- Email: [email protected]
Dependence of nanoscale friction and adhesion properties of articular cartilage on contact load
Journal of Biomechanics 44 (2011) 1340–1345
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Dependence of nanoscale friction and adhesion properties of articular cartilage on contact load S.M.T. Chan a, C.P. Neu a,1, K. Komvopoulos b,n, A.H. Reddi a a b
Center for Tissue Regeneration and Repair, University of California, Davis, Medical Center, Sacramento, CA 95817, USA Department of Mechanical Engineering, University of California, Berkeley, CA 94720, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Accepted 5 January 2011
Boundary lubrication of articular cartilage by conformal, molecularly thin films reduces friction and adhesion between asperities at the cartilage–cartilage contact interface when the contact conditions are not conducive to fluid film lubrication. In this study, the nanoscale friction and adhesion properties of articular cartilage from typical load-bearing and non-load-bearing joint regions were studied in the boundary lubrication regime under a range of physiological contact pressures using an atomic force microscope (AFM). Adhesion of load-bearing cartilage was found to be much lower than that of nonload-bearing cartilage. In addition, load-bearing cartilage demonstrated steady and low friction coefficient through the entire load range examined, whereas non-load-bearing cartilage showed higher friction coefficient that decreased nonlinearly with increasing normal load. AFM imaging and roughness calculations indicated that the above trends in the nanotribological properties of cartilage are not due to topographical (roughness) differences. However, immunohistochemistry revealed consistently higher surface concentration of boundary lubricant at load-bearing joint regions. The results of this study suggest that under contact conditions leading to joint starvation from fluid lubrication, the higher content of boundary lubricant at load-bearing cartilage sites preserves synovial joint function by minimizing adhesion and wear at asperity microcontacts, which are precursors for tissue degeneration. & 2011 Elsevier Ltd. All rights reserved.
Keywords: Adhesion Articular cartilage Boundary film Contact load Friction Lubrication SZP/lubricin/PRG4
1. Introduction The lubrication function of articular cartilage is essential for maintaining tissue health and preventing disease and damage such as osteoarthritis. While articular cartilage is lubricated by mixed modes of lubrication (Lewis and McCutchen, 1959; Walker et al., 1968; Unsworth, 1991; Gleghorn and Bonassar, 2008) in the presence of synovial fluid (Schmidt et al., 2007), boundary lubrication occurs under contact conditions of high load, low sliding speed, and/ or reduced fluid viscosity. Under such interfacial conditions, the asperities on the opposing surfaces are separated only by a molecularly thin lubricant layer (Gleghorn and Bonassar, 2008), hereafter referred to as a boundary film. In most sliding systems operating under a range of loads and speeds, such as synovial joints where mixed modes of lubrication are encountered during a walking cycle, a sufficiently thick fluid film cannot be continuously maintained between the articulating surfaces. As a consequence, contact occurs at the higher summits of the countersurfaces, known as asperities. In the absence of fluid film lubrication, for example, when the sliding n
Corresponding author. Tel.: +1 510 642 2563; fax: + 1 510 643 5599. E-mail address: [email protected] (K. Komvopoulos). 1 Present address: Weldon School of Biomedical Engineering, Purdue University, IN 47907, USA. 0021-9290/$ - see front matter & 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jbiomech.2011.01.003
speed decreases to zero for motion reversal, the tribological behavior of reciprocating systems is controlled by asperity contact interactions (Bronzino, 2006; Neu et al., 2008; Winter, 2009). Thus, boundary lubrication conditions are unavoidable in synovial joints. The formation of a lubricious boundary film can be viewed as a last line of surface protection against solid-solid contact, which is precursor for high friction and excessive wear. Indeed, precocious joint degeneration has been observed in the absence of an effective boundary lubricant film (Marcelino et al., 1999; Jay et al., 2007). The study of cartilage at the nano/microscale is important for understanding the role of molecular boundary films in cartilage tribology. However, the effect of the removal of boundary films and other proteins during sliding on the friction properties of cartilage has largely been ignored in macroscale friction studies. Although the presence of boundary films is an important factor in measuring friction at any scale (Jay et al., 2007; Gleghorn et al., 2009; Chan et al., 2010), continuous sliding of cartilage in macroscale friction tests removes not only molecularly thin boundary films, but also other proteins including collagen (Stachowiak et al., 1994; Gleghorn et al., 2010). Furthermore, in friction studies performed with microprobe-based instruments such as the atomic force microscope (AFM), the cartilage surface is not subjected continuously to normal and shear loading, whereas in macroscale studies traditionally performed with a tribometer, the cartilage surface is continuously
S.M.T. Chan et al. / Journal of Biomechanics 44 (2011) 1340–1345
loaded and sheared. Previous studies have indeed demonstrated that migrating cartilage contacts yield significantly lower friction coefficients than continuously loaded cartilage contacts (Caligaris and Ateshian, 2008; Shi et al., 2011). However, these studies again focus on the role of interstitial fluid pressurization in macroscale lubrication, where exuded fluid may be trapped and contained in the surface valleys, neglecting the effects of molecular boundary films at those instances that cartilage is starved from fluid film lubrication. Nanoscale friction testing enables bio-tribological studies to be performed at the scale of molecular boundary films and asperity microcontacts without the confounding effects of fluid exudation. Therefore, the present study specifically probes the friction properties of boundary lubricant films at cartilage surfaces, and provides insight into the role of boundary film lubrication on the macroscale friction behavior of articular cartilage. In view of the molecular thickness of boundary films, measurement of their mechanical and tribological properties can best be accomplished with microprobe-based instruments, such as the AFM and its derivatives. These instruments enable local material investigations under very low normal forces and extremely small displacement increments of sub-Angstrom resolution. The AFM has been used to study the nanotribological properties of cartilage (Coles et al., 2008; Chan et al., 2010) and the evolution of osteoarthritis in articular cartilage (Stolz et al., 2009). This technique is suitable for probing the nanotribological response of cartilage under boundary lubrication conditions because it can simulate surface interactions at the asperity level and the applied contact pressure is sufficiently high to prevent the formation of an intervening fluid film (Krishnan et al., 2004) as the bulk tissue is compressed by the probe tip. In addition, the AFM could be used to track in situ changes in the surface morphology and to study the evolution of different friction mechanisms, such as stickslip, adhesion, and plowing, which may dominate the tribological response of rough and/or poorly lubricated sliding surfaces (Urbakh et al., 2004; Coles et al., 2008). Basic understanding of nanoscale friction and adhesion mechanisms at the articular cartilage contact interface under boundary lubrication conditions is essential for studying the tissue response under extreme loading (contact) conditions, leading to tissue degradation, wear, and, ultimately, osteoarthritis. Therefore, the main objective of this study was to examine the nanoscale friction and adhesion behavior of cartilage, and to investigate the dependence of articular cartilage friction on applied normal load under boundary lubrication sliding conditions. Friction and adhesion force measurements and surface topography images (roughness) obtained with an AFM and immunohistochemistry results are presented to elucidate the effect of externally applied normal load on boundary film formation and its implications in cartilage friction during sliding in the boundary lubrication regime.
2. Materials and methods 2.1. Specimens Osteochondral explants from 1–3 week old bovine femoral condyles were harvested aseptically as previously described (DuRaine et al., 2009; Chan et al., 2010). For each condyle, one explant was obtained from a medial anterior location and one from a medial posterior location. For comparison, explants were harvested from articular joint locations of typically high (anterior) and low (posterior) physiological contact pressures (Neu et al., 2007), hereafter referred to as the load-bearing (M1) and non-load-bearing (M4) joint locations, respectively. Explants were stored at 37 1C and 5% CO2 in culture medium for 12 h, and trimmed to 1.5 mm in height with a custom cutting jig and razor prior to testing.
