Highly Crosslinked vs Conventional Polyethylene Particles—An In Vitro Comparison of Biologic Activities

The Journal of Arthroplasty Vol. 23 No. 5 2008

Highly Crosslinked vs Conventional Polyethylene Particles—An In Vitro Comparison of Biologic Activities Richard Lynn Illgen II, MD,* Todd M. Forsythe, BS,* J. Wesley Pike, PhD,y Michel P. Laurent, PhD, z and Cheryl R. Blanchard, PhD z

Abstract: Highly cross-linked polyethylenes (HXPEs) have been introduced to reduce wear after hip arthroplasty. The improved wear characteristics of HXPEs are well documented, but the relative biologic activity of HXPE and conventional polyethylene (CPE) particles remains unclear. Longevity (Zimmer, Warsaw, Ind; HXPE) and GUR 1050 (Zimmer; CPE) particles were isolated and characterized from a hip simulator and their in vitro inflammatory responses (tissue necrosis factor α, interleukin 1α, and vascular endothelial growth factor levels) were compared using macrophages. The average diameter of Longevity particles (0.111 μm) was smaller than CPE particles (0.196 μm), and both were predominantly round (granular appearance). The inflammatory response to HXPE and CPE was concentrationdependent. No statistically significant differences were noted at low (0.1 surface area ratio [SAR]) and intermediate (0.75 SAR) doses. At the highest dose tested (2.5 SAR), HXPE was significantly more inflammatory than CPE based on relative tissue necrosis factor α and vascular endothelial growth factor secretion levels. Further study is needed to determine if similar findings would be noted in vivo over a broad concentration range. Key words: osteolysis, wear, debris, UHMWPE, in vitro, cytokine, cross-linked. © 2008 Elsevier Inc. All rights reserved.

Periprosthetic osteolysis is the most common cause of implant failure after total hip arthroplasty and represents a substantial obstacle for the long-term survival of total knee, elbow, and shoulder arthroplasties [1,2]. The National Center for Health Statistics estimated that approximately 160 000

total hip arthroplasties and 250,000 total knee arthroplasties are performed annually in the United States [3]. Unfortunately, significant failure rates have limited the long-term survival of these implants [4], often resulting in the need for revision surgery with substantially higher rates of complication and worse long-term clinical performance compared with primary procedures at similar follow-up intervals [5]. These challenges have prompted investigators to introduce novel alternative bearing materials in an effort to improve wear performance and reduce revision rates. Recent efforts to reduce the incidence of osteolysis after total hip arthroplasty include the introduction of highly cross-linked polyethylenes (HXPEs). These new biomaterials have improved wear characteristics compared with conventional polyethylenes (CPEs) both in vivo [6,7] and when tested in modern hip simulators [8-10]. Although

From the *Department of Orthopedics and Rehabilitation, School of Medicine, and Public Health, University of Wisconsin, Madison, Wisconsin; yDepartment of Biochemistry, University of Wisconsin, Madison, Wisconsin; and zZimmer, Incorporated, Warsaw, Indiana. Submitted September 1, 2006; accepted May 28, 2007. This study was funded by a Department of Orthopedics and Rehabilitation Research Grant (University of Wisconsin School of Medicine and Public Health, Madison) and an unrestricted research grant from Zimmer, Inc (Warsaw, Ind). Reprint requests: Richard Illgen II, MD, University of Wisconsin Hospitals and Clinics, 600 Highland Avenue, Clinical Science Center K4-739, Madison, WI 53792. © 2008 Elsevier Inc. All rights reserved. 0883-5403/08/2305-0013$34.00/0 doi:10.1016/j.arth.2007.05.043

