ORIGINAL ARTICLE

Tenosynovial giant cell tumour (pigmented villonodularsynovitis-)-like changes in periprosthetic interface membranes

Stephan Söder1 & Stefan Sesselmann2& Thomas Aigner3 & Stephan Oehler4 &

Abbas Agaimy1

Received: 9 August 2015 /Revised: 2 October 2015 /Accepted: 21 October 2015# Springer-Verlag Berlin Heidelberg 2015

Abstract Tenosynovial giant cell tumour (TSGCT; synonym,pigmented villonodular synovitis (PVNS)) is a rare low-grademesenchymal neoplasm of either intra-articular or extra-articular origin. The etiopathogenesis of TSGCT is still un-certain, but recent studies showed a translocation involvingcolony-stimulating factor 1 (CSF-1) gene in a subset ofcases. Histological features mimicking TSGCT can some-times be encountered in periprosthetic interface membranes.To investigate the frequency and morphologic spectrum ofthis phenomenon, we conducted a systematic analysis of477 periprosthetic interface membranes and performed im-munohistochemical analysis on a subset of lesions com-pared to genuine TSGCT. In 26 of 477 periprosthetic mem-brane samples (5 %), at least some TSGCT-like featureswere found and 18 cases (4 %) strongly resembled it. Wearparticles were detected in 100 % of the TSGCT-like lesionsbut only in 63.3 % of the whole cohort of periprostheticmembranes (p value <0.001). Immunohistochemistry com-paring true TSGCT and TSGCT-like membranes showedsimilar inflammatory infiltrates with slightly elevatedCD3+/CD8+ T lymphocytes and a slightly higher

proliferation index in TSGCT samples. In conclusion,TSGCT-like changes in periprosthetic membranes likelyrepresent exuberant fibrohistiocytic inflammatory responseinduced by wear particles and should be distinguished fromgenuine (neoplastic) TSGCT. Although TSGCT andTSGCT-like periprosthetic membranes represent differententities, their comparable morphology might reflect analo-gous morphogenesis.

Keywords PVNS .PVNS-like .TSGCT .Tenosynovial giantcell tumour .Wear particles . Periprosthetic interfacemembranes

Introduction

Tenosynovial giant cell tumour (TSGCT; synonym,pigmented villonodular synovitis (PVNS)) is an uncommonlow-grade mesenchymal lesion affecting both intra-articularand extra-articular sites. The intra-articular form ischaracterised macroscopically by massively thickened,brown-reddish synovial villi and microscopically by nodularproliferations of synovial cells, multi-nucleated osteoclast-likegiant cells, macrophages and foamy cells as well as prominentintracellular iron deposits [1, 2]. TSGCT can occur in a local-ised or a diffuse form and may ultimately lead to the destruc-tion of the affected joint, which is most commonly the knee [3,4]. In isolated cases, malignant transformation of TSGCT hasbeen reported [5–8].

Typically, individuals in their 3rd or 4th decades are affect-ed with a frequency of 1.8 cases/106 per year [3, 9–11]. Theexact etiopathogenesis of TSGCT remained fully obscure un-til recently [10]. In the past, two main theories have beenproposed by different authors: one regarding TSGCT(PVNS) as a reactive lesion resulting from chronic

* Abbas Agaimyabbas.agaimy@uk-erlangen.de

1 Institute of Pathology, Friedrich-Alexander-University,91054 Erlangen, Germany

2 Department of Orthopaedic Surgery,Friedrich-Alexander-University, 91054 Erlangen, Germany

3 Institute of Pathology, Hospital of Coburg, 96450 Coburg, Germany4 Department of Orthopaedics, Hospital of Rummelsberg,

90592 Schwarzenbruck, Germany

Virchows ArchDOI 10.1007/s00428-015-1874-9

inflammatory changes [12, 13] and the other considering it atrue neoplasm [14–16]. Recent studies are in favour of theneoplastic theory, based on a specific translocation involvingthe CSF-1 gene locus detected in some cases [15, 17].

There have been infrequent reports on the occurrence ofTSGCT after total endoprosthesis (TEP) implantation [18,19]. Most reported patients were in their 6th or 7th decades,and the presence of wear particles such as polyethylene debrisor polymethylmethacrylate (PMMA) was described [18]. Afew other reports described nodular proliferations in so-called periprosthetic interface membranes. For example, Karlet al. [20] reported a case of a non-pigmented villonodularsynovitis after arthroplasty.

