Biomaterials 23 (2002) 4833–4838

Nitric oxide synthase in tissues around failed hip prostheses

S. Steaa,*, M. Visentina, M.E. Donatib, D. Granchic, G. Ciapettic, A. Sudanesed, A. Tonia,d

aLaboratory of Medical Technology, Istituti Ortopedici Rizzoli, Via di Barbiano 1/10, Bologna, ItalybLaboratory for Biocompatibility Research on Implant Materials, Istituti Ortopedici Rizzoli, Bologna, Italy

cLaboratory of Pathophysiology of Orthopaedic Implants, Istituti Ortopedici Rizzoli, Bologna, Italyd1st Orthopaedic Department, Istituti Ortopedici Rizzoli, Bologna, Italy

Received 28 January 2002; accepted 10 May 2002

Abstract

Nineteen patients who had undergone hip revision surgery for aseptic loosening of joint prostheses were studied. Tissue samples

were harvested at the interface between bone and implant, either at the stem or at the cotyle level. Immunohistochemistry was

performed on tissue sections to detect nitric oxide synthase (NOS), the enzyme which enables the synthesis of nitric oxide (NO), a

molecule which can activate bone resorption. Quantitative analysis of the positive cells and correlation with the presence of

particulate wear debris and radiological data were performed.

The authors observed a trend towards a moderate increase in positive cells due to inducible NOS in tissues containing particulate

wear debris, especially of a plastic material. This increase, however, did not achieve statistical significance. On the contrary, there

was a statistical correlation between iNOS (inducible NOS) and the severity of osteolysis around the prosthetic implant.

Pharmacological control of the biosynthesis of NO may be considered in the prevention or treatment of loosening.

r 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Nitric oxide synthase; Hip joint prosthesis; Loosening; Immunohistochemistry; Bone/implant interface

1. Introduction

Aseptic loosening is the major complication leading torevision surgery of total hip replacement. The produc-tion of particulate wear debris contributes to localosteolysis and bone resorption at the bone–prosthesisinterface. Cells involved in this phenomenon are mainlymacrophages, activated through phagocytosis, thatrelease mediators responsible for bone resorption, suchas Interleukin 1 alpha (IL1 a), Interleukin 1 beta (IL1 b),Interleukin 6 (IL6), Tumor Necrosis Factor alpha(TNF a) [1–4]. Proinflammatory cytokines, such asInterleukin 1, Tumor Necrosis Factor and Interferongamma have potent effects on bone remodelling and arealso capable of stimulating nitric oxide (NO) productionin bone cells. The mechanism for this appears to be acytokine-mediated activation of the transcription factorNFkb in macrophages, which leads to an upregulationof the transcription of inducible NOS, the presence of

iNOS in human macrophages can thus be correlatedwith the synthesis of NO.

NO is a small molecule that is generated by the nitricoxide synthase enzymes (NOS) from molecular oxygenand the terminal guanidino nitrogen of the amino acidl-arginine, yielding l-citrulline as a co-product [5,6].

NO can also be generated nonenzymatically fromnitrite in the acid environment of the stomach [7] andpharmacologically by compounds such as organicnitrates and sodium nitroprusside [5], which are usedclinically as vasodilators. Three isoforms of NOS havebeen identified: a neuronal form (nNOS or NOS I) [8],and endothelial form (eNOS or NOS III) [9], and aninducible form (iNOS or NOS II) [10]. Both eNOS andnNOS are constitutively expressed at low levels in theirtissues of origin and their activity is mainly regulated bychanges in free intracellular Ca2+ concentration. Reg-ulation of iNOS mainly takes place at the level of genetranscription. Transcription of the iNOS gene isactivated by pro-inflammatory cytokines such as inter-leukin-1 (IL-1), TNF a, gamma interferon (g IFN) andendotoxin, whereas glucocorticoids and anti-inflamma-tory cytokines IL-4, IL-10 and TGF b are inhibitory

*Corresponding author. Tel.: +39-51-6366-861; fax: +39-51-6366-

863.

E-mail address: stea@tecno.ior.it (S. Stea).

0142-9612/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved.

PII: S 0 1 4 2 - 9 6 1 2 ( 0 2 ) 0 0 2 3 6 - 3

[11]. Different cell types differ in their ability to produceNO after cytokine stimulation. Human primary osteo-blast and hepatocyte cultures require combination oftwo or three cytokines for a significant induction of NOsynthesis [12] whereas human chondrocytes can beinduced by single cytokines such as IL-1 and TNF toproduce NO [13]. In all cell types combinations ofcytokines are generally more potent inducers of NOthan single cytokines [14].

