Acta Biomaterialia 6 (2010) 1107–1118
Contents lists available at ScienceDirect
Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat
Interactions of bacteria with specific biomaterial surface chemistries under flow conditions M.G. Katsikogianni, Y.F. Missirlis * Laboratory of Biomechanics and Biomedical Engineering, Department of Mechanical Engineering and Aeronautics, University of Patras, Rion, 26504 Patras, Greece
a r t i c l e
i n f o
Article history: Received 31 March 2009 Received in revised form 16 July 2009 Accepted 4 August 2009 Available online 9 August 2009 Keywords: Bacterial adhesion Self-assembled monolayers Surface characterization Shear rate
a b s t r a c t The effect of specific chemical functionalities on the adhesion of two Staphylococcus epidermidis strains under flow was investigated by using surfaces prepared by self-assembly of alkyl silane monolayers on glass. Terminal methyl (CH3) and amino (NH2) groups were formed in solution and by chemical vapor deposition of silanes, at elevated temperature. Hydroxyl (OH)-terminated glass was used as control. Surface modification was verified by contact angle and zeta potential measurements, atomic force microscopy and X-ray photoelectron spectroscopy. A parallel plate flow chamber was used to evaluate bacterial adhesion at various shear rates. The effect of the solution’s ionic strength on adhesion was also studied. Adhesion was found to be dependent on the monolayer’s terminal functionality. It was higher on the CH3 followed by the NH2 and minimal on the OH-terminated glass for both strains. The increase in the ionic strength significantly enhanced adhesion to the various substrates, in accordance with the Derjaguin–Landau–Verwey–Overbeek (DLVO) theory. The extended DLVO theory explained well the combined effects of surface and solution properties on bacterial adhesion under low shear rates. However, the increase in the shear rate restricted the predictability of the theory and revealed macromolecular interactions between bacteria and NH2-terminated surfaces. Ó 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction Microbial adhesion and subsequent biofilm formation occur in many medical applications, where implantation of artificial organs and medical devices, and therefore the use of biomaterials, has become indispensable. In spite of non-septic conditions during the surgical procedure and systematic administration of antibiotics, infection impedes the materials’ long-term use [1,2]. Microbial adhesion to the biomaterial surface is the critical step in the development of infections related to implanted or intravascular devices, and Staphylococcus epidermidis has been identified as a predominant cause of infection in the presence of a medical device due to its abilities to adhere to surfaces, to produce a mucoid substance, more commonly known as slime, and to form multilayered cell clusters, which are protected from antibiotic therapy, physiological shear and possibly host cell-mediated defenses [3–6]. Bacterial adhesion is mediated by interactions between the material and the bacterial surfaces [7]. Both specific (i.e. receptor–ligand), in the case of protein/cell coated surfaces, and nonspecific (i.e. colloidal-type) interactions contribute to the ability of the bacterial cell to attach to the biomaterial surface. However, their relative contributions are not completely understood [8,9]. * Corresponding author. Tel.: +30 2610969460; fax: +30 2610969464. E-mail address:
[email protected] (Y.F. Missirlis).
Moreover, for many of the materials, the surface chemistry is quite complex, and it is difficult to be sure of the chemical composition at the bacteria–biomaterial interface. Many commercially available materials may contain trace impurities and surface-active additives, such as antioxidants or processing aids, which further complicate interpretation of bacteria–material interactions and result in uncertainties concerning the types of functional groups present at the surface [10,11]. Modifications of the surfaces of polymers via plasma-processing techniques in order to improve their biocompatibility, usually produce numerous functional groups and chemical crosslinks on the surface [12–17], while treatments often cause severe degradation of the surface, leading to increased roughness as well as to surface heterogeneity [14,15]. Time-dependent conformational rearrangements of these surfaces may also be observed, in response to environmental conditions and changes [17]. Moreover, antibiotic loading, resistance and release rates over long periods are difficult to control [18]. Recently, much interest has arisen in self-assembled monolayers (SAMs), with the goal of developing molecular-level control over surface properties [19,21]. SAMs formed by the adsorption of terminally functionalized alkanethiols [HS(CH2)n–R] onto gold substrates [19,20] or terminally functionalized alkyltrichlorosilanes [Cl3Si–R] or alkyltriethoxysilanes [EtO3Si–R] onto hydroxylated silicon and glass surfaces [21,22] are structurally the best
1742-7061/$ - see front matter Ó 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2009.08.006
1108
M.G. Katsikogianni, Y.F. Missirlis / Acta Biomaterialia 6 (2010) 1107–1118
ordered interfaces currently available for studying the interaction of cells and proteins with substrates of different surface chemistries [23–25], and provide the capability of circumventing many of the aforementioned experimental uncertainties [24]. However, the preparation of SAMs by the absorption of organosilanes onto hydroxylated substrates in solution is rather tricky as the deposition of aggregated organosilane molecules frequently degrades the quality of the SAMs, increasing roughness and surface heterogeneity [26,27]. Since it has been shown that bacterial adhesion is influenced not only by the surface chemistry but also by roughness and configuration [28,29], we prepared CH3- and NH2-terminated SAMs at the vapor/solid interface in order to produce surface chemistries with similar topography and roughness. This enables the direct examination of the effect of the surface chemistry, energy and charge on bacterial adhesion. Moreover, we investigated how the solution ionic strength (I.S.) influences the bacterial and the material surface charge and consequently the bacterial adhesion, by varying the I.S. from 0.01 to 0.1 M. Furthermore, since the process of bacterial adhesion to indwelling medical devices is associated in most cases with flow of body fluids [30], physical forces such as shear generated by local hemodynamics may modulate the adhesion process. Although there is a lot of experimental work in the literature on the critical shear rate to prevent adhesion and to simulate detachment of already adhering microorganisms from glass and hydrophobic surfaces [31–33], little previous work exists on bacterial adhesion to model SAMs, other than hydrophilic and hydrophobic alkanethiols and protein-coated substrates, under a wide range of shear rates that correspond to the physiological ones [25,34,35]. For this reason, in this work we examined bacterial adhesion to OH-terminated glass, CH3- and NH2-terminated organosilanes under shear rates between 50 and 2000 s1, which correspond to the physiological shear rates for laminar flow in blood vessels. OH-, CH3- and NH2groups are present on metallic, ceramic and polymeric surfaces, as well as on biological substrates, and were chosen because their physicochemical properties are significantly different. The models that have been proposed for the quantitative prediction of bacteria–material physicochemical interactions and the explanation of the observed bacterial attachment are based on colloidal theories and macromolecular binding considerations [36–39]. The most prominent of the colloidal theories are the Derjaguin–Landau–Verwey–Overbeek (DLVO) [36], thermodynamic [37] and extended DLVO (XDLVO) theories [38]. More recent and detailed models account for other interactions such as macromolecular binding [39]. Although Meinders et al. [31] concluded that the deposition efficiency and desorption rates of bacteria at a wall shear rate of 50 s1 were better described by the XDLVO theory, the effect of higher shear rates on the predictability of the colloidal theories has not been extensively described and the fundamental mechanisms governing bacterial adhesion are still poorly understood. In this direction, this study attempts to answer in a fundamental way, by using the simplest possible chemistries, the following questions: does the surface chemistry and solution I.S. influence S. epidermidis adhesion on the SAMs? How does adhesion depend on shear rate and on the relative contribution of physicochemical and hydrodynamic interactions? Does the S. epidermidis adhesion behavior agree with the trends predicted by the thermodynamic, the DLVO and the XDLVO theories? These questions are addressed by quantitative measurement of bacterial adhesion on surfaces in laminar flow as a function of fluid shear rate. In this contribution, we describe the preparation and characterization of hydroxylated glass substrates and SAMs deposited from organosilanes, CH3- and NH2-terminated.