1341
Barbara, CA). AFM scanning was performed with triangular silicon nitride (Si3N4) tips (Veeco Instruments, Santa Barbara, CA) of nominal radius equal to 10 nm and height-to-base aspect ratio of 15.5. The tips were attached to microcantilevers of bending stiffness equal to 0.01 N/m. Linearity of the position-sensitive detector over the voltage range (0–8 V) used in this study was confirmed by the method of Silva and Van Vliet (2006). Surface areas (60 60 mm2) were scanned at a 901 tip angle, using a resolution of 128 128 pixels and a scanning frequency of 1 Hz. The normal stiffness (0.83 nN/V) of the AFM microcantilever was calibrated by performing force calibration on a glass substrate based on the thermal calibration method (Hutter and Bechhoefer, 1993) and Igor Pro (version 5) software. The normal load N was varied between 1.7 and 6.8 nN, resulting in tip normal displacements in the range of 2–5 nm, using the displacement corresponding to the tip engagement with the cartilage surface as a reference point. The mean (Hertzian) contact pressure pm ¼ ½16NE2 =9p3 R2 1=3 , where E ¼ ½ð1v2c Þ=Ec þ ð1v2t Þ=Et 1 , for Ec ¼2 MPa, nc ¼ 0.5, Et ¼ 200 GPa, nt ¼0.2, and N ¼ 1.7 and 6.8 nN is equal to 1.9 and 3.0 MPa, respectively. These pressure values are within the physiological contact pressure range. The lateral stiffness (102 nN/V) of the AFM microcantilever was calibrated by the direct force balance method (Asay and Kim, 2006) using a silicon grating (TGF11, Mikromasch, Wilsonville, OR). Normal and lateral stiffness calibrations performed before each test demonstrated good consistency. After each test, the AFM tip was cleaned with double deionized water and ethanol. The tip was frequently examined with an optical microscope to ensure that measurements in subsequent tests were not biased by possible tip damage and/or tissue transfer during previous scanning. Surface roughness was calculated from acquired AFM images as the root-mean-square of height deviations RS. Cartilage explants were scanned while fully immersed in phosphate buffered saline (Sigma-Aldrich, St. Louis, MO). Scans were acquired from five different surface locations of each explant. At each location, scans were obtained consecutively with increasing normal load to minimize any effects from a previous scan on the friction and adhesion force measurements obtained subsequently (Coles et al., 2008; Chan et al., 2010). The average friction force was determined from the halfwidth of the friction loop generated by the lateral (friction) force signal, i.e., onehalf of the difference between the mean lateral trace and retrace values, also accounting for any offset of the force sensor. The friction force was then averaged over all (5) scanned areas and all (6) samples tested and plotted as a function of applied normal load. The engineering friction coefficient m defined as
m¼
F N
ð1Þ
where F is the measured friction force and N is the externally applied normal load, was obtained as a function of normal load to determine if nanoscale friction of load- and non-load-bearing cartilage sliding in the boundary lubrication regime exhibits a dependence on the normal load. The true normal load applied at the contact interface may be affected by adhesion. The adhesion force acts normal to the contacting surfaces under both static and dynamic (sliding) contact conditions. However, the adhesion force of a static interface, measured at the instant of surface separation in the normal direction (pull-off force) differs from that of a dynamic contact interface subjected to shearing. Thus, to avoid confusion with the adhesion force of a static interface, the normal force measured under a zero external load is referred to as the adhesion force upon shearing. Therefore, to include the contribution of interfacial adhesion to the normal load, Eq. (1) was modified to the following equation: F ¼ m~ ðN þ Nad Þ
ð2Þ
where m~ is the true friction coefficient and Nad is the adhesion force measured during shearing in the absence of an external normal load. In view of Eq. (2), the friction force is expressed as F ¼ F0 þ m~ N
ð3Þ
where F0 ð ¼ m~ Nad Þ is the friction force in the absence of an external normal load, and is determined from the intercept of the line fit through the friction force data with the y-axis. For high normal loads, the effect of adhesion on the friction force becomes secondary, and Eq. (3) reduces to F m~ N, implying that at relatively high normal loads (NcNad ) the engineering friction coefficient approaches the true friction coefficient. Using Eqs. (1) and (3), the engineering friction coefficient can be expressed in terms of the true friction coefficient as follows: 1 m ¼ m~ þ F0 ð4Þ N Eq. (4) indicates that the engineering friction coefficient is inversely proportional to the externally applied normal load and approaches the true friction coefficient with the increase of the external normal load. 2.3. Immunohistochemistry
2.2. Friction, adhesion, and surface roughness measurements The adhesion and friction properties and surface morphology (roughness) of cartilage explants were studied with an AFM (MFP-3D-CF, Asylum Research, Santa
Immunolocalization of the known boundary lubricant in articular cartilage, superficial zone protein (SZP), was performed for M1 and M4 cartilage explants as described previously (Neu et al., 2007; Chan et al., 2010). Briefly, samples were
1342
S.M.T. Chan et al. / Journal of Biomechanics 44 (2011) 1340–1345
fixed in Bouin’s solution overnight, embedded in paraffin, exposed to primary antibody S6.79 (1:5000), and stained with an ABC kit (Vector Laboratories, Burlingame, CA) with secondary antibody mouse IgG.
2.4. Statistical analysis Explants obtained from six different bovine femoral condyles were used in the statistical data analysis (i.e., n¼ 6 for each type of explant (M1 or M4) and given external normal load). Data points shown in the figures presented in the next section represent mean values, whereas error bars indicate one standard error of the mean (SEM) above and below corresponding mean values. Friction force and friction coefficient data were fit with linear and nonlinear equations, respectively, and evaluated for goodness-of-fit by the coefficient of determination R2. Friction coefficient and surface roughness differences between M1 and M4 explants were determined from paired t-tests performed at each applied normal load for a significance level a ¼ 0.05.
A linear variation of the friction force with the applied normal load was observed for both M1 and M4 cartilage (Fig. 2A). Line fits yielded the following force relationships: FM1 ¼0.13N + 0.16 (R2 ¼0.978) and FM4 ¼0.20N + 0.61 (R2 ¼0.827). The slope of each linear fit represents the true friction coefficient of each cartilage region, i.e., m~ M1 ¼ 0:13 and m~ M4 ¼ 0:20. Load-bearing cartilage (M1) produced much lower friction forces than non-load-bearing cartilage (M4) through the entire normal load range. In addition, M1 cartilage yielded a significantly lower adhesion force Nad ð ¼ F0 =m~ Þ than M4 cartilage, i.e., Nad,M1 ¼ 1:21 7 0:4 1nN and Nad,M4 ¼ 2:97 70:39 nN (Fig. 2B). The mean and SEM values of the adhesion force are due to the experimental scatter in the friction force data (Fig. 2A) used to determine F0 and m~ . The engineering friction coefficient m of both load-bearing and non-load-bearing cartilage regions varied nonlinearly with the applied normal load N (Fig. 3). Importantly, this figure reflects the
3. Results Fig. 1 shows typical friction force responses of load-bearing (M1) and non-load-bearing (M4) cartilage for an external normal load of 6.8 nN. Since each friction test comprised back and forth scanning of the AFM tip over the same track under a constant normal load, two friction curves are shown in each plot. The fairly constant friction responses suggest that reciprocating sliding did not affect the dominant friction mechanism(s). The friction force of M1 and M4 cartilage, determined as one-half of the difference between the mean trace and retrace responses shown in Fig. 1, is equal to 1.15 and 1.85 nN, respectively.
Fig. 1. Representative friction force response of (A) load-bearing (M1) and (B) non-load-bearing (M4) cartilage sliding against a Si3N4 AFM tip under an externally applied normal load of 6.8 nN. The trace and retrace curves shown in each plot are due to back and forth scanning of the AFM tip.
Fig. 3. The engineering friction coefficient of cartilage sliding in the boundary lubrication regime decreased nonlinearly with the increase of the externally applied normal load, revealing a decreasing effect of adhesion on cartilage friction with increasing normal load. This trend was more pronounced for non-load-bearing cartilage (M4), whereas for load-bearing cartilage (M1) the friction coefficient remained relatively low and stable through the entire load range. In addition, loadbearing cartilage yielded lower friction coefficients than non-load-bearing cartilage through the entire normal load range (po0.032). Such difference in the nanoscale friction behavior of cartilage may be attributed to the formation of boundary films with higher contents of lubricious proteins, secreted at higher rates from chondrocytes in the superficial zone of load-bearing cartilage due to mechanotransduction effects. Data points represent mean values, while error bars indicate one SEM above and below the corresponding mean values.
Fig. 2. Adhesion force (measured during shearing in the absence of an external normal load) and friction force of load-bearing cartilage (M1) were consistently lower than those of non-load-bearing cartilage (M4), whereas surface roughness was similar. (A) Linear fits to the friction force versus normal load data yielded true friction coefficients (slopes) m~ M1 ¼ 0:13 and m~ M4 ¼ 0:20. (B) The adhesion force during shearing (obtained as the ratio of the friction force at zero external normal load and the true friction coefficient) of M1 and M4 cartilage was found equal to 1.21 70.41 and 2.97 70.39 nN, respectively. (C) Cartilage surface roughness did not show a dependence on joint location or discernible variation due to scanning under different loads (p ¼ 0.27). Data points represent mean values, while error bars indicate one SEM above and below the corresponding mean values.
S.M.T. Chan et al. / Journal of Biomechanics 44 (2011) 1340–1345
1343
Fig. 4. AFM images of (A, B) load-bearing cartilage (M1) and (C, D) non-load-bearing cartilage (M4) obtained after friction testing under (A,C) the lowest (1.7 nN) and (B,D) the highest (6.8 nN) applied normal load. The images do not reveal any discernible features of surface damage or topography changes due to sliding even for the highest normal load, in qualitative agreement with the relatively invariant surface roughness measured after friction testing through the entire normal load range. Corresponding RS values for these areas are (A) 306.03, (B) 317.26, (C) 406.79, and (D) 382.89 nm.
