Modulation of bone ingrowth and tissue differentiation bylocal infusion of interleukin-10 in the presence ofultra-high molecular weight polyethylene (UHMWPE)wear particles

Stuart Goodman,1 Michael Trindade,1 Ting Ma,1 Mel Lee,1 Neal Wang,1 Takashi Ikenou,1 Ippe Matsuura,1

Keita Miyanishi,1 Nora Fox,1 Donald Regula,2 Mark Genovese,3 John Klein,4 Dan Bloch,3 R. Lane Smith1

1Division of Orthopaedic Surgery, Stanford University Medical Center, 300 Pasteur Drive, Stanford, California 943052Department of Pathology, Stanford University Medical Center, Stanford, California 943053Division of Immunology and Rheumatology, Stanford University Medical Center, Stanford, California4Alza Corp., Palo Alto, California

Received 11 January 2002; accepted 19 February 2002

Abstract: Interleukin-10 (IL-10) is a cytokine that plays amajor role in suppressing the inflammatory response, par-ticulary cell-mediated immunity that is characteristic of theTH1 response. The purpose of this study was to determinewhether local infusion of IL-10 could mitigate the suppres-sion of bone ingrowth associated with polyethylene wearparticles. Drug test chambers were implanted in the proxi-mal tibia of 20 mature New Zealand white rabbits. The DTCprovided a continuous 1 × 1 × 5-mm canal for tissue in-growth. After a 6-week period for osseointegration, the DTCwas then connected to an osmotic diffusion pump. IL-10 atdoses of 0.1–100 ng/mL (0.25 �L/h) was infused with orwithout ultra-high molecular weight polyethylene particles(0.5 ± 0.2 �m diameter, 1012 particles/mL) present in thechamber for a 3- or 6-week period. The tissue in the chamberwas harvested after each treatment; sections were stainedwith hematoxylin and eosin for morphometric analysis. Os-teoclast-like cells were identified by immunohistochemicalstaining using a monoclonal antibody directed against thealpha chain of the vitronectin receptor, CD51. Osteoblasts

were identified using alkaline phosphatase staining. Indose–response studies, infusion of 1 ng/mL IL-10 yieldedthe greatest bone ingrowth in the presence of particles. Theaddition of polyethylene particles evoked a marked foreignbody reaction and fibrosis; bone ingrowth was significantlysuppressed (p = 0.0003). Bone ingrowth was increased byover 48% with infusion of IL-10 for the final 3 weeks of a6-week ultra-high molecular weight polyethylene particleexposure compared with particles alone (p = 0.027). IL-10 isa cytokine that plays a major role in suppressing the inflam-matory response, especially cell-mediated immunity that ischaracteristic of the TH1 response. Local infusion of im-mune-modulating cytokines such as IL-10 may prove to beuseful in abating particle-induced periprosthetic osteolysis.© 2003 Wiley Periodicals, Inc. J Biomed Mater Res 65A: 43–50, 2003

Key words: total joint replacement; bone ingrowth; anti-inflammatory cytokine; interleukin-10; polyethylene par-ticles

INTRODUCTION

The main factor limiting long-term survivorship oftotal joint replacements is prosthetic wear, the produc-tion of particulate debris, and subsequent peripros-thetic osteolysis.1 Osteolysis may occur with or with-

out prosthetic loosening and usually is a silent diseaseclinically. Despite improvements in prosthetic designand materials, surgical technique, patient selection,and education, revision surgery is still common andconstitutes 25% or more of the arthritis caseload insome university institutions.1,2

Polyethylene (PE) particles may adversely affectboth initial osseointegration and subsequent bone re-modeling of a total joint replacement.1,3–7 Immediatelyafter implanting a total joint replacement, the bone–prosthesis interface is exposed continuously to PEwear particles specific to the biomaterials implanted.The inflammatory and foreign body reaction to thisdebris may interfere with initial prosthetic osseointe-

Correspondence to: S. B. Goodman; e-mail: goodbone@stanford.edu

Contract grant sponsor: Arthritis FoundationContract grant sponsor: Alza CorpContract grant sponsor: Pharmacia

© 2003 Wiley Periodicals, Inc.

gration, jeopardizing long-term implant stability. Con-tinued exposure to PE particles adversely affects boneremodeling (formation and resorption) and results inperiprosthetic osteolysis.8–12

