Regulation of the Friction Coefficient of Articular Cartilage by TGF-b1 and IL-1b Grayson DuRaine,1 Corey P. Neu,1 Stephanie M.T. Chan,1 Kyriakos Komvopoulos,2 Ronald K. June,1 A. Hari Reddi1 1 Center for Tissue Regeneration and Repair, Department of Orthopaedic Surgery, University of California, Davis, Medical Center, 4635 Second Ave., Sacramento, CA 95817, 2Department of Mechanical Engineering, University of California, Berkeley, CA 94720

Received 19 April 2007; accepted 24 April 2008 Published online 6 August 2008 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jor.20713

ABSTRACT: Articular cartilage functions to provide a low-friction surface for joint movement for many decades of life. Superficial zone protein (SZP) is a glycoprotein secreted by chondrocytes in the superficial layer of articular cartilage that contributes to effective boundary lubrication. In both cell and explant cultures, TGF-b1 and IL-1b have been demonstrated to, respectively, upregulate and downregulate SZP protein levels. It was hypothesized that the friction coefficient of articular cartilage could also be modulated by these cytokines through SZP regulation. The friction coefficient between cartilage explants (both untreated and treated with TGF-b1 or IL-1b) and a smooth glass surface due to sliding in the boundary lubrication regime was measured with a pin-on-disk tribometer. SZP was quantified using an enzyme-linked immunosorbant assay and localized by immunohistochemistry. Both TGF-b1 and IL-1b treatments resulted in the decrease of the friction coefficient of articular cartilage in a location- and time-dependent manner. Changes in the friction coefficient due to the TGF-b1 treatment corresponded to increased depth of SZP staining within the superficial zone, while friction coefficient changes due to the IL-1b treatment were independent of SZP depth of staining. However, the changes induced by the IL-1b treatment corresponded to changes in surface roughness, determined from the analysis of surface images obtained with an atomic force microscope. These findings demonstrate that the low friction of articular cartilage can be modified by TGF-b1 and IL-1b treatment and that the friction coefficient depends on multiple factors, including SZP localization and surface roughness. ß 2008 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 27:249– 256, 2009 Keywords: cartilage; cytokine; growth factors; friction coefficient; SZP/lubricin/PRG4

Articular cartilage is an avascular tissue with limited innate potential for repair and regeneration that provides a low-friction surface for joint movement.1 Lubrication of these surfaces is critical to normal joint function. A boundary lubricant of articular cartilage, referred to as superficial zone protein (SZP), has been identified as a product of the proteoglycan 4 gene (PRG4).2 SZP is a glycoprotein secreted by chondrocytes in the superficial layer of articular cartilage and is homologous to lubricin and megakaryocyte stimulating factor (MSF) precursor.3 The boundary lubricating efficacy of SZP, either separately or in conjunction with other synovial fluid components, has been studied at latex-glass,4–6 cartilage-cartilage,7,8 and cartilageglass9 interfaces. Recent work10 revealed a correlation between SZP expression level at the articular surface and short-term friction coefficient at the cartilage-glass interface. In addition to its function as a boundary lubricant, SZP inhibits integrative cartilage repair and synovial cell overgrowth.11,12 SZP is a significant protein that plays a key role in the normal function of synovial joints, and human and mouse mutants of the PRG4 gene display precocious arthropathy.12,13 Changes in cartilage homeostasis are thought to precede the initiation of osteoarthritis.14 A model of cartilage homeostasis involves the steady-state maintenance of the tissue and the interplay between mechanical and biochemical environmental signals, such as cytokines and their interactions with cells and surrounding extracellular matrix.1 With reference to cytokines, cartilage homeostasis has been modeled as the balance Correspondence to: Corey P. Neu (T: 916-734-3311; F: 916-7345750; E-mail: [email protected]) ß 2008 Orthopaedic Research Society. Published by Wiley Periodicals, Inc.

between the actions of morphogens and growth factors, such as transforming growth factors (TGF-b), bone morphogenetic proteins, and insulin-like growth factor, and the actions of catabolic cytokines, such as interleukins (IL-1b) and tumor necrosis factor alpha. Introduction of inflammatory (catabolic) cytokines, such as IL-1b, to the articular joint reduces the production of matrix molecules and leads to changes in matrix constitution that result in reduced mechanical strength and ability to maintain normal cartilage function.15,16 It is well established that cytokines play an important role in cartilage homeostasis and that SZP is regulated in part by the cytokines of the joint involved in cartilage homeostasis.17,18 Previous research in cell culture demonstrated that the level of SZP secreted into the media can be controlled by applying different cytokines.19 Anabolic cytokines and growth factors (e.g., TGF-b family members) increase the expression of SZP, while catabolic cytokines (e.g., IL-1b) decrease the expression and accumulation of SZP.17,19 Differences in the friction coefficient at regions of the articular cartilage surface with distinct SZP protein levels motivated the investigation of the effect of changes in the SZP expression level due to cytokine treatment on the coefficient of friction. The main objectives of this investigation were to evaluate the friction coefficient of cartilage explants treated with TGF-b1 and IL-1b and examine changes in the friction coefficient in terms of SZP accumulation and explant surface roughness. An unexpected decrease in the friction coefficient was observed after IL-1b treatment, independent of SZP accumulation. Since IL-1b has been previously identified to cause damage of the articular surface,16 this finding was interpreted in the context of articular surface topography (roughness) characteristics. JOURNAL OF ORTHOPAEDIC RESEARCH FEBRUARY 2009