Fig. 5. Representative results of immunolocalization of SZP (a key boundary lubricant of cartilage) using monoclonal antibody S6.79 showed an increased expression at (A) load-bearing cartilage surfaces (M1) compared to (B) non-loadbearing cartilage surfaces (M4). The higher content of SZP at M1 cartilage surfaces correlates with the lower friction and adhesion characteristics of load-bearing cartilage.
Since all tested explants exhibited similar surface roughness (Fig. 2C), i.e., RS,M1 ¼374.24714.79 nm and RS,M2 ¼394.407 30.90 nm (p ¼0.27), the above trends in the nanoscale friction and adhesion properties of articular cartilage cannot be attributed to topographical differences. Moreover, sequential AFM scanning from the lowest to the highest normal load did not yield discernible changes in the surface morphology for any of the tested samples (Fig. 4), implying that adhesion was the dominant friction mechanism. However, histological staining for SZP showed significantly higher SZP content for M1 than M4 surfaces (Fig. 5), suggesting that variations in the nanoscale tribological properties were due to differences in the composition of the boundary films formed on load- and non-load-bearing cartilage surfaces. Similar results were obtained with other explants, confirming the consistently higher expression level of SZP in load-bearing (M1) than non-load-bearing (M4) cartilage surfaces.
4. Discussion decreasing effect of adhesion on friction with increasing normal load. Curve fitting of Eq. (4) through the data shown in Fig. 3 yielded the following relationships for the engineering friction coefficient of load-bearing and non-load-bearing cartilage: mM1 ¼ 0:126ð1=NÞ þ 0:139 (R2 ¼0.838) and mM4 ¼ 0:712ð1=NÞ þ0:178 (R2 ¼0.889). From an uncertainty analysis based on 95% confidence bounds it was found that the coefficients of these equations are within the range of the coefficients of the linear fits of Fig. 2A. Non-load-bearing joint regions exhibited significantly higher friction coefficients at each applied normal load (pr0.032), which decreased at a much faster rate with increasing normal load than the friction coefficients of load-bearing joint regions. The friction coefficients of both M1 and M4 cartilage decreased with the increase of the normal load, approaching asymptotically to steady-state values close to those of the true friction coefficients predicted from the slopes of the linear fits (Fig. 2A).
The main objective of this study was to determine if nanoscale friction of articular cartilage exhibits normal load dependence in the boundary lubrication regime. The presented results show that the engineering friction coefficient decreased nonlinearly with the increase of the applied normal load, with non-load-bearing cartilage demonstrating stronger load dependence than loadbearing cartilage (Fig. 3). At relatively high normal loads, the engineering friction coefficient approached the true friction coefficient (Fig. 3), determined from the slope of a linear fit through the friction force versus normal load data (Fig. 2A). Moreover, load-bearing cartilage exhibited lower friction coefficients than non-load-bearing cartilage through the entire normal load range examined (Fig. 3). Adhesion between the Si3N4 AFM tip and the cartilage surface was significant (Fig. 2B) and directly related to the friction response. The lower friction and adhesion of loadbearing cartilage than non-load-bearing cartilage also correlated
1344
S.M.T. Chan et al. / Journal of Biomechanics 44 (2011) 1340–1345
with the higher expression levels of boundary lubricants, such as SZP, at load-bearing joint regions (Fig. 5). In addition to the observed normal load dependence, nanoscale friction may show a dependence on sliding speed in both fluid film and boundary lubrication regimes (Drummond et al., 2003; Urbakh et al., 2004). The apparent violation of Amonton’s and Coulomb’s laws of friction that hold true at the macroscale can be attributed to scale effects and the dynamic characteristics of asperity microcontacts. In contrast to macroscale sliding contacts, micro/nanoscale contacts are constantly subjected to compressive, shear, and adhesive forces within microseconds (Urbakh et al., 2004). Transient asperity microcontacts generate attractive (adhesive) force fluctuations (on the order of several Pa to GPa) during these instantaneous load-unload cycles which are conducive to stick-slip (Drummond et al., 2003; Naka et al., 2007). Stick-slip can play a particularly important role in reciprocating sliding systems, such as synovial joints, where the reversal of the motion direction at zero-speed instances promotes the development of adhesion forces at asperity microcontacts. Significant adhesion effects on the friction behavior reported in other studies (Crockett et al., 2005; Coles et al., 2008) illustrate the importance of adhesion and stick-slip on the friction behavior of cartilage. Interestingly, an inverse relationship of adhesion and friction between the Si3N4 AFM tip and articular cartilage was previously observed (Chan et al., 2010). However, those adhesion force and hysteresis measurements were obtained with the more traditional approach that uses the force–displacement response obtained during the normal approach and retraction of the AFM tip from the cartilage surface (Crockett et al., 2005; Chan et al., 2010). While the normal approach method is useful for determining the adhesion energy and hysteresis, the lateral direction approach may be more relevant for measuring the adhesion component of friction at the nanoscale. Since the adhesion hysteresis depends on the sliding direction, a normal approach of the AFM tip to the cartilage surface may not fully represent the adhesion force or adhesion hysteresis associated with lateral displacement (sliding). Indeed, significantly higher adhesion forces have been measured in microscale sliding contact compared to adhesive pull-off forces measured during normal contact (Timpe and Komvopoulos, 2006). Necking and bridging forces due to bonding of atoms and/or ions between the contacting surfaces may further contribute to the increase of the adhesion force during normal surface approach (Zappone et al., 2008). One potential limitation to this study is the use of nonfunctionalized AFM probe tips, which are not representative of physiological conditions. However, others have used functionalized AFM tips to mimic cartilage surface chemistry or minimize protein adhesion (Coles et al., 2008, 2010) and obtained friction coefficients and surface roughness values similar to those found in this study. While the non-functionalized, sharp Si3N4 AFM tip may not reproduce the adhesion and plowing forces of a cartilage–cartilage sliding system, the existence of these friction mechanisms at asperity microcontacts persists regardless of surface chemistry and tip shape (Park et al., 2004; Coles et al., 2008). Therefore, AFM probe tips play an important role in understanding the nano/microscale friction behavior of cartilage in the boundary lubrication regime, where extensive wear may occur in the absence of fluid film lubrication. Even though the Si3N4 probe–cartilage interface may not ideally replicate the cartilage– cartilage interface, the current study nevertheless illuminates the relative differences between the tribological properties of loadbearing and non-load-bearing joint regions. The presence of boundary lubricants is essential for separating contacting solid surfaces during reciprocating sliding. Load-bearing cartilage demonstrated lower adhesion forces and friction coefficients than non-load-bearing cartilage (Figs. 1–3). Furthermore,
load-bearing cartilage also showed a weaker dependence of friction on normal load (Fig. 3). Under boundary lubrication sliding conditions, load-bearing cartilage produced nearly constant and low friction coefficient, which also suggests better wear resistance through a range of contact pressures under physiological settings. These results may be attributed to the higher expression levels of boundary lubricants in the load-bearing regions of the joint (Neu et al., 2007; Nugent-Derfus et al., 2007), as evidenced by the immunohistochemistry results (Fig. 5). The lower adhesion of load-bearing cartilage may also be attributed to the anti-adhesive and chondroprotective properties of SZP, which prevent synoviocyte overgrowth and cartilage–cartilage adhesion (Englert et al., 2005; Rhee et al., 2005). The results of the present study demonstrate a persistent dependence of the friction coefficient of articular cartilage on normal load. When joint contact conditions are not conducive to the formation of a sufficiently thick fluid film between the conformal cartilage surfaces, boundary lubricants, such as SZP (also known as lubricin and PRG4), may lower adhesion at the interfaces of asperity microcontacts and, consequently, reduce articular cartilage friction. In summary, the nanoscale friction and adhesion properties of articular cartilage from typical load-bearing and non-load-bearing regions of femoral condyles were examined with an AFM under a range of physiological contact conditions. Load-bearing cartilage demonstrated significantly lower adhesion and friction than nonload-bearing cartilage through the entire normal load range investigated in this study. This trend may be attributed to the higher surface content of boundary lubricants, such as SZP, in the load-bearing regions of the joint, which contribute to the formation of lubricious boundary films. These molecularly thin boundary films reduce friction and adhesion at asperity microcontacts occurring at those instances that the joint is starved from fluid film lubrication. Therefore, effective boundary lubrication of cartilage by continuously replenished boundary films is critical for normal joint function and for preventing tissue degeneration and wear caused by adhering asperities.
Conflict of interest statement The authors confirm that there are no conflicts of interest.