721

722 The Journal of Arthroplasty Vol. 23 No. 5 August 2008 much is known regarding the improved biomaterial properties of HXPEs compared with CPEs, there is little information regarding the relative biologic activity of the wear particles generated. To understand if the substantial reduction in polyethylene wear noted for HXPE compared with CPE will result in the intended biologic effect of reduced rates of periprosthetic osteolysis, more study is needed to quantify the relative biologic activity of the wear particles generated. Many variables affecting the inflammatory response to wear particles have been studied in vitro and in vivo [11]. The inflammatory response to wear particles has been quantified in vitro by measuring the release of cytokines associated with bone resorption. These cytokines include tissue necrosis factor α (TNF-α), [11-13] vascular endothelial growth factor (VEGF), [14-16] and interleukin 1α (IL-1α). [12,17-19] Important wear particle characteristics influencing the host biologic response include particle size, particle concentration, particle shape, and material type. The effect of cross-linking polyethylene on the biologic response to wear particles has not been studied in detail. Adequate supply of cross-linked polyethylene particles has been a factor limiting the in vitro and in vivo analysis of the inflammatory response to these wear particles. Such study requires careful particle characterization to validate that the isolated particles are actually polyethylene, that the particles are of clinically relevant size, and that contamination has been avoided. Some investigators have substituted commercially available high-density polyethylene that is available in large quantity for analysis [17,20]. Other investigators have used pinon-disk methods to generate ultra-high molecular weight polyethylene (UHMWPE) particles [21]. Currently, the optimal method for particle generation remains the subject of debate. In the current study, particles were obtained by isolating UHMWPE particles from the lubricating serum of a 12-station Boston Hip Simulator (AMTI, Boston, Mass). This hip simulator follows a wear path that is very similar to that observed in vivo. This method of particle generation was used to closely approximate the types of particles that would be generated after total hip arthroplasty. The goals of the present study were to characterize the generated wear particles (HXPE and CPE) by scanning electron microscopy (SEM) and to compare the relative in vitro macrophage inflammatory response induced by these 2 different types of polyethylene particles. Particle characterization was performed by SEM to visually confirm homogeneity of samples (no contamination), to determine the

particle morphologies, and to establish the size distribution profiles of the isolated particles. This study specifically compared the relative in vitro inflammatory responses to one type of HXPE currently in clinical use (Longevity, Zimmer, Warsaw, Ind) with the inflammatory response to one CPE (GUR 1050 UHMWPE, Zimmer). The hypothesis tested was that similar dose-response inflammatory profiles will be noted for HXPE and CPE particles such that the substantial reduction in wear rates observed for HXPE compared with CPE would have the intended biologic effect of reducing the rate of periprosthetic osteolysis after total hip arthroplasty.

Materials and Methods Wear Particles Undiluted bovine calf lubricating serum was collected from a multistation wear simulator (AMTI-Boston 12-Station Hip Simulator [HS2-121000], AMTI) testing of commercially available Food and Drug Administration–approved HXPE (Longevity, 10 Mrad electron-beam irradiated, Zimmer) and CPE (GUR 1050 UHMWPE, 3.7 Mrad γ-sterilized in nitrogen, Zimmer). The HXPE and CPE liners were machined from the same base lot of compression-molded GUR 1050 UHMWPE. Before wear testing, the CPE liners were artificially aged in oxygen at 73 psi, 70°C, for 14 days, according to ASTM Standard F2003, Method B [22]. The wear particles were isolated based on modifications to the method described by Scott et al [23]. The diluted digest was filtered through a 0.1-μm polycarbonate filter membrane and then sonicated in methanol to remove the polyethylene particles. Particles were retrieved from methanol by ultracentrifugation, followed by washing and subsequent ultracentrifugation in ethanol. The ethanol was removed and the particles were opsonized with pooled human type AB serum (H-1513, Sigma, St Louis, Mo) using a previously described technique [24]. Particle opsonization is an important step to optimize particle-macrophage interaction under these in vitro testing conditions [24]. Particles were thoroughly washed with phosphate-buffered saline and tested for endotoxin by limulus assay (E-Toxate, Sigma) before finally being resuspended in serum-free Dulbecco's modification of eagle's medium (DMEM) stock solutions. The particle concentrations and morphologies were characterized by SEM at ×10 000 (15 kV, Cambridge S 360 SEM, LEO Electronics, Thornwood, NY) and quantified with image analysis software (Image Pro Plus analysis software, MediaCybernetics, Carlsbad,

Highly Cross-Linked vs Conventional Polyethylene Particles
Illgen et al

Calif) [25]. Surface area ratios (SARs) were used as previously described by Shanbhag et al [26] to quantify the dose of highly cross-linked (HXPE) and CPE. In our study, the dose of HXPE and CPE particles ranged from 0.02× to 2.5× SAR. The upper range of the particle testing dose was limited by the reduced supply of HXPE particles found in the hip wear simulator fluid. In addition to the 2 UHMWPE particle species described, 2 controls were used: (1) nonstimulated macrophages (negative control) and (2) commercially pure titanium particles, with a mean diameter of b3 μm (98%) (Cerac, Milwaukee, Wis) at 2 concentrations (0.14× and 1.41× SAR) as positive controls for cytokine response. The titanium particles were used as a qualitative rather than quantitative control because previous studies have shown titanium to be substantially more inflammatory than polyethylene particles. In this study, the titanium particle treatment groups were used to establish that our culture conditions were adequate to generate a positive response.