249

250

DURAINE ET AL.

EXPERIMENTAL PROCEDURES Materials Articular cartilage was harvested from 1–3-month-old bovine stifle (tibiofemoral) joints obtained from a local abattoir within 6 h of sacrifice. DMEM/F12 and antibiotic solution (both from Invitrogen, Grand Island, NY), cell culture components and reagents (from Sigma, St. Louis, MO or Fisher Scientific, Tustin, CA), VECTASTAIN elite ABC Reagent (PK-6105) and DAB Substrate Kit (both from Vector Laboratories, Burlingame, CA), and human recombinant TGF-b1 or IL-1b (both from R&D Systems, Minneapolis, MN) were used in this study. Tissue Acquisition and Culture Bovine calf stifle joints were opened using an aseptic technique, and two side-by-side osteochondral explants were harvested from M1 (anterior and load bearing) and M4 (posterior and non-load bearing) regions of the femoral medial condyles10 (Fig. 1). Hereafter, these samples will be referred to as M1 and M4 explants. Four total explants were collected from each joint, specifically two explants from location M1 and two from location M4. Only one joint per animal was used in an experimental group. Explants were matched per joint during treatment (e.g., control and treated explants from region M1 of the same joint). The designation as control or treated was assigned randomly per pair. A 5-mm coring reamer was used to extract the osteochondral explants, whereas an adjustable custom-made jig was used to remove subchondral bone and trim the explants to lengths of 4 mm. The explants were allowed to submerse and equilibrate in 2 mL of serum- and cytokine-free culture medium consisting of DMEM/F12, 0.2% bovine serum albumin (BSA), 1% penicillin/streptomycin, 50 mg/mL ascorbic acid 2-phosphate, and 5% CO2 at 378C for 24 h. The medium was replaced after the equilibration period. Subsequently, the explants were incubated in a medium supplemented with 10 ng/mL of TGF-b1 or 10 ng/mL of IL-1b for 2 or 5 days. Control explants were treated with vehicle alone, i.e., 5 mM HCl and 0.1% BSA for the TGF-b1 group and phosphate-buffered saline (PBS) and 0.1% BSA for the IL-1b group (1 mL per mL of media). The media were changed at day 2 for the 5-day experimental groups, and fresh cytokine or control vehicle was added. Concentrations of TGF-b1 and IL-1b were chosen based on previous changes in SZP protein expression in culture.19,20 A decrease in SZP protein expression has been observed for IL-1a treatment of articular cartilage7,21,22 and for IL-1b in mandibular condyle chondrocytes23 at similar concentrations. Friction Testing To determine the effect of growth factors or cytokines on the friction coefficient of the articular surface, cartilage explants treated with either TGF-b1 or IL-1b were tested using a