Interleukin-10 (IL-10) is a cytokine that plays a ma-jor role in suppressing the inflammatory response, es-pecially cell-mediated immunity that is characteristicof the TH1 response.13,14 IL-10 down-regulates thespontaneous and cytokine-induced expression of HLAclass II antigens on human monocytes. IL-10 decreasesthe secretion of the pro-inflammatory cytokines IL-1,IL-6, IL-8, tumor necrosis factor-�, and Granulocyte-Macrophage-Colony Stimulating Factor (GM-CSF) bymonocytes/macrophages, increases the production ofthe antiinflammatory cytokine IL-1ra, and inhibits theproduction of PGE2 and reactive oxygen (H2O2) andnitrogen (NO) intermediates after activation of mono-cytes/macrophages. IL-10 directly inhibits T cell pro-liferation by blocking IL-2 secretion and inhibits mi-gration inhibitory factor secretion by T cells. IL-10augments B cell proliferation and differentiation. Itsability to suppress the cell-mediated immune responseand pro-inflammatory cytokine release from macro-phages makes IL-10 a good candidate for use in modu-lating the chronic inflammatory and foreign body re-sponse associated with wear debris from joint replace-ments.15,16 Human IL-10 has been shown to haveactivity on cells from lower and higher vertebratesbecause of the highly preserved DNA and amino acidhomology of IL-10, as well as the similarity in speci-ficity of binding of IL-10 to its receptor.12,13

The purpose of this study was to modulate the ad-verse biological response to wear particles in vivo withlocal infusion of the anti-inflammatory cytokine IL-10.

MATERIALS AND METHODS

DTC

The DTC is a titanium device that is implanted at the levelof the cortex in the proximal medial tibial metaphysis ofrabbits, similar to the bone harvest chamber (Fig. 1).17 Theremovable core of the DTC is provided with a 10-�L cavityacting as a fluid reservoir linked to the 1 × 1 × 5 mm in-growth canal via a 1-mm long 0.25-mm diameter diffusioncapillary. For continuous infusion, two tubes are connectedto the reservoir. One tube is connected to a subcutaneousosmotic minipump (Alza Corp, Palo Alto, CA). The othertube is open to the subcutis, thereby functioning as a drainfor the reservoir. This allows a slow continuous flow of thesolution through the reservoir, creating a concentration gra-dient along the diffusion capillary. The percolation of thesolvent through the diffusion capillary to the bone is mini-mal due to the reservoir being drained to the subcutaneoustissue. Thus, the ingrowing tissue can be locally treated witha solution without disturbance of solvent flow. After eachtreatment, through a small 1-cm incision, the chamber is

exposed and disassembled; the tissue in the chamber is har-vested for subsequent analysis. At the same time, throughanother small incision, the minipump and tubing (contain-ing the new solution to be infused) are replaced.

PE

The particles were a generous gift from Dr. Pat Campbellof the Joint Replacement Institute in Los Angeles. The ultra-high molecular weight polyethylene (UHMWPE) particleswere retrieved from joint simulator tests performed at theJoint Replacement Institute by Dr. Harry McKellop and col-leagues. Scanning electron microscopy revealed that the par-ticles averaged 0.5 ± 0.2 �m in diameter. The particles werewashed with 5N sodium hydroxide and isopropanol repeat-edly, sterilized with ethylene oxide, and deemed endotoxinfree using a high-sensitivity limulus amebocyte lysate test

Figure 1. Illustration of the DTC. This CP titanium deviceis implanted at the level of the cortex in the proximal medialtibial metaphysis of rabbits. The removable core of thechamber contains a 10-�L fluid reservoir linked to the 1 × 1× 5-mm tissue ingrowth canal via a 1-mm long 0.25-mmdiameter diffusion capillary. For continuous infusion of IL-10, one tube connects the reservoir in the chamber with asubcutaneous osmotic minipump. The other tube, whichfunctions as a drain, connects the reservoir with the subcu-tis. This allows a slow continuous flow of the solution in thepump through the reservoir, creating a concentration gradi-ent along the diffusion capillary. After each treatment, thechamber is disassembled and the tissue in the chamber isharvested for subsequent analysis. The osmotic minipump isalso re-exposed and filled with the appropriate solution.(Reprinted from the Journal of Biomedical Engineering, Vol-ume 10, Aspenberg P, Albrektsson T, Lohmander LS, andThorngren KG, Drug test chamber: a titanium implant foradministration of biochemical agents to a standardized bonecallus in situ, pages 70–73, Copyright 1988, with permissionfrom Elsevier Science.

44 GOODMAN ET AL.

(QCL-1000, Biowhittaker Inc.).18,19 The particles were resus-pended in 1% sodium hyaluronate (Healon, Pharmacia,Sweden) at a concentration of 1.7 × 1012 particles/mL. Thissuspension was introduced into the canal of the chamber viaa spatula. These PE particles are clinically relevant,20 and thedosages were based on previous experiments performed inour laboratory using both smaller and larger PE particles.7,21

Furthermore, these dosages are comparable to those in stud-ies using retrieved human tissues by Margevicius et al. ifcorrections are made for particle size and volume, lossessecondary to the isolation process, and the heterogeneity inthe characteristics of the retrieved tissue is considered.22

IL-10

In experiment 1, human recombinant IL-10 (R & D Sys-tems) at concentrations of 0.1, 1, 10, and 100 ng/mL wasinfused continuously for 3 weeks at 0.25 �L/h via Alzetpump 2004 (Alza Corp, Mountain View, CA). In experiment2, a dose of 1 ng/mL was infused for 3 or 6 weeks (seebelow). The carrier for the IL-10 was 0.1% bovine serumalbumin in phosphate-buffered saline.