In general, periprosthetic membranes consist of connectivetissue formed at the interface between the endoprosthesis andthe adjacent bone and neosynovia. The development ofperiprosthetic interface membranes can lead to loosening ofthe implant [21]. Because of differences in age distributionand in clinical presentation and the close association with anendoprosthesis, it seems likely that a lesion that resemblesTSGCT arising in joints post-TEP implantation is a specialtype of periprosthetic interface membrane with TSGCT-likehistological features rather than a classic (neoplastic) TSGCTor PVNS.

We performed a retrospective study of periprosthetic inter-face membranes to describe in detail the occurrence and char-acteristics of TSGCT-like changes following TEP implanta-tion. We also performed immunohistochemical analysis of theinflammatory infiltrates in comparison with genuine TSGCTcases.

Materials and methods

All periprosthetic membrane specimens submitted to the In-stitute of Pathology of the University Hospital of Erlangen-Nuremberg in the year 2009 were retrieved from the archives.The slides from routine pathological workup (477 cases) weresystematically screened for histological features of TSGCT(PVNS). Cases that showed distinct villous hyperplasia, fibro-sis, hemosiderin deposits and a mixed cellular infiltrate com-posed of mononuclear cells, fibroblasts, foamy histiocytes andmultinucleated giant cells were considered TSGCT-like. Forthe purpose of this study, the cases were further classified intotwo groups: high-grade if they closely resembled TSGCTandlow-grade when the histological features were less pro-nounced. As controls, 10 cases of genuine TSGCT from oursurgical pathology files were used for comparison of the cel-lular constituents of both groups by immunohistochemistry.Clinical data including follow-up information were providedby the treating doctors. Hematoxylin-eosin stained 3 μm sec-tions were used for morphological evaluation of the samplesand the classic Prussian blue reaction was used for the

detection of iron deposits. Wear particles were detected byconventional and polarisation microscopy (Leitz DM RB,Leica Microsystems, Wetzlar, Germany). Particle identifica-tion was carried out according to the histopathological particlealgorithm suggested by Krenn et al. [22]. All cases were crit-ically evaluated by two pathologists (S.S., A. A.).

For immunohistochemical analysis, 3 μm paraffin sectionsof 10 cases of TSGCTand 11 cases of TSGCT-like interphasemembranes from joints with TEPwere stained with antibodiesagainst CD3, CD4, CD8, CD20, CD34, CD68, CD138, mastcell tryptase (MCT), Ki-67 and CSF-1 (clone 5A1, 1:100000,Gene Tex, Irvine, USA) using a Ventana automated system(Ventana Medical Systems, Tucson, AR, USA) or werestained manually. All stains were performed according to thestandard conditions established for routine diagnostics at ourinstitution (Table 1). For each immunostain, at least 100 cellswere counted in three different areas and then the averagevalue was used for final statistical comparison.

Statistical analysis was performed with SPSS (SPSS Inc.,Chicago, IL, USA) using the Chi-square test and the Mann–Whitney U test. A p value <0.05 was considered statisticallysignificant.

Results

Histological evaluation

Based on criteria defined for subtyping periprosthetic mem-branes suggested by Morawietz et al. [21] and a revised con-sensus classification by Krenn et al. [23], systematic screeningof 477 periprosthetic interface membranes showed 224 cases(47 %) to be of wear particle-induced type, 116 cases (24 %)of infectious type, 78 cases (16 %) of combined type (wearparticles and infection) and 59 cases (12 %) of fibrous type.This distribution differs only slightly from the data publishedby Morawietz et al. [21], the proportions of membranes of theinfectious and the combined type being slightly higher in ourseries. Of our samples, 271 (56.8 %) were derived from thehip joint, 203 (42.6 %) from the knee, 1 from the shoulderjoint and in 2 cases clinical data regarding the localisationwere not available. Among the 477 reviewed periprostheticmembranes, 26 cases (5 %) showed similarity to TSGCTand were thus considered to be TSGCT-like membranes asdefined above (Figs. 1, 2 and 3). Of these 26 cases, 18 qual-ified as high-grade lesions. We found a higher rate of TSGCT-like features in samples from the knee (7 %), versus samplesfrom the hip (4 %), but this difference was not statisticallysignificant (p=0.24).