723

archived at −80°C. All assays were repeated in duplicate. After removal of media, treated cells were washed twice with phosphate-buffered saline and qualitatively inspected to insure viability under the above-noted testing conditions. Cytokine Assays of Conditioned Media In vitro macrophage activation was quantified by measuring the secreted levels of 3 pro-inflammatory cytokines (TNF-α, IL-1α, and VEGF) using commercially available enzyme-linked immunosorbent assay testing kits (R & D Systems, Minneapolis, Minn). Tissue necrosis factor α [11-13], IL-1α [12,17,18], and VEGF [14-16] were chosen for analysis because they have been shown to play significant roles in the pathophysiology of bone resorption and particulate-induced osteolysis. Tissue necrosis factor α is widely recognized as an important

Cell Culture The RAW 264.7 murine monocytic/macrophagic cell line (ATCC, Manassas, Va) was used in these experiments as an in vitro model of the macrophage inflammatory cytokine response to wear particles. This cell line has been used by previous investigators performing similar assays [16,27] and has the advantages of being a robust, stable, established cell line, whose cytokine responses are predictably uniform. These RAW cells were grown, maintained, and assayed in DMEM + 10% heat-inactivated fetal bovine serum (HI-FBS). Cells were seeded in 24-well tissue culture plate (Costar, Corning, Lowell, Mass) at a density of 2.5 × 105 cells/mL (1 mL/well). After a 48-hour incubation (37°C, 5% CO2 in humidified air), 8 wells were sacrificed to trypsinize and count viable cells, using trypan blue dye exclusion with a hemacytometer. Cells were 80% to 90% confluent and had a total of 1.5 × 106 cells per well. The remaining wells were then challenged by carefully removing old media, and adding 250 μL of fresh DMEM + HI-FBS, followed by 250 μL of prepared treatment media (serum-free DMEM and particle species to be tested). After a 4-hour incubation, 500 μL of DMEM + HI-FBS was added to adjust the final volume to 1 mL, with a concentration of 1.5 × 106 cells/mL. Treated cells were cultured for a total of 25 hours, before the conditioned media were collected. The conditioned media were stored at 4°C for 4 to 6 hours, until cytokine assays were performed, and were then

Fig. 1. A, Highly cross-linked polyethylene particles. Scanning electron micrograph at magnification ×10 000 (15 kV, Cambridge S 360 SEM, LEO Electronics). B, Conventional polyethylene particles. Scanning electron micrograph (SEM) at magnification ×10 000 (15 kV, Cambridge S 360 SEM, LEO Electronics).

724 The Journal of Arthroplasty Vol. 23 No. 5 August 2008 cytokine involved in inflammatory pathways and may increase the potency of the receptor activator of nuclear factor κB/receptor activator of nuclear factor κB ligand signaling cascade for wear particleinduced osteolysis [13]. Both TNF-α [11-13] and IL-1 [12,17] are potent activators of osteoclasts. Interleukin 1 modulates osteoclasts and osteoblasts in both stimulatory and inhibitory fashions [11]. The important role of IL-1 in bone remodeling and osteolysis has also been demonstrated by previous investigators [12,17]. Interleukin 1α was chosen based on previous reports demonstrating up-regulation as high as 16-fold, when comparing aseptically loose uncemented hips to control synovial tissue [19]. Finally, VEGF has been shown to play a significant role in osteoclast differentiation and function [15]. Vascular endothelial growth factor expression has been shown to be up-regulated by

macrophages [14] when exposed to wear particles, and VEGF-neutralizing antibody is capable of nearly complete inhibition of particulate-induced bone resorption [16]. Dose-response curves were generated for HXPE and CPE by measuring the secreted cytokine levels for macrophages stimulated with particles and for unstimulated macrophages over the concentration range tested (0.02×-2.5× SAR). Testing was conducted in accordance with the manufacturer's protocols, and the colorimetric results were measured at 450 nm in a plate reader (ELx800, Bio-Tek, Winooski, Vt). Statistical Methods The optical densities determined by absorbance at 450 nm were converted to concentrations based on

Fig. 2. A, Highly cross-linked polyethylene particle size distribution profile. B, Conventional polyethylene particle size distribution profile.

Highly Cross-Linked vs Conventional Polyethylene Particles
Illgen et al

a standard curve. Average cytokine concentrations (TNF-α, IL-1α, and VEGF) were calculated for each particle type (HXPE or CPE) over the concentration range tested (0.02×-2.5× SAR). The baseline level of secreted cytokines was also calculated for unstimulated macrophages. An analysis of covariance was then used to estimate the slope and intercept of the cytokine concentration for each particle type over the concentration range tested. Slopes and intercepts were compared between particle types within the analysis of covariance framework using a level of significance of .05. Unpaired t tests were used to analyze the difference between particle types at an individual dose and to report the difference in mean particle diameters. All analyses were performed using SAS statistical software (SAS Institute Inc, Cary, NC).

