. Then fresh Mueller-Hinton broth (MHB) with 0, 2, 8, 32, and 128 μg/ml ciprofloxacin (Sigma-Aldrich, St. Louis, MO) was added to the wells containing wafers, and wells were incubated at 37°C for 48 h, with MHB medium with the appropriate concentration of ciprofloxacin being changed every 12 h. The wafers were washed four times with PBS and then put into tubes with PBS. The biofilms on wafers were detached and disaggregated with an ultrasonic bath treatment. The number of bacterial cells was determined using the spread plate technique. The experiment was performed three times to obtain means and standard errors of means.

Statistics.

Topographical and contact angle measurement comparisons were performed using one-way analysis of variance (ANOVA). Differences between control and TMS-coated wafers were compared by two-tailed unpaired Student's t test. Experiments were repeated independently a minimum of three times using three random samples for each repeat. Differences with a P value of less than 0.05 among the groups were considered to be statistically significant.

RESULTS

TMS coating thickness.

With ellipsometric measurements using an AutoEL-II automatic ellipsometer, the thickness of the resultant TMS plasma coatings deposited on Si wafers was found to be 20 to 30 nm, with a refractive index of about 1.62. This calculated refractive index of TMS coatings is in a good agreement with results reported elsewhere (5, 27). This ellipsometry method has been widely used for thickness measurement of thin coating layers (5, 104, 107). Of note, the thickness of the oxide layer on actual SS or Ti surfaces generated with O₂ plasma pretreatment prior to TMS plasma coating is estimated to be 3 to 4 nm based on our previous studies (13) and therefore could be considered negligible compared to the thickness of TMS coatings (20 to 30 nm).

Surface chemistry.

XPS is a quantitative spectroscopic technique that measures the elemental composition of the surface, usually from the top 1 to 10 nm of the material. As a result, for the wafers with 20- to 30-nm TMS plasma coatings, XPS detected only the elements within the coating material rather than the underlying substrate material, i.e., SS or Ti.

The elemental composition of wafers with and without TMS plasma coating is listed in Tables 1 and 2. These data were calculated from survey scans of substrate surfaces. The major elements on the SS control (no plasma coating) surface were C, O, Fe, Cr, and Si and trace amounts of N, Mo, Ni, and F, which were expected considering the chemical composition of bare SS (Table 1). The presence of C was mainly due to the contaminant organic species adsorbed to the metal surface. Oxygen can be attributed to the protective oxide layer that always forms on SS surfaces. With the TMS plasma coating deposited on the SS surface, more C and Si were detected, which is reasonable because the TMS monomer [(CH₃)₃-SiH] contains three C atoms and one Si atom. The presence of oxygen on the coating surface was likely due to the surface oxidation that occurred when the coating material was exposed to air. The theoretical atomic composition of pure TMS excluding hydrogen was also included in Tables 1 and 2 for direct comparison to the TMS plasma coating deposited on the SS or Ti surface.

Table 1.

Surface elemental composition of SS (316L) as determined by XPS survey scan and theoretical composition of pure TMS (excluding hydrogen) for direct comparison to TMS plasma coating

Sample	Atomic fraction for elements (%)
	C	N	Si	O	Fe	Cr	Mo	Ni	F
Uncoated SS (control)	44.43	1.15	2.09	35.60	10.04	4.32	0.53	0.46	1.38
TMS-coated SS	56.18	0.00	26.57	17.25	0.00	0.00	0.00	0.00	0.00
Pure TMS (theoretical)	75.00	0.00	25.00	0.00	0.00	0.00	0.00	0.00	0.00

Table 2.