Polyethylene wear particles were detected in all 26 (100%)of the cases with TSGCT-like features (Fig. 4) but in only 302(63.3 %) of the whole cohort, which is statistically highlysignificant (p<0.001). In addition to polyethylene wear

Virchows Arch

particles, black pigment (metallic non-ferrous wear particlesor X-ray opacifiers like barium sulfide) was detected in 15cases with TSGCT-like features. In 14 cases, PMMA couldbe identified. No ceramic wear particles or examples of parti-cle corrosion could be found. Of the 26 cases with TSGCT-

like features, 25 were of the wear particle-induced type and 1of the combined type, but none of infectious or fibrous type.No polyethylene wear particles or other forms of wear parti-cles were found in classical TSGCTcases. Regarding the age,no statistically significant difference was found between caseswith and without TSGCT-like features (69.0±11.4 and 67.5±11.3, respectively). The elapsed time between primaryarthroplasty and revision surgery was 9.7±6.5 years in ourseries, which is consistent with the time frame reported forperiprosthetic membranes of the wear type [21].

Immunohistochemical evaluation

Immunohistochemistry using a CSF-1 antibody showed weakto moderate nuclear staining of unknown significance in al-most all samples from both groups, but specific cytoplasmicstaining was not detected (data not shown).

The distribution and prevalence of different cell popula-tions was evaluated by immunohistochemistry in 10 cases ofgenuine TSGCT and 11 cases of TSGCT-like mem-branes. The staining results are summarised in Table 2and shown in Fig. 5.

We found both groups to be strongly positive forCD68; however, expression of CD68 was stronger inthe TSGCT-like membranes as high numbers of CD68-positive cells were observed in addition to large multi-nucleated cells, foamy cells and macrophage-like syno-vial cells (so-called type A-synovial cells) in this groupwith an average of 63 % positive cells (Fig. 5f, o). Thefibroblast-like synovial cells (so-called type B-synovialcells) displayed much weaker CD68 staining, typicallywith less than 10 % positive cells.

A mixed inflammatory cell population was observed inboth TSGCT and their mimics, showing a similar distributionpattern in both categories. A similar number of CD3-positiveT cells were found between synovial and stromal cells with aslightly higher number in the TSGCT group (p=0.022). The

Table 1 Sources of antibodies and conditions for immunohistochemical stains

Antigen Detecting Species Pretreatment Dilution Source

CD3 T cells R h 1:100 Medac, Wedel, Germany

CD4 Thelper/inducer cells M h 1:50 Novocastra, Newcastle, UK

CD8 Tcytotoxic/suppressor cells M h 1:100 Dako, Hamburg, Germany

CD20 B cells M h 1:500 Dako

CD34 Endothelia M n 1:1000 Beckman Coulter, Krefeld, Germany

CD68 Histiocytes M pn 1:200 Dako

CD138 Plasma cells M h 1:500 Zytomed, Berlin, Germany

Tryptase (MCT) Mast cells M h 1:100 Dako

Mib-1 Proliferation M h 1:100 Dako

R rabbit, M mouse, h heat, n no pretreatment, pn pronase

Fig. 1 Both TSGCT (a) and TSGCT-like membranes (b) showed asimilar nodular proliferation pattern at low magnification

Virchows Arch

number of CD4-positive helper T cells and CD8-positivecytotoxic/suppressor T lymphocytes was higher in theTSGCT group than in the TSGCT-like membranes (for CD4and CD8, p=0.026 and 0.024, respectively). In general, theproportion of T lymphocytes ranged 1–5% and did not exceed10 %. No statistically significant difference was found regard-ing CD20-positive B lymphocytes and CD138-positive plas-ma cells or mast cells. CD34 immunohistochemistry showedincreased stromal vascularisation in both TSGCT andTSGCT-like membranes with no significant difference (datanot shown). True TSGCT displayed a slightly higher prolifer-ative activity on (Ki-67) than the group of TSGCT-like lesions

(mean Ki-67 index 4.5 vs. 2.5 %, respectively) but with mar-ginal statistical significance (p=0.026) (Fig. 5i, r).