Results Particle Morphology The HXPE wear particles retrieved from hip simulator fluid in this study were predominantly granular (87%), with a small proportion of fibrils (6%) and flakes (7%) (Fig. 1B). The CPE wear particles contained submicrometer-sized granules (78%), elongated fibrils (14%), and a small proportion of flakes (8%) (Fig. 1A).

725

Particle Size Distribution The average diameter of HXPE particles (0.111 μm) retrieved from hip simulator fluid in this study was smaller than those noted for CPE (0.196 μm; P b .0001) (Fig. 2A-B). The diameters of nearly all the particles in both groups were less than 1.0 μm; more than 95% of particles in both groups had diameters less than 0.55 μm, with 70% at or less than 0.2 μm and 60% at or less than 0.15 μm (Fig. 2A-B). Endotoxin Testing The HXPE and the CPE particles both tested positive for endotoxin according to the limulus assay. Inflammatory Response to HXPE and CPE Wear Particles Trypan blue dye exclusion assay confirmed that all cultures were maintained at 89.8% viability (SD, 3.6), with no bias due to particle type nor dosage. In the current study, similar levels of VEGF expression were documented in response to titanium and polyethylene particles (Fig. 3). However, both IL-1α and TNF-α were expressed at much higher levels in response to titanium (175 pg/mL and 11 250 pg/mL, respectively) than in response to polyethylene (Figs. 4 and 5). Titanium levels were not shown in Figs. 4 and 5 because expression levels were off the scale.

Fig. 3. Relative inflammatory cytokine profile (VEGF)—HXPE vs CPE. Error bars represent the SE.

726 The Journal of Arthroplasty Vol. 23 No. 5 August 2008

Fig. 4. Relative inflammatory cytokine profile (IL-1α)—HXPE vs CPE. Interleukin 1α levels in response to titanium stimulation are not shown because of very high and off-scale levels of expression. Error bars represent the SE.

The VEGF cytokine assay yielded results of 819.18, 836.79, and 1189.08 pg/mL at HXPE doses of 0.08, 0.84, and 2.52 SAR, respectively. Conven-

tional polyethylene elicited similar levels of 703.72, 736.49, 845.48, and 1007.46 pg/mL, at particulate doses of 0.02, 0.24, 0.72, and 2.41 SAR, respectively

Fig. 5. Relative inflammatory cytokine profile (TNF-α)—HXPE vs CPE. Tumor necrosis factor α levels in response to titanium stimulation are not shown because of very high and off-scale levels of expression. Error bars represent the SE.

Highly Cross-Linked vs Conventional Polyethylene Particles
Illgen et al

(Fig. 3). The levels of secreted VEGF in response to HXPE and CPE particles were not different at low (0.02 SAR) and intermediate (0.84 SAR) doses. However, at the highest dose tested (approximately 2.5× SAR), levels of VEGF were 18% higher in the HXPE group (P = .04, Fig. 3). The assay for IL-1α demonstrated HXPE results of −2.54, −1.96, and 1.96 pg/mL, at doses of 0.08, 0.84, and 2.52 SAR, respectively. Conventional polyethylene elicited IL-1α levels of −1.61, −1.22, −1.25, and 2.59 pg/mL, at respective doses of 0.02, 0.24, 0.72, and 2.41 SAR (Fig. 4). There were no significant differences noted between the levels of secreted IL-1α in response to HXPE and CPE particles over the concentration range tested (0.02-2.41 SAR, P N .05, Fig. 4). Highly cross-linked polyethylene doses of 0.08, 0.84, and 2.52 SAR elicited TNF-α levels of 316.22, 502.21, and 1582.27 pg/mL, respectively. Tissue necrosis factor α levels in response to CPE doses of 0.02, 0.24, 0.72, and 2.41 SAR, were 319.45, 378.80, 507.82, and 1058.33 pg/mL, respectively (Fig. 5). There were no significant differences noted between the levels of secreted TNF-α in response to HXPE and CPE particles at equivalent low and intermediate SAR dosing levels (0.1× and 0.75× SAR, both P N .05) (Fig. 5). However, at the highest concentration tested (approximately 2.5× SAR), the levels of secreted TNF-α were approximately 50% higher for HXPE compared with CPE (P b .001, Fig. 5).