2. Materials and methods 2.1. Preparation of substrata 2.1.1. Chemicals Glass slides were purchased from Knittel Gläser, octadecyltriethoxysilane [ODS; H3C(CH2)17Si(OCH2CH3)3] from Gelest Inc.; aminopropyltriethoxysilane [APTES; H2N(CH2)3Si(OCH2CH3)3], hexane and methylene iodide from Sigma–Aldrich; toluene, HNO3, H2SO4, H2O2, Na2HPO4 and KH2PO4 from Merck; glycerol anhydrous from Fluka; and ethanol absolute and NaOH from Carlo Erba. 2.1.2. Pretreatment of glass slides Glass slides were hydrolyzed by immersion in NaOH aqueous solution 5 M for 1 h. Afterwards, they were soaked in fresh piranha solution (3:1 sulfuric acid 98%/hydrogen peroxide 30%) for 1 h. The piranha solution converted all groups to silanol groups (Si–OH) [40]. The hydroxyl (OH)-terminated glass slides were washed with Milli-Q water, ethanol, and hexane in the case of CH3-terminated monolayers, and dried for 120 min at 120 °C. The OH-terminated glass substrates were prepared immediately prior to silanization or kept under water till use in order to prevent the ageing effect [17]. 2.1.3. Preparation of self-assembled monolayers The OH-terminated glass slides were coated with self-assembled monolayers terminated by CH3 or NH2, following the vapor phase method [27]. In particular, the CH3-terminated monolayers were formed by placing the glass slides together with a glass cup filled with 0.2 ml of ODS into a 65 cm3 Teflon container. The container was sealed and placed in an oven maintained at 150 °C for 3 h. Afterwards, the glass substrates were rinsed twice with hexane and dried for 30 min at 80 °C. The NH2-terminated monolayers were also formed by placing a glass cup filled with 0.1 ml of APTES diluted with 0.7 ml of toluene together with the cleaned glass substrates into the Teflon container, under a dry nitrogen atmosphere in order to avoid APTES polymerization. The container was sealed and placed in an oven maintained at 90 °C for 1 h. Subsequently, the glass substrates treated with APTES were sonicated for 20 min successively in dehydrated ethanol and dehydrated toluene. The samples were then sonicated further in NaOH (1 mM) and HNO3 (1 mM) to remove any excess absorbed APTES molecules. Finally, the samples were rinsed with Milli-Q water and blown dry with a nitrogen gas stream. 2.2. Bacterial strains and culture conditions The bacterial strains used in this study were the reference type culture S. epidermidis ATCC 35984, which is slime producing, and ATCC 12228, which is not slime producing. Microorganisms were kept at 70 °C, in a solution containing 70% tryptic soy broth (TSB; Difco Laboratories) and 30% diluted glycerol (glycerol/water: 1/1). Before each experiment, 10 ll of the thawed bacterial suspension was subcultured onto tryptic soy agar (TSA; Difco Laboratories) for 24 h at 37 °C. Stationary-phase cells were obtained by incubating two or three colonies from the TSA in 5 ml of TSB for 18 h at 37 °C in a rotary shaker at 120 rpm. Cells were harvested by centrifugation at a centrifugal force of 2683g, at 4 °C for 10 min, washed twice with 0.01 or 0.1 M phosphate-buffered saline (PBS), pH 7.4, and finally resuspended in 0.01 or 0.1 M PBS at a concentration of 3 108 colony-forming units (CFU) ml–1, according to the McFarland standard (BioMerieux).
M.G. Katsikogianni, Y.F. Missirlis / Acta Biomaterialia 6 (2010) 1107–1118
2.3. Material and bacterial surface characterization 2.3.1. Surface topography and roughness The topography of the various surfaces was examined by means of a Multimode AFM (Nanoscope III, Veeco). Contact mode cantilevers and integrated silicon nitride tips (Veeco) were used. Height data from the 5 lm2 area images were processed using the first order flattening option and the average surface roughness (Ra) was evaluated using the Nanoscope software (Veeco). 2.3.2. Contact angle measurements and surface free energy evaluation The wettability of the various surfaces was determined by measuring the contact angles of three probe liquids with different polarities, using ultrapure water, methylene iodide and glycerol as the wetting agents. Measurements were made at room temperature and ambient humidity using the sessile drop technique [41]. The deposition of 3 ll droplets on each substrate was recorded and analyzed to obtain the contact angles, using the CAM 100 goniometer and the KSV 100 software (KSV Instruments Ltd.). In the case of the bacterial cells, the measurements were performed on bacterial layers deposited on membrane filters according to the method described by Busscher et al. [41]. Briefly, a suspension of each S. epidermidis strain in PBS 0.1 M (after washing) was deposited onto 0.2 lm filter in order to obtain a thick lawn of bacteria after filtration. The lawn of bacteria was then air-dried for 3 h, until the socalled ‘‘dried plateau” was obtained. At this time the moisture among the cellular exterior evaporated while the cells were not dehydrated and contact angles of each probe liquid could be measured. Three random locations were examined per slide and three independently prepared slides were analyzed for each substrate or bacteria. Measured contact angles of the three probe liquids were converted into surface free energies and their components, according to the Lifshitz–van der Waals acid–base (LW-AB) approach of the thermodynamic theory [37], using the equation:
ð1 þ cos hÞcL ¼ 2
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
LW þ cLW S cL
qffiffiffiffiffiffiffiffiffiffiffi
cþS cL þ
qffiffiffiffiffiffiffiffiffiffiffi
cS cþL
ð1Þ
LW S
in which c is the Lifshitz–van der Waals component of the surface free energy, cþ S is the electron acceptor and cS the electron donor parameters of the acid–base component of the surface free energy, where S is the substratum or the bacterial surface. For the liquids þ used, these parameters cLW L , cL ; cL , are known [37,42]. 2.3.3. Zeta potential measurements To determine the zeta potential of the substrata, 1 lm diameter colloidal beads of glass were treated with the CH3- and NH2-terminated silanes at elevated temperature, in the same way as the glass substrates, using the vapor phase method. The OH-, CH3- and NH2terminated glass beads were suspended in 0.01 or 0.1 M PBS. The concentration of the beads was adjusted to 1 107 beads ml–1 by measuring the optical density of their suspensions at 550 nm, according to the McFarland standard. In the case of the bacterial cells, bacteria were suspended in 0.01 or 0.1 M PBS to a concentration of 1 107 cells ml–1. Then the electrophoretic mobilities (l) of the glass or the bacterial suspensions were obtained from the velocity that the particles (beads or bacteria) acquired at 150 V using a Laser Zeta Meter (Malvern Instruments Ltd.). Afterwards, the bacterium and substrate surface zeta potentials, fB ; fS , respectively, were related to l by using the Helmholtz–Smoluchowsky Eq. (2), and evaluated:
f¼
4p n
e
l
ð2Þ
where n is the viscosity and e is the dielectric permittivity of the suspending liquid [43,44].
1109
All the experiments were carried out at room temperature and repeated for three independently prepared substrates and cultured bacterial strains, three times for each one. 2.3.4. X-ray photoelectron spectroscopy and surface chemical composition determination The surface chemical composition of bare and silane modified glass substrates was determined by X-ray photoelectron spectroscopy (XPS). XPS data were obtained with a LHS-10 spectrometer (SPECS Scientific Instruments, Inc.), using an Al Ka non-monochromatized X-ray source. Survey scans (0–1000 eV) were performed at 97 eV pass energy to determine the surface composition. SPECS data analysis software was used to calculate the elemental compositions from peak areas and to fit the peaks of the high-resolution spectra. Elemental surface compositions were expressed in percentages of carbon, oxygen, silicon and nitrogen. Two separate measurements were taken on different spots, for each substrate, for two separately prepared surfaces. XPS was also used in order to calculate the thickness of the silane monolayers, as described in the Supplementary Information (Eq. (S1)) [45]. 2.4. Dynamic bacterial adhesion assays For evaluating bacterial adhesion under flow conditions, the parallel plate flow chamber (PPFC) described by Stavridi et al. [46] was used. The configuration of the chamber is such that the sample is sandwiched between two Plexiglas plates in such a way that a parallel plate flow chamber is formed. Four syringes were placed in an automated syringe pump and connected to four different chambers. The pump was programmed to travel the pistons back and forth every 60 s. This cycle repeated itself for 120 min. All experiments were carried out at 37 °C. The experimental setup was placed inside a thermostated box (INFORS HT). The shear rate (S) was calculated by the following formula:
S¼
6Q 2
Wh
ð3Þ
where Q is the flow rate, W (width of the chamber) = 0.015 m and h (height of the chamber) = 0.35 103 m. Four shear rates were used: 50, 500, 1000 and 2000 s1. Pulsation was avoided by examining bacterial adhesion in the central area of the substrates and away from the flow entrances. Each experiment was performed three times. For good statistics, each time, the bacterial suspension used was from a different bacterial culture and the substrates were from different silane modified glass slides. 2.5. Quantification of bacterial adhesion 2.5.1. Colony-forming units counting method After the adhesion experiments, each sample was gently rinsed with 0.01 or 0.1 M PBS to remove non-adherent or loosely adherent bacteria. During the rinses, care was taken to avoid the formation of an air–liquid interface over the bacteria-covered surface, therefore preserving their position. Afterwards, the samples were placed in a tube with 5 ml of fresh sterile 0.01 M PBS. All the tubes were maintained in ice and sonicated for 10 min in an ultrasonic cleaner at 25 kHz; then 10-fold serial dilutions of the sonicated solutions were inoculated onto TSA plates, and the numbers of adherent bacterial colonies were counted after 18 h of incubation at 37 °C. Material samples were also plated on TSA plates after the sonication procedure in order to check if all bacteria were removed by sonication. In the cases where there were still bacteria on the samples, the bacterial colonies were counted and added to the PBS
1110
M.G. Katsikogianni, Y.F. Missirlis / Acta Biomaterialia 6 (2010) 1107–1118
bacterial counts [47]. This method did not cause bacterial death, due to the rigidity of the Gram-positive outer cell wall. 2.5.2. Scanning electron microscopy After the adhesion experiments, each sample was gently rinsed with 0.01 or 0.1 M PBS to remove non-adherent or loosely adherent bacteria and then fixed for 20 min with 2.5% glutaraldehyde (Sigma) in PBS [47]. After fixation, the samples were dehydrated by several passages in ethanol–water solutions for 20 min each, using increasing concentrations of ethanol up to 100%. After sputter coating with gold, the samples were investigated by scanning electron microscopy (SEM; JEOL-JSM 6300). Adherent bacteria were counted in three fields for each shear rate value and for each sample (three samples) in order to eliminate the possible uneven distribution of bacteria, while magnifications of 2000 were used. The total numbers of adherent bacteria were counted in each field by using the Image Pro Plus Analysis Software (Media Cybernetics), to give the density of bacteria per cm2 of the substratum surface.