Acknowledgments The authors acknowledge the Spectral Imaging Facility, University of California, Davis, for the use of the AFM, and T.M. Schmid (Rush University, Chicago, IL) for the donation of the S6.79 antibody. References Asay, D.B., Kim, S.H., 2006. Direct force balance method for atomic force microscopy lateral force calibration. Rev. Sci. Instrum. 77, 043903. Bronzino, J.D., 2006. Biomedical Engineering Fundamentals. CRC Press, Boca Raton, LA. Caligaris, M., Ateshian, G.A., 2008. Effects of sustained interstitial fluid pressurization under migrating contact area, and boundary lubrication by synovial fluid, on cartilage friction. Osteoarthritis Cartilage 16, 1220–1227. Chan, S.M.T., Neu, C.P., DuRaine, G., Komvopoulos, K., Reddi, A.H., 2010. Atomic force microscope investigation of the boundary-lubricant layer in articular cartilage. Osteoarthritis Cartilage 18, 956–963. Coles, J.M., Blum, J.J., Jay, G.D., Darling, E.M., Guilak, F., Zauscher, S., 2008. In situ friction measurement on murine cartilage by atomic force microscopy. J. Biomech. 41, 541–548. Coles, J.M., Zhang, L., Blum, J.J., Warman, M.L., Jay, G.D., Guilak, F., Zauscher, S., 2010. Loss of cartilage structure, stiffness, and frictional properties in mice lacking PRG4. Arthritis Rheum. 62, 1666–1674. Crockett, R., Roos, S., Rossbach, P., Dora, C., Born, W., Troxler, H., 2005. Imaging of the surface of human and bovine articular cartilage with ESEM and AFM. Tribol. Lett. 19, 311–317.
S.M.T. Chan et al. / Journal of Biomechanics 44 (2011) 1340–1345
Drummond, C., Israelachvili, J., Richetti, P., 2003. Friction between two weakly adhering boundary lubricated surfaces in water. Phys. Rev. E Stat. Nonlin. Soft Matter Phys. 67, 066110. DuRaine, G., Neu, C.P., Chan, S.M.T., Komvopoulos, K., June, R.K., Reddi, A.H., 2009. Regulation of the friction coefficient of articular cartilage by TGF-b1 and IL-1b. J. Orthop. Res. 27, 249–256. Englert, C., McGowan, K.B., Klein, T.J., Giurea, A., Schumacher, B.L., Sah, R.L., 2005. Inhibition of integrative cartilage repair by proteoglycan 4 in synovial fluid. Arthritis Rheum. 52, 1091–1099. Gleghorn, J.P., Bonassar, L.J., 2008. Lubrication mode analysis of articular cartilage using Stribeck surfaces. J. Biomech. 41, 1910–1918. Gleghorn, J.P., Jones, A.R.C., Flannery, C.R., Bonassar, L.J., 2009. Boundary mode lubrication of articular cartilage by recombinant human lubricin. J. Orthop. Res. 27, 771–777. Gleghorn, J.P., Doty, S.B., Warren, R.F., Wright, T.M., Maher, S.A., Bonassar, L.J., 2010. Analysis of frictional behavior and changes in morphology resulting from cartilage articulation with porous polyurethane foams. J. Orthop. Res. 28, 1292–1299. Hutter, J.L., Bechhoefer, J., 1993. Calibration of atomic-force microscope tips. Rev. Sci. Instrum. 64, 1868–1873. Jay, G.D., Torres, J.R., Rhee, D.K., Helminen, H.J., Hytinnen, M.M., Cha, C.J., Elsaid, K., Kim, K.S., Cui, Y., Warman, M.L., 2007. Association between friction and wear in diarthrodial joints lacking lubricin. Arthritis Rheum. 56, 3662–3669. Krishnan, R., Kopacz, M., Ateshian, G.A., 2004. Experimental verification of the role of interstitial fluid pressurization in cartilage lubrication. J. Orthop. Res. 22, 565–570. Lewis, P.R., McCutchen, C.W., 1959. Mechanism of animal joints: experimental evidence for weeping lubrication in mammalian joints. Nature (London) 184, 1285. Marcelino, J., Carpten, J.D., Suwairi, W.M., Gutierrez, O.M., Schwartz, S., Robbins, C., Sood, R., Makalowska, I., Baxevanis, A., Johnstone, B., Laxer, R.M., Zemel, L., Kim, C.A., Herd, J.K., Ihle, J., Williams, C., Johnson, M., Raman, V., Alonso, L.G., Brunoni, D., Gerstein, A., Papadopoulos, N., Bahabri, S.A., Trent, J.M., Warman, M.L., 1999. CACP, encoding a secreted proteoglycan, is mutated in camptodactyly-arthropathy-coxa vara-pericarditis syndrome. Nat. Genet. 23, 319–322. Naka, M.H., Hasuo, M., Fuwa, Y., Ikeuchi, K., 2007. Correlation between friction of articular cartilage and reflectance intensity from superficial images. Tribol. Int. 40, 200–207. Neu, C.P., Khalafi, A., Komvopoulos, K., Schmid, T.M., Reddi, A.H., 2007. Mechanotransduction of bovine articular cartilage superficial zone protein by transforming growth factor b signaling. Arthritis Rheum. 56, 3706–3714. Neu, C.P., Komvopoulos, K., Reddi, A.H., 2008. The interface of functional biotribology and regenerative medicine in synovial joints. Tissue Eng. Part B Rev. 14, 235–247.
1345
¨ Nugent-Derfus, G.E., Takara, T., O’Neill, J.K., Cahill, S.B., Gortz, S., Pong, T., Inoue, H., Aneloski, N.M., Wang, W.W., Vega, K.I., Klein, T.J., Hsieh-Bonassera, N.D., Bae, W.C., Burke, J.D., Bugbee, W.D., Sah, R.L., 2007. Continuous passive motion applied to whole joints stimulates chondrocyte biosynthesis of PRG4. Osteoarthritis Cartilage 15, 566–574. Park, S., Costa, K.D., Ateshian, G.A., 2004. Microscale frictional response of bovine articular cartilage from atomic force microscopy. J. Biomech. 37, 1679–1687. Rhee, D.K., Marcelino, J., Baker, M., Gong, Y., Smits, P., Lefebvre, V., Jay, G.D., Stewart, M., Wang, H., Warman, M.L., Carpten, J.D., 2005. The secreted glycoprotein lubricin protects cartilage surfaces and inhibits synovial cell overgrowth. J. Clin. Invest. 115, 622–631. Schmidt, T.A., Gastelum, N.S., Nguyen, Q.T., Schumacher, B.L., Sah, R.L., 2007. Boundary lubrication of articular cartilage: role of synovial fluid constituents. Arthritis Rheum. 56, 882–891. Shi, L., Sikavitsas, V.I., Striolo, A., 2011. Experimental friction coefficients for bovine cartilage measured with a pin-on-disk tribometer: testing configuration and lubricant effects. Ann. Biomed. Eng. 39, 132–146. Silva, E.C.C.M., Van Vliet, K.J., 2006. Robust approach to maximize the range and accuracy of force application in atomic force microscopes with nonlinear position-sensitive detectors. Nanotechnology 17, 5525–5530. Stachowiak, G.W., Batchelor, A.W., Griffiths, L.J., 1994. Friction and wear changes in synovial joints. Wear 171, 135–142. Stolz, M., Gottardi, R., Raiteri, R., Miot, S., Martin, I., Imer, R., Staufer, U., Raducanu, ¨ A., Duggelin, M., Baschong, W., Daniels, A.U., Friederich, N.F., Aszodi, A., Aebi, U., 2009. Early detection of aging cartilage and osteoarthritis in mice and patient samples using atomic force microscopy. Nature Nanotechnol. 4, 186–192. Timpe, S.J., Komvopoulos, K., 2006. The effect of adhesion on the static friction properties of sidewall contact interfaces of microelectromechanical devices. J. Microelectromech. Syst. 15, 1612–1621. Unsworth, A., 1991. Tribology of human and artificial joints. Proc. Inst. Mech. Eng. H 205, 163–172. Urbakh, M., Klafter, J., Gourdon, D., Israelachvili, J., 2004. The nonlinear nature of friction. Nature 430, 525–528. Walker, P.S., Dowson, D., Longfield, M.D., Wright, V., 1968. ’’Boosted lubrication’’ in synovial joints by fluid entrapment and enrichment. Ann. Rheum. Dis. 27, 512–520. Winter, D.A., 2009. Biomechanics and Motor Control of Human Movement 4th ed. Wiley, Hoboken, NJ. Zappone, B., Thurner, P.J., Adams, J., Fantner, G.E., Hansma, P.K., 2008. Effect of Ca2 + ions on the adhesion and mechanical properties of adsorbed layers of human osteopontin. Biophys. J. 95, 2939–2950.