pin-on-disk tribometer operated in the boundary lubrication regime in reciprocating sliding mode.24 Explants were harvested from locations M1 and M4 as described previously. The number of samples (n ¼ 9–14) depended on the specific group tested (i.e., n ¼ 9 for 2-day TGF-b1 treatment of M1 and M4 explants; n ¼ 11 and 12 for 5-day TGF-b1 treatment of M1 and M4 explants, respectively; n ¼ 10 for 2-day IL-1b treatment of M1 and M4 explants; and n ¼ 13 and 14 for 5-day IL-1b treatment of M1 and M4 explants, respectively). The pin specimens consisted of explants affixed to acrylic pins by ethyl cyanoacrylate. The articular surface was brought into contact with a polished glass disk and was fully immersed in PBS. The choice of PBS for the current work was made to prevent confounded interactions between native cartilage surface molecules (e.g., SZP) and those in a test solution, such as whole synovial fluid. The glass disk was repositioned or ultrasonically cleaned after every fourth specimen to ensure a fresh surface. In all of the tests, the sliding speed was 0.5 mm/s, the radius of the wear track was 5 mm, and the normal load was 1.8 N, resulting in an average contact pressure of 0.1 MPa. These load and speed conditions have been shown to yield sliding in the boundary lubrication regime.10 Prior to the initiation of each friction test, the sample was allowed to equilibrate under the applied load (i.e., 0.1 MPa average pressure) in an unconstrained test configuration for 2 min to minimize any fluid effects during testing.10 Test parameters, such as normal load, sliding speed, and equilibrium conditions, were determined from analytical and experimental analyses to allow for sufficient interstitial fluid depressurization and to obtain reproducible coefficient of friction measurements in the boundary lubrication regime.10,25 The test duration of each friction experiment was fixed at 60 s. Data were collected at 0.1 s intervals with a data acquisition system (Labview, National Instruments, Austin, TX) and processed using a standard software package (Microsoft, Seattle, WA). Friction force data from each entire test were used to compute the mean and standard deviation values of the coefficient of friction for each femoral condyle location and treatment. These experimental conditions have been previously used to investigate the effect of SZP expression level on the coefficient of friction of articular cartilage.10 SZP Quantification and Immunohistochemistry Samples from the treatment groups that resulted in significantly different changes in the friction coefficient were used for SZP quantification and localization. SZP accumulation in the media was quantified by enzyme-linked immunosorbent assay (ELISA), using SZP purified from bovine synovial fluid and cultured synovium and articular cartilage as a standard.26 The SZP protein was quantified in media collected from a subset of

Figure 1. Dependence of immunolocalization of the SZP protein on femoral condyle location at the articular cartilage surface. The SZP expression at anterior harvest location M1 extended several cell layers into the tissue compared to the posterior location M4. JOURNAL OF ORTHOPAEDIC RESEARCH FEBRUARY 2009

FRICTION COEFFICIENT OF ARTICULAR CARTILAGE

cartilage explants of standardized volume (n ¼ 9 for each group tested) following treatment with TGF-b1 or IL-1b. The SZP content was determined using a standard method from samples that were serially diluted with PBS and reacted with S6.79 (1:5,000) as the primary antibody and anti-mouse horseradish peroxidase-conjugated goat (1:3,000, Vector Laboratories) as the secondary antibody. For immunohistochemistry, tissues were fixed in Bouin’s solution overnight, followed by paraffin embedding and sectioning. Immunostaining was performed following a standard method with S6.79 (1:1,000) as the primary antibody and an ABC kit (Vector Laboratories) with mouse IgG as the secondary antibody for signal detection. Another set of consecutive sections was stained with 1% toluidine blue for histological examination to verify expected patterns of proteoglycan content (not shown here for brevity). Images were obtained with an optical microscope (LSM 510, Carl Zeiss, Jena, Germany) at 100 magnification. Surface Roughness Measurement Treatment groups that showed significant differences in the friction characteristics were further examined in additional samples using an atomic force microscope (AFM) (model MFP3D-CF, Asylum Research, Santa Barbara, CA) with an extended z-range of 28 mm. Triangular silicon nitride tips (model MSCT-AUNM, Veeco, Santa Barbara, CA) with a nominal tip radius of 10 nm and spring constant of 0.01 N/m were used in all of the AFM scans. AFM images of 60  60 mm2 scan areas were obtained in contact mode using 128  128 pixels, 1 Hz scan rate, and 2.5 V set point. Force calibration performed with a glass substrate, using Igor Pro software (version 5, Wavemetrics, Lake Oswego, OR) and based on the thermal calibration method, showed that this set point corresponded to an applied normal contact force of 2.39  0.02 nN. Samples sectioned with a razor blade to a thickness of 1.5 mm and affixed with ethyl cyanoacrylate to a custommade sample holder were immersed in PBS throughout imaging to maintain tissue hydration. Sample groups from each joint comprised treated and control pairs from both M1 and M4 joint locations (Fig. 1), i.e., four cartilage explants per joint. Five pairs of M1 explants and six pairs of M4 explants treated with TGF-b1, and eight pairs of M1 explants and seven pairs of M4 explants treated with IL-1b were used in the AFM analysis. Images from five different locations of each explant surface were used to calculate the mean and standard deviation values of the root-mean-square surface roughness Rq given by " #1=2 N 1X 2 Rq ¼ z ; ð1Þ N i¼1 i where zi is the deviation of the ith measured height from the mean surface plane and N is the number of data points in each surface scan. Mechanical Testing Compression experiments were performed to evaluate possible chemical treatment and time-dependent load effects on the bulk material properties. For this purpose, pairs of articular cartilage explants harvested from bovine stifle joints were used as control and treated [5-day treatment with mediasupplemented IL-1b (10 ng/mL), as described previously] samples, for a total of n ¼ 6, to perform compression experiments. After measuring the sample dimensions with digital