2.3.3, k (m1) is the inverse Debye length, calculated as k ¼ 0:328 1010 ðIÞ1=2 (m1), where I is the ionic strength of the solution, and d is the separation distance between the bacterium and the surface. The way the LW and the EL interaction energies between a bacterium and a substratum surface immersed in liquid are calculated is presented in the Supplementary Information S2. 2.6.3. XDLVO theory According to the XDLVO theory [38], the total energy of bacterial adhesion is expressed as the sum of LW, EL and AB interaction EL AB energies, U LW BLS , U BLS and U BLS , respectively. The way the AB interaction energy between a bacterium and a substratum surface immersed in liquid is calculated is described in the Supplementary Information S3. The total interaction energy between a bacterium and a substratum surface as a function of distance, according to the XDLVO theory, is calculated by the equation: 2
2.6. Colloidal theories, bacteria–material interaction energy and force calculations To understand the associated forces of adhesion, the bacteria were modelled as non-living colloidal particles. Three theoretical approaches were used – the thermodynamic, DLVO and XDLVO – which have been described extensively elsewhere [48] and are presented concisely here. 2.6.1. Thermodynamic theory, Gibbs free energy change According to the LW-AB approach of the thermodynamic theory [38,42], the tendency of bacterial adhesion is expressed by the ) (J m–2)] of the process, when Gibbs free energy change [(DGLW-AB adh the separation distance (d) between the bacterium (B) and the surface (S) immersed in liquid (L) tends to zero [49]. According to this approach, the total free energy of adhesion is AB the sum of the LW and AB adhesion energies, DGLW d0 and DGd0 , respectively, and is calculated by using the following equation:
qffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiqffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffi cLW cLW cLW cLW B L L S qffiffiffiffiffiffi ffi qffiffiffiffiffiffi qffiffiffiffiffiffi pffiffiffiffiffiffi pffiffiffiffiffiffi pffiffiffiffiffiffi pffiffiffiffiffiffi qffiffiffiffiffi cþL cB þ cS cL þ cL cþB þ cþS cþL þ2 qffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffi cþB cS cB cþS
DGLW-AB ¼2 adh
2pd0 DGLW LW EL AB d0 R U DLVO þ peR BLS ¼ U BLS þ U BLS þ U BLS ¼ d 2 1 þ ekd 2 2kd ln 1 e þ f 2fB fS ln þ f B S 1 ekd d0 d þ 2pRkDGAB d0 exp k1
ð6Þ
where k1 is the characteristic decay of the AB interactions in water (k1 1 nm). DGAB d0 is the AB component of the free energy of adhesion of a bacterium to a substrate, at the closest separation distance (d0 ) [49]. All the other symbols have been described in the Section 2.6.2 By differentiating Eq. (5) or Eq. (6), the total interaction force between bacteria and substrates can be evaluated, as a function of distance. 2.7. Statistical analysis The effects of the surface free energy and flow conditions on bacterial adhesion were statistically analyzed using the SPSS package for windows. One-way analysis of variance (ANOVA) and in particular post hoc comparisons of all possible combinations of group means was performed using the Scheffe significant difference test. In all cases p < 0.05 was chosen to denote the significance level. Moreover, regression analysis and correlation coefficients (R2) were obtained by using SPSS. Correlations were taken as significant for p < 0.01.
ð4Þ 3. Results 2.6.2. DLVO theory According to the DLVO theory [36], the total interaction energy [U DLVO BLS (J)] of the bacterial adhesion process is expressed as the sum of the LW and the electrostatic (EL) interaction energy terms, U LW BLS and U EL BLS , respectively, and is calculated by the equation: 2
2pd0 DGLW LW EL d0 R U DLVO þ peR BLS ¼ U BLS þ U BLS ¼ d 2 1 þ ekd 2 2kd lnð1 e þ f 2fB fS ln þ f Þ B S 1 ekd
ð5Þ
where R is the bacterial radius (0.5 lm is the average bacterial radius, as measured in SEM images), d is the distance between the surfaces of the bacterium and the substrate, d0 is usually defined as 0.158 nm, and may be regarded as the distance between the outer electron shells [49], DGLW d0 is the LW component of the total free energy of adhesion, e ¼ 80 8:854 1012 (C2 J1 m1) is the dielectric permittivity of water, fB ; fS (V) are the bacterium and substratum surface zeta potentials, evaluated as described in Section
3.1. Surface morphology and roughness The AFM analysis of the NaOH-piranha-treated glass slides revealed that glass presents a relatively smooth surface with average surface roughness (Ra) of 0.9 ± 0.2 nm, for a scanned area of 25 lm2 (Fig. 1a). Fig. 1b and c present the surface morphology and Ra of the NH2- and CH3-terminated glass slides, respectively. The results show that the deposition of the specific silanes at the vapor/solid interface did not significantly influence the surface morphology and Ra of the glass slides. 3.2. Contact angle measurements and surface free energy components Table 1 presents the mean values of the experimentally measured water, methylene iodide (CH2I2) and glycerol contact angles h (deg) of the ATCC 35984 and ATCC 12228 bacterial cells and the OH-, NH2- and CH3-terminated glass slides. The influence of Ra on the measured contact angles is considered negligible since all the
1111
M.G. Katsikogianni, Y.F. Missirlis / Acta Biomaterialia 6 (2010) 1107–1118
Fig. 1. AFM micrographs and average surface roughness of (a) glass after 1 h treatment in NaOH-piranha; (b) glass-NH2, prepared at the vapor/solid phase after 1 h deposition time of APTES in toluene followed by sonication and (c) glass-CH3, prepared at the vapor/solid phase after 3 h deposition time of ODS in hexane.