Clinical “follow-up”

During a follow-up period of 5 years, 11 additionalperiprosthetic membranes were submitted from 5 of the 26patients with TSGCT-like periprosthetic membranes. One ofthese patients had an additional TSGCT-like membrane sub-mitted 5 years later and another periprosthetic membrane withsimilar morphology and polyethylene wear submitted 9 yearsearlier. Thus, this patient had TSGCT-like changes on three

Fig. 2 TSGCT displayed amixed cellular infiltrate mainlycomposed of mononuclear cellsand fibroblasts (a, b); at highmagnification, multinucleatedgiant cells (c) and foamyhistiocytes (d) are seen. TSGCTshows a strong staining withclassic Prussian blue stain (e).(Panels a at ×10; b, e at ×20 and c,d at ×63)

Fig. 3 TSGCT-like membranesdisplayed a similar mixed cellularinfiltrate composed ofmononuclear cells and fibroblasts(a, b) as seen in TSGCT. Highermagnification of multinucleatedgiant cells (c) and foamyhistiocytes (d). Strong stainingwith classic Prussian blue stain(e). (Panels a at ×10; b, e at ×20and c, d at ×63)

Virchows Arch

different occasions during a period of 14 years. None of theother 25 patients had previous or subsequent additionalTSGCT-like periprosthetic membranes.

Discussion

In this large-scale study on periprosthetic membranes, wefound TSGCT-like reaction/changes in 5 % of samples de-rived from hip and knee joints with endoprosthesis. The over-all frequency of TSGCT-like changes in periprosthetic

membranes in our series is obviously much higher than wouldbe expected for sporadic cases of true TSGCT. In a study of1388 primary arthroplasties, Lawrence and colleagues foundonly a single case of TSGCT, indicating a frequency of 0.07%[24]. Given the rarity of typical cases of TSGCT, the differentage distribution of TSGCT-like lesions compared to classicalTSGCT and the uniform association with wear particles, it isvery likely that TSGCT-like (PVNS-like) periprosthetic mem-branes are reactive conditions and not “true” neoplasticTSGCT/PVNS in the classical sense. Based on these observa-tions, the episodic reports of PVNS after total endoprosthesis[18, 19] might at least in part have represented PVNS-likechanges in periprosthetic membranes in agreement with ourfindings, especially as wear particles or at least black pigmentwas reported in most of these “PVNS” cases [18, 19]. Even ifthe different clinical situation of TSGCT/PVNS and TSGCT-like periprosthetic membranes strongly suggest that both rep-resent different entities, their similar morphology might be theresult of similarities in pathogenesis independent from theiretiology.

A significant histological finding in our current study is thedetection of wear particles in 100 % of the TSGCT-likeperiprosthetic membranes. There is mounting evidence thatwear particles can induce chronic inflammatory reactions, in-cluding increased expression of proinflammatory cytokinesthat ultimately may lead to osteolysis and aseptic looseningof the endoprosthesis (reviewed in Lin et al. 2014) [25–27].Interestingly, Lin et al. demonstrated that polyethylene wearparticles can induce cytokine expression [28]. In our study,CD8-positive T lymphocytes were found in both TSGCTand TSGCT-like periprosthetic membranes. An important roleof CD8-positive lymphocytes in the development of TSGCT/PVNS has been reported earlier [13]. As yet, it has not beenpossible to establish definitely whether all cases of classicalPVNS do represent reactive lesions [12, 13] or true neoplasms[14–16]. Numerous genetic alterations have been describedfor TSGCT/PVNS [6, 29, 30], but most were only found ina small subset of cases. One exception to this is the transloca-tion at 1p13 chromosome, which leads to a fusion ofCSF-1 tothe COL6A3 gene [15]. This specific translocation leads toactivation of CSF-1 expression and higher levels of CSF1R.It has been detected in 61% of TSGCTcases. However, only aminority of the lesional cells (2–16 %) carry this translocation[16]. Furthermore, elevated CSF-1 and CSF1R expressionlevels are not limited to cases with CSF-1/COL6A3 fusionand have been consistently demonstrated in cases without thistranslocation [16]. CSF-1 activates and attracts monocytes andmacrophages, and is involved in the formation of osteoclast-like giant cells [31], but increased expression of CSF-1 hasalso been reported in rheumatoid arthritis and reactive syno-vitis [16], as well as in periprosthetic membranes [32]. Thesimilarity of the mononuclear inflammatory infiltrate and thepresence of CSF-1 in both TSGCT and TSGCT-like