The iNOS pathway is thought to be capable ofgenerating much larger quantities of NO (nanomolarrange) over a more prolonged time frame than theconstitutive NOS (cNOS) enzymes (picomolar concen-trations) [15]. The role of NO as physiologic mediatorhas been extensively reviewed as it is rather unique. It ishighly diffusible and soluble in aqueous and hydro-phobic environments, but its activity is limited by bloodvessels, as it is inactivated by haemoglobin. It does nothave a specific cell receptor, and it diffuses through cellmembranes. It regulates blood pressure, platelet aggre-gation, wound healing and apoptosis but it is alsoinvolved in allograft response, colitis and gastrointest-inal mobility disorders.

In addition, it is implicated in a number of conditionsof orthopaedic interest, including inflammation, arthri-tis, osteoporosis, sepsis, ligament healing and asepticloosening of joint prostheses [6]. There is good evidenceto suggest that NO has biphasic effects on osteoclasticbone resorption. Low concentrations of NO have beenshown to potentiate IL-1-induced bone resorption.Accumulating evidence suggests that the iNOS pathwayplays an important role in cytokine- and inflammation-induced bone loss. Inflammation-induced osteoporosishas been shown to be mediated in part by activation ofthe iNOS pathway [16] and recent studies have showsthat activation of the iNOS pathway is essential for IL-1-stimulated bone resorption, both in vivo and in vitro[17]. High concentrations of NO inhibit osteoclastformation and activity. Experiments using cell andorgan cultures have shown that IFN g in combinationwith IL-1 and/or TNF a induces iNOS expression,leading to very high levels of NO that inhibit boneresorption [18,19]. The inhibitory effect of high level ofNO appear to be due to the inhibition of both osteoclastformation and activity [20] and NO-induced apoptosisof osteoclast progenitors [20]. NO appears to havebiphasic effects on osteoblast activity: small amounts ofNO which are produced constitutively by osteoblastsmay act as stimulator of growth and cytokine produc-tion [21] while high concentrations of NO, such as thoseobserved after stimulation with pro-inflammatory cyto-kines, have potent inhibitory effect on osteoblast growthand differentiation [22]. The aim of this study was toverify the presence of NO in cases of hip prosthesisloosening and, at the same time, explore its possiblecorrelation with the presence of particles released by the

wear of joint surfaces. To estimate the amount of NOpresent in the tissues in vivo, the percentage of cellscontaining iNOS was determined both in the interfacialpsuedomembranes and in the newly formed jointcapsule collected during revision surgery of loosenedhip prostheses.

2. Materials and methods

2.1. Materials

In the present study nineteen patients were examined.All were affected by loosening of hip prosthesis. Theirmean age was 64.3 yr (range 26–88 yr). The mean age ofthe implant was 109 months (range 8–240 months).Twelve of the loosened prostheses comprised a metaland polyethylene articulating components. Five wereceramic–ceramic and 1 was an endoprostheses. Inaddition, 1 undefined arthroprosthesis was studied. Ofthe nineteen prostheses, 8 had either a cemented stem orcotyle.

Prosthesis stability was evaluated by comparingantero-posterior and lateral view radiographs beforeand after surgery. The results were classified intothree groups: stable, at risk or unstable, depending onthe presence of bone contact or radiolucency around theimplant, according to Engh’s method [23]. This methodinvolves studying the progression of bone/prosthesiscontact compared to what was observed on the firstradiograph after surgery.

Interface membranes were collected from the acet-abular and proximal periprosthetic region of the femoralshaft after removal of the loosened prosthesis. Samplesfrom newly formed joint capsule were taken as well. Theorientation of the tissue was preserved by putting amark on the surface facing the prosthesis. Altogether 29biopsies were examined, 11 of which were taken at theinterface with the stem, 11 at the interface with thecotyle and 7 from the newly formed joint capsule.