Table 1 Mean values and standard deviations of water, methylene iodide (CH2I2) and glycerol contact angles (h) and zeta potential (f) for solutions with I.S.s of 0.01 and 0.1 M for S. epidermidis strains ATCC 35984 and ATCC 12228, and the various materials (three samples, three measurements for each one). Sample
h water (deg)
h CH2I2 (deg)
h glycerol (deg)
f (mV), 0.01 M
f (mV), 0.1 M
ATCC 35984 ATCC 12228 Glass Glass-NH2 Glass-CH3
23.1 ± 3.2 112.0 ± 2.3 10.0 ± 2.1 49.1 ± 3.3 93.0 ± 3.2
64.5 ± 4.1 78.2 ± 4.3 34.5 ± 2.4 36.8 ± 3.1 55.2 ± 2.1
24.2 ± 2.9 93.8 ± 2.3 19.0 ± 1.8 47.1 ± 2.1 71.6 ± 2.3
48.3 ± 3.2 45.8 ± 2.5 57.5 ± 3.1 25.4 ± 2.4 20.7 ± 6.9
25.8 ± 3.1 23.2 ± 2.8 30.6 ± 2.7 5.1 ± 4.5 11.3 ± 8.2
substrates present quite a small Ra and therefore the real area of the surface is not significantly different from the geometric one [50]. The results presented in Table 1 show that the ATCC 35984 bacteria have significantly lower water, CH2I2 and glycerol contact angles in comparison to the ATCC 12228 bacteria, revealing important differences in cell membrane characteristics. Moreover, the NH2- and especially the CH3-terminated groups significantly increased the measured contact angles of all the liquids, in comparison to the OH-terminated glass. Since the standard deviations of the contact angle measurements are relatively low, the mean values are used, for computaLW tional reasons, to calculate the LW (cLW S ; cB ) and the AB þ ; c ; c ; c ) components of the total free energy of the bacte(cþ S B B S ), according to the LW-AB ria and the substratum surfaces (cLW-AB S approach, which is presented in Section 2.6.1. The results presented in Table 2 show that the ATCC 35984 bacteria and the
than OH-terminated glass appear to be polar, with higher cAB S cLW S , whereas the ATCC 12228 bacteria and the CH3-terminated than cAB SAMs are rather hydrophobic, with much higher cLW S S . The NH2-terminated SAMs is moderately hydrophobic, with the AB component lower than that of the OH-terminated glass but much higher than that of CH3-terminated SAMs. The LW component of the various bacteria and material surface free energy does not vary as much as the AB component. Therefore, the increase in the free energy of the various samples is mainly due to the significantly enhanced polar component, indicating that the ATCC 35984 bacteria, the OH and to a less extent the NH2-terminated surface have a polar character. Moreover, the cS of the ATCC 35984 bacterial cells, the OH-terminated and to a less extent that of the NH2-terminated glass is much greater than the cþ S . This may suggest that these surfaces have a strongly monopolar surface or that they favor electron-donating or Lewis base
1112
M.G. Katsikogianni, Y.F. Missirlis / Acta Biomaterialia 6 (2010) 1107–1118
Table 2 The bacterial (S. epidermidis strains ATCC 35984 and ATCC 12228) and material total AB ), its apolar (cLW surface free energy (cLW-AB S S ) and polar (cS ) components, and its þ electron donor (c ) and electron acceptor ( c ) characters, as calculated according to S S the LW-AB approach. Sample
ATCC 35984 ATCC 12228 Glass Glass-NH2 Glass-CH3
cS
cLW S
cþS
(mJ m–2)
(mJ m–2)
(mJ m–2)
26.0 17.9 24.7 28.0 31.7
5.7 0.01 6.4 2.6 1.2
45.3 0.5 51.8 29.4 0.03
cAB S
cLW-AB S
32.2 0.08 36.3 17.5 1.19
58.2 17.9 60.9 45.5 32.9
(mJ m–2)
(mJ m–2)
þ properties. In contrast, the cAB S , cS and cS of the ATCC 12228 bacteria and the CH3-terminated SAMs appear low, reflecting their apolar character.
3.3. Zeta potential measurements Zeta potential measurements (f) of both bacterial strains and the three substrates, in 0.01 or 0.1 M PBS, are also presented in Table 1. The results show that f is negative for both bacterial strains, and not significantly different, in both I.S.S. Moreover, the OH-terminated glass is the most negatively charged substrate in solutions with low I.S., whereas the NH2 and the CH3 SAMs present low charge. The increase in the solution I.S. decreases the bacterial and substratum surface zeta potential, due to the neutralization of their negative charge by the cations of the high I.S. solution. 3.4. X-ray photoelectron spectroscopy, surface chemical composition and layer thickness XPS confirmed the presence of CH3- and NH2-groups on the silane-coated glass surfaces. In particular, the XPS spectra of the deposited films are shown in Fig. S1 of the Supplementary Information. The C1s spectrum of the film prepared from ODS (Fig. S1a) consists almost of a single peak centered at 284.6 eV, characteristic of the internal units of a polymethylene chain (–CH2–CH2–CH2. . .–), indicating that a hydrocarbon film, corresponding to its precursors, was formed. On the other hand, the C1s spectrum of the film, prepared from APTES (Fig. S1b) could be resolved into more features centered at binding energies of 284.6, 286.6 and 289.2 eV, which correspond to C–C, C–O and C–N, respectively. The ether peak at 286.6 eV was also observed for the OH-terminated glass. Furthermore, as shown in Fig. S1c, the presence of the N1s XPS peak is an indication that the glass slides were successfully modified. This peak consists of one chemical component with a binding energy of 399.5 eV, which is assigned to NH2 groups, indicating that they are not protonated. This was also confirmed by the zeta potential measurements. Table 3 contains the elemental percentages of C, N, O and Si. The results show that the amount of carbon increases for the NH2- and CH3-terminated SAMs, in comparison to the OH-terminated glass slides, due to the polymethylene chains of the APTES and ODS. The amount of carbon in the OH-terminated glass can be attributed to surface contamination. As anticipated, the amount of oxygen
Table 3 XPS elemental composition for the various substrates, and thickness of the monolayer in nm (two samples). Sample
C (%)
N (%)
O (%)
Si (%)
Thickness (nm)
Glass Glass-NH2 Glass-CH3
8.9 ± 0.1 14.6 ± 0.2 34.8 ± 0.4
0 1 ± 0.1 0
64.4 ± 0.1 58.4 ± 0.4 44.5 ± 0.2
26.7 ± 0 26.0 ± 0.1 20.7 ± 0.1
– 1 ± 0.3 2.5 ± 0.2
decreases in both the NH2 and CH3 SAMs, in comparison to the OH-terminated glass, and this explains the decreased polar character observed by the contact angle measurements and the surface free energy calculations for the NH2 and CH3 SAMs. The only substrate that presents nitrogen is the NH2-terminated one. Furthermore, since the AFM images (Fig. 1) revealed that the surface of the silanes was smooth and homogeneous, the thicknesses of the NH2- and CH3-terminated monolayers were assessed and are also presented in Table 3. The thickness of the NH2-terminated SAM was found to be 1 ± 0.3 nm, which is slightly higher than the calculated molecular length of APTES, which is 0.58 nm [27]. The thickness of the CH3-terminated monolayer was found to be 2.5 ± 0.2 nm, which is in good agreement with the theoretical length of 2.6 nm found for the ODS molecule, if oriented perpendicular to the surface [27,51]. 3.5. Bacterial adhesion Fig. 2 shows the combined effect of the surface chemistry, the solution’s I.S. and the shear rate on bacterial adhesion, as quantified by SEM, for the ATCC 35984 (Fig. 2a) and the ATCC 12228 (Fig. 2b) strains. The results show that the two bacterial strains present similar adhesion trends, but the ATCC 12228 strain is more adherent than the ATCC 35984 (p < 0.01 for all possible combinations). Moreover, both bacterial strains adhered the most to the CH3, followed by the NH2 SAM, and only minimally to the OH-terminated glass, for both I.S.s (p < 0.01). The increase in I.S. enhanced bacterial adhesion to all the substrates, and especially to the OH-terminated glass, due to the minimization of the repulsive EL interactions, for both strains (p < 0.01). A decrease in the number of adherent bacteria, for all the materials and both bacterial strains, was observed when the shear rate increased from 50 to 500 or 1000 s1, and especially when it reached 2000 s1. This decrease was significantly different (p < 0.01) for all substrates when the shear rate increased from 50 to 2000 s1. Moreover, the highest decrease in bacterial adhesion with the increase in shear rate was observed in the low I.S. solution for the ATCC 35984 strain to the OH-terminated glass (60% in 0.01 M and 43% in 0.1 M), followed by the CH3 SAM, (41% in 0.01 M and 39% in 0.1 M). The lowest decrease in bacterial adhesion with the increase in shear rate was observed in the high I.S. solution for the ATCC 12228 strain to the NH2 SAM (20% for the ATCC 12228 and 25% for the ATCC 35984, in 0.1 M). The SEM results were in good agreement with those obtained by counting the CFUs of the viable organisms, meaning that the adherent bacteria observed by SEM were in an active form before fixation and dehydration. Moreover, the SEM images revealed that the ATCC 35984 S. epidermidis strain, which is slime producing, produced slime only when it was attached to the CH3 SAM, under all flow rates. The produced slime covered the adherent bacteria and not the substrate. Isolated bacteria were observed on the OH-terminated glass and small aggregates on the NH2 SAM, meaning that the surface chemistry influences not only the number but also the pattern of adhesion and the slime production as well (images not presented). 3.6. Correlations between surface free energy components, zeta potential and number of adherent bacteria Trying to explain the adhesion of the two bacterial strains to the various substrates, we considered that the changes in the chemical structure that took place during the organosilane deposition, together with the subsequent decrease in the OH-terminated glass wettability, its surface free energy and zeta potential by the NH2, and especially the CH3 SAM, have to be the important parameters. The implementation of the LW-AB thermodynamic approach allowed for the investigation of how the number of adherent bacteria
1113
M.G. Katsikogianni, Y.F. Missirlis / Acta Biomaterialia 6 (2010) 1107–1118
Fig. 2. Number of adherent (a) ATCC 35984 and (b) ATCC 12228 bacteria to the various substrates in solution with 0.01 and 0.1 M I.S. under various shear rates, as quantified by SEM.