Surface elemental composition of Ti (Ti₆A₁₄V) as determined by XPS survey scan and theoretical composition of pure TMS (excluding hydrogen) for direct comparison to TMS plasma coating

Sample	Atomic fraction for elements (%)
	C	Zn	Si	O	Ti	Al	Na
Uncoated Ti (control)	39.19	0.37	0.00	43.20	13.38	3.86	0.00
TMS-coated Ti	56.38	0.00	25.79	17.62	0.00	0.00	0.21
Pure TMS (theoretical)	75.00	0.00	25.00	0.00	0.00	0.00	0.00

The data in Table 2 indicate that on the surfaces of Ti control wafers (without TMS plasma coating), besides the Ti and Al from the bulk material, a large percentage of carbon was present, primarily due to organic contaminants on the surface. Oxygen was believed to come from the protective oxide layer that typically forms on the metal surface. The presence of Zn in a trace amount may be due to contamination during the sample-handling process. As discussed above, with TMS coating, it is expected that more carbon and silicon will be present on the TMS-coated Ti surface, because monomer TMS [(CH₃)₃-SiH] contains three carbon atoms and one Si atom. Again, the oxygen on the TMS coating could be attributed to the oxidation of the coating material after its exposure to the atmosphere, which has been reported to be observed on the surface of many other plasma-deposited coatings (14, 27, 39, 77, 101).

In order to analyze the chemical state of carbon on the substrate surface, the convolution of high-resolution peaks of C1s was conducted for the four types of samples, i.e., control SS, SS with TMS plasma coating, control Ti, and Ti with TMS plasma coating. Figure 3 shows the high-resolution peaks and the subpeaks under each peak generated from curve fitting. Compared to uncoated SS (Fig. 3), the surface with TMS coating exhibits more C component, with a binding energy of 284.5 eV, indicating a large amount of CH₃ formed on the surface. A similar phenomenon was noticed on the TMS-coated Ti surface as well, indicative of functional CH₃ groups generated at the surface regardless of the underlying bulk material (Fig. 3). Sabata et al. also found C—C/C—H, C—O, and CO on the surfaces of TMS plasma coating using XPS and similar atomic concentrations of C, Si, and O (77). Barranco et al. reported similar atomic concentrations of C, Si, and O on TMS coating surfaces as well (5).

Fig 3.

High-resolution spectra of C1s for SS (316L) and Ti (Ti₆A₁₄V) substrates before and after TMS coating. (A) C1s of SS without TMS coating; (B) C1s of SS with TMS coating; (C) C1s of Ti without TMS coating; (D) C1s of Ti with TMS coating.

Contact angle.

Contact angle measurements of uncoated controls and wafers with TMS plasma coatings indicated that the TMS coating rendered the surfaces of both 316L SS and Ti hydrophobic, as reflected by the contact angle increase from about 70° to around 100° (71° ± 5° and 66° ± 7° for uncoated SS and Ti, respectively; 106° ± 1° and 99° ± 6° for coated SS and Ti, respectively). Surface contact angles of TMS plasma coatings were reported in a previous work (80), and they are in agreement with our data.

Surface roughness.

The optical images of the surface topography of SS and Ti substrates before and after TMS plasma coating are presented in Fig. 4, and the corresponding surface roughness data calculated by the optical profilometer are summarized in Table 3. There was no significant difference in surface roughness for both materials, characterized by the average roughness (R_a) and the root mean square roughness (R_q) (Table 3), between TMS-coated and as-received substrates without plasma coating. This observation was supported by the statistical analysis result, a P value of >0.05, obtained using one-way ANOVA. In other words, the surface roughness was largely determined by the bare SS and Ti substrates, and a thin layer of TMS plasma coating in the thickness of 20 to 30 nm did not substantially change the underlying substrate roughness at a level of 200 to 450 nm, even though the standard deviations of surface roughness for TMS-coated SS or Ti substrates appeared to be smaller than those for uncoated counterparts. It was also noted that the as-received Ti substrates had rougher surfaces than as-received SS.

Fig 4.

Typical surface topography of commercial SS (316L) and Ti (Ti₆A_l4V) as received and the TMS plasma coatings on top of them, obtained with a Wyko NT 9100 profiler. Optical images for scanning areas of approximately 125 μm by 94 μm are shown in panels A1 to D1, and the corresponding surface profiles are in panels A2 to D2.

Table 3.