2.2. Histological methods

All testing procedures used in our Laboratory arevalidated according to the quality assurance criteria ofthe E.U., the accepted standard for the operation oftesting laboratories (EN 45001 standard). Tissue speci-mens were coded and blinded at reception. Samples wereincluded in optimum cutting temperature media (OCT)(Miles, Elkhart, IN) snap-frozen in liquid nitrogen andstored at �70751C until processed. A few serial sections6 mm thick were cut, using a cryostat, from each tissuefragment. They were air-dried and fixed in cold acetonefor immunohistochemical analysis. Whenever possible,part of the tissue sample was paraffin embeddedfollowing standard methods and Hematoxylin–eosin

S. Stea et al. / Biomaterials 23 (2002) 4833–48384834

staining was performed. Detection of particulate weardebris was carried out in standard and polarized lightmicroscopy and classified with a score from 0 (no weardebris) to 5 (copious wear debris, inducing necrosis). Asemiquantitative evaluation of particulate wear debriswas carried out using � 500 magnification following theguidelines of a previously published method [24]. Thedetection limit of the method was obviously set up bythe resolving power of light microscopy; therefore,submicron particles were not detected.

2.3. Immunohistochemical analysis

Slides were air-dried and rehydrated in phosphate-buffered saline for 10min. Endogenous peroxidaseactivity was inhibited with 1% hydrogen peroxide for1min at 41C, followed by rinsing in two changes ofphosphate-buffered saline for 5min. The slides werethen incubated with monoclonal anti-iNOS antibody,isotype IgG2a, clone 6 diluted 1:100 (TransductionLaboratories, Lexington, KY, USA) in a humidifiedchamber for 2 h at room temperature. Negative controlsto check for nonspecific binding included replacing theprimary antibody with buffer. Positive controls includedbiopsied specimens from synovia obtained from rheu-matic patients. Cell positivity was detected by using astreptavidin–biotin–peroxidase system. Slides wererinsed in two changes of phosphate-buffered saline andincubated for 60min at room temperature with abiotinylated secondary antibody (Sigma, St. Louis, MO,USA) followed by a rinse in two changes of phosphate-buffered saline and incubation with streptavidin–perox-idase complex for 60min at room temperature. Thesections were then exposed, to 3-amino-9-ethylcarbazole(AEC) and diluted hydrogen peroxide for 10mins,counterstained in hematoxylin and mounted in glycer-ol–gelatin. They were stored in the dark at +41C untilexamined.

2.4. Image analysis

Computerized image analysis was used to analysestained sections in order to obtain the percentage of cellspositive to cytokine immunohistochemistry. The imageanalysis system consisted of a video camera, a monitor,a graphic tablet and software (Immagini e Computer,Milano, Italy).

Using a 40� microscope objective, 8 randomlyselected fields were analysed for each section, withthe interposition of a monochromatic filter at 530 nmto better identify positive cells. Selected fieldswere digitized and the computer images obtained wereelaborated to enhance positive cell identification. Thetotal number of cells and the number of positive cellswere finally calculated with the help of a computerizedcounter. The result was expressed as a percentage of

positive cells. Reproducibility of multiple sections fromthe same site was checked through the analysis ofduplicates in a blind manner.

2.5. Statistical methods

Analysis of variance was carried out to determine thevariance between the samples within each patient as wellas the variance between different patients. Variance wasanalysed by means of the Kruskal–Wallis test. Correla-tion was tested by applying the Fisher test. The Studentt-test and nonparametric U-test of Mann Whitney wereutilized as well. The critical p value was always set at0.05.

3. Results

3.1. Histological analysis of wear debris

Table 1 shows patient data, type and score ofparticulate wear debris present in the tissues, whichwere collected during revision surgery, and the semi-quantitative results of the radiological analysis ofloosening. The scores shown for wear debris are thehighest among those found in the single biopsies of eachpatient. In 15 biopsies (52%) only plastic wear debriswas present, in 2 (7%) only metallic wear debris, in 2(7%) only ceramic, in 4 (14%) mixed wear debris,mainly metallic–plastic. In the remaining 6 (20%)biopsies no type of wear debris could be detected bythe optic microscope.

The histological pattern was the usual one, with aninfiltration of histiocytes phagocyting the small particlesand extracellular clusters of bigger particles, especiallyof ceramic material. Giant cell reactions were observedonly when there were large plastic particles.

3.2. Immunohistochemical analysis

The quantitative results of immunohistochemical testsare shown in Table 2. There is a tendency towards anincrease in positivity in biopsies with plastic or metalwear debris compared to biopsies in which wear debriswas absent. The increase, however, is not statisticallysignificant. The distribution of iNOS positive cellswithin the microscopic fields was not uniform becausecells phagocyting wear particles, or surrounding largeclusters, were positive.