(N), for both bacterial strains, is correlated not only with the cLW-AB , S AB and its apolar (cLW S ) and polar (cS ) components, but also with the þ electron donor (c S ) and electron acceptor (cS ) characters of the substratum surfaces. Regression analysis of these data revealed that N was negatively and its polar (cAB correlated with the cLW-AB S S ) component (p < 0.001), whereas it was not significantly correlated with its apolar (cLW S ) component. þ Concerning the c S and cS characters of the substratum surfaces, the regression analysis revealed that N was negatively correlated þ with c S (p < 0.001) but not significantly correlated with cS for both bacterial strains. Since the ATCC 35984 strain and the substratum surfaces, except CH3 SAM, have a much higher electron donor than electron acceptor character, we assume that the electron donor character of the substratum surface is one of the material properties that control bacterial adhesion. In particular, an increase in c S decreases N (p < 0.001). Moreover, the zeta potential of the bacterial and the substratum surfaces is another parameter that significantly influences bacterial adhesion. Since both strains appeared highly negatively charged, especially in low I.S. solutions, N was negatively correlated with materials’ zeta potential. For this reason, adhesion was found to be lowest onto the OH-terminated glass that appeared highly negatively charged, whereas the increase in I.S. enhanced adhesion due to the minimization of the repulsive EL interactions. Since both bacterial strains had the same adhesion trend to the various substrates, although the one is polar and the other is not, the above-mentioned correlations were observed for both strains. From all the above it becomes clear that bacterial adhesion is influenced by a combination of interactions between the bacteria and the substrata, such as EL, LW and AB, which depend on the physicochemical characteristics of the bacterium, the fluid interface and the substratum. In order to examine the effect of these parameters on adhesion in a more quantitative manner, the thermodynamic, DLVO and XDLVO theories were implemented.
3.7. Correlations between free energy of adhesion, interaction energy, interaction–hydrodynamic forces and number of adherent bacteria 3.7.1. Thermodynamic theory – number of adherent bacteria Table 4 summarizes the Gibbs free energy changes of adhe) of the two S. epidermidis strains interacting with sion (DGLW-AB d0 the various substrates, as they are calculated according to the LW-AB approach of the thermodynamic theory (Section 2.6.1). DGLW-AB is decoupled in each case to its components: DGLW d0 d0 and DGAB . d0 values for the ATCC This approach results in negative DGLW-AB adh 35984 strain interacting with the CH3 SAM, whereas for the ATCC values are calculated for both the 12228 strain negative DGLW-AB adh NH2- and CH3-terminated SAMs. values presented in Table Therefore, according to the DGLW-AB d0 4, ATCC 35984 adhesion should be favored to the CH3 SAM, whereas ATCC 12228 should be favored to both the NH2 and CH3 SAMs. In order to examine if the pronounced effect of the total free en) and its acid–base (cAB ergy of the substratum surfaces (cLW-AB S S ) component on bacterial adhesion can be explained by the LW-AB thermodynamic approach, we plotted N as a function of the total free energy of adhesion (DGadh ) for both bacterial strains to all substrata, for both I.S., and a shear rate of 50 s1 (Fig. 3a). Exponential regression analysis of these data revealed that N was negatively (p < 0.001). Moreover, the correlation correlated with DGLW-AB d0 was better for the solution with high I.S., because the thermodynamic theory does not account for electrostatic interactions, which become more significant in low I.S. solutions. Furthermore, as observed in Table 4, the driving force for the negative or positive LW DGLW-AB values is the DGAB adh d0 component, since the values of DGd0 AB are low in comparison to the DGd0 ones, for all the possible combinations. Therefore, N is negatively correlated to DGAB d0 , whereas it is not significantly associated with DGLW d0 . These results indicate that the thermodynamic theory predicts the observed bacterial adhesion in a qualitative manner.
Table 4 AB ) and its apolar (DGLW The Gibbs free energy of S. epidermidis strains ATCC 35984 and ATCC 12228 adhesion (DGLW-AB d0 d0 ) and polar (DGd0 ) components, as calculated according to the LW-AB approach. Sample
–2 DGLW d0 (mJ m ), ATCC 35984
–2 DGAB d0 (mJ m ), ATCC 35984
DGLW-AB (mJ m–2), d0 ATCC 35984
–2 DGLW d0 (mJ m ), ATCC 12228
–2 DGAB d0 (mJ m ), ATCC 12228
DGLW-AB (mJ m–2), d0 ATCC 12228
Glass Glass-NH2 Glass-CH3
0.3 0.5 0.9
20.1 13.9 9.7
19.8 13.4 10.5
0.3 0.6 0.8
0 25.2 77.6
0.3 25.8 78.4
1114
M.G. Katsikogianni, Y.F. Missirlis / Acta Biomaterialia 6 (2010) 1107–1118
Fig. 3. Number of adherent bacteria per cm2 (N), for both S. epidermidis strains, correlated with: (a) the total adhesion free energy (DGadh), for shear rate 50 s1; (b) the total interaction energy as this was evaluated by the DLVO theory, for shear rate 50 s1; (c) the total interaction energy as this was evaluated by the XDLVO theory, for shear rate 50 s1 and (d) the total interaction energy as this was evaluated by the XDLVO theory, for shear rate 2000 s1.
3.7.2. DLVO theory – number of adherent bacteria XDLVO Table 5 presents the total interaction energies (U DLVO d¼1 nm ; U d¼1 nm ) evaluated according to the DLVO and the XDLVO theories, respecEL AB tively, and their LW (U LW d¼1 nm ), EL (U d¼1 nm ) and AB (U d¼1 nm ) components, between the ATCC 35984 S. epidermidis strain and the various substrates for distance d = 1 nm and for solutions with I.S.s of 0.01 M and 0.1 M. The same interaction energies for the ATCC 12228 S. epidermidis strain interacting with the various substrates are listed in Table 6. The DLVO theory predicts that both bacterial strains should be repelled by all the substrates at a small separation distance (d = 1 nm), for both I.S.s, because the EL interactions are repellent and predominate over the LW interactions at small separation distances. Moreover, the DLVO theory predicts that the increase in I.S. decreases bacterial repellence. Furthermore, it is observed that the EL interactions between the two bacterial strains and the various substrates are not significantly different, because the zeta potential for the two bacterial strains is not significantly different either. The predictability of the DLVO theory was examined by plotting N as a function of the total interaction energy (U DLVO d¼1 nm ) between the two bacterial strains and the various substrates for distance
d = 1 nm and for both I.S. solutions under a shear rate of 50 s1. The results presented in Fig. 3b reveal that the DLVO theory predicts the effect of the surface chemistry on bacterial adhesion for the solution with low I.S. better than the thermodynamic theory. The DLVO accounts for the EL interactions, which become predominant under low I.S. conditions, whereas under high I.S. the AB interactions are the predominant ones, but these are not encountered by the DLVO theory. Therefore, the DLVO theory qualitatively predicts the inhibition in bacterial adhesion to highly negatively charged substrates under low I.S. and the enhancement in adhesion with the increase in I.S., whereas it does not do so for the effect of surface chemistry on bacterial adhesion for high I.S., as observed by the lower correlation coefficient in Fig. 3b. 3.7.3. XDLVO theory – number of adherent bacteria According to the XDLVO theory and the results presented in Tables 5 and 6, the ATCC 35984 strain should be attracted to the CH3 SAM, whereas the ATCC 12228 strain should be attracted to both the NH2 and CH3 SAMs, due to the attractive AB interactions that are the predominant ones at close distance. Moreover, although the EL and LW interactions are not significantly different between
Table 5 XDLVO LW Total interaction energies (U DLVO d¼1 nm and U d¼1 nm ), evaluated according to the DLVO and the XDLVO theories, respectively, and their Lifshitz–van der Waals (U d¼1 nm ), electrostatic AB (U EL d¼1 nm ) and acid–base (U d¼1 nm ) components, between the ATCC 35984 S. epidermidis strain and the various substrates, for distance d = 1 nm and for solutions with I.S.s of 0.01 and 0.1 M. Sample
U LW d¼1 nm (J), ATCC 35984
U EL d¼1 nm (J), 0.01 M, ATCC 35984
U DLVO d¼1 nm (J), 0.01 M, ATCC 35984
U EL d¼1 nm (J), 0.1 M, ATCC 35984
U DLVO d¼1 nm (J), 0.1 M, ATCC 35984
U AB d¼1 nm (J), ATCC 35984
U XDLVO d¼1 nm (J), 0.01 M, ATCC 35984
U XDLVO d¼1 nm (J), 0.1 M, ATCC 35984
Glass Glass-NH2 Glass-CH3
8.20E-21 1.73E-20 2.66E-20
6.62E-18 2.53E-18 1.79E-18
6.62E-18 2.52E-18 1.77E-18
1.06E-18 1.42E-19 3.62E-19
1.05E-18 1.25E-19 3.35E-19
2.71E-17 1.88E-17 1.31E-17
3.37E-17 2.13E-17 1.14E-17
2.81E-17 1.89E-17 1.28E-17
1115
M.G. Katsikogianni, Y.F. Missirlis / Acta Biomaterialia 6 (2010) 1107–1118
Table 6 XDLVO LW Total interaction energies (U DLVO d¼1 nm and U d¼1 nm ), evaluated according to the DLVO and the XDLVO theories, respectively, and their Lifshitz–van der Waals (U d¼1 nm ), electrostatic AB (U EL d¼1 nm ) and acid–base (U d¼1 nm ) components, between the ATCC 12228 S. epidermidis strain and the various substrates, for distance d = 1 nm and for solutions with I.S.s of 0.01 and 0.1 M. Sample
U LW d¼1 nm (J), ATCC 12228
U EL d¼1 nm (J), 0.01 M, ATCC 12228
U DLVO d¼1 nm (J), 0.01 M, ATCC 12228
U EL d¼1 nm (J), 0.1 M, ATCC 12228
U DLVO d¼1 nm (J), 0.1 M, ATCC 12228
U AB d¼1 nm (J), ATCC 12228