Surface roughness analysis of SS and Ti without and with TMS coating

Sample	Surface roughness (nm)a
	Ra	Rq
Uncoated SS (control)	229.14 ± 56.56	312.44 ± 79.38
TMS-coated SS	208.44 ± 20.57	292.14 ± 32.96
Uncoated Ti (control)	336.22 ± 43.85	423.73 ± 52.72
TMS-coated Ti	340.68 ± 25.06	431.03 ± 28.63

^aMean ± standard deviation (n = 5).

Biofilm formation on uncoated and TMS-coated SS and Ti wafers.

A crystal violet staining assay was performed to measure biofilm formation on both SS and Ti wafers. Crystal violet staining is used widely to measure biofilm formation because it is a less labor-intensive and faster way to assess biofilm formation than viable cell counting (69).

A significant reduction of crystal violet staining was observed on TMS-coated wafers compared to uncoated controls (Fig. 5A). More than 98% reduction in biofilm staining was observed on both TMS-coated SS (P < 0.001) and Ti wafers (P < 0.001) (Fig. 5A). The slime production was also significantly reduced on TMS-coated surfaces compared to uncoated controls, with a 97.8% reduction on SS wafers (P < 0.001) and a 97.5% reduction on Ti wafers (P < 0.001) (Fig. 5B).

Fig 5.

S. epidermidis biofilm formation on TMS-coated and uncoated surfaces. (A) Crystal violet staining of biofilm formation on four groups of wafers (uncoated SS, TMS-coated SS, uncoated Ti, and TMS-coated Ti) was quantitated by OD₅₉₅ reading of stain solubilized with ethanol. (B) Slime production of biofilms on the four groups of wafers. (C) Counts of viable bacterial cells from biofilms on the four groups of wafers. Data were pooled from nine samples (three independent experiments with triplicates) and are presented as means ± standard errors of the means. **, P < 0.01.

Similar results were observed (>99.5% reduction of biofilm formation) by counting bacterial cells recovered from biofilms on wafers with or without human plasma pretreatment (Fig. 5C). No significant difference in biofilm formation was observed between wafers treated with human plasma and untreated wafers (Fig. 5C).

For SS wafers without human plasma pretreatment prior to biofilm testing, 4.2 × 10⁸ ± 0.6 × 10⁸ CFU bacteria were recovered from biofilms on uncoated SS surfaces, while only 1.1 × 10⁶ ± 0.2 × 10⁶ CFU were recovered from biofilms on TMS-coated SS surfaces (P < 0.001). When human plasma pretreatment was applied, 4.2 × 10⁸ ± 0.8 × 10⁸ CFU were recovered from biofilms on uncoated SS surfaces, while only 1.6 × 10⁶ ± 0.2 × 10⁶ CFU were recovered from biofilms on TMS-coated SS surfaces (P < 0.001). Similarly, for Ti surfaces without human plasma pretreatment before biofilm testing, 4.9 × 10⁸ ± 0.7 × 10⁸ CFU were recovered from biofilms on uncoated wafers while 2.0 × 10⁶ ± 0.5 × 10⁶ CFU were recovered from TMS-coated wafers (P < 0.001). When human plasma pretreatment was used, 4.6 × 10⁸ ± 0.7 × 10⁸ CFU were recovered from biofilms on uncoated Ti surfaces while only 1.6 × 10⁶ ± 0.3 × 10⁶ CFU were recovered from biofilms on TMS-coated Ti surfaces (P < 0.001) (Fig. 5C).

Washing intensity in the methods of assessing biofilm formation varied among different investigators (3, 21, 25, 56). In order to assess the impact of washing intensity on estimation of bacterial load in biofilms, additional experiments were performed in which the biofilms on the wafers were washed with PBS for eight times. The number of bacteria recovered from the biofilms decreased with increasing washing intensity. A total of 1.7 × 10⁸ ± 0.3 × 10⁸ CFU were recovered from uncoated SS, while 3.9 × 10⁵ ± 0.6 × 10⁵ CFU were recovered from TMS-coated SS. A total of 1.8 × 10⁸ ± 0.5 × 10⁸ CFU were recovered from uncoated Ti, and 4.8 × 10⁵ ± 0.9 × 10⁵ CFU were recovered from TMS-coated Ti with higher washing stringency (Fig. 5C). Nevertheless, the inhibition of biofilm formation on TMS-coated surfaces was similar for different washing stringencies (>99.6%), suggesting that with proper control, similar results can be obtained under different washing conditions in biofilm assays.