3.3. Immunohistochemical and loosening analysis

The results of immunohistochemical tests werecompared to the severity of prosthesis loosening, andit was found that in cases of severe osteolysis, the

S. Stea et al. / Biomaterials 23 (2002) 4833–4838 4835

incidence of iNOS positive cells was higher than in casesof mild or moderate osteolysis (Table 3).

The increase is significant in the stem specimens(p ¼ 0:0417), while it is not in cotyle specimens.

4. Discussion

This study analyses the presence of iNOS in tissuespecimens from around the prosthesis in cases ofrevision surgery due to hip prosthesis loosening. It wasobserved that the presence of the enzyme dedicated tothe synthesis of NO is associated with the presence ofparticulate wear debris. The significance to be attributedto the presence of this enzyme is not clear. In fact, therole the NO molecule plays in bone metabolism is stillcontroversial. Its very peculiar chemical characteristicsand such a brief half-life may provide an explanation tothe conflicting evaluations carried out until now on theaction of NO. In fact, it is currently hypothesised thatNO has two different types of effect. Direct effects aredefined as those reactions fast enough to occur betweenNO and specific biological molecules. Indirect effects donot involve NO, but rather are mediated by reactivenitrogen oxide species (RNOS) formed from thereaction of NO either with oxygen or superoxide [25].It seems certain, however, that NO exerts biphasiceffects on bone cell activity: high concentrations of NOinhibit bone resorption by inhibiting osteoclast forma-tion and by inhibiting the resorptive function of matureosteoclasts, and, at the same time, inhibiting growth anddifferentiation of osteoblasts. Lower NO concentrationspotentiate cytokine-induced bone resorption and may beessential for normal osteoclast function and osteoblastgrowth [26].

The presence of NO in low doses, therefore, wouldseem to be an unfavourable element in the prognosis of

Table 2

Percentage of iNOS-positive cells according to the presence of

particulate wear debris in the biopsy

No. of

biopsies

Percentage of iNOS-

positive cells

Mean7Standard

deviation

No wear debris 6 30732

Plastic wear debris 15 49730

Metal wear debris 2 4577

Ceramic wear debris 2 11716

Mixed type wear debris 4 33741

Table 3

Percentage of iNOS-positive cells found in interface tissues according

to the degree of implant loosening

No. of

biopsies

Percentage of iNOS-

positive cells

Mean7standard

deviation (median)

Mild or modest cotyle

loosening

15 35732 (20)

Severe cotyle loosening 12 48732 (50)

Mild or modest stem

loosening

12 26726 (20)

Severe stem loosening 17 50732 (50)

Table 1

Summary of personal data of the patients, implant duration in months, articular coupling

Patient Sex Age Implant

duration

Coupling Plastic

wear

debris

Metal

wear

debris

Ceramic

wear

debris

Degree of

osteolysis

around

cotyle

Degree of

osteolysis

around

stem

V. V. F 63 24 M/pl 2 1 0 m m

M.C. F 76 48 M/pl 5 0 0 s s

S.L. M 75 82 M/pl 0 1 0 m m

K.H. F 88 168 M/pl 3 0 0 s s

Q. C. M 26 8 Cer/cer 0 4 0 m m

S.L. F 69 48 Un 0 0 0 m m

F.G F 64 156 M/pl 4 0 0 s s

F. F 69 72 Cer–cer 0 0 0 m m

D S.A. F 53 72 Cer–cer 0 0 3 s s

C.A. F 73 132 Cer–cer 2 0 2 m m

P.M. F 64 12 M/pl 1 1 0 s s

C.L. F 77 144 M/pl 4 0 0 m s

D. S. F 71 156 M/pl 4 0 0 m m

C.G. M 37 216 M/pl 4 1 0 m m

L. M. F 74 24 Endo 4 0 0 n.v. s

S.G. F 76 72 M/pl 0 1 0 s s

A.V. F 75 108 M/pl 5 1 0 m m

Q.A. M 64 144 M/pl 4 0 0 s m

L.G. F 60 240 M/pl 4 0 0 m s

M=metal, Pl=plastic, Cer=ceramic, Un=Unknown, n.v.=nonvaluable, m=mild or modest s=severe.