Surgical procedure

Institutional guidelines for the care and use of laboratoryanimals were strictly followed. DTCs were implanted in 10mature male New Zealand white rabbits ages 6–12 monthsand weighing 3.5–4.2 kg under general endotracheal anes-thesia. One dose of cephalosporin antibiotic was given asprophylaxis for infection. After a 6-week period for osseo-integration, the contents of the chamber were harvested anddiscarded. The DTC was then connected, via polyvinyl tub-ing, to an Alzet osmotic diffusion pump (Model �2004;Alza). The diffusion pump contains 240 �L of fluid and hasa delivery rate of 0.25 �L/h. The carrier solution alone wasinfused for 3 or 6 weeks, followed by a harvest. IL-10 wasthen infused with or without particles present in the cham-ber for 3- or 6-week periods. The contents of the chamber

were harvested after each treatment, and a new diffusionpump and tubing with the appropriate solution were placedin the rabbit.

In experiment 1, IL-10 was infused at dosages of 0, 0.1, 1,10, or 100 ng/day for 3-week periods in the absence of par-ticles (see Table I). In the latter treatments, IL-10 was infusedat the above dosages in the presence of PE particles.

In experiment 2, treatments were extended to 6 weeks todetermine the effects of IL-10 infusion for the entire 6 weeksversus the final 3 of 6 weeks, after the foreign body andchronic inflammatory reaction to UHMWPE particles hadalready been established for 3 weeks (see Table II). A dose ofIL-10 of 1 ng/day was used for experiment 2 because thisdose resulted in the greatest percentage of bone ingrowth inthe presence of particles.

Tissue processing

Harvested specimens were placed in optimal cutting me-dium and snap frozen in liquid nitrogen and stored at −70°Cuntil processed. Serial 6-�m sections were cut using a cryo-stat (Cambridge Instruments, Buffalo, NY) and mounted onmicroscope slides. Five or more sections from the center ofthe specimen were stained with hematoxylin and eosin forgeneral morphological characterization. Sequential sectionson both sides of the central sections above were used forimmunoperoxidase staining using monoclonal antibodies toidentify cells expressing the vitronectin receptor (osteoclast-like cells). Osteoblasts were identified using alkaline phos-phatase staining.

Immunohistochemistry

Frozen sections were fixed in absolute acetone at 4°C for10 min, air dried, and the slides treated with the acetone foran additional 10 min. The slides were then dried for 45 min,placed in an airtight container, and stored at −70°C untilprocessed. After thawing, the sections were incubated atroom temperature with monoclonal antibodies to identifyosteoclast-like cells exhibiting the vitronectin receptor

Table IDose–Response Protocol for Experiment 1 Using the

DCT in Rabbits

Week Treatment

0 Osseointegration6 No particles, infuse carrier alone9 UHMWPE particles, infuse carrier alone

12 No particles, IL-10: dose 1: 0.1 ng/mL15 No particles, IL-10: dose 2: 1.0 ng/mL18 No particles, IL-10: dose 3: 10 ng/mL21 No particles, IL-10: dose 4: 100 ng/mL24 UHMWPE particles, IL-10: dose 1: 0.1 ng/mL27 UHMWPE particles, IL-10: dose 2: 1.0 ng/mL30 UHMWPE particles, IL-10: dose 3: 10 ng/mL33 UHMWPE particles, IL-10: dose 4: 100 ng/mL36 UHMWPE particles, infuse carrier alone39 No particles, infuse carrier alone42 Euthanize

Table IIProtocol for Experiment 2 Using the DCT in Rabbits:

6-Week Studies

Week Treatment

0 Osseointegration6 No particles, infuse carrier alone

12 UHMWPE particles, infuse carrier alone18 No particles, IL-10 1.0 ng/mL infused during last 3

of 6 weeks24 No particles, IL-10 1.0 ng/mL during all 6 weeks30 UHMWPE particles, IL-10 1.0 ng/mL infused during

last 3 of 6 weeks36 UHMWPE particles, IL-10 1.0 ng/mL infused during

all 6 weeks42 No particles, infuse carrier alone48 Euthanize

45IL-10 AND BONE INGROWTH

(CD51, mouse anti-human monoclonal antibody to integrin�V�3, Chemicon, Temecula, CA). This monoclonal antibody(although not specific for osteoclasts) has been used previ-ously to identify osteoclast-like cells and correlates stronglywith tartrate resistant acid phosphatase staining.23 Further-more, CD51-positive cells demonstrate more concentratedstaining and are easier to count than cells stained with tar-trate resistant acid phosphatase. The sections were then in-cubated with biotinylated goat anti-mouse immunoglobu-lins and ABC solution (Vector, Burlingame, CA). The sec-tions were washed in phosphate buffered saline after eachantibody step. Positively stained cells were visualized withdiaminobenzidine (SIGMA FAST DAB tablet; Sigma Chemi-cal Co., St Louis, MO) dissolved in 5 mL of deionized water.The slides were counterstained with hematoxylin, dehy-drated, and mounted.