Discussion Osteolysis is a complex process that is modulated by multifactorial influences [2,11,26] including the dose and biologic activity of generated wear particles. If modern hip simulators accurately reflect in vivo polyethylene wear rates, the dose of Longevity HXPE particles should be substantially less than for CPE at similar wear intervals. This study now provides relevant information regarding the size distribution profiles of Longevity HXPE and CPE particles retrieved from hip simulators and the doseresponse behavior of macrophages exposed to these particles under similar in vitro testing conditions. The physiologically relevant size of polyethylene particles remains controversial. The reported size of polyethylene particles retrieved from hip arthroplasties varies in the literature. Reported mean particle diameters range from less than 0.1 to 0.63 μm and size ranges vary from submicrometer to more than 50 μm [28,29]. This range of reported retrieved particle sizes creates uncertainty regarding

727

which size represents the most clinically relevant for analysis. It is important to realize that the size of retrieved particles is affected by the method of particle isolation (ie, the pore size of the filter) and the mode of characterization (ie, light microscopy vs SEM). A growing body of evidence suggests that with improved isolation techniques (smaller filter pore sizes) and sophisticated particle characterization (SEM), the micrometer and submicrometer particles represent a higher percentage of particles generated than had been previously appreciated [21]. In addition, previous studies of particle retrieval have focused on isolation of particles from periprosthetic membranes. Particle size present in these membranes reflects only a subset of the particles generated during the normal wear process after total hip arthroplasty. Wear particles are also present in the synovial fluid and some particles are removed by the lymphatic or circulatory systems [30]. Studying only the particles isolated from the periprosthetic membranes potentially excludes substantial numbers of physiologically relevant particle types and sizes. In this study, HXPE and CPE particles were isolated from hip simulators using small filter sizes (pore size = 0.1 μm) and characterized using SEM. The morphology and size distribution profiles for HXPE and CPE particles are similar to those noted in these previous studies of retrieved CPE particles from failed total hip arthroplasties (Fig. 2) [21,28,29]. In this study, most HXPE and CPE particles were granular (Fig. 1), and 95% of both types of particles had diameters of less than 0.55 μm (Fig. 2A and B). The average diameter of HXPE particles retrieved from hip simulator fluid in our study was somewhat smaller than those noted for CPE (0.111 vs 0.196 μm, respectively, P b .0001) (Fig. 2). Some investigators have demonstrated that smaller particles tend to be more inflammatory than larger particles, particularly at the micrometer and submicrometer level [31]. However, the biologic response to wear particles will be affected by many factors including the particle size as well as the number of particles generated (ie, the total particulate load). Surface area ratio dosing was chosen in this study to normalize for particles of differing sizes. Previous investigators have recognized that the magnitude of macrophage–foreign debris interactions are primarily dictated by surface-to-surface contact, and some have adopted SAR dosing to normalize dosing according to the surface area of the particles to be studied [26]. The current study adopted a similar strategy to isolate the variable of particle type and degree of crosslinking for independent analysis.

728 The Journal of Arthroplasty Vol. 23 No. 5 August 2008 This study also parallels previous work by using titanium particles—well established as a potent promoter of particle-induced osteolysis—as the positive control. It is known that titanium particles are substantially more inflammatory than polyethylene particles in vitro [18,27,32]. Therefore, titanium was used as a qualitative rather than a quantitative positive control. Titanium particles were tested to establish that our experimental system was capable of generating a positive response. Although the cytokine levels in response to titanium were significantly greater than those seen in response to polyethylene in vitro, it is important to note that the quantity of polyethylene particles generated in vivo with traditional metal on polyethylene total hip arthroplasty is disproportionate to titanium particles. At present, most investigators agree that polyethylene particles represent one of the most common causes of periprosthetic osteolysis [1,11,20]. For this reason, in our analysis, polyethylene particles were studied and titanium served as a positive control. Based on both in vivo [6,7,33,34] and hip simulator data [8,10], substantial reductions in volumetric wear are predicted for HXPE compared with CPE. A previous hip simulator study demonstrated approximately 98% reduction in wear at 30 million cycles when comparing Longevity HXPE with CPE [10]. Other investigators have demonstrated a 96% reduction in the rate of steady-state wear at 2-year follow-up when comparing Longevity HXPE to CPE in vivo (0.007 and 0.174 mm/y, respectively) [6]. The predicted amount of wear reduction depends on the specific type of HXPE and ranges between 60% and 98% compared with CPE at equivalent wear intervals [6,7,10,33,34]. These observations suggest that although the HXPE particles tended to be smaller than CPE particles, one would expect a much smaller volume of HXPE particles compared with CPE. In our study, the small volume of available particles isolated from the hip simulator with HXPE limited the concentration range of particles tested. Over the concentration range tested in our study, the relative inflammatory response to HXPE and CPE was concentration-dependent (P b .02). At low and intermediate dose levels, no differences were noted in the inflammatory potential of HXPE and CPE particles (Figs. 3-5). However, at the highest concentration tested (2.5× SAR), HXPE was approximately 50% more inflammatory than CPE based on evaluation of the secreted levels of TNF-α (P b .001, Fig. 5) and 18% more inflammatory based on the secreted levels of VEGF (Fig. 3, P = .04). Cytokine levels of IL-1α