S. Stea et al. / Biomaterials 23 (2002) 4833–48384836

joint prostheses. The osteolysis observed in asepticloosening may be partially due to the direct action ofsome proinflammatory cytokine on bone cells andpartially mediated by NO.

In fact, many of the iNOS-positive macrophageswithin the interfacial membrane had phagocytosed largeamounts of wear debris, suggesting a role for phagocyticstimuli in inducing iNOS in human macrophages. Thesefindings confirm in vitro and in vivo data recentlyreported by other authors [27,28].

In particular, it has been demonstrated in vitro thatmurine macrophage cells (RAW 264.7 cell line) stimu-lated with particulate debris are capable of releasingcopious amounts of NO and that the release isdependent upon the period of culture and the type anddosage of the challenging particles. Titanium alloyparticles are the most stimulatory, followed by titaniumand polymethylmethacrylate [29].

Other authors found that besides iNOS production,peroxynitrite-induced cellular damage is evident inperimplant tissue. That means high levels of NO cancause macrophage cell death and this would result in therelease of phagocytosed wear debris into the extracel-lular matrix. A detrimental cycle of events would then beestablished with further phagocytosis by newly recruitedinflammatory cells and subsequent NO synthesis. SinceNO has been implicated in the induction and main-tenance of chronic inflammation with resulting loss ofbone, and peroxynitrite in the pathogenesis of celldamage, they both may be central to the pathogenesis ofaseptic loosening [30].

5. Conclusions

We conclude that iNOS is expressed by perimplantmacrophages in vivo and that NO may be involved inaseptic loosening. Phagocytosis of wear debris maypromote the induction of iNOS and subsequent NObiosynthesis. The severity of periprosthetic osteolysis isrelated to the amount of iNOS present in interfacialtissue. Pharmacological control of NO biosynthesis maybe considered in the prevention or treatment of loosen-ing.

Acknowledgements

We are grateful to Giuseppina Sivieri for her technicalhelp.

References

[1] al Saffar N, Revell PA. Pathology of the bone-implant interfaces.

J Long-Term Effects Med Implants 1999;9(4):319–47.

[2] Goodman SB, Knoblich G, O’Connor M, Song Y, Huie P, Sibley

R. Heterogeneity in cellular and cytokine profiles from multiple

samples of tissue surrounding revised hip prostheses. J Biomed

Mater Res 1996;31:421–8.

[3] Jiranek WA, Machado M, Jasty M, Jevsevar JD, Wolfe H,

Goldring SR, Goldberg MJ, Harris WH. Production of cytokines

around loosened cemented acetabular components. Analysis with

immunohistochemical techniques and in situ hybridization.

J Bone Jt Surg Am 1993;75:863–79.

[4] Stea S, Visentin M, Granchi D, Melchiorri C, Soldati S, Sudanese

A, Toni A, Montanaro L, Pizzoferrato A. Wear debris and

cytokines production in the interface membrane of loosened

prostheses. J Biomater Sci, Polym Ed 1999;10:247–58.

[5] Feelisch M, Stamler JS. Methods in nitric oxide research.

Chichester: Wiley, 1996.

[6] Evans CH, Stefanovic-Racic M, Lancaster J. Nitric oxide and

its role in orthopaedic disease. Clin Orthop Rel Res 1995;312:

275–94.

[7] Benjamin N, O’Driscoll F, Dougall H, Duncan C, Smith L,

Golden M, Mc Kenzie H. Sthomac NO synthesis. Nature 1994;

368:502.

[8] Bredt DS, Hwang PM, Glatt CE, Lowenstein C, Reed RR,

Snyder SH. Cloned and expressed nitric oxide synthase structu-

rally resembles cytochrome P-450 reductase. Nature 1991;351:

714–8.

[9] Robinson LJ, Weremowicz S, Morton CC, Michel T. Isolation

and chromosomal localization of the human endothelial nitic

oxide synthase (nos3) gene. Genomics 1994;19:350–7.

[10] Lowenstein CJ, Glatt CS, Bredt DS, Snyder SH. Cloned and

expressed macrophage nitric oxide synthase contrasts with the

brain enzyme. Proc Natl Acad Sci USA 1992;89:6711–5.

[11] Moncada S, Higgs A. The l-Arginine nitric oxide pathway.

N Engl J Med 1993;329:2002–12.

[12] Ralston SH, Todd D, Helfrich MH, Benjamin N, Grabowski PS.