251

calipers, the samples were placed in a custom-made chamber of a materials testing system (Enduratec 3200, Eden Prairie, MN) where they were compressed by polished and nonporous stainless steel platens at a rate of 50 mm/s to an average pressure of 12.5 kPa (preload). Subsequently, a 5% nominal compression was applied at a rate of 10 mm/s, resulting in a peak average pressure of 361.0  0.1 kPa, and data were collected over a period of 30 min at a sampling frequency of 180 Hz. The average (engineering) stress was determined from the initial cross-sectional area calculated from the measured diameter. The initial sample height used to calculate the engineering strain was measured at the equilibrium platen position after a 10 min equilibration time. All of the samples were incubated and tested at 378C. The bulk stiffness was obtained as the ratio of the engineering stress to the engineering strain. Statistical Analysis For each treatment (TGF-b1 or IL-1b), location (M1 or M4), and treatment time (2 or 5 days) combination, a paired t-test was performed to determine any significant differences among the friction coefficients of the treated and untreated control samples using a standard software package (SAS Institute, Cary, NC). Differences in SZP expression following cytokine treatment were evaluated using a paired t-test. Average surface roughness measurements were compared using a one-way nested ANOVA for four treatment levels (M1 control, M1 treated, M4 control, and M4 treated) and both 5-day IL-1b treatment and 2-day TGF-b1 treatment groups. A nested analysis was used to account for the multiple (five) measurements obtained from each explant. A significance level of p < 0.05 was used to determine differences between groups. A paired t-test was also used in the statistical analysis of the bulk stiffness of the control and treated samples used in the compression experiments. Results are presented in the form of data points representing mean values  one standard error of the mean.

RESULTS Coefficient of Friction The friction coefficient demonstrated a dependence on anatomical location, cytokine, and treatment duration. The 2-day TGF-b1 treatment of posterior M4 explants produced a decrease in friction coefficient compared to the untreated explants (p ¼ 0.035; Fig. 2A). The friction coefficients of the explants from other locations did not change (statistically) following 2-day TGF-b1 or IL-1b treatment compared to controls (p > 0.445). Five-day IL-1b treatment of M1 explants resulted in lower friction coefficient than that of the controls (p ¼ 0.015; Fig. 2B). The friction coefficients of explants from other locations did not change following 5-day TGF-b1 or IL-1b treatment compared to controls (p > 0.396). SZP Quantification and Immunohistochemistry SZP accumulation in media did not vary statistically following either 2-day TGF-b1 treatment or 5-day IL-1b treatment of M1 explants compared to untreated controls (p > 0.149; Fig. 3A, B). The SZP depth of staining in M4 explants increased after a 2-day TGFb1 treatment compared to untreated controls (Fig. 3C). Furthermore, SZP localization after TGF-b1 treatment JOURNAL OF ORTHOPAEDIC RESEARCH FEBRUARY 2009

252

DURAINE ET AL.

Figure 2. Effects of femoral condyle (explant) location and type and duration of treatment on the friction coefficient of the cartilage surface. (A) Two-day treatment with TGF-b1 of explants from location M4 yielded a decrease in friction coefficient over untreated controls, while 2-day treatment with IL-1b produced insignificant differences compared to that of controls. (B) Five-day treatment with IL-1b of explants from location M1 yielded a decrease in friction coefficient over untreated controls, while 5-day treatment with TGF-b1 or IL-1b of all the other explants did not produce statistically different friction coefficients. (The bars with the asterisks correspond to p < 0.035.)

of the M4 explants was similar to that of the untreated M1 explants. The SZP staining in all of the cartilage explants did not show discernible changes after a 5-day IL-1b treatment (Fig. 3D). Surface Roughness Figure 4 shows surface roughness results and representative topography images of treated and control explants obtained from cartilage locations M1 and M4. The surface roughness [Equation (1)] of the M1 explants

increased as a result of the 5-day IL-1b treatment (Rq ¼ 465.96  63.53 nm) compared to that of the controls (Rq ¼ 271.41  22.92 nm) (p < 0.0001), while other differences between treatments with TGF-b1 and IL-1b and explants from locations M1 and M4 were insignificant (p > 0.05) (Fig. 4A, B). The AFM images shown in Figure 4C and D demonstrate that the surface topographies of the M1 explants contained fibrillated structures and fine-scale irregularities, while those of the M4 explants were essentially featureless. It appears

Figure 3. Dependence of SZP staining depth in cartilage explants on location and type of treatment: (A) 2-day treatment with TGF-b1; (B) 5-day treatment with IL-1b. (C) and (D) show corresponding immunolocalization SZP results. JOURNAL OF ORTHOPAEDIC RESEARCH FEBRUARY 2009