Assessment and statistical analysis

Histomorphometric analysis was performed of multipleserial, longitudinal sections, which enabled the entire lengthof the specimen to be visualized on one slide. A computer-ized, interactive image analysis system using NIH Imageinterfaced with a MacIntosh computer (Apple, Cupertino,CA) and a real-time video light microscope system was usedto perform the morphometric analysis. One section from thecenter of the specimen stained with hematoxylin and eosinwas used for the morphometric analysis. The areas or pointsfor measurement on the entire section were outlined manu-ally using a mouse pad, which was connected to the com-puter and a television screen containing the histological im-age. All measurements were performed blindly, withoutknowledge of the treatment.

Histomorphometric analyses were performed to deter-mine total tissue (section) area, total bone area, ratio of bonearea divided by total tissue area and expressed as a percent-age, ratio of the area of alkaline phosphatase staining di-vided by tissue area and the total number of vitronectinreceptor-positive cells per section area.

Statistical analysis was accomplished using the StatView

software program (Abacus Concepts, Berkeley, CA). Para-metric statistics were used including a one-way analysis ofvariance, and post-hoc paired t tests. A significance level of0.05 was used.

RESULTS

Experiment 1

The number of animals included in each of the treat-ments in experiment 1 is enumerated in Table III. Oneanimal died on induction of anesthesia after the sixthharvest. Four additional animals demonstrated localsigns of infection or inflammation around the chamberat some time during the experiment (after the fourth,sixth, seventh, and seventh harvests), and the animalscould not be salvaged. One other animal had all 12treatments except the final one, secondary to poorwound healing. In two animals who underwent alltreatments, the tubing was found to be disconnectedfrom the chamber on one occasion, and these singleharvests were excluded from the analysis.

When carrier alone was infused, the chamber wasfilled with longitudinally oriented woven bone in afibrovascular stroma (Fig. 2). The addition of PE par-ticles significantly suppressed bone ingrowth as re-flected by percentage of bone area (p = 0.001, see TableIII). Infusion of IL-10 alone had no significant effectson the above parameters. PE particles were associatedwith a 50% reduction in alkaline phosphatase expres-sion. Infusion of IL-10 in the presence of PE particlesreversed, in part, the depressive effects of PE particleson bone ingrowth. When polyethylene particles wereplaced in the chamber, the dose of IL-10 of 1 ng/mLincreased bone ingrowth by 39.2% compared to par-ticles alone. This dose of IL-10 increased boneingrowth the greatest amount in the presence of

Table IIIResults for Experiment 1 Using the Drug Test Chamber in Rabbits

Treatment Percent Bone Areaa Samples

No particles, infuse carrier alone 24.0 ± 2.7 10UHMWPE particles, infuse carrier alone 16.1 ± 2.3** 9No particles, IL-10: dose 1: 0.1 ng/mL 23.4 ± 2.6 8No particles, IL-10: dose 2: 1.0 ng/mL 22.4 ± 2.5 8No particles, IL-10: dose 3: 10 ng/mL 19.9 ± 2.0 8No particles, IL-10: dose 4: 100 ng/mL 26.1 ± 3.4 8UHMWPE particles, IL-10: dose 1: 0.1 ng/mL 21.2 ± 3.0 5UHMWPE particles, IL-10: dose 2: 1.0 ng/mL 25.2 ± 1.2 5UHMWPE particles, IL-10: dose 3: 10 ng/mL 19.8 ± 3.1 5UHMWPE particles, IL-10: dose 4: 100 ng/mL 21.7 ± 5.1 5UHMWPE particles, infuse carrier alone 15.1 ± 1.6* 5No particles, infuse carrier alone 21.7 ± 6.1 4

aThe amount of bone as a percentage of total tissue in the section.**p = 0.0014 compared to the first treatment with no particles and carrier infusion.*p = 0.055 compared to the first treatment with no particles and carrier infusion.

46 GOODMAN ET AL.

UHMWPE particles and was subsequently chosen forexperiment 2. In the presence of particles, expressionof the osteoblast marker alkaline phosphatase was sig-nificantly decreased compared to carrier alone (p =0.03); this effect could not be reversed with the addi-tion of IL-10 at any dose. Although the numbers ofosteoclasts per section area was increased in the pres-ence of particles, this did not reach statistical signifi-cance (Table IV). However, when expressed as thenumber of osteoclasts per bone perimeter, particleswere associated with increased numbers of osteoclas-tic cells (p = 0.05). In the presence of IL-10 at 1 ng/mLwith and without particles, osteoclast numbers wereincreased compared to particles alone (p < 0.05).