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Dependence of nanoscale friction and adhesion properties of articular cartilage on contact load S.M.T. Chan a, C.P. Neu a,1, K. Komvopoulos b,n, A.H. Reddi a a b
Center for Tissue Regeneration and Repair, University of California, Davis, Medical Center, Sacramento, CA 95817, USA Department of Mechanical Engineering, University of California, Berkeley, CA 94720, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Accepted 5 January 2011
Boundary lubrication of articular cartilage by conformal, molecularly thin films reduces friction and adhesion between asperities at the cartilage–cartilage contact interface when the contact conditions are not conducive to fluid film lubrication. In this study, the nanoscale friction and adhesion properties of articular cartilage from typical load-bearing and non-load-bearing joint regions were studied in the boundary lubrication regime under a range of physiological contact pressures using an atomic force microscope (AFM). Adhesion of load-bearing cartilage was found to be much lower than that of nonload-bearing cartilage. In addition, load-bearing cartilage demonstrated steady and low friction coefficient through the entire load range examined, whereas non-load-bearing cartilage showed higher friction coefficient that decreased nonlinearly with increasing normal load. AFM imaging and roughness calculations indicated that the above trends in the nanotribological properties of cartilage are not due to topographical (roughness) differences. However, immunohistochemistry revealed consistently higher surface concentration of boundary lubricant at load-bearing joint regions. The results of this study suggest that under contact conditions leading to joint starvation from fluid lubrication, the higher content of boundary lubricant at load-bearing cartilage sites preserves synovial joint function by minimizing adhesion and wear at asperity microcontacts, which are precursors for tissue degeneration. & 2011 Elsevier Ltd. All rights reserved.
Keywords: Adhesion Articular cartilage Boundary film Contact load Friction Lubrication SZP/lubricin/PRG4
1. Introduction The lubrication function of articular cartilage is essential for maintaining tissue health and preventing disease and damage such as osteoarthritis. While articular cartilage is lubricated by mixed modes of lubrication (Lewis and McCutchen, 1959; Walker et al., 1968; Unsworth, 1991; Gleghorn and Bonassar, 2008) in the presence of synovial fluid (Schmidt et al., 2007), boundary lubrication occurs under contact conditions of high load, low sliding speed, and/ or reduced fluid viscosity. Under such interfacial conditions, the asperities on the opposing surfaces are separated only by a molecularly thin lubricant layer (Gleghorn and Bonassar, 2008), hereafter referred to as a boundary film. In most sliding systems operating under a range of loads and speeds, such as synovial joints where mixed modes of lubrication are encountered during a walking cycle, a sufficiently thick fluid film cannot be continuously maintained between the articulating surfaces. As a consequence, contact occurs at the higher summits of the countersurfaces, known as asperities. In the absence of fluid film lubrication, for example, when the sliding n
Corresponding author. Tel.: +1 510 642 2563; fax: + 1 510 643 5599. E-mail address: [email protected] (K. Komvopoulos). 1 Present address: Weldon School of Biomedical Engineering, Purdue University, IN 47907, USA. 0021-9290/$ - see front matter & 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jbiomech.2011.01.003
speed decreases to zero for motion reversal, the tribological behavior of reciprocating systems is controlled by asperity contact interactions (Bronzino, 2006; Neu et al., 2008; Winter, 2009). Thus, boundary lubrication conditions are unavoidable in synovial joints. The formation of a lubricious boundary film can be viewed as a last line of surface protection against solid-solid contact, which is precursor for high friction and excessive wear. Indeed, precocious joint degeneration has been observed in the absence of an effective boundary lubricant film (Marcelino et al., 1999; Jay et al., 2007). The study of cartilage at the nano/microscale is important for understanding the role of molecular boundary films in cartilage tribology. However, the effect of the removal of boundary films and other proteins during sliding on the friction properties of cartilage has largely been ignored in macroscale friction studies. Although the presence of boundary films is an important factor in measuring friction at any scale (Jay et al., 2007; Gleghorn et al., 2009; Chan et al., 2010), continuous sliding of cartilage in macroscale friction tests removes not only molecularly thin boundary films, but also other proteins including collagen (Stachowiak et al., 1994; Gleghorn et al., 2010). Furthermore, in friction studies performed with microprobe-based instruments such as the atomic force microscope (AFM), the cartilage surface is not subjected continuously to normal and shear loading, whereas in macroscale studies traditionally performed with a tribometer, the cartilage surface is continuously
S.M.T. Chan et al. / Journal of Biomechanics 44 (2011) 1340–1345
loaded and sheared. Previous studies have indeed demonstrated that migrating cartilage contacts yield significantly lower friction coefficients than continuously loaded cartilage contacts (Caligaris and Ateshian, 2008; Shi et al., 2011). However, these studies again focus on the role of interstitial fluid pressurization in macroscale lubrication, where exuded fluid may be trapped and contained in the surface valleys, neglecting the effects of molecular boundary films at those instances that cartilage is starved from fluid film lubrication. Nanoscale friction testing enables bio-tribological studies to be performed at the scale of molecular boundary films and asperity microcontacts without the confounding effects of fluid exudation. Therefore, the present study specifically probes the friction properties of boundary lubricant films at cartilage surfaces, and provides insight into the role of boundary film lubrication on the macroscale friction behavior of articular cartilage. In view of the molecular thickness of boundary films, measurement of their mechanical and tribological properties can best be accomplished with microprobe-based instruments, such as the AFM and its derivatives. These instruments enable local material investigations under very low normal forces and extremely small displacement increments of sub-Angstrom resolution. The AFM has been used to study the nanotribological properties of cartilage (Coles et al., 2008; Chan et al., 2010) and the evolution of osteoarthritis in articular cartilage (Stolz et al., 2009). This technique is suitable for probing the nanotribological response of cartilage under boundary lubrication conditions because it can simulate surface interactions at the asperity level and the applied contact pressure is sufficiently high to prevent the formation of an intervening fluid film (Krishnan et al., 2004) as the bulk tissue is compressed by the probe tip. In addition, the AFM could be used to track in situ changes in the surface morphology and to study the evolution of different friction mechanisms, such as stickslip, adhesion, and plowing, which may dominate the tribological response of rough and/or poorly lubricated sliding surfaces (Urbakh et al., 2004; Coles et al., 2008). Basic understanding of nanoscale friction and adhesion mechanisms at the articular cartilage contact interface under boundary lubrication conditions is essential for studying the tissue response under extreme loading (contact) conditions, leading to tissue degradation, wear, and, ultimately, osteoarthritis. Therefore, the main objective of this study was to examine the nanoscale friction and adhesion behavior of cartilage, and to investigate the dependence of articular cartilage friction on applied normal load under boundary lubrication sliding conditions. Friction and adhesion force measurements and surface topography images (roughness) obtained with an AFM and immunohistochemistry results are presented to elucidate the effect of externally applied normal load on boundary film formation and its implications in cartilage friction during sliding in the boundary lubrication regime.
2. Materials and methods 2.1. Specimens Osteochondral explants from 1–3 week old bovine femoral condyles were harvested aseptically as previously described (DuRaine et al., 2009; Chan et al., 2010). For each condyle, one explant was obtained from a medial anterior location and one from a medial posterior location. For comparison, explants were harvested from articular joint locations of typically high (anterior) and low (posterior) physiological contact pressures (Neu et al., 2007), hereafter referred to as the load-bearing (M1) and non-load-bearing (M4) joint locations, respectively. Explants were stored at 37 1C and 5% CO2 in culture medium for 12 h, and trimmed to 1.5 mm in height with a custom cutting jig and razor prior to testing.