FRICTION COEFFICIENT OF ARTICULAR CARTILAGE

253

Figure 4. Surface roughness of cartilage explants from locations M1 and M4 treated for 5 days with (A) TGF-b1 and (B) IL-1b. Representative AFM images of explant surface topographies after (C) 2-day treatment with TGF-b1 and (D) 5-day treatment with IL-1b. Results from corresponding untreated (control) explants are shown as a reference. (The bar with the asterisk corresponds to p < 0.0001.)

that the 5-day IL-1b treatment resulted in loss of the fibrillated structures (especially for M1 explants) and the development of larger wavelength surface waviness. Mechanical Behavior The bulk properties of the explants were not influenced by the IL-1b treatment. In particular, insignificant differences (p > 0.512) in the stiffness of control and IL-1b treated explants were found after 5, 120, and 1,800 s (i.e., equilibrium) in stress relaxation experiments (Fig. 5). In addition, the stiffness was constant for both control and treated samples after 120 s.

DISCUSSION The presented results indicate a dependence of the friction coefficient of articular cartilage on explant location and cytokine (TGF-b1 or IL-1b) treatment. It was found that the SZP staining depth in the cartilage explants was influenced by the type and duration of the TGF-b1 treatment. These experimental findings demonstrate that the addition of these cytokines has direct implications on the friction coefficient of the articular surface. The decrease in the friction coefficient following the 2-day TGF-b1 treatment corresponded to increased SZP staining depth in the explants. The 5-day IL-1b treatment resulted in significant surface roughening of the explants harvested from the anterior cartilage location. The decrease in the friction coefficient of the M4 explants following a 2-day TGF-b1 treatment (Fig. 2A) was attributed to the increased depth of staining of SZP in the near-surface region (Fig. 3A, C). SZP immunolocalization to a depth within the cartilage has been shown previously.19,20,27 The SZP level and the decrease in the

friction coefficient of these explants were similar to those of the untreated M1 explants. An increase in the depth of SZP staining of the M1 explants was also detected after a 2-day TGF-b1 treatment, although the friction coefficient did not differ significantly from that of the untreated M1 explants (p > 0.05). The lack of further reduction in the friction coefficient of the TGF-b1 treated M1 explants is presumed to be a consequence of an articular surface with saturated SZP (Fig. 3C). A comparison of SZP accumulation in the media of TGFb1 treated and untreated M1 explants (p > 0.05, Fig. 3A) suggests that this treatment did not result in a significant increase in SZP released into the medium, despite the increased staining depth indicated by the SZP immunolocalization results (Fig. 3C). Although SZP

Figure 5. Stiffness of untreated and treated (5-day treatment with IL-1b) cartilage explants obtained after 5, 120, and 1800 s (equilibrium) under 5% compression. The data show insignificant differences (p > 0.512) in stiffness between treated and untreated samples for a given time. JOURNAL OF ORTHOPAEDIC RESEARCH FEBRUARY 2009

254

DURAINE ET AL.

was released in the media of the 2-day culture model, the former finding suggests that an available pool of SZP still remained within the explant, depending on sample location and cytokine treatment. The 5-day TGF-b1 treatment did not produce a decrease in the friction coefficient of the M4 explants (Fig. 2B). It is believed that this resulted from the loss of SZP under the longer culture conditions (even with the TGF-b1 treatment). The SZP (PRG4) concentration is commonly expressed in mg/cm2. Based on the data shown in Figure 3A and the 5-mm explant diameter and 2-mL media volume, the SZP expression levels of M1 and M4 explants are in the range of 100–200 mg/cm2. Recently published data of SZP released into media using slightly smaller (i.e., 3 mm diameter) explants from the patellofemoral groove have shown SZP expression levels of 30 and 40 mg/cm2 per 24 h after incubation for 1 and 2 days, respectively.27 The decrease in the friction coefficient following a 5-day IL-1b treatment of the M1 explants (Fig. 2B) indicated that a compensatory mechanism might have increased the SZP in the matrix similar to that observed for the 2-day TGF-b1 treatment. An increase in SZP that would account for the decrease of the friction coefficient did not occur during the IL-1b treatment (Fig. 3). However, alterations in GAG content,28 which was not measured in this study, have been observed to change the friction coefficient. The friction coefficients of the 2-day IL-1b treated M1 and M4 explants were similar (Fig. 2B). Differences in the SZP staining in the controls of the 2-day and 5-day immunohistochemistry data (Fig. 3C, D) were attributed to culture conditions (i.e., duration of treatment) and not to differences in the vehicle controls, since both would have been diluted at a ratio of 1:1,000. Although changes in the cartilage coefficient of friction that resulted from the 2-day TGF-b1 treatment corresponded to the increase of the SZP staining depth in the superficial zone, in the case of the 5-day IL-1b treatment, the friction coefficient did not correspond to the SZP staining depth. Consequently, it was hypothesized that the IL-1b could have had a destructive effect on the articular cartilage surface. The destruction of the articular cartilage matrix by IL-1b is a well-known effect that produces characteristic damage features attributable to surface erosion.15,29 IL-1b has been partly implicated in cartilage matrix degradation by enhancing the expression of matrix metalloproteinases (MMPs)14,30 and aggrecanases.29 Furthermore, there is evidence that IL-1b and TNF-a colocalize with MMPs in the superficial layer of the arthritic cartilage, revealing a key role of this layer in the pathogenesis of arthritic diseases.15 Cartilage surface roughening was only observed with M1 explants following 5-day IL-1b treatment (Fig. 4B). Surface topography images (Fig. 4C, D) showed a generally fibrous structure with feature heights as large as 1 mm. The disruption of the distinct fibrous structure on the surfaces of the M1 explants after 5-day IL-1b treatment may be responsible for both the increased surface roughness and the higher scatter in the roughJOURNAL OF ORTHOPAEDIC RESEARCH FEBRUARY 2009