were not found to be significantly different between HXPE and CPE. However, the current study simply confirmed previous reports noting low levels of IL-1α in response to CPE particles at the concentrations tested (Fig. 4) [18,26]. The clinical relevance of these findings will be dependent on how the doses tested in this in vitro study compare to particle doses present in the hip joint after total hip arthroplasty. Further study in this area is needed. We believe that differences in inflammatory response will be overshadowed by the substantial reduction in particle dose for HXPE compared with CPE because of its improved wear characteristics. The available in vivo [6,7,33,34] and hip simulator data [8,10] demonstrate that wear rates for many HXPEs tested have been less than 0.1 mm/y [6,7,33,34]. Such low wear rates may be below the so-called osteolytic threshold. Although the concept of an osteolytic threshold remains controversial [35], improved wear rates may translate into lower rates of osteolysis for HXPE compared with CPE at similar wear intervals. This prediction requires confirmation based on radiographic and clinical outcome at long-term follow-up intervals. Other investigators have also studied the relative inflammatory potential of HXPE and CPE wear particles [21]. In our study, particles were isolated from hip simulator fluid and dosed according to SARs. This SAR dosing scheme was chosen because previous investigators have suggested that it represents the most appropriate manner to quantify the biologic response to wear particles [26]. Other investigators have used a pin-on-disk method to generate HXPE and CPE particles [21]. This group noted a higher percentage of HXPE particles in the submicrometer range compared with CPE and also noted greater inflammatory potential for HXPE compared with CPE based on in vitro testing [21]. It should be noted that this group used a different type of HXPE, generated the particles with a different method, and used a different dosing scheme compared with our study. We believe that particles isolated from a hip simulator may more accurately reflect a physiologically relevant size distribution profile compared with particles generated by a pin-on-disk method. In addition, significant biomaterial differences exist between the various forms of HXPE currently in clinical use. These differences may result in significant alterations to the size, shape, and number of particles generated. Such differences may have a substantial impact on the biologic activity of the generated particles and the long-term rates of osteolysis observed with these materials.

Highly Cross-Linked vs Conventional Polyethylene Particles
Illgen et al

Several limitations should be noted regarding the current study. The particle concentration range tested was limited by the small amount of HXPE particles generated in hip simulators. It is unknown if the inflammatory profiles for HXPE and CPE would be similar over a wider concentration range. It should also be noted that there are many different types of HXPE available today. Many commercially available HXPEs are manufactured with different degrees of cross-linking, use alternative modes of sterilization, have different levels of oxidation present, and are cross-linked using a variety of techniques. Such differences may have significant effects on the polyethylene wear rates and the types of particles generated. The current study focused on determining the biologic activity of wear particles generated from one of these HXPEs currently in clinical use (Longevity) and comparing it to one CPE (GUR 1050, UHMWPE) obtained from the same company (Zimmer). Further study is needed to determine if similar findings would be noted with other types of commercially available HXPEs and CPEs. This study used an immortalized, adherent cell line to test biologic responses. This particular line has been used in previous studies, and there is a clear precedent for using adherent lines in studies of this nature [26]. Although these techniques frequently have been used, their behavior may deviate from the phagocytic behavior of nonadherent macrophages in vivo. Further analysis of phagocytic activity was not documented in this study. The RAW 264.7 murine monocytic/macrophagic cell line was selected over human cells because this cell line has been used by previous investigators performing similar assays [16,27] and has the advantages of being a robust, stable, established cell line, whose cytokine responses are predictably uniform. Previous studies using human cells have used peripheral blood mononuclear cells [15,17,18,31], which would likely exhibit less robust, uniform, predictable behavior, and likely contain a mixed cell population—not solely macrophages. It is acknowledged, however, that human cells may respond differently than the murine cells tested here. Some inhibition of IL-1α expression was noted in this study. Previous investigators have suggested that the polystyrene cell culture dishes may cause a confounding effect of IL-1α secretion by macrophages [18]. In the current study, no stimulatory effect was demonstrated at low (0.08 SAR) and intermediate dosing (0.84 SAR) levels (Fig. 4). Stimulation of levels of IL-1α greater than zero was only noted at the highest dose tested (2.5 SAR, Fig. 4). These data support the hypothesis that IL-1α