Fig. 4 With hematoxylin-eosin (a) and polarisation microscopy (b)polyethylene wear particles (arrow heads) are detected in TSGCT-likemembranes. (Panels a and b at ×63)

Table 2 Results of immunostaining in TSGCT and TSGCT-likemembranes

Antigen TSGCT-like TSGCT p value

CD3 (T cells) 5.0±3.8 8.2±4.0 0.022

CD4 (Thelper/inducer) 3.8±2.5 6.7±2.9 0.026

CD8 (Tcytotoxic/suppressor) 4.1±2.4 6.6±2.0 0.024

CD20 (B cells) 1.9±2.8 2.0±1.6 >0.05

CD68 (A synoviocytes) 63.2±9.6 56.0±8.4 >0.05

CD68 (B synoviocytes) 7.7±2.6 5.7±1.6 >0.05

CD138 (plasma cells) 1.0±2.0 1.5±2.5 >0.05

Tryptase (mast cells ) 1.9±0.9 1.5±0.8 >0.05

Mib-1 (Proliferation) 2.5±0.8 4.5±2.6 0.026

B and T cell markers, CD68 CD138 and MCT are given in percent

Virchows Arch

Fig. 5 By immunohistochemistry, TSGCT (a–i) and TSGCT-likemembranes (j–r) display similar staining patterns with the formertypically showing a stronger inflammatory infiltrate. Both TSGCT andTSGCT-like lesions stained for T cell markers CD3 (a, j), CD4 (b, k),CD8 (c, l) and CD20 (d, m) as B cell marker. The vascularisation is

visualised with CD34 (e, n), and the synovial cells with CD68 (f, o). Inboth groups, only a few plasma cells were detected with CD138 (g, p).Mast cells are visualised with mast cell tryptase (h, q). Mib-1 as aproliferation marker (i, r) show the proliferation in both groups. (Allpanels are at ×10)

Virchows Arch

periprosthetic membranes make it likely that there are similar,albeit complex, interactions between different inflammatorycell populations involving cytokines that lead to the observedphenotype shared by the two groups of lesions. A furtherfactor, or at least a co-factor, in the development of bothTSGCT and TSGCT-like periprosthetic membranes might beiron particles in the form of haemosiderin, which is found inaggregates in both conditions. The accumulation of iron inclassical TSGCT and in TSGCT-like periprosthetic mem-branes is likely to be a consequence of chronic inflammationand/or increased susceptibility to haemorrhage [13], and itresults in cell damage which could in turn induce effectslike cell activation and proliferation. As it is known thatp r o t e a s e s a r e i n vo l v ed i n t h e l oo s en i ng o fendoprosthesis in periprosthetic membranes and in jointdestruction in TSGCT, it is of interest to explore factorsinvolved in the regulation of proteases [33].

From a diagnostic point of view, our study indicates that thediagnosis of TSGCT in a periprosthetic membrane should bemade with great caution, as up to 5 % of periprosthetic mem-branes can display features closely similar to genuine TSGCT,especially if wear particles are detectable. Careful histologicalassessment should facilitate recognition of these reactivetumour-like lesions (Table 3). TSGCT-like lesions found inthe neosynovia of periprosthetic membranes probablyrepresent an exuberant fibrohistiocytic inflammatory re-action, likely induced by wear particles. Their develop-ment might be governed by the same pathological

principles responsible for the pathogenesis of classicalTSGCT in native synovial membranes, probably involv-ing chronic inflammatory/irritation-induced synovial hy-perplasia and cytokines. Our findings therefore providea potential clue (and possibly a model) for better under-standing of the pathophysiological and pathogeneticmechanisms responsible for the initiation and growthof classical TSGCT. As reproducible molecular tech-niques become available, genetic studies including inparticular CSF-1 alterations are necessary to explore and com-pare the pathogenesis of these architecturally similar but clin-ically and biologically distinctive groups of lesions.

Acknowledgments We are grateful to Claudia Giedl (Immunohisto-chemistry Laboratory, Erlangen) and Ingrid Mons (Department of Exper-imental Tumour Pathology, Erlangen) for technical assistance.

Compliance with ethical standards

Conflict of interest The authors declare no conflict of interest.