ness measurements. Surface roughening decreases the real contact area, resulting in lower friction.31 This effect is bimodal since large asperities may increase friction due to the enhancement of asperity interlocking, while very smooth surfaces yield larger real contact areas and, in turn, higher friction coefficients. Alternatively, early destruction of the cartilage matrix components by IL-1b actions may lead to a loosely bound and short-lived layer of molecules. While the low shear strength of this layer could contribute to a decrease in the coefficient of friction, it would be expected to also increase the wear rate dramatically, producing a detrimental effect on the longterm durability of the joint. Therefore, the higher surface roughness of articular cartilage after 5-day IL-1b treatment may be attributed to the degradation of molecules in the lamina splendens resulting from enzymatic degradation cascades induced by the IL-1b. However, it is not clear whether surface roughening was due to an overall increased fibrillation of small to moderately sized topographical maxima and minima occurring in high frequency at the surface or the formation of large but infrequent extremes along an otherwise smooth surface. Insight into the specific roughening mechanism during IL-1b treatment would elucidate the pattern and distribution of the degradation of surface and matrix molecules and their role in boundary lubrication of articular joints. Additionally, representative images (Fig. 4) were chosen based on the collective difference between the average roughness of all measurements and those of a single joint. The surface roughness of the chosen sample was within 33 nm of respective average roughness values. AFM images revealed variability in the surface structure between joints, which has also been observed in other studies.32,33 In the present work, AFM images were acquired using an ultra-sharp (i.e., 10 nm radius) tip attached to a flexible microcantilever. Although an ultra-sharp tip is desirable for imaging fine surface features, it may alter the surface topography by scratching. To minimize the contact force exerted on the cartilage surface by the sharp tip while maintaining the capability to detect extreme height variations, a microcantilever of very low stiffness was used in all of the AFM scans. Similarly sharp probe tips have been used to image articular cartilage, and the obtained blurred images were attributed to the high viscosity of the surface.33 However, the probe stiffness in that study was two times higher than that used in this investigation. Except for some evidence of surface smearing, a microcantilever of stiffness equal to 0.01 N/m (as opposed to 0.022 N/m) mitigated the problem of smearing in the present AFM scans. The surface roughness values of the control samples (typically 271  23 nm) (Fig. 4A, B) are comparable with the roughness of articular cartilage surfaces measured with micrometer-sized probe tips (379  83 nm)34 and blunt probe tips (462  216 nm).32 The decrease in the friction coefficient of the articular cartilage treated with IL-1b is likely culture condition dependent. To avoid confounding results of other