U XDLVO d¼1 nm (J), 0.01 M, ATCC 12228
U XDLVO d¼1 nm (J), 0.1 M, ATCC 12228
Glass Glass-NH2 Glass-CH3
8.35E-21 1.76E-20 2.71E-20
6.24E-18 2.47E-18 1.78E-18
6.25E-18 2.49E-18 1.80E-18
9.49E-19 1.36E-19 3.32E-19
9.58E-19 1.54E-19 3.60E-19
1.16E-20 3.40E-17 1.05E-16
6.26E-18 3.16E-17 1.03E-16
9.69E-19 3.39E-17 1.04E-16
the two bacterial strains, the AB interactions are, due to the significantly different water and glycerol contact angles. Therefore, according to the XDLVO theory, the ATCC 12228 strain should be more adherent than the ATCC 35984 strain, due to the differences in their polarity and consequently the differences in the AB interactions. The predictability of the XDLVO theory was also examined, by plotting N as a function of the total interaction energy (U XDLVO d¼1 nm ) between the two bacterial strains and the various substrates for distance d = 1 nm and for both I.S. solutions under a shear rate of 50 s1. The results presented in Fig. 3c reveal that the XDLVO theory is better than the thermodynamic theory at predicting the effect of the surface chemistry on bacterial adhesion for the solution with low I.S., because the XDLVO accounts not only for the LW and AB interactions, as the thermodynamic does, but also for the EL, which become predominant under low I.S. conditions. Moreover, the XDLVO describes the adhesion results under high I.S. better than the DLVO theory, because the XDLVO encounters the AB interactions as well, which are the predominant ones under high I.S. solutions and are not encountered by the DLVO theory. Therefore, the XDLVO theory qualitatively predicts the combined effect of surface chemistry and solutions’ I.S. on bacterial adhesion and the enhancement in adhesion with the increase in I.S., and is the most suitable for the qualitative prediction of bacterial adhesion when compared to the thermodynamic and DLVO theories. 3.7.4. Interaction forces and shear rate The correlations presented in Fig. 3a–c concern the number of adherent bacteria to the various substrates for shear rate 50 s1. To examine the predictability of the XDLVO theory for shear rate 2000 s1, N was plotted as a function of the total interaction energy (U XDLVO d¼1 nm ) between the two bacterial strains and the various substrates for distance d = 1 nm and for both I.S. solutions. It was observed that the correlation coefficients were lower for the higher shear rate than for the lower one (Fig. 3d). For this reason it is useful to investigate the interaction forces between the bacteria and the substrates as a function of distance. Using the XDLVO theory, and in particular by differentiating Eq. (6), the total interaction force between the ATCC 35984 bacteria and the CH3 SAM for I.S. = 0.1 M was evaluated as a function of distance and is presented in Fig. S2a of the Supplementary Information. Fig. S2b of the Supplementary Information presents the total interaction force between the ATCC 35984 bacteria and the NH2 SAM for I.S. = 0.1 M as a function of distance. The results show that, although bacteria are attracted to the CH3 SAM for distances of less than 5 nm, there is an energy barrier between 5 and 9 nm that the bacteria must overcome in order to attach to the surface. This probably explains the large decrease in adhesion to the CH3 SAM under high flow conditions. In the case of NH2 SAM, although bacteria are repelled by the surface at small separation distances, there is an interaction minimum at distances between 25 and 40 nm. This attractive interaction, due to the balance of the repulsive EL interactions by the LW ones, seems to enable the bacterial attachment to the NH2 SAM. However, this attractive interaction is
low, meaning that, apart from the interactions that the XDLVO theory accounts for, bacteria interact with the NH2 SAM via macromolecular bonds, as Ma and Dickinson [39] proposed. CH3 SAM does not seem to favor this kind of interaction since it is an inert surface, in contrast to the NH2 one. In the case of the OH-terminated glass, the highly repulsive EL interactions do not seem to favor the formation of macromolecular binding, explaining the large decrease in bacterial adhesion with increasing shear rate. 4. Discussion In this study we investigated the effect of the surface chemistry and solution I.S. on two S. epidermidis strains (one producing slime, the other not) adhesion, and how the adhesion depends on the shear rate and on the relative contribution of physicochemical and hydrodynamic interactions. Moreover, we examined whether the S. epidermidis adhesion behavior agrees with the trends predicted by the thermodynamic, DLVO and XDLVO theories. These questions were addressed by quantitative measurement of bacterial adhesion on surfaces in laminar flow as a function of fluid shear rate. In this contribution, we described the preparation and characterization of hydroxylated glass substrates and SAMs deposited from organosilanes, CH3- and NH2-terminated. The adhesion of the two reference bacterial strains, a slime-positive one (ATCC 35984) and a slime-negative one (ATCC 12228), showed similar trends, but the second strain, although not producing slime, exhibited the best adhesion ability, due probably to its surface characteristics, and in particular its low polarity. The bacterial adhesion depended on the material surface chemistry. Adhesion was highest on the CH3-terminated glass, followed by the NH2-terminated glass, and minimal on the OH-terminated glass, for both strains and all flow conditions. This indicated that the number of adherent bacteria N was negatively correlated with ) and its polar (cAB the substratum surface free energy (cLW-AB S S ) component, whereas it was not significantly correlated with its apolar (cLW S ) component. Moreover, N was negatively correlated with the electron donor character of the substratum surface (c S ), but was not significantly correlated with the electron acceptor (cþ S ). In particular, an increase in c S decreases N. Our literature survey revealed controversies concerning the effect of the surface free energy and its polar component on adhesion. In our previous studies [15,16] we observed that the adhesion of S. epidermidis to He- and He/O2-treated poly(ethylene terephthalate) (PET) was reduced, in comparison to the adhesion to PET, due to the increase in the surface free energy and polar component, whereas the ageing time and the consequent decrease in the surface free energy and polar component favored bacterial adhesion [17]. Moreover, Gallardo-Moreno et al. [13] observed that ultraviolet irradiation of titanium alloy increased its surface free energy and reduced S. epidermidis adhesion, whereas, in another study [52], adhesion was increased when polystyrene was plasma treated with O2. With regard to the predictability of the LW-AB approach of thermodynamic theory, we observed that it predicts bacterial
1116
M.G. Katsikogianni, Y.F. Missirlis / Acta Biomaterialia 6 (2010) 1107–1118
adhesion in a qualitative manner, particularly at high I.S. solutions, where the EL interactions are negligible. In particular, it was found and DGAB that N was negatively correlated with DGLW-AB d0 d0 but was LW not significantly correlated with DGd0 at shear rates of both 50 and 2000 s1. These results are in agreement with our previous studies [16,17]. The zeta potential of the bacterial and substratum surfaces also influences bacterial adhesion. Since both strains appeared to be highly negatively charged, N was negatively correlated with the materials’ zeta potential, whereas the increase in I.S. enhanced adhesion, due to the minimization of the repulsive EL interactions. Furthermore, the DLVO theory qualitatively predicts the inhibition in bacterial adhesion to highly negatively charged substrates under low I.S. and the enhancement in adhesion with the increase in I.S. However, it does not predict well the effect of the surface chemistry on bacterial adhesion for high I.S., because then the AB interactions are the predominant ones and are not accounted for by the DLVO theory. These results are in agreement with those of Sharp and Dickinson [53] and Rijnaaarts et al. [54]. With regard to the predictability of the XDLVO theory, the results revealed that N was negatively correlated with the total interaction energy (U XDLVO d¼1 nm ) and therefore it qualitatively predicts the combined effect of the surface chemistry and solutions’ I.S. on bacterial adhesion and the enhancement in adhesion with the increase in I.S., for small separation distances. Since the XDLVO theory is a combination of the DLVO and thermodynamic theories, it predicts the effect of the surface chemistry on bacterial adhesion for the solution with low I.S. better than the thermodynamic theory, and it predicts the adhesion results under high I.S. better than the DLVO theory. These results are in agreement with observations of Meinders et al. [31]. The flow conditions strongly influence the number of attached bacteria. In particular, it was observed that the increase in shear rate decreased the number of adherent bacteria. The highest decrease was observed when the shear rate was increased from 50 to 2000 s1 in the low I.S. solution for the ATCC 35984 strain adhering to the OH-terminated glass, followed by the CH3 SAM. The lowest decrease in bacterial adhesion with increasing shear rate was observed in the high I.S. solution for both bacterial strains adhering to the NH2 SAM. Therefore, although the XDLVO predicted strong attractive interactions between both bacterial strains and the CH3 SAM for small separation distances, the increase