Experiment 2

The number of animals included in each of the treat-ments in experiment 2 is enumerated in Table V. Twoanimals demonstrated local signs of infection or in-flammation around the chamber at the fifth harvestand were euthanized. These harvests were excluded.One other animal failed to thrive after the fourth har-vest and demonstrated a systemic neurological abnor-mality and was euthanized. One animal was excludedfor two treatments because of delayed wound healing,which subsequently resolved. On five occasions, nosample was found in the chamber upon harvest.

In experiment 2, the implanted particles demon-strated more clumping than in experiment 1. The par-ticles and surrounding reaction formed a more promi-nent portion of the sections than previously noted inexperiment 1. Infusion of IL-10 alone for 3 or 6 weekshad no significant effects on bone ingrowth comparedto controls (Table V). Infusion of IL-10 during the final3 of 6 weeks reversed in part the suppressive effects ofUHMWPE particles. Bone ingrowth was increased byover 48% with infusion of IL-10 for the last 3 of 6weeks in the presence of UHMWPE particles com-pared to particles alone and this reached statisticalsignificance (p = 0.027). However, the increase in boneingrowth in the 3 of 6-week IL-10 plus particle groupwas still significantly less than controls receiving car-rier alone without particles (p = 0.027). Compared toinfusion of the first carrier, the other significant treat-ments included particles plus IL-10 for 3 of 6 weeks (p= 0.027) or 6 weeks (p = 0.0039) and the second carrier(p = 0.0054). IL-10 plus particles for 3 of 6 weeks de-creased bone ingrowth less than IL-10 plus particlesfor a 6-week period (p = 0.0085).

Particles not only suppressed bone ingrowth com-pared to the initial carrier control but also decreasedthe number of osteoclasts per total section area (p =0.004) and the osteoblast area per total section area (p= 0.003) (Table VI). However, infusion of IL-10 for 3 or

Figure 2. Histology of the tissue retrieved from the DTC:Experiment 1. (a) no particles, infuse carrier alone for 3weeks (hematoxylin and eosin stain); (b) UHMWPE par-ticles, infuse carrier alone for 3 weeks (hematoxylin and eo-sin stain); (c) UHMWPE particles and infuse IL-10 for 3weeks (hematoxylin and eosin stain). When particles wereplaced in the chamber for a 3-week period, bone ingrowthwas suppressed (b) compared to the treatment without par-ticles (a). Bone ingrowth was significantly increased by in-fusion of IL-10 for 3 weeks (c) in comparison to infusion ofcarrier alone in the presence of particles (b). Hematoxylinand eosin staining, 100× magnification. [Color figure can beviewed in the online issue, which is available at www.inter-science.wiley.com]

47IL-10 AND BONE INGROWTH

6 weeks in the presence of particles had no statisticallysignificant effects on osteoclast or osteoblast num-bers/areas compared to the particle treatment alone.

DISCUSSION

Wear of components for joint replacement and par-ticulate disease, the adverse local and systemic bio-logical reaction to wear particles, and their byproductsare currently the most common causes of revision totaljoint replacement.1 Indeed, it has been stated that themajority of revision hip and knee replacements cur-rently performed are related to wear and its sequelae.1

These operations are more difficult than primary ar-throplasty, have a higher complication rate, and havea poorer outcome. Few conservative methods of man-agement are currently available. By understanding thebiological processes associated with osseointegration,loosening, and osteolysis of joint replacements, it ishoped that more conservative, minimally invasive

treatments may be instituted to prolong the lifetime ofa joint arthroplasty.

IL-10 is a cytokine that is implicated in the suppres-sion of the inflammatory response, especially cell-mediated immunity that is characteristic of the TH1response.13,14 IL-10 down-regulates the spontaneousand cytokine-induced expression of HLA class II an-tigens on human monocytes. IL-10 decreases the se-cretion of the pro-inflammatory cytokines IL-1, IL-6,IL-8, tumor necrosis factor-�, and GM-CSF by mono-cytes/macrophages, increases the production of theanti-inflammatory cytokine IL-1ra, and inhibits theproduction of PGE2 and reactive oxygen (H2O2) andnitrogen intermediates after activation of monocytes/macrophages. The effects of IL-10 are pleiotropicwith respect to dose and time. As the above pro-in-flammatory mediators have all been implicated in par-ticle associated osteolysis, local infusion of IL-10 pro-vided a potential approach to mitigate the adversebiological effects of wear debris.15,16,24–28

In experiment 1, the addition of phagocytosable PEparticles decreased bone ingrowth into the DTC cham-

Table VResults for Experiment 2 Using the DTC in Rabbits

Treatment Percent Bone Areaa Samples

No particles, infuse carrier alone 37.9 ± 4.1 10UHMWPE particles, infuse carrier alone 17.2 ± 2.3** 8No particles, IL-10 last 3 of 6 weeks 27.5 ± 5.4 9No particles, IL-10 1.0 ng/mL during all 6 weeks 30.4 ± 5.4 9UHMWPE particles, IL-10 last 3 of 6 weeks 25.6 ± 5*# 6UHMWPE particles, IL-10 all 6 weeks 13.6 ± 3.6 6No particles, infuse carrier alone 27.1 ± 3.8 6

aAmount of bone as a percentage of total tissue in the section.**p = 0.0003 compared to carrier alone.*p = 0.027 compared to carrier alone.�p = 0.027 compared to particles.