References

1. Darling JM, Goldring SR, Harada Y, Handel ML, Glowacki J,Gravallese EM (1997) Multinucleated cells in pigmentedvillonodular synovitis and giant cell tumor of tendon sheath expressfeatures of osteoclasts. Am J Pathol 150:1383–1393

Table 3 Common and divergent features of TSGCT-like periprosthetic membranes and TSGCT

Feature TSGCT-like periprosthetic membranes Genuine TSGCT

Clinical setting Uniformly related to joint endoprosthesis De novo neoplasm unrelated to endoprosthesis

Age Typically older people affected (>60 years) Typically young to mid-aged (<40 years)

Tissue of origin Periprosthetic tissue (neosynovia) Tendosynovial tissue

Relation to joints Uniformly periprosthetic location May be either intra-articular or extra-articular

PE wear particles Uniformly detectable Unrelated to PE wear particles

histological architecture Mainly villous pattern with brown-reddish villi Variable combination of villous pattern andsolid-nodular areas

Predominant cell type Large foamy macrophages with bluish cytoplasmichinge containing foreign material

Mononuclear ovoid to elongated fibroblast-likecells

Giant cells Variable, may be prominent, mainly of foreignbody type

Highly variable, can be scare, both of osteoclasticand histiocytoid type

intracellular iron deposits Variable, may be absent or minimal Frequently present

Predilection Knee joints Knee joints

Behaviour Can persist or recur over long time, but rarely High local recurrence rate

Metastatic potential None Rare

Etiology Exuberant fibrohistiocytic inflammatory responseinduced by wear particles

Related to CSF-1 gene fusions at least in asubset of cases

Overall frequency in periprosthetic membranes 5 % (current study) 0.07 % (ref. [24])

Virchows Arch

2. Neale SD, Kristelly R, Gundle R, Quinn JM, Athanasou NA (1997)Giant cells in pigmented villo nodular synovitis express an osteo-clast phenotype. J Clin Pathol 50:605–608

3. Durr HR, Stabler A, Maier M, Refior HJ (2001) Pigmentedvillonodular synovitis. Review of 20 cases. J Rheumatol28:1620–1630

4. Miller WE (1982) Villonodular synovitis: pigmented andnonpigmented variations. South Med J 75:1084–1086

5. Imakiire N, Fujino T, Morii T et al (2011) Malignant pigmentedvillonodular synovitis in the knee—report of a case with rapid clin-ical progression. Open Orthop J 5:13–16

6. Layfield LJ, Meloni-Ehrig A, Liu K, Shepard R, Harrelson JM(2000) Malignant giant cell tumor of synovium (malignantpigmented villonodular synovitis). Arch Pathol Lab Med 124:1636–1641

7. Li CF, Wang JW, Huang WW et al (2008) Malignant diffuse-typetenosynovial giant cell tumors: a series of 7 cases comparing with24 benign lesions with review of the literature. Am J Surg Pathol32:587–599

8. Nielsen AL, Kiaer T (1989) Malignant giant cell tumor ofsynovium and locally destructive pigmented villonodular synovitis:ultrastructural and immunohistochemical study and review of theliterature. Hum Pathol 20:765–771

9. Chin KR, Brick GW (2002) Extraarticular pigmented villonodularsynovitis: a cause for failed knee arthroscopy. Clin Orthop RelatRes 404:330–338

10. Mankin H, Trahan C, Hornicek F (2011) Pigmented villonodularsynovitis of joints. J Surg Oncol 103:386–389

11. Ma X, Shi G, Xia C, Liu H, He J, Jin W (2013) Pigmentedvillonodular synovitis: a retrospective study of seventy five cases(eighty one joints). Int Orthop 37:1165–1170

12. Aigner T, Oehler S, Niedobitek HG, Fassbender G, Kirchner T(1998) Iron deposits, cell populations and proliferative activity inpigmented villonodular synovitis of the knee joint. Verh Dtsch GesPathol 82:327–331

13. Oehler S, Fassbender HG, Neureiter D, Meyer-Scholten C,Kirchner T, Aigner T (2000) Cell populations involved inpigmented villonodular synovitis of the knee. J Rheumatol 27:463–470

14. Berger I, Rieker R, Ehemann V, Schmitz W, Autschbach F,Weckauf H (2005) Analysis of chromosomal imbalances by com-parative genomic hybridisation of pigmented villonodular synovi-tis. Cancer Lett 220:231–236