FRICTION COEFFICIENT OF ARTICULAR CARTILAGE

cytokines on SZP expression during TGF-b1 or IL-1b treatment, all tissues were maintained in serum-free chemically defined media. Furthermore, the explants remained in an unloaded free-swelling condition. This culture technique does not exactly mimic the in vivo situation of a rich cytokine environment and associated mechanical loading, both of which can regulate SZP expression.17,19,21,35,36 This may further explain the loss of SZP in the controls of the 5-day culture media and tissue (Fig. 3B, D). The effects of potential confounding factors, such as interstitial fluid pressurization and/or bulk changes in the matrix due to chemical treatment, on the obtained results were found to be minimal. Considering the contact parameters used in this study, particularly the low average contact pressure (0.1 MPa), and the stiffness results obtained from compression experiments (Fig. 5), fluid depressurization occurred within 2 min under load. In addition, increasing the equilibration time up to 10 min did not produce any changes in the coefficient of friction measured upon subsequent sliding in the boundary lubrication regime.10 In view of the results shown in Figure 5, it may be interpreted that the bulk material properties were not altered significantly by the cytokine treatment under the experimental conditions of this study. It should be realized that time effects in the stress-relaxation experiment do not directly resemble those in a creep experiment (as was used in the friction testing). Additionally, while the effects of interstitial fluid depressurization cannot be completely ruled out (although believed to be minimal based on analytical and experimental results10,25), any influence of depressurization would similarly influence all samples tested. Therefore, at a minimum, relative differences in the friction coefficients of the control and treated samples are relevant and important for comparison. From a perspective of applying these results to tissue engineering and regenerative medicine, producing a construct that replicates the lubricating function of articular cartilage depends on regulating the friction coefficient of that construct, potentially by using TGF-b1 or other growth factors and morphogens. The results of this investigation illustrate the need for multiple functional assessments (including biochemical and tribological assays), which can ascertain the long-term lubricating ability of a construct. To produce functional constructs for the regeneration of articular cartilage, a set of criteria, including the presence of boundary lubricants (i.e., SZP) and characterization of the surface roughness, must be satisfied to ensure a construct that would provide a long-lasting repair.

ACKNOWLEDGMENTS The authors are grateful to Dr. J. Zhou (University of California, Berkeley) for technical support and Dr. T. Schmid (Rush University, Chicago, IL) for the generous gift of the S6.79 primary antibody. This research was funded by the National Institutes of Health under Grant No. NIBIB 1F32 EB003371-01A1, the Lawrence J. Ellison Endowed Chair, and

255

the National Science Foundation under Grant No. CMS0528506.

REFERENCES 1. Reddi AH. 2003. Cartilage morphogenetic proteins: role in joint development, homoeostasis, and regeneration. Ann Rheum Dis 62(Suppl 2):ii73–ii78. 2. Ikegawa S, Sano M, Koshizuka Y, et al. 2000. Isolation, characterization and mapping of the mouse and human PRG4 (proteoglycan 4) genes. Cytogenet Cell Genet 90:291–297. 3. Schumacher BL, Block JA, Schmid TM, et al. 1994. A novel proteoglycan synthesized and secreted by chondrocytes of the superficial zone of articular cartilage. Arch Biochem Biophys 311:144–152. 4. Jay GD, Haberstroh K, Cha CJ. 1998. Comparison of the boundary-lubricating ability of bovine synovial fluid, lubricin, and Healon. J Biomed Mater Res 40:414–418. 5. Jay GD, Harris DA, Cha CJ. 2001. Boundary lubrication by lubricin is mediated by O-linked beta(1-3)Gal-GalNAc oligosaccharides. Glycoconj J 18:807–815. 6. Jay GD. 1992. Characterization of a bovine synovial fluid lubricating factor. I. Chemical, surface activity and lubricating properties. Connect Tissue Res 28:71–88. 7. Schmidt TA, Gastelum NS, Nguyen QT, et al. 2007. Boundary lubrication of articular cartilage: role of synovial fluid constituents. Arthritis Rheum 56:882–891. 8. Murakami T, Higaki H, Sawae Y, et al. 1998. Adaptive multimode lubrication in natural synovial joints and artificial joints. Proc Inst Mech Eng [H] 212:23–35. 9. Swann DA, Hendren RB, Radin EL, et al. 1981. The lubricating activity of synovial fluid glycoproteins. Arthritis Rheum 24:22–30. 10. Neu CP, Khalafi A, Komvopoulos K, et al. 2007. Mechanotransduction of articular cartilage superficial zone protein by TGF-b signaling. Arthritis Rheum 56:3706–3714. 11. Englert C, McGowan KB, Klein TJ, et al. 2005. Inhibition of integrative cartilage repair by proteoglycan 4 in synovial fluid. Arthritis Rheum 52:1091–1099. 12. Rhee DK, Marcelino J, Baker M, et al. 2005. The secreted glycoprotein lubricin protects cartilage surfaces and inhibits synovial cell overgrowth. J Clin Invest 115:622–631. 13. Marcelino J, Carpten JD, Suwairi WM, et al. 1999. CACP, encoding a secreted proteoglycan, is mutated in camptodactylyarthropathy-coxa vara-pericarditis syndrome. Nat Genet 23: 319–322. 14. Moos V, Fickert S, Muller B, et al. 1999. Immunohistological analysis of cytokine expression in human osteoarthritic and healthy cartilage. J Rheumatol 26:870–879. 15. Tetlow LC, Adlam DJ, Woolley DE. 2001. Matrix metalloproteinase and proinflammatory cytokine production by chondrocytes of human osteoarthritic cartilage: associations with degenerative changes. Arthritis Rheum 44:585–594. 16. van den Berg WB, Joosten LA, Kollias G, et al. 1999. Role of tumour necrosis factor alpha in experimental arthritis: separate activity of interleukin 1beta in chronicity and cartilage destruction. Ann Rheum Dis 58(Suppl 1):I40–I48. 17. Flannery CR, Hughes CE, Schumacher BL, et al. 1999. Articular cartilage superficial zone protein (SZP) is homologous to megakaryocyte stimulating factor precursor and is a multifunctional proteoglycan with potential growth-promoting, cytoprotective, and lubricating properties in cartilage metabolism. Biochem Biophys Res Commun 254: 535–541. 18. Jay GD, Tantravahi U, Britt DE, et al. 2001. Homology of lubricin and superficial zone protein (SZP): products of megakaryocyte stimulating factor (MSF) gene expression by human synovial fibroblasts and articular chondrocytes localized to chromosome 1q25. J Orthop Res 19:677–687. JOURNAL OF ORTHOPAEDIC RESEARCH FEBRUARY 2009