729

level expression may be inhibited by the in vitro testing conditions used in our study. The role of endotoxin in the inflammatory response to wear particles also remains controversial. The HXPE and CPE particles in our study tested positive for endotoxin. We chose not to remove this endotoxin for the following reasons: endotoxin is likely present in vivo on wear particles [36]; the chemical means to remove endotoxin could alter the surface chemistry of the polyethylene [36]; the procedure to completely remove endotoxin and verify its presence has been questioned; and the role of endotoxin in the inflammatory response to wear particles remains controversial [11,36,37]. The limited supply of HXPE particles also precluded the analysis of endotoxin-negative particles. Further study is necessary to determine how the response to HXPE and CPE particles would be altered in the absence of endotoxin both in vitro and in vivo. The introduction of alternative bearing materials such as HXPE for use in total hip arthroplasty provides the potential for significantly improved wear performance compared with CPE. However, it is important to recognize that the goal of improved clinical performance and longer implant survival will be influenced not only by wear rates but also the relative inflammatory potential of the particles that are generated. This study provides evidence that at some concentrations, crosslinking modestly increases the inflammatory response compared with CPE. We believe that the relatively small increase in inflammatory response noted for this HXPE (Figs. 3 and 5) will be overshadowed by the more significant reduction in wear compared with CPE. Clinical confirmation of this prediction is required, and longterm studies comparing HXPE and CPE are needed to determine if crosslinking will reduce in vivo wear rates and have the intended biologic effect of reducing the rate of osteolysis and improving the clinical outcome after total hip arthroplasty.

Acknowledgments The assistance of M Watanuki, SK Bhambri, WK Overton, D Schmucker, F Jones, and Glen Leverson is greatly appreciated.

References 1. Harris WH. The problem is osteolysis. Clin Orthop 1995:46. 2. Willert HG, Semlitsch M. Reactions of the articular capsule to wear products of artificial joint prostheses. J Biomed Mater Res 1977;11:157.

730 The Journal of Arthroplasty Vol. 23 No. 5 August 2008 3. Popovic JR, Kozak LJ. National Hospital Discharge Survey: Annual Summary, 1998. National Center for Health Statistics. Vital Health Stat 2000;13:148. 4. Malchau H, Herberts P, Ahnfelt L. Prognosis of total hip replacement in Sweden. Follow-up of 92,675 operations performed 1978-1990. Acta Orthop Scand 1993;64:497. 5. Stiehl JB, Saluja R, Diener T. Reconstruction of major column defects and pelvic discontinuity in revision total hip arthroplasty. J Arthroplasty 2000; 15:849. 6. Manning DW, Chiang PP, Martell JM, et al. In vivo comparative wear study of traditional and highly cross-linked polyethylene in total hip arthroplasty. J Arthroplasty 2005;20:880. 7. Heisel C, Silva M, Schmalzried TP. In vivo wear of bilateral total hip replacements: conventional versus crosslinked polyethylene. Arch Orthop Trauma Surg 2005;125:555. 8. Muratoglu OK, Bragdon CR, O'Connor DO, et al. A novel method of cross-linking ultra-high-molecularweight polyethylene to improve wear, reduce oxidation, and retain mechanical properties. Recipient of the 1999 HAP Paul Award. J Arthroplasty 2001;16:149. 9. Campbell P, Shen FW, McKellop H. Biologic and tribologic considerations of alternative bearing surfaces. Clin Orthop Relat Res 2004:98. 10. Laurent MP, Yao JQ, Bhambri SK, et al. High cycle wear of highly crosslinked UHMWPE acetabular liners evaluated in a hip simulator. Transactions of the 6th World Biomaterials Congress; 2000. 11. Illgen R, Shanbhag AS, Jacobs JJ. Periprosthetic osteolysis: pathophysiology and medical management. Semin Arthroplasty 2002;13:238. 12. Wei S, Kitaura H, Zhou P, et al. IL-1 mediates TNFinduced osteoclastogenesis. J Clin Invest 2005;115: 282. 13. Schwarz EM, Lu AP, Goater JJ, et al. Tumor necrosis factor-alpha/nuclear transcription factor-kappaB signaling in periprosthetic osteolysis. J Orthop Res 2000;18:472. 14. Miyanishi K, Trindade MC, Ma T, et al. Periprosthetic osteolysis: induction of vascular endothelial growth factor from human monocyte/macrophages by orthopaedic biomaterial particles. J Bone Miner Res 2003;18:1573. 15. Aldridge SE, Lennard TW, Williams JR, et al. Vascular endothelial growth factor receptors in osteoclast differentiation and function. Biochem Biophys Res Commun 2005;335:793. 16. Illgen R, Shanbhag AS, Deshmukh AV, et al. In vitro analysis of novel, clinically relevant osteoclast inhibitors. Transactions of the Orthopedic Research Society 2001;26:134. 17. Matthews JB, Besong AA, Green TR, et al. Evaluation of the response of primary human peripheral blood mononuclear phagocytes to challenge with in vitro generated clinically relevant UHMWPE particles of