Human osteoblast-like cells produce nitric oxide and express

inducible nitric oxide synthase. Endocrinology 1994;135:330–6.

[13] Grabowski PS, MacPherson H, Ralston SH. Nitric oxide

production in cells derived from the human joint. Br J Rheum

1996;35:207–12.

[14] Ralston SH, Grabowski PS. Mechanisms of cytokine induced

bone resorption: role of nitric oxide cyclic guanoside monopho-

sphate and prostaglandins. Bone 1996;19:29–33.

[15] Palmer RMJ. The l-arginine nitric oxide pathway. Curr Opin

Nephrol Hyperten 1993;2:122–8.

[16] Armour KE, Van’t Hof RJ, Grabowski PS, Reid DM, Ralston

SH. Evidence for a pathogenic role of nitric oxide in inflamma-

tion-induced osteoporosis. J Bone Miner Res 1999;14:2137–42.

[17] van’ t Hof RJ, Armour KJ, Smith LM, Amrmour KE, Wei XQ,

Liew FY, Ralston SH. Requirement of the inducible nitric oxide

synthase pathway for IL-1 induced osteoclastic bone resorption.

Proc Natl Acad Sci USA 2000;97:7993–8.

[18] Brandi ML, Hukkanen M, Umeda T, Moradi-Bidhendi N,

Bianchi S, Gross SS, Polak JM, MacIntyre I. Bidirectional

regulation of osteoclast function by nitric oxide synthase iso-

forms. Proc Natl Acad Sci USA 1995;92:2954–8.

[19] Ralston SH, Ho LP, Helfrich MH, Grabowski PS, Johnston PW,

Benjamin N. Nitric oxide: a cytokine-induced regulator of bone

resorption. J Bone Miner Res 1995;10:1040–9.

[20] van’t Hof RJ, Ralston SH. Cytokine-induced nitric oxide inhibits

bone resorption by inducing apoptosis of osteoclast progenitors

and suppressing osteoclast activity. J Bone Miner Res 1997;12:

1797–804.

[21] Riancho JA, Salas E, Zarrabeitia MT, Olmos JM, Amado JA,

Fernandez-Luna JL, Gonzales-Macias J. Expression and func-

tional role of nitric oxide synthase in osteoblast-like cells. J Bone

Miner Res 1995;10:439–46.

S. Stea et al. / Biomaterials 23 (2002) 4833–4838 4837

[22] Damoulis PD, Hauschka PV. Cytokines induce nitric oxide

production in mouse osteoblast. Biochem Biophys Res Comm

1994;201:924–31.

[23] Engh CA, Bobyn JD, Glassman AH. Porous-coated hip replace-

ment. The factors governing bone ingrowth stress shielding, and

clinical results. J Bone Jt Surg 1987;69/B:45–55.

[24] Pizzoferrato A, Savarino L, Stea S, Tarabusi C. Results of

histological grading on 100 cases of hip prosthesis failure.

Biomaterials 1988;9:314–8.

[25] Wink DA, Mitchell JB. Chemical biology of nitric oxide: insights

into regulatory, cytotoxic and cytoprotective mechanisms of nitric

oxide. Free Radic Biol Med 1998;25:434–56.

[26] Ralston SH. Nitric oxide and bone: what a gas! Br J Rheumatol

1997;36, 831–8.

[27] Watkins SC, Macaulay W, Turner D, Kang R, Rubash HE,

Evans CH. Identification of inducible nitric oxide synthase in

human macrophages surrounding loosened hip prostheses. Am

J Pathol 1997;150:1199–206.

[28] Moilanen E, Moilanen T, Knowles R, Charles I, Kadoya Y, Al-

Saffar N, Revell PA, Moncada S. Nitric oxide synthase in

expressed in human macrophages during foreign body inflamma-

tion. Am J Pathol 1997;150:881–7.

[29] Shanbhag AS, Macaulay W, Stefanovic-Racic M, H Rubash.

Nitric oxide release by macrophages in response to particulate

wear debris. J Biomed Mater Res 1998;41:497–503.

[30] HukkanenM, Corbett SA, Batten J, Konttinen YT, McCarthy ID,

Maclouf J, Santavirta S, Hughes SP, Polak JM. Aseptic loosening

of total hip replacement. Am J Pathol 1997;150:1199–206.

S. Stea et al. / Biomaterials 23 (2002) 4833–48384838