1341
Barbara, CA). AFM scanning was performed with triangular silicon nitride (Si3N4) tips (Veeco Instruments, Santa Barbara, CA) of nominal radius equal to 10 nm and height-to-base aspect ratio of 15.5. The tips were attached to microcantilevers of bending stiffness equal to 0.01 N/m. Linearity of the position-sensitive detector over the voltage range (0–8 V) used in this study was confirmed by the method of Silva and Van Vliet (2006). Surface areas (60 60 mm2) were scanned at a 901 tip angle, using a resolution of 128 128 pixels and a scanning frequency of 1 Hz. The normal stiffness (0.83 nN/V) of the AFM microcantilever was calibrated by performing force calibration on a glass substrate based on the thermal calibration method (Hutter and Bechhoefer, 1993) and Igor Pro (version 5) software. The normal load N was varied between 1.7 and 6.8 nN, resulting in tip normal displacements in the range of 2–5 nm, using the displacement corresponding to the tip engagement with the cartilage surface as a reference point. The mean (Hertzian) contact pressure pm ¼ ½16NE2 =9p3 R2 1=3 , where E ¼ ½ð1v2c Þ=Ec þ ð1v2t Þ=Et 1 , for Ec ¼2 MPa, nc ¼ 0.5, Et ¼ 200 GPa, nt ¼0.2, and N ¼ 1.7 and 6.8 nN is equal to 1.9 and 3.0 MPa, respectively. These pressure values are within the physiological contact pressure range. The lateral stiffness (102 nN/V) of the AFM microcantilever was calibrated by the direct force balance method (Asay and Kim, 2006) using a silicon grating (TGF11, Mikromasch, Wilsonville, OR). Normal and lateral stiffness calibrations performed before each test demonstrated good consistency. After each test, the AFM tip was cleaned with double deionized water and ethanol. The tip was frequently examined with an optical microscope to ensure that measurements in subsequent tests were not biased by possible tip damage and/or tissue transfer during previous scanning. Surface roughness was calculated from acquired AFM images as the root-mean-square of height deviations RS. Cartilage explants were scanned while fully immersed in phosphate buffered saline (Sigma-Aldrich, St. Louis, MO). Scans were acquired from five different surface locations of each explant. At each location, scans were obtained consecutively with increasing normal load to minimize any effects from a previous scan on the friction and adhesion force measurements obtained subsequently (Coles et al., 2008; Chan et al., 2010). The average friction force was determined from the halfwidth of the friction loop generated by the lateral (friction) force signal, i.e., onehalf of the difference between the mean lateral trace and retrace values, also accounting for any offset of the force sensor. The friction force was then averaged over all (5) scanned areas and all (6) samples tested and plotted as a function of applied normal load. The engineering friction coefficient m defined as
m¼
F N
ð1Þ
where F is the measured friction force and N is the externally applied normal load, was obtained as a function of normal load to determine if nanoscale friction of load- and non-load-bearing cartilage sliding in the boundary lubrication regime exhibits a dependence on the normal load. The true normal load applied at the contact interface may be affected by adhesion. The adhesion force acts normal to the contacting surfaces under both static and dynamic (sliding) contact conditions. However, the adhesion force of a static interface, measured at the instant of surface separation in the normal direction (pull-off force) differs from that of a dynamic contact interface subjected to shearing. Thus, to avoid confusion with the adhesion force of a static interface, the normal force measured under a zero external load is referred to as the adhesion force upon shearing. Therefore, to include the contribution of interfacial adhesion to the normal load, Eq. (1) was modified to the following equation: F ¼ m~ ðN þ Nad Þ
ð2Þ
where m~ is the true friction coefficient and Nad is the adhesion force measured during shearing in the absence of an external normal load. In view of Eq. (2), the friction force is expressed as F ¼ F0 þ m~ N
ð3Þ
where F0 ð ¼ m~ Nad Þ is the friction force in the absence of an external normal load, and is determined from the intercept of the line fit through the friction force data with the y-axis. For high normal loads, the effect of adhesion on the friction force becomes secondary, and Eq. (3) reduces to F m~ N, implying that at relatively high normal loads (NcNad ) the engineering friction coefficient approaches the true friction coefficient. Using Eqs. (1) and (3), the engineering friction coefficient can be expressed in terms of the true friction coefficient as follows: 1 m ¼ m~ þ F0 ð4Þ N Eq. (4) indicates that the engineering friction coefficient is inversely proportional to the externally applied normal load and approaches the true friction coefficient with the increase of the external normal load. 2.3. Immunohistochemistry
2.2. Friction, adhesion, and surface roughness measurements The adhesion and friction properties and surface morphology (roughness) of cartilage explants were studied with an AFM (MFP-3D-CF, Asylum Research, Santa
Immunolocalization of the known boundary lubricant in articular cartilage, superficial zone protein (SZP), was performed for M1 and M4 cartilage explants as described previously (Neu et al., 2007; Chan et al., 2010). Briefly, samples were
1342
S.M.T. Chan et al. / Journal of Biomechanics 44 (2011) 1340–1345
fixed in Bouin’s solution overnight, embedded in paraffin, exposed to primary antibody S6.79 (1:5000), and stained with an ABC kit (Vector Laboratories, Burlingame, CA) with secondary antibody mouse IgG.
2.4. Statistical analysis Explants obtained from six different bovine femoral condyles were used in the statistical data analysis (i.e., n¼ 6 for each type of explant (M1 or M4) and given external normal load). Data points shown in the figures presented in the next section represent mean values, whereas error bars indicate one standard error of the mean (SEM) above and below corresponding mean values. Friction force and friction coefficient data were fit with linear and nonlinear equations, respectively, and evaluated for goodness-of-fit by the coefficient of determination R2. Friction coefficient and surface roughness differences between M1 and M4 explants were determined from paired t-tests performed at each applied normal load for a significance level a ¼ 0.05.
A linear variation of the friction force with the applied normal load was observed for both M1 and M4 cartilage (Fig. 2A). Line fits yielded the following force relationships: FM1 ¼0.13N + 0.16 (R2 ¼0.978) and FM4 ¼0.20N + 0.61 (R2 ¼0.827). The slope of each linear fit represents the true friction coefficient of each cartilage region, i.e., m~ M1 ¼ 0:13 and m~ M4 ¼ 0:20. Load-bearing cartilage (M1) produced much lower friction forces than non-load-bearing cartilage (M4) through the entire normal load range. In addition, M1 cartilage yielded a significantly lower adhesion force Nad ð ¼ F0 =m~ Þ than M4 cartilage, i.e., Nad,M1 ¼ 1:21 7 0:4 1nN and Nad,M4 ¼ 2:97 70:39 nN (Fig. 2B). The mean and SEM values of the adhesion force are due to the experimental scatter in the friction force data (Fig. 2A) used to determine F0 and m~ . The engineering friction coefficient m of both load-bearing and non-load-bearing cartilage regions varied nonlinearly with the applied normal load N (Fig. 3). Importantly, this figure reflects the
3. Results Fig. 1 shows typical friction force responses of load-bearing (M1) and non-load-bearing (M4) cartilage for an external normal load of 6.8 nN. Since each friction test comprised back and forth scanning of the AFM tip over the same track under a constant normal load, two friction curves are shown in each plot. The fairly constant friction responses suggest that reciprocating sliding did not affect the dominant friction mechanism(s). The friction force of M1 and M4 cartilage, determined as one-half of the difference between the mean trace and retrace responses shown in Fig. 1, is equal to 1.15 and 1.85 nN, respectively.
Fig. 1. Representative friction force response of (A) load-bearing (M1) and (B) non-load-bearing (M4) cartilage sliding against a Si3N4 AFM tip under an externally applied normal load of 6.8 nN. The trace and retrace curves shown in each plot are due to back and forth scanning of the AFM tip.
Fig. 3. The engineering friction coefficient of cartilage sliding in the boundary lubrication regime decreased nonlinearly with the increase of the externally applied normal load, revealing a decreasing effect of adhesion on cartilage friction with increasing normal load. This trend was more pronounced for non-load-bearing cartilage (M4), whereas for load-bearing cartilage (M1) the friction coefficient remained relatively low and stable through the entire load range. In addition, loadbearing cartilage yielded lower friction coefficients than non-load-bearing cartilage through the entire normal load range (po0.032). Such difference in the nanoscale friction behavior of cartilage may be attributed to the formation of boundary films with higher contents of lubricious proteins, secreted at higher rates from chondrocytes in the superficial zone of load-bearing cartilage due to mechanotransduction effects. Data points represent mean values, while error bars indicate one SEM above and below the corresponding mean values.
Fig. 2. Adhesion force (measured during shearing in the absence of an external normal load) and friction force of load-bearing cartilage (M1) were consistently lower than those of non-load-bearing cartilage (M4), whereas surface roughness was similar. (A) Linear fits to the friction force versus normal load data yielded true friction coefficients (slopes) m~ M1 ¼ 0:13 and m~ M4 ¼ 0:20. (B) The adhesion force during shearing (obtained as the ratio of the friction force at zero external normal load and the true friction coefficient) of M1 and M4 cartilage was found equal to 1.21 70.41 and 2.97 70.39 nN, respectively. (C) Cartilage surface roughness did not show a dependence on joint location or discernible variation due to scanning under different loads (p ¼ 0.27). Data points represent mean values, while error bars indicate one SEM above and below the corresponding mean values.
S.M.T. Chan et al. / Journal of Biomechanics 44 (2011) 1340–1345
1343
Fig. 4. AFM images of (A, B) load-bearing cartilage (M1) and (C, D) non-load-bearing cartilage (M4) obtained after friction testing under (A,C) the lowest (1.7 nN) and (B,D) the highest (6.8 nN) applied normal load. The images do not reveal any discernible features of surface damage or topography changes due to sliding even for the highest normal load, in qualitative agreement with the relatively invariant surface roughness measured after friction testing through the entire normal load range. Corresponding RS values for these areas are (A) 306.03, (B) 317.26, (C) 406.79, and (D) 382.89 nm.