Table IVNumber of Cells Positive for the Vitronectin Receptor per Section Area and Alkaline Phosphatase Staining per

Section Area Harvested from the DTC Containing IL-10 and Polyethylene Particles

Treatment Total VR+/SAa Alk Phosph/SAb

No particles, infuse carrier alone 55.0 ± 4.9 0.076 ± 0.009+ particles, infuse carrier alone 62.5 ± 9.6 0.043 ± 0.007No particles, IL-10: dose 1: 0.1 ng/mL 47.5 ± 6.6 0.038 ± 0.008No particles, IL-10: dose 2: 1.0 ng/mL 95.5 ± 10.6* 0.027 ± 0.004No particles, IL-10: dose 3: 10 ng/mL 89.2 ± 16.7 0.044 ± 0.006No particles, IL-10: dose 4: 100ng/mL 80.6 ± 6.8 0.052 ± 0.009+ particles, IL-10: dose 1: 0.1 ng/mL 67.1 ± 11.1 0.023 ± 0.006+ particles, IL-10: dose 2: 1 ng/mL 89.8 ± 17.6# 0.034 ± 0.002+ particles, IL-10: dose 3: 10 ng/mL 46.7 ± 7.0 0.036 ± 0.008+ particles, IL-10: dose 4: 100 ng/mL 64.0 ± 16 0.037 ± 0.006+ particles, infuse carrier alone 75.8 ± 15.4 0.075 ± 0.016No particles, infuse carrier alone 63.7 ± 21.8 0.068 ± 0.012

aTotal number of vitronectin receptor-positive cells per section area.bTotal area of staining for alkaline phosphatase per section area.*p = 0.038 compared to particles alone.#p = 0.034 compared to particles alone.

48 GOODMAN ET AL.

ber. Furthermore, the osteoblast marker alkaline phos-phatase was significantly decreased in the particletreatment compared to the carrier alone. The numbersof osteoclasts per bone perimeter were also increasedby the addition of particles to the chamber. These find-ings are consistent with previous studies using thismodel and others and implicate both decreased boneformation and increased bone resorption leading todecreased net bone formation in the presence of PEparticles.4,7,23

Locally infused IL-10 appeared to reverse, in part,the suppressive effects of PE particles on bone in-growth in the DTC model. In experiment 1, a 1 ng/mLdose of IL-10 yielded the greatest amount of bone in-growth in the presence of particles, and this dose wassubsequently chosen for experiment 2. However, inexperiment 1, the percentage of bone ingrowth in theparticle plus 1 ng/mL of IL-10 treatment was notfound to be statistically different from the treatment ofparticles alone. This may be due to the small numberof animals that underwent both treatments due to at-trition. As well, the administration of IL-10 immedi-ately concurrent with placement of the particles mayhave suppressed the initial inflammatory responsethat is part of the process of fracture healing and boneingrowth. Interestingly, the addition of 1.0 ng/mL ofIL-10 (in the presence or absence of PE particles) wasassociated with significantly increased osteoclastnumbers; alkaline phosphatase staining, though de-creased was not significantly different from the treat-ment with particles alone. These findings were unex-pected and may be due to model-specific effects or thepleiotropic nature of IL-10.

In experiment 2, the addition of particles for a6-week period significantly decreased bone ingrowth,similar to that seen at 3 weeks in experiment 1. Theaddition of IL-10 for a 6-week period in the presenceof particles depressed bone ingrowth, probably as aresult of the persistent effects of the particles over a

6-week period and the prolonged suppression of theinitial inflammatory reaction necessary for fracturehealing and bone ingrowth. However, when IL-10 wasinfused for the last 3 of 6 weeks with particles in thechamber for 6 weeks, bone ingrowth was significantlyincreased compared to the particle treatment. In thiscase, IL-10 did not appear to interfere with the initialphases of inflammation and bone ingrowth (becauseIL-10 infusion was instituted after 3 weeks), yet IL-10infusion favorably modulated the foreign body andinflammatory reaction to particulate polyethylene.Subsequent analysis did not help distinguish whetherthese effects of IL-10 in the presence of particles werea result of modulation of osteoblasts, osteoclasts orboth. A limitation of the present experiments is thefact that the analysis performed documents changes incell numbers or areas rather than function. Furtherexperiments using the present model, combining invivo and in vitro studies may further delineate the un-derlying mechanisms of bone metabolism, in thepresence of particles and IL-10.29 One potentialmechanism influencing the changes in bone ingrowthobserved here may be inhibition by IL-10 of the pro-inflammatory cytokines IL-6, tumor necrosis factor-�and IL-1�, all of which are up-regulated in the BHC inthe presence of particles.16,29

The PE particle sizes and dosages used in the cur-rent experiments were derived from human retrievaland animal studies, and are therefore clinically rel-evant.7,20–22 Even higher concentrations of PE particlesmay be present at specific areas of the bone-implantinterface, such as in areas of ballooning osteolysis.1

Based on mathematical calculations, the concentrationof 1.7 × 1012 particles/mL used in the present experi-ments would take up less than 10% of the volume forbone ingrowth within the chamber. This would haveno effect on comparisons between treatments withparticles (±IL-10), and negligible effects on compari-sons with the control treatments.