256

DURAINE ET AL.

19. Khalafi A, Schmid TM, Neu C, et al. 2007. Increased accumulation of superficial zone protein (SZP) in articular cartilage in response to bone morphogenetic protein-7 and growth factors. J Orthop Res 25:293–303. 20. Niikura T, Reddi AH. 2007. Differential regulation of lubricin/ superficial zone protein by transforming growth factor beta/ bone morphogenetic protein superfamily members in articular chondrocytes and synoviocytes. Arthritis Rheum 56:2312– 2321. 21. Schmidt TA, Gastelum NS, Han EH, et al. 2007. Differential regulation of proteoglycan 4 metabolism in cartilage by IL-1alpha, IGF-I, and TGF-beta1. Osteoarthritis Cartilage 16:90–97. 22. Jones AR, Gleghorn JP, Hughes CE, et al. 2007. Binding and localization of recombinant lubricin to articular cartilage surfaces. J Orthop Res 25:283–292. 23. Ohno S, Schmid T, Tanne Y, et al. 2006. Expression of superficial zone protein in mandibular condyle cartilage. Osteoarthritis Cartilage 14:807–813. 24. Klapperich C, Komvopoulos K, Pruitt L. 1999. Tribological properties and microstructure evolution of ultra-high molecular weight polyethylene. J Tribol 121:394–402. 25. Dowson D, Higginson GR. 1977. Elastohydrodynamic lubrication. New York: Pergamon Press. 26. Su JL, Schumacher BL, Lindley KM, et al. 2001. Detection of superficial zone protein in human and animal body fluids by cross-species monoclonal antibodies specific to superficial zone protein. Hybridoma 20:149–157. 27. Nugent GE, Schmidt TA, Schumacher BL, et al. 2005. Biomechanical regulation of chondrocyte synthesis of PRG4. Trans Orthop Res Soc 30:81.

JOURNAL OF ORTHOPAEDIC RESEARCH FEBRUARY 2009

28. Basalo IM, Chen FH, Hung CT, et al. 2006. Frictional response of bovine articular cartilage under creep loading following proteoglycan digestion with chondroitinase ABC. J Biomech Eng 128:131–134. 29. Stanton H, Rogerson FM, East CJ, et al. 2005. ADAMTS5 is the major aggrecanase in mouse cartilage in vivo and in vitro. Nature 434:648–652. 30. Middleton J, Manthey A, Tyler J. 1996. Insulin-like growth factor (IGF) receptor, IGF-I, interleukin-1 beta (IL-1 beta), and IL-6 mRNA expression in osteoarthritic and normal human cartilage. J Histochem Cytochem 44: 133–141. 31. Rabinowicz E. 1995. Friction and wear of materials. New York: Wiley; p xv, 315. 32. Park S, Costa KD, Ateshian GA. 2004. Microscale frictional response of bovine articular cartilage from atomic force microscopy. J Biomech 37:1679–1687. 33. Jurvelin JS, Muller DJ, Wong M, et al. 1996. Surface and subsurface morphology of bovine humeral articular cartilage as assessed by atomic force and transmission electron microscopy. J Struct Biol 117:45–54. 34. Krishnan R, Caligaris M, Mauck RL, et al. 2004. Removal of the superficial zone of bovine articular cartilage does not increase its frictional coefficient. Osteoarthritis Cartilage 12:947–955. 35. Nugent GE, Aneloski NM, Schmidt TA, et al. 2006. Dynamic shear stimulation of bovine cartilage biosynthesis of proteoglycan 4. Arthritis Rheum 54:1888–1896. 36. Nugent GE, Schmidt TA, Schumacher BL, et al. 2006. Static and dynamic compression regulate cartilage metabolism of PRoteoGlycan 4 (PRG4). Biorheology 43:191–200.