Fig. 5. Representative results of immunolocalization of SZP (a key boundary lubricant of cartilage) using monoclonal antibody S6.79 showed an increased expression at (A) load-bearing cartilage surfaces (M1) compared to (B) non-loadbearing cartilage surfaces (M4). The higher content of SZP at M1 cartilage surfaces correlates with the lower friction and adhesion characteristics of load-bearing cartilage.
Since all tested explants exhibited similar surface roughness (Fig. 2C), i.e., RS,M1 ¼374.24714.79 nm and RS,M2 ¼394.407 30.90 nm (p ¼0.27), the above trends in the nanoscale friction and adhesion properties of articular cartilage cannot be attributed to topographical differences. Moreover, sequential AFM scanning from the lowest to the highest normal load did not yield discernible changes in the surface morphology for any of the tested samples (Fig. 4), implying that adhesion was the dominant friction mechanism. However, histological staining for SZP showed significantly higher SZP content for M1 than M4 surfaces (Fig. 5), suggesting that variations in the nanoscale tribological properties were due to differences in the composition of the boundary films formed on load- and non-load-bearing cartilage surfaces. Similar results were obtained with other explants, confirming the consistently higher expression level of SZP in load-bearing (M1) than non-load-bearing (M4) cartilage surfaces.
4. Discussion decreasing effect of adhesion on friction with increasing normal load. Curve fitting of Eq. (4) through the data shown in Fig. 3 yielded the following relationships for the engineering friction coefficient of load-bearing and non-load-bearing cartilage: mM1 ¼ 0:126ð1=NÞ þ 0:139 (R2 ¼0.838) and mM4 ¼ 0:712ð1=NÞ þ0:178 (R2 ¼0.889). From an uncertainty analysis based on 95% confidence bounds it was found that the coefficients of these equations are within the range of the coefficients of the linear fits of Fig. 2A. Non-load-bearing joint regions exhibited significantly higher friction coefficients at each applied normal load (pr0.032), which decreased at a much faster rate with increasing normal load than the friction coefficients of load-bearing joint regions. The friction coefficients of both M1 and M4 cartilage decreased with the increase of the normal load, approaching asymptotically to steady-state values close to those of the true friction coefficients predicted from the slopes of the linear fits (Fig. 2A).
The main objective of this study was to determine if nanoscale friction of articular cartilage exhibits normal load dependence in the boundary lubrication regime. The presented results show that the engineering friction coefficient decreased nonlinearly with the increase of the applied normal load, with non-load-bearing cartilage demonstrating stronger load dependence than loadbearing cartilage (Fig. 3). At relatively high normal loads, the engineering friction coefficient approached the true friction coefficient (Fig. 3), determined from the slope of a linear fit through the friction force versus normal load data (Fig. 2A). Moreover, load-bearing cartilage exhibited lower friction coefficients than non-load-bearing cartilage through the entire normal load range examined (Fig. 3). Adhesion between the Si3N4 AFM tip and the cartilage surface was significant (Fig. 2B) and directly related to the friction response. The lower friction and adhesion of loadbearing cartilage than non-load-bearing cartilage also correlated
1344
S.M.T. Chan et al. / Journal of Biomechanics 44 (2011) 1340–1345
with the higher expression levels of boundary lubricants, such as SZP, at load-bearing joint regions (Fig. 5). In addition to the observed normal load dependence, nanoscale friction may show a dependence on sliding speed in both fluid film and boundary lubrication regimes (Drummond et al., 2003; Urbakh et al., 2004). The apparent violation of Amonton’s and Coulomb’s laws of friction that hold true at the macroscale can be attributed to scale effects and the dynamic characteristics of asperity microcontacts. In contrast to macroscale sliding contacts, micro/nanoscale contacts are constantly subjected to compressive, shear, and adhesive forces within microseconds (Urbakh et al., 2004). Transient asperity microcontacts generate attractive (adhesive) force fluctuations (on the order of several Pa to GPa) during these instantaneous load-unload cycles which are conducive to stick-slip (Drummond et al., 2003; Naka et al., 2007). Stick-slip can play a particularly important role in reciprocating sliding systems, such as synovial joints, where the reversal of the motion direction at zero-speed instances promotes the development of adhesion forces at asperity microcontacts. Significant adhesion effects on the friction behavior reported in other studies (Crockett et al., 2005; Coles et al., 2008) illustrate the importance of adhesion and stick-slip on the friction behavior of cartilage. Interestingly, an inverse relationship of adhesion and friction between the Si3N4 AFM tip and articular cartilage was previously observed (Chan et al., 2010). However, those adhesion force and hysteresis measurements were obtained with the more traditional approach that uses the force–displacement response obtained during the normal approach and retraction of the AFM tip from the cartilage surface (Crockett et al., 2005; Chan et al., 2010). While the normal approach method is useful for determining the adhesion energy and hysteresis, the lateral direction approach may be more relevant for measuring the adhesion component of friction at the nanoscale. Since the adhesion hysteresis depends on the sliding direction, a normal approach of the AFM tip to the cartilage surface may not fully represent the adhesion force or adhesion hysteresis associated with lateral displacement (sliding). Indeed, significantly higher adhesion forces have been measured in microscale sliding contact compared to adhesive pull-off forces measured during normal contact (Timpe and Komvopoulos, 2006). Necking and bridging forces due to bonding of atoms and/or ions between the contacting surfaces may further contribute to the increase of the adhesion force during normal surface approach (Zappone et al., 2008). One potential limitation to this study is the use of nonfunctionalized AFM probe tips, which are not representative of physiological conditions. However, others have used functionalized AFM tips to mimic cartilage surface chemistry or minimize protein adhesion (Coles et al., 2008, 2010) and obtained friction coefficients and surface roughness values similar to those found in this study. While the non-functionalized, sharp Si3N4 AFM tip may not reproduce the adhesion and plowing forces of a cartilage–cartilage sliding system, the existence of these friction mechanisms at asperity microcontacts persists regardless of surface chemistry and tip shape (Park et al., 2004; Coles et al., 2008). Therefore, AFM probe tips play an important role in understanding the nano/microscale friction behavior of cartilage in the boundary lubrication regime, where extensive wear may occur in the absence of fluid film lubrication. Even though the Si3N4 probe–cartilage interface may not ideally replicate the cartilage– cartilage interface, the current study nevertheless illuminates the relative differences between the tribological properties of loadbearing and non-load-bearing joint regions. The presence of boundary lubricants is essential for separating contacting solid surfaces during reciprocating sliding. Load-bearing cartilage demonstrated lower adhesion forces and friction coefficients than non-load-bearing cartilage (Figs. 1–3). Furthermore,
load-bearing cartilage also showed a weaker dependence of friction on normal load (Fig. 3). Under boundary lubrication sliding conditions, load-bearing cartilage produced nearly constant and low friction coefficient, which also suggests better wear resistance through a range of contact pressures under physiological settings. These results may be attributed to the higher expression levels of boundary lubricants in the load-bearing regions of the joint (Neu et al., 2007; Nugent-Derfus et al., 2007), as evidenced by the immunohistochemistry results (Fig. 5). The lower adhesion of load-bearing cartilage may also be attributed to the anti-adhesive and chondroprotective properties of SZP, which prevent synoviocyte overgrowth and cartilage–cartilage adhesion (Englert et al., 2005; Rhee et al., 2005). The results of the present study demonstrate a persistent dependence of the friction coefficient of articular cartilage on normal load. When joint contact conditions are not conducive to the formation of a sufficiently thick fluid film between the conformal cartilage surfaces, boundary lubricants, such as SZP (also known as lubricin and PRG4), may lower adhesion at the interfaces of asperity microcontacts and, consequently, reduce articular cartilage friction. In summary, the nanoscale friction and adhesion properties of articular cartilage from typical load-bearing and non-load-bearing regions of femoral condyles were examined with an AFM under a range of physiological contact conditions. Load-bearing cartilage demonstrated significantly lower adhesion and friction than nonload-bearing cartilage through the entire normal load range investigated in this study. This trend may be attributed to the higher surface content of boundary lubricants, such as SZP, in the load-bearing regions of the joint, which contribute to the formation of lubricious boundary films. These molecularly thin boundary films reduce friction and adhesion at asperity microcontacts occurring at those instances that the joint is starved from fluid film lubrication. Therefore, effective boundary lubrication of cartilage by continuously replenished boundary films is critical for normal joint function and for preventing tissue degeneration and wear caused by adhering asperities.
Conflict of interest statement The authors confirm that there are no conflicts of interest.