15. West RB, Rubin BP, Miller MA et al (2006) A landscape effect intenosynovial giant-cell tumor from activation of csf1 expression bya translocation in a minority of tumor cells. Proc Natl Acad Sci U SA 103:690–695

16. Cupp JS, Miller MA, Montgomery KD et al (2007) Translocationand expression of csf1 in pigmented villonodular synovitis,tenosynovial giant cell tumor, rheumatoid arthritis and other reac-tive synovitides. Am J Surg Pathol 31:970–976

17. Rubin BP (2007) Tenosynovial giant cell tumor and pigmentedvillonodular synovitis: a proposal for unification of these clinicallydistinct but histologically and genetically identical lesions. SkeletRadiol 36:267–268

18. Ma XM, Xia CY, Fu PL, Liu HM, Yu HY, He J (2014) Unusualcases of pigmented villonodular synovitis after arthroplasty. Int JClin Exp Med 7:1150–1154

19. Chung BJ, Park YB (2011) Pigmented villonodular synovitis afterTKA associated with tibial component loosening. Orthopedics 34:e418–e420

20. Karl LA, Sundstrom WR (1991) Prosthesis-induced synovitis sim-ulating villonodular synovitis. Wis Med J 90:165–168

21. Morawietz L, Classen RA, Schroder JH et al (2006) Proposal for ahistopathological consensus classification of the periprosthetic in-terface membrane. J Clin Pathol 59:591–597

22. Krenn V, Thomas P, Thomsen M et al (2014) Histopathologicalparticle algorithm. Particle identification in the synovia and theslim. Z Rheumatol 73:639–649

23. Krenn V, Morawietz L, Perino G et al (2014) Revised histopatho-logical consensus classification of joint implant related pathology.Pathol Res Pract 210:779–786

24. Lawrence T, Moskal JT, Diduch DR (1999) Analysis of routinehistological evaluation of tissues removed during primary hip andknee arthroplasty. J Bone Joint Surg: Am Vol 81:926–931

25. Lin TH, Yao Z, Sato T et al (2014) Suppression of wear-particle-induced pro-inflammatory cytokine and chemokine production inmacrophages via NF-kappaB decoy oligodeoxynucleotide: a pre-liminary report. Acta Biomater 10:3747–3755

26. PaulusAC, Frenzel J, Ficklscherer A et al (2014) Polyethylenewearparticles induce TLR 2 upregulation in the synovial layer of mice. JMater Sci Mater Med 25:507–513

27. Yao Z, Keeney M, Lin TH et al (2014) Mutant monocytechemoattractant protein 1 protein attenuates migration of and in-flammatory cytokine release by macrophages exposed to orthope-dic implant wear particles. J Biomed Mater Res Part A 102:3291–3297

28. Lin TH, Kao S, Sato T et al (2015) Exposure of polyethylene par-ticles induces interferon-gamma expression in a natural killer Tlymphocyte and dendritic cell coculture system in vitro: a prelimi-nary study. J Biomed Mater Res Part A 103:71–75

29. Dahlen A, Broberg K, Domanski HA et al (2001) Analysis of thedistribution and frequency of trisomy 7 in vivo in synovia frompatients with osteoarthritis and pigmented villonodular synovitis.Cancer Genet Cytogenet 131:19–24

30. Ohjimi Y, Iwasaki H, Ishiguro M et al (1996) Short arm of chro-mosome 1 aberration recurrently found in pigmented villonodularsynovitis. Cancer Genet Cytogenet 90:80–85

31. Fujikawa Y, Sabokbar A, Neale S, Athanasou NA (1996) Humanosteoclast formation and bone resorption by monocytes andsynovial macrophages in rheumatoid arthritis. Ann RheumDis 55:816–822

32. Al-Saffar N, Revell PA (2000) Differential expression oftransforming growth factor-alpha and macrophage colony-stimulating factor/colony-stimulating factor-1R (c-fins) by multi-nucleated giant cells involved in pathological bone resorption atthe site of orthopaedic implants. J Orthop Res 18:800–807

33. Uchibori M, Nishida Y, Tabata I et al (2004) Expression of matrixmetalloproteinases and tissue inhibitors of metalloproteinases inpigmented villonodular synovitis suggests their potential role forjoint destruction. J Rheumatol 31:110–119

Virchows Arch