Table VINumber of Cells Positive for the Vitronectin Receptor in the Entire Section and Ratio of Alkaline Phosphatase

Staining Area Divided by Section Area in Specimens Harvested from the DTC Containing IL-10 and PolyethyleneParticles in Experiment 2

Treatment Total VR+/SAa Alk Phosph/SAb

No particles, infuse carrier alone 47.7 ± 4.9** 0.150 ± 0.015#

UHMWPE particles, infuse carrier alone 33.9 ± 3.6 0.076 ± 0.008No particles, IL-10 last 3 of 6 weeks 46.8 ± 6.8 0.100 ± 0.012No particles, IL-10 1.0 ng/mL during all 6 weeks 46.7 ± 4.9* 0.111 ± 0.010^UHMWPE particles, IL-10 last 3 of 6 weeks 36.4 ± 7 0.070 ± 0.010UHMWPE particles, IL-10 all 6 weeks 50.9 ± 7.4 0.084 ± 0.012No particles, infuse carrier alone 49.7 ± 8.9 0.070 ± 0.011

aTotal number of vitronectin receptor-positive cells per section area.bTotal area of staining for alkaline phosphatase per section area.**p = 0.004 compared to particles alone.*p = 0.037 compared to particles alone.�p = 0.003 compared to particles alone.^p = 0.028 compared to particles alone.

49IL-10 AND BONE INGROWTH

Wear particles are continuously generated from ar-ticulating and non-articulating interfaces of total jointreplacements. Although advances are being made inbearing materials, prosthetic design and surgical tech-nique, wear debris will always be produced. If oste-olysis is diagnosed in its early stages and is veryslowly progressive (as diagnosed clinically and radio-graphically) and the prosthesis and periprostheticbone are not irrevocably affected, temporizing mea-sures may be indicated to prolong the life of the pros-thesis and delay revision surgery. In this respect, localinfusion of immune modulating cytokines such as IL-10 may prove to be useful in mitigating particle-induced peri-prosthetic osteolysis and provide a mini-mally invasive alternative to revision surgery.

References

1. Wright TM, Goodman SB. Implant wear in total joint replace-ment: clinical and biologic issues, material and design consid-erations. Rosemont, IL: American Academy of OrthopaedicSurgeons; 2001. p 1–224.

2. Amstutz HC, Ma SM, Jinnah RH, Mai L. Revision of asepticloose total hip arthroplasties. Clin Orthop 1982;170:21–33.

3. Kodoya Y, Revell PA, Al-Saffar N, Kobayashi A, Scott G, Free-man MAR. Bone formation and bone resorption in failed totaljoint arthroplasties: histomorphometric analysis with histo-chemical and immunohistochemical technique. J Orthop Res1996;14:473–482.

4. Dean DD, Schwartz Z, Liu Y, Blanchard CR, Agrawal C,Mabrey JD, Sylvia VL, Lohmann CH, Boyan BD. The effect ofultra-high molecular weight polyethylene wear debris onMG63 osteosarcoma cells in vitro. J Bone Joint Surg 1999;81-A:452–461.

5. Yao J, Cs-Szab G, Jacobs JJ, Kuettner KE, Glant TT. Suppressionof osteoblast function by titanium particles. J Bone Joint Surg1997;79A:107–112.

6. Allen MJ, Myer BJ, Millet PJ, Rushton N. The effects of par-ticulate cobalt, chromium and cobalt-chromium alloy on hu-man osteoblast-like cells in vitro. J Bone Joint Surg 1997;79-B:475–482.

7. Goodman SB. The effects of micromotion and particulate ma-terials on tissue differentiation. Bone chamber studies in rab-bits. Acta Orthop Scand 1994;65(Supp 258):1–43.

8. Schmalzried TP, Kwong LM, Jasty M, Sedlacek RC, Haire TC,O’Connor DO, Bragdon CR, Kabo JM, Malcolm AJ, Harris WH.The mechanism of loosening of cemented acetabular compo-nents in total hip arthroplasty. Analysis of specimens retrievedat autopsy. Clin Orthop 1992;274:60–78.

9. Vornasier VL, Wright J, Seligman J. The histomorphologic andmorphometric study of asymptomatic hip arthroplasty—Apost-mortem study. Clin Orthop 1991;271:272–282.

10. Maloney WJ, Jasty M, Harris WH, Galante JO, Callaghan JJ.Endosteal erosion in association with stable uncemented femo-ral components. J Bone Joint Surg 1990;72-A:1026–1034.