Estimating the numbers of bacterial in biofilms involves removal of the bacteria from the surfaces by sonication. There are also various methods of using ultrasonication to dislodge biofilms (49, 50, 68, 88, 91, 92). Kobayashi et al. reported that 1 to 3 min of sonication helped release bacteria, while a longer period of sonication lysed bacteria (49, 50). As a result, we sonicated metal wafers with biofilms for either 1 or 3 min to dislodge biofilms. As a quality control, we isolated genomic DNA from bacteria on the wafers without sonication and compared it to genomic DNA from the wafers after sonication. qPCR was used to estimate how much residual DNA was recovered from wafers after sonication compared to wafers with biofilms without sonication by estimating the amount of 16S rRNA gene DNA in each DNA sample. The bacterial DNA isolated from the wafers after 1 min of sonication was approximately 7.1 ± 3.3% of DNA isolated from the wafers without sonication. The bacterial DNA isolated from the wafers after 3 min sonication was only 0.9 ± 0.2% of DNA on the wafers without sonication. The amount of DNA isolated from the biofilms on the wafers without sonication was not significantly different from that of bacteria in the sonication solution for both 1- and 3-min sonication treatments, suggesting that sonication was very effective at dislodging biofilms from the surfaces.

The structures of biofilms and bacterial colonization on metal wafers were visualized by CLSM. Bacterial cells were stained by the Live/Dead bacterial viability kit. Live cells with intact cell membranes would be stained with SYTO9 and emit a green fluorescence, whereas dead cells with damaged cell membranes would be stained with propidium iodide and emit a red fluorescence. Multilayer biofilms were formed on both uncoated SS and Ti wafers (Fig. 6) with tower structures (29, 86, 97). Strikingly, only sporadic cells or cell clusters were observed on TMS-coated wafers of both materials (Fig. 6). Similar results were obtained by SEM analysis (Fig. 7). Of note, there appeared to be fewer bacterial cells on the TMS-coated surface than expected from the viable cell counting estimation (Fig. 5C), suggesting that the additional steps in processing the samples for CLSM or SEM may decrease the numbers of bacteria detected on the surfaces comparing to CFU counting. Nevertheless, multiple methods of assessing biofilm formation all demonstrated that S. epidermidis was significantly less prone to develop biofilms on TMS-coated surfaces than on uncoated surfaces.

Fig 6.

Confocal laser scanning microscope studies of biofilm formation. One representative picture of three is presented. A) Biofilm formation on uncoated and TMS-coated SS wafers. (B) Biofilm formation on uncoated and TMS-coated Ti wafers.

Fig 7.

Scanned electron microscope studies of biofilm formation. One representative picture of three is presented. (A) Biofilm formation on uncoated and TMS-coated SS wafers. (B) Biofilm formation on uncoated and TMS-coated Ti wafers.

Bacterial attachment on uncoated and TMS-coated wafers.

A primary-attachment assay over a 4-h time course was performed to monitor the effect of TMS coating on the initial phase of biofilm formation (Fig. 8 and 9). At the 1-h time point, the number of bacterial cells adhering to TMS-coated wafers was only 4.7% of that adhering to uncoated SS wafers (1.3 × 10⁵ ± 0.6 × 10⁵ versus 2.8 × 10⁶ ± 0.5 × 10⁶ CFU; P < 0.002), and 2.4% of uncoated Ti wafers (6.1 × 10⁴ ± 3 × 10⁴ versus 2.5 × 10⁶ ± 0.6 × 10⁶ CFU; P < 0.003). At 2 h, the number of bacterial cells on TMS-coated wafers was 1.2% that of uncoated SS wafers (2.6 × 10⁵ ± 1.1 × 10⁵ versus 2.1 × 10⁷ ± 0.4 × 10⁷ CFU; P < 0.001) and 0.8% that of uncoated Ti wafers (1.8 × 10⁵ ± 0.8 × 10⁵ versus 2.2× 10⁷ ± 0.8 × 10⁷ CFU; P < 0.02). At 4 h, the number of bacterial cells on TMS-coated wafers was less than 1% that of uncoated controls (Fig. 8).