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

known size and dose. J Biomed Mater Res 2000;52: 296. Sethi RK, Neavyn MJ, Rubash HE, et al. Macrophage response to cross-linked and conventional UHMWPE. Biomaterials 2003;24:2561. Shanbhag AS, Jacobs JJ, Black J, et al. Cellular mediators secreted by interfacial membranes obtained at revision total hip arthroplasty. J Arthroplasty 1995;10:498. von Knoch M, Jewison DE, Sibonga JD, et al. The effectiveness of polyethylene versus titanium particles in inducing osteolysis in vivo. J Orthop Res 2004; 22:237. Ingram JH, Stone M, Fisher J, et al. The influence of molecular weight, crosslinking and counterface roughness on TNF-alpha production by macrophages in response to ultra high molecular weight polyethylene particles. Biomaterials 2004;25:3511. ASTM F2003-00 Standard Guide for Accelerated Aging of Ultra-High Molecular Weight Polyethylene. West Conshohocken, PA: ASTM International; 2001. Scott M, Widding K, Jani S. Do current wear particle isolation procedures underestimate the number of particles generated by prosthetic bearing components? Wear 2001;250:1213. Shanbhag AS, Jacobs JJ, Black J, et al. Effects of particles on fibroblast proliferation and bone resorption in vitro. Clin Orthop Relat Res 1997:205. ASTM F1877-98 Standard Practice for Characterization of Particles. West Conshohocken, PA: ASTM International; 2003. Shanbhag AS, Jacobs JJ, Black J, et al. Macrophage/ particle interactions: effect of size, composition and surface area. J Biomed Mater Res 1994;28:81. Glant TT, Jacobs JJ, Molnar G, et al. Bone resorption activity of particulate-stimulated macrophages. J Biomed Mater Res 1993;8:1071. Campbell P, Ma S, Yeom B, et al. Isolation of predominantly submicron-sized UHMWPE wear particles from periprosthetic tissues. J Biomed Mater Res 1995;29:127. Shanbhag AS, Jacobs JJ, Glant TT, et al. Composition and morphology of wear debris in failed uncemented total hip replacement. J Bone Joint Surg Br 1994;76:60. Urban RM, Jacobs JJ, Tomlinson MJ, et al. Dissemination of wear particles to the liver, spleen, and abdominal lymph nodes of patients with hip or knee replacement. J Bone Joint Surg 2000;82:457. Ingram J, Matthews JB, Tipper J, et al. Comparison of the biological activity of grade GUR 1120 and GUR 415HP UHMWPE wear debris. Biomed Mater Eng 2002;12:177. Taki N, Tatro JM, Nalepka JL, et al. Polyethylene and titanium particles induce osteolysis by similar, lymphocyte-independent, mechanisms. J Orthop Res 2005;23:376. D'Antonio JA, Manley MT, Capello WN, et al. Fiveyear experience with Crossfire highly cross-linked polyethylene. Clin Orthop Relat Res 2005;441:143.

Highly Cross-Linked vs Conventional Polyethylene Particles
Illgen et al 34. Dorr LD, Wan Z, Shahrdar C, et al. Clinical performance of a Durasul highly cross-linked polyethylene acetabular liner for total hip arthroplasty at five years. J Bone Joint Surg 2005;87: 1816. 35. Harris WH. “The lysis threshold”: an erroneous and perhaps misleading concept? J Arthroplasty 2003; 18:506.

731

36. Ragab AA, VanDeMotter R, Lavish SA, et al. Measurement and removal of adherent endotoxin from titanium particles and implant surfaces. J Orthop Res 1999;17:803. 37. Greenfield EM, Bi Y, Ragab AA, et al. Does endotoxin contribute to aseptic loosening of orthopedic implants? J Biomed Mater Res B Appl Biomater 2005;72:179.