11. Schmalzried TP, Guttmann D, Grecula M, Amstutz HC. Therelationship between design, position, and articular wear ofacetabular components inserted without cement and the de-velopment of pelvic osteolysis. J Bone Joint Surg 1994;76-A:677–688.

12. Willert HG, Bertram H, Buchhorn GH. Osteolysis in alloarthro-plasty of the hip. The role of ultra-high molecular weight poly-ethylene wear particles. Clin Orthop 1990;258:95–107.

13. Banchereau J, Briere F, Caux C, Fluckiger A-C, Merville P,Rousset F. Interleukin-10. In: Henderson B, Bodmer MW, edi-tors. Therapeutic modulation of cytokines. Boca Raton, FL:CRC Press; 1996. p 317–331.

14. Di Giovine FS, Mee JR, Duff GW. Immunoregulatory cyto-kines. In: Henderson B, Bodmer MW, editors. Therapeuticmodulation of cytokines. Boca Raton, FL: CRC Press; 1996.p 37–65.

15. Pollice PF, Hsu J, Hicks DG, Bukata S, Rosier RN, Reynolds PR,Puzas JE, O’Keefe RJ. Interleukin-10 inhibits cytokine synthesisin monocytes stimulated by titanium particles: evidence of ananti-inflammatory regulatory pathway. J Orthop Res 1998;16:697–704.

16. Trindade MC, Lind M, Nakashima Y, Sun D, Goodman SB,Schurman DJ, Smith RL. Interleukin-10 inhibits polymethyl-methacrylate particle induced interleukin-6 and tumor necro-sis factor-alpha release by human monocyte/macrophages invitro. Biomaterials 2001;22:2067–2073.

17. Aspenberg P, Albrektsson T, Lohmander LS, Thorngren KG.Drug test chamber: a titanium implant for administration ofbiochemical agents to a standardized bone callus in situ.J Biomed Eng 1988;10:70–73.

18. Ragab AA, Bi Y, Lavish SA, Van De Motter RR, Goldberg VM,Greenfield EM. Stimulation of cytokine production by titaniumwear particles is due to adherent endotoxin. Trans OrthopaedRes Soc 1988;23:355.

19. Ragab AA, Van De Motter R, Lavish SA, Goldberg VM,Ninomiya JT, Carlin CR, Greenfield EM. Measurement andremoval of adherent endotoxin from titanium particles andimplant surfaces. J Orthop Res 1999;17:803–809.

20. Campbell P, Ma S, Yeom B, McKellop H, Schmalzried TP,Amstutz HC. Isolation of predominantly submicron-sizedUHMWPE wear particles from periprosthetic tissues. J BiomedMater Res 1995;29:127–131.

21. Goodman SB, Song Y, Chun L, Regula D, Aspenberg P. Effectsof TGF beta on bone ingrowth in the presence of polyethyleneparticles. J Bone Joint Surg 1999;81-B:1069–1075.

22. Margevicius KJ, Bauer TW, McMahon JT, Brown SA, Merritt K.Isolation and characterization of debris in membranes aroundtotal joint prostheses. J Bone Joint Surg 1994;76:1664–1675.

23. Chun L, Yoon J, Song Y, Huie P, Regula D, Goodman SB. Thecharacterization of macrophages and osteoclasts in tissues har-vested from revised total hip prostheses. J Biomed Mater Res1999;48:899–903.

24. Shanbhag AS, Jacobs JJ; Black J, Galante JO, Glant TT. Cellularmediators secreted by interfacial membranes obtained at revi-sion total hip arthroplasty. J Arthroplasty 1995;10:498–506.

25. Horikoshi M, Macaulay W, Booth RE, Crossett LS, Rubash HE.Comparison of interface membranes from failed cemented andcementless hip and knee prostheses. Clin Orthop 1994;309:69–87.

26. Jiranek WA, Machado M, Jasty M, Jevsevar D, Wolfe HJ,Goldring SR, Goldberg MJ, Harris WH. Production of cyto-kines around loosened cemented acetabular components.Analysis with immunohistochemical techniques and in situ hy-bridization. J Bone Joint Surg 1993;75-A:863–879.

27. Goodman SB, Huie P, Song Y, Schurman D, Maloney W, Wool-son S, Sibley R. Cellular profile and cytokine production atprosthetic interfaces. J Bone Joint Surg 1998;80-B:531–539.

28. Takei I, Takagi M, Santavirta S, Ida H, Ishii M, Ogino T, AinolaM, Konttinen YT. Messenger ribonucleic acid expression of 16matrix metalloproteinases in bone-implant interface tissues ofloose artificial hip joints. J Biomed Mater Res 2000;52:613–620.

29. Trindade MC, Song Y, Aspenberg P, Smith RL, Goodman SB.Proinflammatory mediator release in response to particle chal-lenge: studies using the bone harvest chamber. J Biomed MaterRes 1999;48:434–439.

50 GOODMAN ET AL.