Fig 8.

Primary attachment assay over a 4-h time course. (A) Biofilm formation on uncoated and TMS-coated SS wafers. (B) Biofilm formation on uncoated and TMS-coated Ti wafers. Data were pooled from nine samples (three independent experiments with triplicates) and are presented as means ± standard errors of the means. *, P < 0.05; **, P < 0.01.

Fig 9.

Confocal laser scanning microscope studies of biofilm attachment assays. One representative picture of three is presented. Biofilm formation at 1 h (A), 2 h (B), and 4 h (C) on uncoated and TMS-coated wafers is shown.

As shown in Fig. 9, on the uncoated control wafers (both Ti and SS), there were scattered bacteria on the surfaces at 1 h. At 2 h, bacterial cells increased on the surface and started to form multilayer clusters. At 4 h, multilayer biofilms with complicated 3-D structures were formed. On the other hand, only sporadic bacterial colonies were observed on TMS-coated wafers at 1 h, 2 h, and 4 h.

Susceptibility of bacteria on uncoated and TMS-coated wafers to antibiotics.

The susceptibilities of bacterial cells on both TMS-coated and control wafers to antibiotic treatment were tested with ciprofloxacin. Bacteria in biofilms were treated with different concentrations of ciprofloxacin ranging from 2 to 128 μg/ml. The minimal bactericidal concentration (MBC) for RP62A biofilms, which achieves 99.9% killing of bacteria in biofilms, is 128 μg/ml (73) or more (72).

Ciprofloxacin treatment reduced bacterial counts on both TMS-coated and uncoated wafers in a dose-dependent manner (Fig. 10). Ciprofloxacin was able to decrease the number of bacteria recovered from biofilms on uncoated surfaces by more than 99% (2 log₁₀ units) (99.6% ± 0.07% on SS surfaces and 99.3% ± 0.1% on Ti surfaces) at 128 μg/ml, but the decrease in bacterial number did not reach 99.9% (3 log₁₀ units). On the other hand, bacterial cells on TMS-coated wafers were markedly more susceptible to ciprofloxacin treatment (Fig. 10). The number of bacteria on TMS-coated surfaces treated with 8 μg/ml ciprofloxacin was less than 0.01% (−4 log₁₀ units) of bacteria on the bare surfaces with neither TMS coating nor ciprofloxacin treatment. Numbers of bacteria on TMS-coated surfaces treated with 32 μg/ml ciprofloxacin were less than 0.1% (−3 log₁₀ units) of the numbers on TMS-coated surfaces without ciprofloxacin treatment, indicating that MBCs of ciprofloxacin for biofilms on TMS-coated surfaces were less than 32 μg/ml. Few bacteria were recovered from the TMS-coated surfaces (177 ± 80 on TMS-coated SS surfaces and 444 ± 217 on TMS-coated Ti surfaces) treated with 128 μg/ml ciprofloxacin.

Fig 10.

Biofilm response to ciprofloxacin treatment. (A) Viable cell counts from biofilms on TMS-coated and uncoated SS wafers treated with different concentrations of ciprofloxacin. (B) Viable cell counts from biofilms on TMS-coated and uncoated Ti wafers treated with different concentrations of ciprofloxacin. Data were pooled from nine samples (three independent experiments with triplicates) and are presented as means ± standard errors of the means.