in shear rate reduced bacterial adhesion, presumably by the inability of the bacteria to make contact with the surface at the high shear rate, thus interfering with the XDLVO. The bacteria did not manage to overcome the energy barrier that existed between 5 and 9 nm in the case of CH3 SAM under high shear rates. This explains the large decrease in the number of adherent bacteria. In the case of NH2 SAM, although the ATCC 35984 bacteria should be repelled by the surface at small separation distances according to all the colloidal theories, the XDLVO theory revealed an interaction minimum for distances between 25 and 40 nm. Although this interaction was low, a large number of adherent bacteria were observed even at the high shear rate. This implies that the bacteria are attracted to the NH2 SAM and withstand the detachment hydrodynamic forces by interaction forces that are not predicted by the colloidal theories. These interaction forces may be attributed to macromolecular binding between the bacterial surface appendages of the stationary-phase cells that were used [55] and the active surfaces [39], such as the NH2-terminated surface. The inert CH3 SAM does not seem to favor this kind of interaction. In the case of the OH-terminated glass, the strong repulsive colloidal forces seem to inhibit the formation of macromolecular binding. These results are in agreement with those of Finlay et al. [56], who observed that although the highest number of Enteromorpha
zoospores adhered to the surface of the less polar SAMs, at high shear stress zoospores detached more easily from the less polar SAMs than from the polar ones. In contrast, Bayoudh et al. [57] observed that bacterial adhesion strength measurements were in agreement with the adhesion free energy calculations. Moreover, Gomez-Suarez et al. [32], by using the passage of air–liquid interfaces to induce bacterial detachment from substratum surfaces, observed that detachment was highest at the lowest velocity of the passing air bubble, regardless of the bacterial strain involved and the polar character of the substratum, whereas at the highest air bubble velocity all bacterial strains detached more easily from hydrophilic glass than from hydrophobic dimethyldichlorosilanecoated glass. According to all the above, bacterial adhesion seems to be the result of a complicated interplay of LW, EL, AB, hydrodynamic and macromolecular forces. The high shear rate conditions not only significantly decreased bacterial adhesion, but also allowed for the observation of other than colloidal interactions between the bacteria and the NH2-terminated surface. However, the shear does not allow for direct and exact evaluation of the macromolecular interactions since these are directed perpendicular to the surface. According to Chang and Hammer [58], the magnitude of the force at which a cell relents is quoted as the strength of adhesion; however, the character or direction of the force, which depends on the adhesion assay, can affect the result, even if forces of precisely the same magnitude are applied in the different assays. Therefore, the estimation of the macromolecular binding forces is not such an easy process and requires not only bacterial adhesion experiments, even under shear, but detachment experiments as well. COOH-terminated surfaces can also be used for further studies.
5. Concluding remarks We have demonstrated that the material and bacterial surface free energy and charge, as well as the fluid phase properties and shear conditions, significantly influence the adhesion of two S. epidermidis strains to SAMs. The results were qualitatively predicted by the colloidal theories, with the XDLVO theory best explaining the combined effect of the surface chemistry and the solution’s I.S. on bacterial adhesion. However, simulated hemodynamic shear conditions identified limitations to the colloidal theories. Higher shear rates indicated the presence of other than the colloidal interactions between the bacteria and the substrates. Therefore, the driving forces for the adhesion of S. epidermidis to biomaterials may be considered a combination of interactions governed by physicochemical–macromolecular and physical forces dominated by shear. The macromolecular interactions apparently stem from the highly dynamic surface of the bacteria and their response to the environmental changes, and is the subject of further investigation.
Acknowledgements The authors thank Professor I. Spiliopoulou, from the Department of Microbiology, School of Medicine, University of Patras, for providing the bacterial strains and laboratory equipment; Professor P.G. Koutsoukos, from the Department of Chemical Engineering, University of Patras, for allowing the use of the Laser Zeta Meter; Dr. B. Seifert, from GKSS, Berlin, for valuable discussions on the preparation of SAMs and Dr. E. Siokou, from the FORTH/ICE-HT, Patras, for the XPS analysis.
M.G. Katsikogianni, Y.F. Missirlis / Acta Biomaterialia 6 (2010) 1107–1118
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.actbio.2009.08.006. Appendix B. Figures with essential colour discrimination Certain figures in this article, particularly Figures 1, 2 and 3, are difficult to interpret in black and white. The full colour images can be found in the on-line version, at doi:10.1016/j.actbio.2009. 08.006. References [1] Von Eiff C, Peters G, Heilmann C. Pathogenesis of infections due to coagulasenegative staphylococci. Lancet Infect Dis 2002;2:677–85. [2] Vincent J-L. Nosocomial infections in adult intensive-care units. Lancet 2003;361:2068–77. [3] Vuong C, Otto M. Staphylococcus epidermidis infections. Microbes Infect 2002;4:481–9. [4] Costerton JW, Stewart PS, Greenberg EP. Bacterial biofilms: a common cause of persistent infections. Science 1999;284:1318–22. [5] Gray ED, Peters G, Verstegen M, Regelmann WE. Effect of extracellular slime substance from Staphylococcus epidermidis adhesion on human cellular immune response. Lancet 1984;18:365–7. [6] Duguid IG, Evans E, Brown MR, Gilbert P. Effect of biofilm culture upon the susceptibility of Staphylococcus epidermidis to tobramycin. J Antimicrob Chemother 1992;30:803–10. [7] Hermansson M. The DLVO theory in microbial adhesion. Colloids Surf B: Biointerf 1999;14:105–19. [8] Morra M, Cassinelli C. Bacterial adhesion to polymer surfaces: a critical review of surface thermodynamic approaches. J Biomater Sci Polymer Ed 1997;9:55–74. [9] Heilmann C, Schweitzer O, Gerke C, Vanittanakom N, Mack D, Gotz F. Molecular basis of intercellular adhesion in the biofilm-forming Staphylococcus epidermidis. Mol. Microbiol 1996;20:1083–91. [10] Tyler BJ, Ratner BD, Castner DG, Briggs D. Variations between BiomerTM lots. I. Significant differences in the surface chemistry of two lots of a commercial poly(urethane). J Biomed Mater Res 1992;26:273–89. [11] Silver JH, Lewis KB, Ratner BD, Cooper SL. The effect of polyol type on the surface structure of sulfonate-containing polyurethanes. J Biomed Mater Res 1993;27:735–45. [12] Wang IW, Danilich M, Anderson J, Marchant RE. Adhesion of Staphylococcus epidermidis to biomedical polymers: contributions of surface thermodynamics and hemodynamic shear conditions. J Biomed Mater Res 1995;29:485–93. [13] Gallardo-Moreno AM, Pacha-Olivenza MA, Saldana L, Pérez-Giraldo C, Bruque JM, Vilaboa N, et al. In vitro biocompatibility and bacterial adhesion of physico-chemically modified Ti6Al4V surface by means of UV irradiation. Acta Biomater. 2009;5:181–92. [14] Katsikogianni M, Spiliopoulou I, Dowling DP, Missirlis YF. Adhesion of slime producing Staphylococcus epidermidis strains to PVC and diamond-like carbon/ silver/fluorinated coatings. J Mater Sci Mater Med 2006;17:679–89. [15] Amanatides E, Mataras D, Katsikogianni M, Missirlis YF. Plasma surface treatment of polyethylene terephthalate films for bacterial repellence. Surf Coat Technol 2006;200:6331–5. [16] Katsikogianni MG, Syndrevelis CS, Amanatides E, Mataras DS, Missirlis YF. Plasma treated and a-C:H coated PET performance in inhibiting bacterial adhesion. Plasma Processes Polym 2007;4:S1046–51. [17] Katsikogianni M, Amanatides E, Mataras DS, Missirlis YF. Staphylococcus epidermidis adhesion to He, He/O2 plasma treated PET films and aged materials: contributions of surface free energy and shear rate. Colloids Surf B: Biointerf 2008;65(2):257–68. [18] Popat KC, Eltgroth M, LaTempa TJ, Grimes CA, Desai TA. Decreased Staphylococcus epidermidis adhesion and increased osteoblast functionality on antibiotic-loading titania nanotubes. Biomaterials 2007;28:4880–8. [19] Bain CD, Whitesides GM. Molecular-level control over surface order in selfassembled monolayer films of thiols on gold. Science 1988;240:62–3. [20] Troughton EB, Bain CD, Whitesides GM, Nuzzo RG, Allara DL, Porter MD. Monolayer films prepared by the spontaneous self-assembly of symmetrical and unsymmetrical dialkyl sulfides from solution onto gold substrates: structure, properties and reactivity of constituent functional groups. Langmuir 1988;4:365–85. [21] Wasserman SR, Tao Y-T, Whitesides GM. Structure and reactivity of alkysiloxane monolayer formed by reaction of alkyl trichlorosilane on silicon substrate. Langmuir 1989;5:1074. [22] Dulcey CS, Georger JH, Krauthamer Jr V, Stenger DA, Fare TL, Calvert JM. Deep UV photochemistry of chemisorbed monolayers: patterned coplanar molecular assemblies. Science 1991;252:551–4. [23] Margel S, Vogler EA, Firment L, Watt T, Haynie S, Sogah DY. Peptide, protein and cellular interactions with self-assembled monolayer model surfaces. J Biomed Mater Res 1993;27:1463–76.