Acknowledgments The authors acknowledge the Spectral Imaging Facility, University of California, Davis, for the use of the AFM, and T.M. Schmid (Rush University, Chicago, IL) for the donation of the S6.79 antibody. References Asay, D.B., Kim, S.H., 2006. Direct force balance method for atomic force microscopy lateral force calibration. Rev. Sci. Instrum. 77, 043903. Bronzino, J.D., 2006. Biomedical Engineering Fundamentals. CRC Press, Boca Raton, LA. Caligaris, M., Ateshian, G.A., 2008. Effects of sustained interstitial fluid pressurization under migrating contact area, and boundary lubrication by synovial fluid, on cartilage friction. Osteoarthritis Cartilage 16, 1220–1227. Chan, S.M.T., Neu, C.P., DuRaine, G., Komvopoulos, K., Reddi, A.H., 2010. Atomic force microscope investigation of the boundary-lubricant layer in articular cartilage. Osteoarthritis Cartilage 18, 956–963. Coles, J.M., Blum, J.J., Jay, G.D., Darling, E.M., Guilak, F., Zauscher, S., 2008. In situ friction measurement on murine cartilage by atomic force microscopy. J. Biomech. 41, 541–548. Coles, J.M., Zhang, L., Blum, J.J., Warman, M.L., Jay, G.D., Guilak, F., Zauscher, S., 2010. Loss of cartilage structure, stiffness, and frictional properties in mice lacking PRG4. Arthritis Rheum. 62, 1666–1674. Crockett, R., Roos, S., Rossbach, P., Dora, C., Born, W., Troxler, H., 2005. Imaging of the surface of human and bovine articular cartilage with ESEM and AFM. Tribol. Lett. 19, 311–317.
S.M.T. Chan et al. / Journal of Biomechanics 44 (2011) 1340–1345
Drummond, C., Israelachvili, J., Richetti, P., 2003. Friction between two weakly adhering boundary lubricated surfaces in water. Phys. Rev. E Stat. Nonlin. Soft Matter Phys. 67, 066110. DuRaine, G., Neu, C.P., Chan, S.M.T., Komvopoulos, K., June, R.K., Reddi, A.H., 2009. Regulation of the friction coefficient of articular cartilage by TGF-b1 and IL-1b. J. Orthop. Res. 27, 249–256. Englert, C., McGowan, K.B., Klein, T.J., Giurea, A., Schumacher, B.L., Sah, R.L., 2005. Inhibition of integrative cartilage repair by proteoglycan 4 in synovial fluid. Arthritis Rheum. 52, 1091–1099. Gleghorn, J.P., Bonassar, L.J., 2008. Lubrication mode analysis of articular cartilage using Stribeck surfaces. J. Biomech. 41, 1910–1918. Gleghorn, J.P., Jones, A.R.C., Flannery, C.R., Bonassar, L.J., 2009. Boundary mode lubrication of articular cartilage by recombinant human lubricin. J. Orthop. Res. 27, 771–777. Gleghorn, J.P., Doty, S.B., Warren, R.F., Wright, T.M., Maher, S.A., Bonassar, L.J., 2010. Analysis of frictional behavior and changes in morphology resulting from cartilage articulation with porous polyurethane foams. J. Orthop. Res. 28, 1292–1299. Hutter, J.L., Bechhoefer, J., 1993. Calibration of atomic-force microscope tips. Rev. Sci. Instrum. 64, 1868–1873. Jay, G.D., Torres, J.R., Rhee, D.K., Helminen, H.J., Hytinnen, M.M., Cha, C.J., Elsaid, K., Kim, K.S., Cui, Y., Warman, M.L., 2007. Association between friction and wear in diarthrodial joints lacking lubricin. Arthritis Rheum. 56, 3662–3669. Krishnan, R., Kopacz, M., Ateshian, G.A., 2004. Experimental verification of the role of interstitial fluid pressurization in cartilage lubrication. J. Orthop. Res. 22, 565–570. Lewis, P.R., McCutchen, C.W., 1959. Mechanism of animal joints: experimental evidence for weeping lubrication in mammalian joints. Nature (London) 184, 1285. Marcelino, J., Carpten, J.D., Suwairi, W.M., Gutierrez, O.M., Schwartz, S., Robbins, C., Sood, R., Makalowska, I., Baxevanis, A., Johnstone, B., Laxer, R.M., Zemel, L., Kim, C.A., Herd, J.K., Ihle, J., Williams, C., Johnson, M., Raman, V., Alonso, L.G., Brunoni, D., Gerstein, A., Papadopoulos, N., Bahabri, S.A., Trent, J.M., Warman, M.L., 1999. CACP, encoding a secreted proteoglycan, is mutated in camptodactyly-arthropathy-coxa vara-pericarditis syndrome. Nat. Genet. 23, 319–322. Naka, M.H., Hasuo, M., Fuwa, Y., Ikeuchi, K., 2007. Correlation between friction of articular cartilage and reflectance intensity from superficial images. Tribol. Int. 40, 200–207. Neu, C.P., Khalafi, A., Komvopoulos, K., Schmid, T.M., Reddi, A.H., 2007. Mechanotransduction of bovine articular cartilage superficial zone protein by transforming growth factor b signaling. Arthritis Rheum. 56, 3706–3714. Neu, C.P., Komvopoulos, K., Reddi, A.H., 2008. The interface of functional biotribology and regenerative medicine in synovial joints. Tissue Eng. Part B Rev. 14, 235–247.
1345
¨ Nugent-Derfus, G.E., Takara, T., O’Neill, J.K., Cahill, S.B., Gortz, S., Pong, T., Inoue, H., Aneloski, N.M., Wang, W.W., Vega, K.I., Klein, T.J., Hsieh-Bonassera, N.D., Bae, W.C., Burke, J.D., Bugbee, W.D., Sah, R.L., 2007. Continuous passive motion applied to whole joints stimulates chondrocyte biosynthesis of PRG4. Osteoarthritis Cartilage 15, 566–574. Park, S., Costa, K.D., Ateshian, G.A., 2004. Microscale frictional response of bovine articular cartilage from atomic force microscopy. J. Biomech. 37, 1679–1687. Rhee, D.K., Marcelino, J., Baker, M., Gong, Y., Smits, P., Lefebvre, V., Jay, G.D., Stewart, M., Wang, H., Warman, M.L., Carpten, J.D., 2005. The secreted glycoprotein lubricin protects cartilage surfaces and inhibits synovial cell overgrowth. J. Clin. Invest. 115, 622–631. Schmidt, T.A., Gastelum, N.S., Nguyen, Q.T., Schumacher, B.L., Sah, R.L., 2007. Boundary lubrication of articular cartilage: role of synovial fluid constituents. Arthritis Rheum. 56, 882–891. Shi, L., Sikavitsas, V.I., Striolo, A., 2011. Experimental friction coefficients for bovine cartilage measured with a pin-on-disk tribometer: testing configuration and lubricant effects. Ann. Biomed. Eng. 39, 132–146. Silva, E.C.C.M., Van Vliet, K.J., 2006. Robust approach to maximize the range and accuracy of force application in atomic force microscopes with nonlinear position-sensitive detectors. Nanotechnology 17, 5525–5530. Stachowiak, G.W., Batchelor, A.W., Griffiths, L.J., 1994. Friction and wear changes in synovial joints. Wear 171, 135–142. Stolz, M., Gottardi, R., Raiteri, R., Miot, S., Martin, I., Imer, R., Staufer, U., Raducanu, ¨ A., Duggelin, M., Baschong, W., Daniels, A.U., Friederich, N.F., Aszodi, A., Aebi, U., 2009. Early detection of aging cartilage and osteoarthritis in mice and patient samples using atomic force microscopy. Nature Nanotechnol. 4, 186–192. Timpe, S.J., Komvopoulos, K., 2006. The effect of adhesion on the static friction properties of sidewall contact interfaces of microelectromechanical devices. J. Microelectromech. Syst. 15, 1612–1621. Unsworth, A., 1991. Tribology of human and artificial joints. Proc. Inst. Mech. Eng. H 205, 163–172. Urbakh, M., Klafter, J., Gourdon, D., Israelachvili, J., 2004. The nonlinear nature of friction. Nature 430, 525–528. Walker, P.S., Dowson, D., Longfield, M.D., Wright, V., 1968. ’’Boosted lubrication’’ in synovial joints by fluid entrapment and enrichment. Ann. Rheum. Dis. 27, 512–520. Winter, D.A., 2009. Biomechanics and Motor Control of Human Movement 4th ed. Wiley, Hoboken, NJ. Zappone, B., Thurner, P.J., Adams, J., Fantner, G.E., Hansma, P.K., 2008. Effect of Ca2 + ions on the adhesion and mechanical properties of adsorbed layers of human osteopontin. Biophys. J. 95, 2939–2950.