DISCUSSION

Monomer TMS [(CH₃)₃-SiH] can be activated and deposited rapidly onto substrate surfaces with good adhesion through a low-temperature plasma coating process. Plasma-deposited organosilicon coatings not only exhibit dense film as conventional plasma coatings but also provide a certain level of abrasion resistance for medical implant surfaces due to their inorganic Si-Si and Si-C-Si backbones. Strong adhesion can be established through the formation of a chemical bond between the plasma-deposited layer and the substrate surface. A thin layer of TMS plasma coating has been widely investigated for corrosion protection of aluminum alloys (102, 105). In the present work, we demonstrated that the TMS plasma coating could significantly decrease S. epidermidis biofilm formation. S. epidermidis is one of the major pathogens responsible for biofilm-related infections (34, 62, 63, 93). Prevention of S. epidermidis biofilm formation on medical devices thus is of great interest to the health care community.

Analysis by SEM and CLSM demonstrated that S. epidermidis could form a multilayer biofilm with a complicated 3-D structure on uncoated materials. However, only scattered bacterial cells were observed on TMS-coated wafers. Slime production was also markedly diminished due to decreased biofilm formation. Exopolysaccharide slime is a critical component of the biofilm matrix and an important virulence factor (46, 64, 94, 95). The reduction of slime production was likely due to a reduced number of bacteria on TMS-coated surfaces. Decreased slime production will also likely make bacteria in biofilms more susceptible to the host immune system (46, 64, 94, 95), which could potentially diminish the virulence of biofilm infections. Attachment and colonization of S. epidermidis were substantially inhibited on TMS-coated surfaces. The bacteria initially attached to TMS-coated wafers failed to progress to later stages of biofilm formation with mature biofilm architecture (29, 86).

Bacteria in biofilms are highly resistant to antibiotics and host immunity (24, 26, 29). Biofilm-associated bacteria are particularly resistant to antibiotic treatment compared to planktonic organisms, probably due to the unique structure of biofilms, which prevents antibiotics from reaching the bacteria, or the altered biofilm microenvironment, which could inactivate antibiotics (62). Furthermore, antibiotics mainly target active cell processes, leading to limited efficacy against bacteria in biofilms, which are different from planktonic bacteria physiologically. Depletion of nutrition and accumulation of waste within biofilms could induce a slow-growing or starved state in which the bacteria are resistant to antibiotics. Additionally, some bacteria may adopt a distinct biofilm phenotype in response to growing on surfaces, which also decreases their sensitivity to antibiotics (20, 30, 62, 84). Based on the observation that S. epidermidis failed to develop mature multilayer biofilm structure on TMS-coated materials, we hypothesized that S. epidermidis cells on TMS-coated wafers would be more susceptible to antibiotic treatment than their counterparts on uncoated wafers.

While bacteria in biofilms exhibited great tolerance to ciprofloxacin, with MBCs of more than 128 μg/ml, bacteria on the surfaces of TMS-coated materials demonstrated significantly increased susceptibility. The result suggests that bacteria on the surfaces of TMS-coated materials had yet to adopt the distinct biofilm phenotype that would make them highly resistant to antibiotic treatment. Another possible explanation for the increased susceptibility of bacteria on TMS-coated materials could be attributed to the structure of the bacterial community. While S. epidermidis developed a multilayer biofilm structure that could prevent penetration of antibiotics or could inactivate antibiotics on uncoated materials, S. epidermidis on TMS-coated materials consisted of mostly scattered cells easily accessible to antibiotics. This result is in good agreement with a previous study showing that S. epidermidis cells in monolayer biofilms had a comparatively low tolerance to antibiotics than bacterial cells in multilayer biofilms (73). Reduced slime production also contributed to increased susceptibility of bacteria in biofilms to antibiotics, since the exopolysaccharide biofilm matrix has been shown to hamper penetration of antibiotics into biofilms (20).