1117
[24] Mrksich M. Using self-assembled monolayers to understand the biomaterials interface. Curr Opin Colloid Int Sci 1997;2:83–8. [25] Tegoulia VA, Cooper SL. Staphylococcus aureus adhesion to self-assembled monolayers: effect of surface chemistry and fibrogen presence. Colloids Surf B: Biointerf 2002;24:217–28. [26] Faucheux N, Schweiss R, Lutzow K, Werner C, Groth T. Self-assembled monolayers with different terminating groups as model substrates for cell adhesion studies. Biomaterials 2004;25:2721–30. [27] Sugimura H, Hozumi A, Kameyama T, Takai O. Organosilane self-assembled monolayers formed at the vapour/solid interface. Surf Interface Anal 2002;34:550–4. [28] Scheuerman TR, Camper AK, Hamilton MA. Effects of substratum topography on bacterial adhesion. J Colloid Interface Sci 1998;208:23–33. [29] Edwards KJ, Rutenberg AD. Microbial response to surface microtopography: the role of metabolism in localized mineral dissolution. Chem Geol 2001;180:19–32. [30] Goldsmith HL, Turitto VT. Reological aspects of thrombosis and haemostasis: basic principles and applications. Thromb Haemostasis 1986;55:415–35. [31] Meinders JM, van der Mei HC, Buccher HJ. Deposition efficiency and reversibility of bacterial adhesion under flow. J Colloid Interface Sci 1995;176:329–41. [32] Gomez-Suarez C, Busscher HJ, van der Mei HC. Analysis of bacterial detachment from substratum surfaces by the passage of air–liquid interfaces. Appl Environ Microbiol 2001;67:2531–7. [33] Boks NP, Norde W, van der Mei HC, Busscher HJ. Forces involved in bacterial adhesion to hydrophilic and hydrophobic surfaces. Microbiology 2008;154:3122–33. [34] Wiencek KM, Fletcher M. Bacterial adhesion to hydroxyl- and methylterminated alkanethiol self-assembled monolayers. J Bacteriol 1995;177:1959–66. [35] Ista LK, Fan H, Baca O, López GP. Attachment of bacteria to model solid surfaces: oligo(ethylene glycol) surfaces inhibit bacterial attachment. FEMS Microb Lett 1996;142:59–63. [36] Derjaguin BW, Landau L. Theory of the stability of strongly charged lyophobic sols and of the adhesion of strongly charged particles in solutions of electrolytes. Acta Physiochim URSS 1941;14:633–62. [37] Absolom DR, Lamberti FV, Policova Z, Zingg W, van Oss CJ, Neumann AW. Surface thermodynamics of bacterial adhesion. Appl Environ Microbiol 1983;46:90–7. [38] van Oss CJ, Good RJ, Chaudhury MK. The role of van der Waals forces and hydrogen bonds in hydrophobic interactions between biopolymers and low energy surfaces. J Colloid Interface Sci 1986;111:378–90. [39] Ma H, Dickinson RB. Kinetic analysis of the attachment of a biological particle to a surface by macromolecular binding. J Theor Biol 2004;226:237–50. [40] Cunliffe D, Smart CA, Alexander C, Vulfson EN. Bacterial adhesion at synthetic surfaces. Appl Environ Microbiol 1999;65:4995–5002. [41] Busscher HJ, Weerkamp AH, van der Mei HC, van Pelt AWJ, De Jong HP, Arends J. Measurement of the surface free energy of bacterial cell surfaces and its relevance for adhesion. Appl Environ Microbiol 1984;48:980–3. [42] van Oss CJ. Hydrophobicity of biosurfaces – origin, quantitative determination and interaction energies. Colloids Surf B: Biointerf 1995;5:91–110. [43] Duval JFL, Busscher HJ, van de Belt-Gritter B, van der Mei HC, Norde W. Analysis of the interfacial properties of fibrillated and nonfibrillated oral streptococcal strains from electrophoretic mobility and titration measurements: evidence for the shortcomings of the ‘‘classical soft-particle approach”. Langmuir 2005;21:11268–82. [44] Li B, Logan BE. Bacterial adhesion to glass and metal-oxide surfaces. Colloids Surf B: Biointerf 2004;36:81–90. [45] Tanuma S, Powell CJ, Penn DR. Calculations of electron inelastic mean free paths. II. Data for 27 elements over the 50–2000 eV range. Surf Interface Anal 1991;17:911–26. [46] Stavridi M, Katsikogianni M, Missirlis YF. The influence of surface patterning and/or sterilization on the haemocompatibility of polycaprolactones. Mater Sci Eng C 2003;23:359–65. [47] An YH, Friedman RJ. Laboratory methods for studies of bacterial adhesion. J Microbiol Methods 1997;30:141–52. [48] Missirlis YF, Katsikogianni M. Theoretical and experimental approaches of bacteria–biomaterial interactions. Mater Sci Eng Technol (Mat-wiss u Werkstofftech) 2007;38:983–94. [49] van Oss CJ, Good RJ. The ‘‘equilibrium distance” between two bodies immersed in a liquid. Colloids Surf 1984;8:373–81. [50] Kamusewitz H, Possart W. Wetting and scanning force microscopy on rough polymer surfaces: Wenzel’s roughness factor and the thermodynamic contact angle. Appl Phys A 2003;76:899–902. [51] Reiniger M, Basnar B, Friedbacher G, Schleberger M. Atomic Force Microscopy of thin organic films on silicon in ultrahigh vacuum and under ambient conditions. Surf Interface Anal 2002;33:85–8. [52] Morra M, Cassinelli C. Staphylococcus epidermidis adhesion to films deposited from hydroxyethylmethacrylate plasma. J Biomed Mater Res 1996;31: 149–55. [53] Sharp JM, Dickinson RB. Direct evaluation of DLVO theory for predicting longrange forces between a yeast cell and a surface. Langmuir 2005;21: 8198–203. [54] Rijnaarts HHM, Norde W, Lyklema J, Zehnder AJB. DLVO and steric contributions to bacterial deposition in media of different ionic strength. J Colloid Interface Sci 1995;14:179–95.
1118
M.G. Katsikogianni, Y.F. Missirlis / Acta Biomaterialia 6 (2010) 1107–1118
[55] Walker SL, Hill JE, Redman JA, Elimelech M. Influence of growth phase on adhesion kinetics of Escherichia coli D21g. Appl Environ Microbiol 2005;71:3093–9. [56] Finlay JA, Callow ME, Ista LK, Lopez GP, Callow JA. The influence of surface wettability on the adhesion strength of settled spores of the green alga Enteromorpha and the diatom Amphora. Integr Comp Biol 2002;42:1116–22.
[57] Bayoudh S, Ponsonnet L, Ben Ouada H, Bakhrouf A, Othmane A. Bacterial detachment from hydrophilic and hydrophobic surfaces using a microjet impingement. Coll Surf A: Physicochem Eng Aspects 2005;266:160–7. [58] Chang KC, Hammer DA. Influence of direction and type of applied force on the detachment of macromolecularly-bound particles from surfaces. Langmuir 1996;12:2271–82.