The surface contact data indicated that the TMS plasma coating rendered the SS and Ti substrate surface more hydrophobic because of the CH₃ functional groups and Si-based structure contained in the plasma coating. Biomaterials with hydrophobic surfaces might play a role in bacterial adhesion to the material surface. Sousa et al. reported that in bacterial adhesion assays performed in saline solution, the adhesion of S. epidermidis to the less hydrophobic acrylic was significantly lower than that to the more hydrophobic silicone (82). However, when there are proteins adsorbed to the surfaces of materials, the interaction of those proteins with bacteria could be more critical to bacterial adhesion, because the surface-adsorbed protein layer could essentially mask direct contact between the bacterium membrane and the material surface. Hydrophilic biomaterials like polyethylene glycol (PEG)-based polymers are found to exhibit good resistance to bacterial adhesion. Such anti-adhesive properties have been attributed to their highly hydrated layer, which serves as a physical and energy barrier effective in preventing microorganisms and proteins from approaching the surface (22, 76, 78, 79). The surface hydrophobicity or hydrophilicity is one of the factors influencing the surface antibacterial effects, but not the most critical one (98).

Bacterial attachment to a solid surface is also highly dependent on other surface properties of the material, such as its chemical composition and reactivity, electronegativity, surface roughness, and porosity (8, 98). The surface chemistry of materials also affects the adsorption of plasma proteins (75), which mediates attachment of bacteria to surfaces. It was reported that some types of bacteria, such as Listeria monocytogenes and Staphylococcus aureus, adhered to and formed biofilms on the surface of polydimethylsiloxane (PDMS) in bacterial attachment assays conducted in protein-containing nutrient broth medium or Trypticase soy broth (32, 45). PDMS is a silicone rubber that is widely used for medical devices and implants. However, S. epidermidis showed only limited adherence to PDMS (32). As expected from the molecular structure, the surface chemistry analysis using XPS confirmed the presence of C-Si and Si-O-Si at the surface of PDMS (45). But it is still unclear why certain bacteria adhere strongly to PDMS (51). The inhibition of bacterial adhesion observed on our TMS plasma-coated surface could be due to surface-bound CH₃ groups, which could result in a significantly different profile of proteins adsorbed to the TMS surfaces than that of proteins adsorbed to the controls, leading to different bacterial interactions.

The surface roughness of biomaterials has been recognized as a very important factor for surface-bacterium interactions. Many studies have shown that the surface roughness of biomaterials strongly influences the degree of bacterial attachment to surfaces (44, 59, 89). For instance, streptococcal adhesion was surface roughness sensitive and increased as the roughness of composite surfaces increased from 20 nm to 150 nm and 350 nm (58). S. epidermidis adhesion and growth were substantially higher on rough titanium surfaces, on a scale comparable to the length of bacterial cells, than on smooth surfaces (99). In contrast, S. aureus cells were found to have a greater propensity for attachment to mechanochemically polished titanium than the as-received titanium, even though the overall polished surfaces are considerably smoother than those of as-received titanium. It was speculated that the Ti surfaces with nanoscale surface features resulting from mechanochemical polishing developed a characteristic pattern that is more suitable for anchoring of spherical S. aureus cells (89). As described above, our data indicated that the surface roughness of uncoated and TMS-coated wafers was in the range of 200 to 450 nm, which is on the same order as the size of S. epidermidis (about 1 μm in diameter) (10) and S. aureus (about 1 μm in diameter as well) (37), and thus could have some effect on bacterial adhesion. However, the antimicrobial activity assay conducted in this work showed that the biofilm formation by S. epidermidis on the surfaces without TMS coating was much greater than that on the TMS-coated surfaces. Therefore, it can be deduced that the surface chemistry created by TMS plasma coatings on substrate surfaces is one of the main factors responsible for discouraging bacterial attachment.

There are also reports demonstrating that human plasma proteins affect bacterial adhesion to biomaterial surfaces. However, reports regarding the effect of human plasma proteins on bacterial attachment have been inconsistent, with some indicating that protein adsorption enhanced bacterial adhesion (7) and others suggesting that surfaces precoated with proteins reduced the adhesion of staphylococci to various biomaterials (11, 65). It is known that staphylococci have fibrinogen-binding proteins, and fibrinogen enhanced their adhesion to biomaterials (16, 67). While S. epidermidis exhibited weak interactions with albumin, the bacteria showed strong adhesion with fibronectin, due to ligand-receptor binding. As a result, surfaces with deposited fibronectin proteins were more favorable for bacterial adhesion than non-protein-coated surfaces (53). Albumin immobilized or physically adsorbed to silicone surfaces significantly reduced S. epidermidis adhesion compared to untreated silicone (87). It was reported that albumin had a much higher binding affinity for the hydrophobic (i.e., having a CH₃ terminus) than the hydrophilic (OH terminus) surface, having a lower degree of ordered structure when adsorbed to the former (75). The contact angles of the hydrophobic and hydrophilic surfaces were found to be 108° and 29°, respectively (4). In contrast, a recent study reported that the attachment of S. aureus to protein-treated superhydrophobic surfaces (with a contact angle of more than 150°) was not significantly different from that to the controls without protein treatment (83). On the other hand, human plasma increased bacterial adhesion to polytetrafluoroethylene but decreased bacterial adhesion to Dacron (106). Taken together, these results show that the overall impact of human plasma on bacterial adhesion is complicated and may vary with different bacteria and materials. In the present study, before the biofilm assay, both uncoated and TMS-coated wafers were treated with human plasma to mimic the in vivo host environment, where human plasma protein adsorption on the surface of medical implants could affect biofilm formation. Interestingly, no significant difference was observed between samples with and without human plasma pretreatment, suggesting that proteins preadsorbed to the TMS-coated surfaces might desorb when in contact with the bacterial medium. To better understand the complicated interaction of proteins with different biomaterial surfaces, future research will be needed to study proteins adsorbed to different surfaces, their dependence on adsorption time, their effect on bacterial adhesion, and their stability on the surfaces.

Currently, there are two major approaches to prevent biofilm formation on implantable medical devices. One is to coat the surface of biomaterials with bactericidal reagents, such as antibiotics. This approach, while effective, risks facilitating selection of antibiotic resistance among pathogens. There were reports that subinhibitory concentrations of bactericidal agents even induced or enhanced biofilm formation (41, 52).

Another approach is to coat the surface with inert reagents to prevent bacterial adhesion, as was done with our TMS plasma coating technology. Unlike the wet chemistry processes of applying antibacterial agents to biomaterials using silane coatings (2, 103), the use of low-temperature plasma for TMS coating deposition on substrates is a dry surface treatment technique to impart desired properties to the surface without affecting the property of the bulk materials. There are no waste chemicals to dispose of, making it an environmentally friendly process involving very few consumables. Therefore, plasma surface modification has become a powerful tool in solving surface preparation problems on biomedical materials (19, 98).

The novel TMS coating technology was able to inhibit bacterial attachment to the surface to a great extent and prevent biofilm maturation. Though TMS coating can reduce the biofilm bacterial count by more than 99%, it has yet to reduce the bacterial load by 3 log units (99.9% inhibition) as defined by biofilm MBC, which limits its potential in clinical application. Nevertheless, TMS coating markedly increased bacterial susceptibility to ciprofloxacin, suggesting that TMS coating could potentiate antibiotic treatment of biofilms. The present study tested biofilm formation only within 48 h, whereas biofilm-related infections in clinics could develop over longer periods. The first 6 h after implantation of medical devices has been reported to be a decisive period for preventing biofilm formation and for the long-term success of an implant (28, 40, 70), suggesting that inhibiting biofilm formation within a short period at an early stage could still be beneficial. Caution should be exercised in interpreting the in vitro data, because the culture conditions of biofilms in the laboratory setting is different from the complicated conditions in vivo. There have been many reports of modifying surface of biomaterials to reduce biofilm formation (2, 3, 6, 15, 38, 48, 71, 85, 100) while fewer reports have documented in vivo efficacy (9, 31, 36, 43, 98). Taken together, these observations show that further improvement and in vivo assessment are needed to explore the clinic relevancy of this coating technology, taking into account host variables in the clinical setting. Nevertheless, the lower MBCs of ciprofloxacin for biofilms on the TMS-coated surfaces indicate that the TMS coating technology could enhance antimicrobial therapy, as the biofilm MBC is considered an in vitro guide to antibiotic use in the clinical setting (72).